<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:wfw="http://wellformedweb.org/CommentAPI/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
	xmlns:slash="http://purl.org/rss/1.0/modules/slash/"
	>
<channel>
	<title>Search results for “income inequality” | Economic Policy Institute</title>
	<atom:link href="https://www.epi.org/search/income+inequality/feed/rss2/" rel="self" type="application/rss+xml" />
	<link>https://www.epi.org</link>
	<description>Research and Ideas for Shared Prosperity</description>
	<lastBuildDate>Fri, 20 Nov 2026 19:37:12 +0000</lastBuildDate>
	<language>en-US</language>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>https://wordpress.org/?v=7.0.9</generator>

<image>
	<url>https://files.epi.org/uploads/cropped-EPI-favicon-32x32.webp</url>
	<title>Search results for “income inequality” | Economic Policy Institute</title>
	<link>https://www.epi.org</link>
	<width>32</width>
	<height>32</height>
</image> 
		<item>
		<title>Policymakers have tried tax cuts to compensate losses in market income growth. That strategy has failed: Why predistribution policies need to be at the forefront of any inequality agenda</title>
		<link>https://www.epi.org/blog/policymakers-have-tried-tax-cuts-to-compensate-losses-in-market-income-growth-that-strategy-has-failed-why-predistribution-policies-need-to-be-at-the-forefront-of-any-inequality-agenda/</link>
		<pubDate>Fri, 20 Nov 2026 16:00:54 +0000</pubDate>
		<dc:creator><![CDATA[Hilary Wething]]></dc:creator>
		<guid isPermaLink="false">https://www.epi.org/?post_type=blog&#038;p=328626</guid>
					<description><![CDATA[The inequality gap is still large, and tax cuts have not narrowed it. Policymakers should return to a predistribution strategy to tackle it.]]></description>
										<content:encoded><![CDATA[<div class="quick-card">
<p><strong><span style="font-family: 'Harriet Display', serif; font-size: 18px; color: #0e5b8b;">Key takeaways</span></strong></p>
<ul>
<li><span style="font-size: 14px;">For the past 45 years, policymakers have attempted to reduce income inequality by relying more on cutting federal taxes rather than boosting transfers like Medicaid.</span></li>
<li><span style="font-size: 14px;">But the tax and transfer system has fallen short of this goal—narrowing the inequality gap from 32% to 20% between 1979 and 2022, only by about one-third of the total gap. At the same time, policies were passed that ripped away any bit of bargaining power or leverage typical workers had in labor markets.</span></li>
<li><span style="font-size: 14px;">Since the data show that we can’t tax and transfer our way out of inequality, the final option remains to change policies that affect the distribution of market income so that inequality doesn’t get too large in the first place.</span></li>
</ul>
</div>
<p>Tax cuts have become the single most namechecked policy in D.C. This has long been the priority of Republican policymaking. Post 1979 though, a nontrivial share of Democrats has also adopted the neoliberal view that markets should be allowed to generate more inequality in the pursuit of economic prosperity. They reasoned that policy could limit the downsides by compensating those on the losing end of higher market-driven inequality by making taxes, social insurance, such as Social Security, and public transfers, such as Medicaid, more equalizing. This strategy is often shorthanded as <a href="https://academic.oup.com/qje/article-abstract/141/3/2087/8571578">“compensating the losers”</a>—those who have been left worse off.</p>
<p>For the past 45 years, to the degree policymakers have aimed to compensate for rising market-based inequality, they have relied more on cutting federal taxes rather than boosting transfers. They cut federal taxes for middle-income families significantly— by about a third. Yet, even with these cuts, middle-class economic security has not substantially improved over time, and much of market-driven inequality has gone uncompensated. Even as support for more universal public benefits has grown in recent years among some Democrats, too many in Washington still have an ingrained reflex of “maybe a tax cut will do it”?</p>
<p>In this post we probe why “compensating the losers” through redistribution failed to reduce economic inequality on its own. While policymakers increased transfers and cut middle-class taxes, the transfers programs were targeted mostly for elderly and low-income households, and the policies that were passed ripped away any bit of bargaining power or leverage typical workers had in labor markets. These policies ultimately supported employers putting a lot less into typical workers’ paychecks, even as the federal government took a bit less out.</p>
<p><span id="more-328626"></span></p>
<h4><strong>The Democratic Party’s shift from predistribution and toward redistribution </strong></h4>
<p>There was once a time when income growth was broadly shared because the federal government supported predistribution—institutions and policies that intervened in labor markets much earlier than at tax withholding time. These institutions and policies include unions, the federal minimum wage, full employment, and high marginal tax rates, which all influenced pretax outcomes. Growth was broadly shared because typical workers’ paychecks kept pace with economywide growth.</p>
<p>But the shift beginning in the 1970s saw a segment of the Democratic Party becoming less supportive of predistribution measures. Today, middle-class families have a $30,676—or 31.8%—gap between what they earned from the market and what their income would have been if it had grown at the same rate as average market income did.</p>
<p>To be clear, Republicans have always been fine to roll back policies that intervened in instances&nbsp;of unequal income trajectories. What changed during this time was a growing number of Democrats who accepted the neoliberal logic that removing predistribution policies to equalize markets was fine, so long as policy sought to “compensate the losers” by making taxes and public transfers more equalizing.&nbsp;In this view, Democrats advocated for tax cuts to the middle class and&nbsp;<a href="https://academic.oup.com/qje/article-abstract/141/3/2087/8571578">stopped introducing predistribution bills</a> to strengthen worker labor standards and pay.</p>
<p>The resulting tax cuts and higher transfer incomes accruing to middle-fifth families blunted only about a third of the income penalty imposed by rising inequality. When the definition of income includes tax and transfer policies, such as social insurance, cash assistance, and tax payments, the gap between what middle-class families earned and what they could have earned if their income grew at the rate of average income is narrower but still remains large at $19,320—or 20.2 % in 2022. In effect, the strategy of “compensating the losers” in the neoliberal game still left middle-class families with a loss of about two-thirds of the income they might have had otherwise.</p>
<p><strong>Figure A</strong> shows how the various components of the tax and transfer system closed the inequality gap from 32% to 20%. We can view this graph as a report card for how policymakers fared in their approach to prioritizing tax and transfer policies as opposed to predistribution policies. Of the nearly 12 percentage point total reduction in inequality for middle-class families, nearly half of the reduction came from social insurance programs (primarily Social Security and Medicare). While this outsize role of social insurance might be a bit counterintuitive given that our measures of inequality focus on the nonelderly population, there has been a <a href="https://www.pewresearch.org/social-trends/2022/03/24/the-demographics-of-multigenerational-households/">growing preponderance of multigenerational families</a> that is likely being picked up here.</p>
<p>Means-tested programs narrowed the 12 percentage point gap by only 2.2 percentage points. These programs are only a small contribution because their design narrowly targets low-income adults. While low-income adults may reside in households with middle-class income, most middle-class households have income too high to qualify for benefits like Medicaid or SNAP/food assistance. This relatively small contribution could grow if policymakers pushed for fiscal resources to allow for universal access to these programs, such as single-payer health care or expanded nutrition and rental assistance programs.</p>


<!-- BEGINNING OF FIGURE -->

<a name=""></a><div class="figure chart-327327 figure-screenshot figure-theme-none" data-chartid="327327" data-anchor=""><div class="figLabel"></div><img decoding="async" src="https://files.epi.org/charts/img/327327-36046-email.png" width="608" alt="" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>The rest of the reduction in inequality came from a reduction in taxes for the middle class. The overall federal tax rate for middle-class families dropped from 19.1% of income in 1979 to 13.1% in 2022. This decline in taxes owed closed about one-third of the reduction of the market-income inequality gap, or 3.9 of the 12 percentage points. Cutting taxes, especially individual income taxes, rivals the reduction in inequality attributed to social insurance programs.</p>
<p>However, tax cutting is a lever that comes with sharp costs, namely the degradation of the public sector’s capacity to do useful things, like ensuring the water system provides potable water or monitoring contagious diseases. Federal tax cuts, and the ensuing degradation of the public sector’s capacity, <a href="https://www.pewresearch.org/politics/2025/12/04/public-trust-in-government-1958-2025/">aligns with the decline in Americans’ trust in the federal government</a>, suggesting that tax cuts are not likely to be a useful lever in the future. Moreover, projections of <a href="https://www.cbo.gov/publication/62105">spending needs over the next 10 years</a> suggest that the U.S. will need more revenue, not less.</p>
<h4><strong>We need predistribution policies that rebalance the labor market</strong></h4>
<p>In total, the scale of fiscal support given to the middle class was nowhere near as large as the scale of what they lost in a labor market that is now rigged against them. The “compensate the losers” strategy of redistribution policies failed to fully close the income gap middle-class families experience from an unbalanced labor market, and the levers in the redistribution toolbox, short of universal programs, are mostly played out. You can see this policy failure not just in the income numbers, but in people’s perceptions of their economic security. The <a href="https://news.gallup.com/poll/1609/consumer-views-economy.aspx">Gallup Economic Confidence Index</a>, which summarizes Americans’ assessments of current economic conditions, is 40% below its level in 1996, and the percentage of Americans who report being satisfied with their personal life has effectively <a href="https://news.gallup.com/poll/655493/new-low-satisfied-personal-life.aspx">remained stagnant since 1979</a>.</p>
<p>In passing policies that ripped away any bit of bargaining power or leverage typical workers had in labor markets, policymakers failed to come close to compensating the losers with fiscal measures. Ultimately, policymakers failed to meaningfully reduce inequality and instead created decades of wage suppression. This wage suppression led to employers putting a lot less <em>into</em> typical workers’ paychecks, even as the federal government took out a bit less.</p>
<p>Since the data show that we can’t tax and transfer our way out of the inequality tax, the final option remains to change policies that affect predistribution income so that inequality doesn’t get too large in the first place.</p>
<hr>
]]></content:encoded>
											
	</item>
		<item>
		<title>Sixteen states will raise their minimum wages on January 1, lifting pay for over 6.6 million workers</title>
		<link>https://www.epi.org/blog/sixteen-states-will-raise-their-minimum-wages-on-january-1-lifting-pay-for-over-6-6-million-workers/</link>
		<pubDate>Fri, 20 Nov 2026 15:35:44 +0000</pubDate>
		<dc:creator><![CDATA[David Cooper, Emma Cohn]]></dc:creator>
		<guid isPermaLink="false">https://www.epi.org/?post_type=blog&#038;p=328650</guid>
					<description><![CDATA[Key More than 6.6 million workers will get a raise starting January 1 as 16 states raise their minimum The increases will range from $0.26 an hour in Nebraska to $1.27 an hour in Michigan will be the latest state to reach at least a $15 minimum wage, joining 17 states and Washington, D.C.]]></description>
										<content:encoded><![CDATA[<div class="quick-card">
<p><span style="font-size: 18px;"><strong><span style="font-family: 'Harriet Display', serif; color: #1372ae;">Key takeaways</span></strong></span></p>
<ul>
<li><span style="font-size: 16px;">More than 6.6 million workers will get a raise starting January 1 as 16 states raise their minimum wages.</span></li>
<li><span style="font-size: 16px;">The increases will range from $0.26 an hour in Nebraska to $1.27 an hour in Michigan.</span></li>
<li><span style="font-size: 16px;">Michigan will be the latest state to reach at least a $15 minimum wage, joining 17 states and Washington, D.C. More workers will live in states with a $15 or greater minimum wage than in states with a $7.25 minimum wage.</span></li>
<li><span style="font-size: 16px;">Once again, localities in Washington state will have the highest minimum wages in the country. Tukwila, for example, will raise its minimum wage to $22.45, while the city of Burien could see a new rate of $22.49 if ongoing legal proceedings are resolved and the current law is upheld.</span></li>
<li><span style="font-size: 16px;">Raising the minimum wage is a simple, yet powerful way to make life more affordable for millions of working people. State and federal policymakers should ensure wage floors meet the needs of all workers.</span></li>
</ul>
</div>
<p>Sixteen states will increase their minimum wages on January 1, boosting earnings for more than 6.6 million workers by a total of nearly $4.3 billion (see <strong>Figure A </strong>and <strong>Table 2</strong>). In addition, 48 cities and counties across the country will raise their minimum wages, lifting pay for even more workers.</p>
<p>These increases are happening because state and local lawmakers, and in some cases voters, took action to help combat affordability challenges for low-wage workers across the country. With this month’s elections dramatically reshaping legislatures across the country, workers in many more states could see their paychecks grow if incoming lawmakers similarly act to raise their state wage floors.</p>
<p><span id="more-328650"></span></p>


<!-- BEGINNING OF FIGURE -->

<a name="Figure-A"></a><div class="figure chart-327536 figure-screenshot figure-theme-none" data-chartid="327536" data-anchor="Figure-A"><div class="figLabel">Figure A</div><img decoding="async" src="https://files.epi.org/charts/img/327536-36064-email.png" width="608" alt="Figure A" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>2026 was the first year that more workers lived in states with a $15 minimum wage than in states with a $7.25 minimum wage. This January, Michigan will be the 18th state to reach the $15-an-hour threshold. The increase from $13.73 an hour is the final scheduled step from a <a href="https://www.michigan.gov/leo/bureaus-agencies/ber/wage-and-hour/min-wage">2025 law</a>; the wage will rise with inflation starting in 2028. <strong>Table 1</strong> shows the size of each state’s minimum wage increase and tipped minimum wage increase, if applicable.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Table-1"></a><div class="figure chart-327539 figure-screenshot figure-theme-none" data-chartid="327539" data-anchor="Table-1"><div class="figLabel">Table 1</div><img decoding="async" src="https://files.epi.org/charts/img/327539-36065-email.png" width="608" alt="Table 1" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>Our estimates account for both those directly and indirectly affected by the new minimum wage. To adjust their wage ladders to the new wage floor, employers often raise wages further up their pay scales, so some workers earning slightly above the minimum wage will also see a boost. We find that impacted workers will see an average increase in annual income of $645 (see Table 2). Affected workers in states with the largest minimum wage hikes, such as Michigan and Rhode Island, will see their pay go up by as much as $900 a year.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Table-2"></a><div class="figure chart-327548 figure-screenshot figure-theme-none" data-chartid="327548" data-anchor="Table-2"><div class="figLabel">Table 2</div><img decoding="async" src="https://files.epi.org/charts/img/327548-36067-email.png" width="608" alt="Table 2" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<h4><strong>Workers in two states won’t see promised raises</strong></h4>
<p>Although millions of workers will benefit from these changes in January, workers in two states with previously scheduled increases will not see their pay go up.</p>
<p>In 2025, Missouri lawmakers passed a bill, <a href="https://documents.house.mo.gov/billtracking/bills251/hlrbillspdf/1616H.01I.pdf">HB 567</a>, which repealed the annual cost-of-living adjustments that voters had approved as part of a 2024 ballot measure, Proposition A. (The same bill <a href="https://www.epi.org/blog/missouri-legislators-repealed-paid-sick-leave-a-bad-policy-decision-that-will-hurt-working-families/">also eliminated the statewide paid sick leave program</a> that was part of Proposition A.) Workers will first feel the consequences of this law in 2027, when the legislature’s repeal of the inflation adjustment will cost an estimated 492,000 workers over $300 million in additional wages they otherwise would have earned.<a href="#_note1" class="footnote-id-ref" data-note_number='1' id="_ref1">1</a></p>
<p>Notably, the Missouri legislature will have a harder time undoing future voter-approved measures, as voters in the state <a href="https://ballotpedia.org/Missouri_Amendment_6,_Initiative_and_Referendum_Powers_and_Legislative_Change_Requirements_Amendment_(2026)">passed a constitutional amendment</a> this November prohibiting the legislature from repealing or amending future citizen-led ballot initiatives without 80% support in both legislative chambers and subsequent voter approval.</p>
<p>In New York, the state’s minimum wage provision requires annual inflation adjustments unless a rise in unemployment or a drop in employment triggers an “<a href="https://www.epi.org/blog/new-yorks-minimum-wage-law-has-a-loophole-that-could-freeze-increases-starting-in-2027-this-off-ramp-provision-must-be-repealed/">off-ramp</a>.” Since the labor market has cooled in 2026, the “off-ramp” will take effect and freeze the minimum wage next year. While minimum wage increases may resume in 2028, this pause will deny much-needed raises <a href="https://www.epi.org/blog/new-yorks-minimum-wage-law-has-a-loophole-that-could-freeze-increases-starting-in-2027-this-off-ramp-provision-must-be-repealed/">to over a million workers</a> in the short term and likely mean permanently lower wages for workers in the state overall in the long run. The rationale for the off-ramp provision was always dubious—poor economic conditions do not reduce the need for higher wages. If anything, they increase the need for strong income supports, whether through social insurance programs or wage increases. Governor Hochul has proposed <a href="https://www.governor.ny.gov/sites/default/files/2026-10/gpb_44_minimum_wage_rates-memo.pdf">a slight change</a> to the law that could allow minimum wage rates to rise with inflation starting March 1, 2027, though it is pending approval by the state legislature.</p>
<h4><strong>New state legislatures bring new chances to address affordability concerns and raise minimum wages</strong></h4>
<p>Although much of the affordability conversation centers around rising costs, state lawmakers have fewer tools at their disposal to control prices, many of which are set through global markets. However, affordability is not just about prices; it’s <a href="https://www.epi.org/blog/wages-inequality-and-the-roots-of-americas-affordability-crisis/">the outcome of a race between prices and <em>incomes</em></a>—and boosting incomes is where state policymakers have straightforward tools at the ready. Raising the minimum wage puts more money directly into the pockets of those who need it the most.</p>
<p>In the recent 2026 elections, Democrats <a href="https://ballotpedia.org/Election_results,_2026:_State_government_trifectas">won new trifecta control </a>of at least six states: Arizona, Michigan, Minnesota, Nevada, Pennsylvania, and Wisconsin.<a href="#_note2" class="footnote-id-ref" data-note_number='2' id="_ref2">2</a> Of these six states, neither <a href="https://www.epi.org/minimum-wage-tracker/">Wisconsin nor Pennsylvania has a minimum wage law above the federal floor</a>, and only Arizona has at least a $15 minimum wage. Lawmakers in these states, and the many others where pro-worker lawmakers have gained power, should prioritize passing substantial minimum wage increases that <a href="https://www.epi.org/publication/setting-high-standards-for-a-federal-minimum-wage-raising-the-wage-to-two-thirds-of-the-national-median-wage-would-lift-pay-for-nearly-40-million-workers/">bring wage floors closer to true living wages</a>.</p>
<p>Legislators should also ensure that low-wage workers’ paychecks at least keep pace with inflation by indexing the minimum wage, which automatically adjusts it each year to reflect changes in prices. <a href="https://www.epi.org/minimum-wage-tracker/#/min_wage">Twenty states and dozens of cities and counties across the country</a> have enacted this straightforward policy. In fact, all but three of the state minimum wage increases taking effect this January are automatic inflation-linked increases.</p>
<p>Lawmakers looking for even more powerful measures to boost incomes over the long run should <a href="https://www.epi.org/publication/setting-high-standards-for-a-federal-minimum-wage-raising-the-wage-to-two-thirds-of-the-national-median-wage-would-lift-pay-for-nearly-40-million-workers/">target ambitious wage levels and then link the minimum wage</a> to average or median wage growth. Because median wages typically outpace prices, wage indexing protects low-wage workers from rising costs and ensures their paychecks grow with improvements in the broader economy.</p>
<p>Connecticut already links its automatic minimum wage adjustments to wage growth, while several other states and many localities maintain strong minimum wages that are indexed to inflation. Four states (California, Connecticut, Rhode Island, and Washington) and the District of Columbia will have minimum wages above $17 an hour this winter, and these minimum wages may reach $20 an hour by 2030.</p>
<p><strong>Table 3</strong> shows that all but one of the localities raising their minimum wages in January have floors above $15 an hour. Additionally, seven localities in Washington and one in California already have or will have minimum wages above $20 an hour. And in Oakland, <a href="https://eastbayinsiders.substack.com/p/oaklands-30-minimum-wage-initiative?mc_cid=a39390cbe7&amp;mc_eid=054d6baea9">advocates have secured enough signatures to land a $30 by 2037 minimum wage initiative</a> on a future ballot.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Table-3"></a><div class="figure chart-327543 figure-screenshot figure-theme-none" data-chartid="327543" data-anchor="Table-3"><div class="figLabel">Table 3</div><img decoding="async" src="https://files.epi.org/charts/img/327543-36066-email.png" width="608" alt="Table 3" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>Decades of economic research <a href="https://www.epi.org/blog/most-minimum-wage-studies-have-found-little-or-no-job-loss/">shows</a> that raising minimum wages is an effective way to boost incomes for low-wage workers, with little if any impact on jobs. As lawmakers seek strategies for addressing households’ affordability concerns, this tried-and-true method should be high on their list.</p>
<h4>Notes</h4>
<p data-note_number='1'><a href="#_ref1" class="footnote-id-foot" id="_note1">1. </a> Authors’ analysis using Economic Policy Institute Minimum Wage Simulation Model; 2021–2024 4-year ACS data pinned to 2025 CPS wage distribution. Employment scaled to match 2026 CPS labor force size. For more details see <a href="https://www.epi.org/publication/minimum-wage-simulation-model-technical-methodology/#:~:text=In%20all%20cases%2C%20the%20EPI,workers%20in%20the%20applicable%20jurisdiction">technical methodology</a>.</p>
<p data-note_number='2'><a href="#_ref2" class="footnote-id-foot" id="_note2">2. </a> As of this writing, North Carolina’s state senate has one race outstanding, the outcome of which will determine control of the chamber. If the Democratic candidate wins, North Carolina will also be a Democratic trifecta.</p>
]]></content:encoded>
											
	</item>
		<item>
		<title>News from EPI › The working class is nearly half people of color, new EPI report finds </title>
		<link>https://www.epi.org/press/the-working-class-is-nearly-half-people-of-color-new-epi-report-finds/</link>
		<pubDate>Thu, 19 Nov 2026 16:38:05 +0000</pubDate>
		<dc:creator><![CDATA[]]></dc:creator>
		<guid isPermaLink="false">https://www.epi.org/?post_type=press&#038;p=328616</guid>
					<description><![CDATA[The working class is more diverse than ever, according to a new Economic Policy Institute report that provides national and state-level The traditional definition of the working class comprises workers with less than a bachelor’s degree.]]></description>
										<content:encoded><![CDATA[<p>The working class is more diverse than ever, according to a <a href="https://www.epi.org/publication/todays-working-class-is-more-diverse-than-ever-a-new-picture-emerges-as-demographic-change-and-economic-inequality-challenge-old-perceptions/">new Economic Policy Institute report</a> that provides national and state-level data.&nbsp;&nbsp;</p>
<p>The traditional definition of the working class comprises workers with less than a bachelor’s degree. Under this definition:&nbsp;</p>
<ul>
<li aria-setsize="-1" data-leveltext='' data-font='Symbol' data-listid='15' data-list-defn-props='{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:[8226],&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}' data-aria-posinset='1' data-aria-level='1'>In 1979, the working class was over 80% white. Today, Black, Hispanic, Asian American and Pacific Islander (AAPI), and Native American people collectively account for nearly half of the working class. &nbsp;</li>
</ul>
<ul>
<li aria-setsize="-1" data-leveltext='' data-font='Symbol' data-listid='15' data-list-defn-props='{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:[8226],&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}' data-aria-posinset='2' data-aria-level='1'>One-fifth of the working class was born outside the United States, and 11.6% are non-U.S. citizens.&nbsp;</li>
</ul>
<p>The report also presents a broader definition of the working class: workers whose family income is in the bottom 75%—below $150,000 per year—regardless of educational level or job type. A more expansive definition should be considered because the wage boost from getting a college degree is lower than it used to be, and underemployment—along with racial and gender pay disparities—can mean the economic status of graduates of color and women is more like some nongraduates.&nbsp;</p>
<p>Under this broader definition:&nbsp;&nbsp;</p>
<ul>
<li aria-setsize="-1" data-leveltext='' data-font='Symbol' data-listid='16' data-list-defn-props='{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:[8226],&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}' data-aria-posinset='1' data-aria-level='1'>The working class remains similarly diverse in terms of race and ethnicity.&nbsp;&nbsp;</li>
</ul>
<ul>
<li aria-setsize="-1" data-leveltext='' data-font='Symbol' data-listid='16' data-list-defn-props='{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:[8226],&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}' data-aria-posinset='2' data-aria-level='1'>31.5% hold a bachelor’s degree or higher, and most of those college graduates hold professional or managerial jobs.&nbsp;&nbsp;</li>
</ul>
<ul>
<li aria-setsize="-1" data-leveltext='' data-font='Symbol' data-listid='16' data-list-defn-props='{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:[8226],&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}' data-aria-posinset='3' data-aria-level='1'>Traditionally women-dominated jobs in teaching, child care, and nursing account for some of the largest shares of working-class managerial and professional occupations.&nbsp;</li>
</ul>
<p>The working-class share varies widely by state, reflecting differences in costs of living and labor force participation. Using family income to define the working class, the five states with the highest working-class population shares are:&nbsp;&nbsp;</p>
<ul>
<li aria-setsize="-1" data-leveltext='' data-font='Symbol' data-listid='13' data-list-defn-props='{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:[8226],&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}' data-aria-posinset='1' data-aria-level='1'>South Dakota (66.7%)&nbsp;</li>
</ul>
<ul>
<li aria-setsize="-1" data-leveltext='' data-font='Symbol' data-listid='13' data-list-defn-props='{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:[8226],&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}' data-aria-posinset='2' data-aria-level='1'>Iowa (64.9%)&nbsp;</li>
</ul>
<ul>
<li aria-setsize="-1" data-leveltext='' data-font='Symbol' data-listid='13' data-list-defn-props='{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:[8226],&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}' data-aria-posinset='3' data-aria-level='1'>Wyoming (64.7%)&nbsp;</li>
</ul>
<ul>
<li aria-setsize="-1" data-leveltext='' data-font='Symbol' data-listid='13' data-list-defn-props='{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:[8226],&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}' data-aria-posinset='4' data-aria-level='1'>Kansas (63.9%)&nbsp;</li>
</ul>
<ul>
<li aria-setsize="-1" data-leveltext='' data-font='Symbol' data-listid='13' data-list-defn-props='{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:[8226],&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}' data-aria-posinset='5' data-aria-level='1'>Nebraska (62.7%)&nbsp;</li>
</ul>
<p>The five states (including Washington, D.C.) with the lowest working-class population shares are:&nbsp;&nbsp;&nbsp;</p>
<ul>
<li aria-setsize="-1" data-leveltext='' data-font='Symbol' data-listid='14' data-list-defn-props='{&quot;335552541&quot;:1,&quot;335559685&quot;:605,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:[8226],&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}' data-aria-posinset='1' data-aria-level='1'>Massachusetts (42.9%)&nbsp;</li>
</ul>
<ul>
<li aria-setsize="-1" data-leveltext='' data-font='Symbol' data-listid='14' data-list-defn-props='{&quot;335552541&quot;:1,&quot;335559685&quot;:605,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:[8226],&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}' data-aria-posinset='2' data-aria-level='1'>Washington, D.C. (42.9%)&nbsp;</li>
</ul>
<ul>
<li aria-setsize="-1" data-leveltext='' data-font='Symbol' data-listid='14' data-list-defn-props='{&quot;335552541&quot;:1,&quot;335559685&quot;:605,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:[8226],&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}' data-aria-posinset='3' data-aria-level='1'>New Jersey (44.9%)&nbsp;</li>
</ul>
<ul>
<li aria-setsize="-1" data-leveltext='' data-font='Symbol' data-listid='14' data-list-defn-props='{&quot;335552541&quot;:1,&quot;335559685&quot;:605,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:[8226],&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}' data-aria-posinset='4' data-aria-level='1'>Maryland (47.8%)&nbsp;</li>
</ul>
<ul>
<li aria-setsize="-1" data-leveltext='' data-font='Symbol' data-listid='14' data-list-defn-props='{&quot;335552541&quot;:1,&quot;335559685&quot;:605,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:[8226],&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}' data-aria-posinset='5' data-aria-level='1'>Washington state (47.9%)&nbsp;</li>
</ul>
<p>“The working class has long outgrown the old image narrowly represented by white tradesmen in hard hats. Shifting trends in U.S. immigration and labor force participation over the last five decades have created a more racially diverse working class than ever before,” said Valerie Wilson, co-author of the report and director of EPI’s Program on Race, Ethnicity, and the Economy. “Improving economic conditions for today’s working class requires policies that meet the needs of all these workers.”&nbsp;</p>
]]></content:encoded>
											
	</item>
		<item>
		<title>Today’s working class is more diverse than ever:  A new picture emerges as demographic change and economic inequality challenge old perceptions</title>
		<link>https://www.epi.org/publication/todays-working-class-is-more-diverse-than-ever-a-new-picture-emerges-as-demographic-change-and-economic-inequality-challenge-old-perceptions/</link>
		<pubDate>Thu, 19 Nov 2026 13:00:03 +0000</pubDate>
		<dc:creator><![CDATA[Kyle K. Moore, Stevie Marvin, Valerie Wilson]]></dc:creator>
		<guid isPermaLink="false">https://www.epi.org/?post_type=publication&#038;p=328439</guid>
					<description><![CDATA[Learn about who is included in the working class, how economic conditions have changed for them, where they work and live, and much more.]]></description>
										<content:encoded><![CDATA[<div class="quick-card">
<p><span style="font-size: 24px;"><strong><span style="font-family: 'Harriet Display', serif; color: #0e5b8b;">Key takeaways</span></strong></span></p>
<h4><span style="color: #0e5b8b;">An expanded view of the working class&nbsp;</span></h4>
<ul>
<li><b>The </b><b>working class is a concept that carries tremendous political weight, but there is no universal definition of what it is.&nbsp;</b>&nbsp;</li>
<li><b>A commonly used definition of working class comprises workers with less than a bachelor’s degree.&nbsp;</b>&nbsp;</li>
<li><b>The working class could reasonably include more college graduates and traditional “white-collar workers” than were previously considered. </b><span style="font-size: 14px;">That’s because the wage boost from getting a college degree is lower than it used to be, and underemployment—along with racial and gender pay disparities—can mean the economic status of graduates of color and women is more like some nongraduates.&nbsp;</span></li>
</ul>
<ul>
<li style="list-style-type: none;">
<ul>
<li><span style="font-size: 14px;"><strong><span class="TextRun Highlight MacChromeBold SCXW110786087 BCX0" data-contrast='auto'><span class="NormalTextRun SCXW110786087 BCX0">Given these factors, a</span><span class="NormalTextRun SCXW110786087 BCX0"> broader </span><span class="NormalTextRun SCXW110786087 BCX0">definition</span><span class="NormalTextRun SCXW110786087 BCX0"> </span><span class="NormalTextRun SCXW110786087 BCX0">of</span><span class="NormalTextRun SCXW110786087 BCX0"> the working class can be </span><span class="NormalTextRun ContextualSpellingAndGrammarErrorV2Themed SCXW110786087 BCX0">considered</span><span class="NormalTextRun SCXW110786087 BCX0"> </span><span class="NormalTextRun SCXW110786087 BCX0">people in the labor force who</span><span class="NormalTextRun SCXW110786087 BCX0">se family income </span><span class="NormalTextRun SCXW110786087 BCX0">is in</span><span class="NormalTextRun SCXW110786087 BCX0"> </span><span class="NormalTextRun SCXW110786087 BCX0">the bottom 75%</span><span class="NormalTextRun SCXW110786087 BCX0">, regardless of education</span><span class="NormalTextRun SCXW110786087 BCX0">al</span><span class="NormalTextRun SCXW110786087 BCX0"> level or job type.</span></span></strong><span class="TextRun Highlight SCXW110786087 BCX0" data-contrast='auto'><span class="NormalTextRun SCXW110786087 BCX0"> This works out to a family income </span></span><strong><span class="TextRun Highlight MacChromeBold SCXW110786087 BCX0" data-contrast='auto'><span class="NormalTextRun SCXW110786087 BCX0">below $150,000</span></span></strong><span class="TextRun Highlight SCXW110786087 BCX0" data-contrast='auto'><span class="NormalTextRun SCXW110786087 BCX0"> per year.</span></span><span class="EOP Selected SCXW110786087 BCX0" data-ccp-props='{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}'>&nbsp;</span></span></li>
</ul>
</li>
</ul>
<ul>
<li style="list-style-type: none;">
<ul>
<li><span style="font-size: 14px;">Under this broader definition, 31.5% of those in the working class have a bachelor&#8217;s degree or higher; a majority of working-class college graduates hold professional or managerial jobs.&nbsp;</span></li>
</ul>
</li>
</ul>
<ul>
<li style="list-style-type: none;">
<ul>
<li><span style="font-size: 14px;">Traditionally women-dominated jobs related to teaching, child care, or nursing make up some of the largest shares of working-class managerial and professional occupations.&nbsp;</span></li>
</ul>
</li>
</ul>
<h4><span style="color: #0e5b8b;">A racially diverse group&nbsp;</span></h4>
<ul>
<li><b>In 1979, the working class (defined by less than a bachelor&#8217;s degree) was 81.5% white.</b></li>
<li><b>Shifting trends in U.S. immigration and labor force participation over the last five decades have created a more racially diverse working class than ever before. Now, the working class is:&nbsp;</b></li>
</ul>
<ul>
<li style="list-style-type: none;">
<ul>
<li>52% white&nbsp;</li>
<li>26.8% Hispanic&nbsp;</li>
<li>15.1% Black</li>
<li>4.7% Asian American</li>
<li>1.4% Native American</li>
</ul>
</li>
<li><b>One-fifth of the working class was born outside the United States, and 11.6% are non-U.S. citizens.&nbsp;</b></li>
</ul>
<ul>
<li><b>The working class is similarly diverse when defined by family income level.&nbsp;</b></li>
</ul>
<h4><span style="color: #0e5b8b;">Differences by state (using family income to define working class)&nbsp;</span></h4>
<ul>
<li><b>There is considerable variation in the working-class composition of states, based on differences in costs of living and labor force participation.&nbsp;</b></li>
<li><b>The five states (including Washington, D.C.) with the lowest working-class population shares are:&nbsp;</b>&nbsp;
<ul>
<li>Massachusetts (42.9%)</li>
<li>Washington, D.C. (42.9%)</li>
<li>New Jersey (44.9%)</li>
<li>Maryland (47.8%)</li>
<li>Washington state (47.9%)&nbsp;</li>
</ul>
</li>
<li><b>The five states with the highest working-class population shares are:&nbsp;</b>
<ul>
<li>South Dakota (66.7%)</li>
<li>Iowa (64.9%)</li>
<li>Wyoming (64.7%)</li>
<li>Kansas (63.9%)</li>
<li>Nebraska (62.7%)</li>
</ul>
</li>
</ul>
</div>
<h2>What is the working class?</h2>
<p><span class="dropped">T</span>he working class remains an often-used rhetorical concept that carries tremendous political weight. Despite having no official definition, “working class voters” are a heavily courted constituency during election years. In the most general sense, the “working class” as a political economic category describes people who rely almost exclusively on wage labor through steady employment to support themselves and their families.</p>
<p>Historically, the term working class has been closely associated with nonmanagerial blue-collar workers and unions, often represented in the popular imagination by white tradesmen in hard hats. This popular image, however, has never represented a complete picture. Due to occupational segregation, Black workers have always made up a significant portion of nonmanagerial laborers commonly identified as working class. Additionally, since the 1970s, increases in women’s labor force participation and immigration have resulted in growing shares of women, Hispanic, and Asian workers in the U.S. labor force. Along with these demographic shifts, there have been changes in the composition of “working class” jobs over time. For example, as traditional blue-collar production and manufacturing employment in the U.S. fell sharply between 1998 and 2019, job opportunities for noncollege graduates were largely replaced by growth in service-sector employment (Scott et al. 2022).</p>
<h3>Who is working class?</h3>
<p>Contemporary views about who is included in the working class range from narrow to extremely broad: those earning very low wages, those without a bachelor’s degree, and anyone who works for a wage or salary, irrespective of their wage or education level. The latter definition, currently being pushed by the growing Democratic Socialist movement, has generated some debate around whether highly educated, relatively high-income wage and salary workers belong in a category with traditional blue-collar, noncollege-educated workers who often live vastly different material lives (Rotondaro 2026).</p>
<h3>What are some factors that complicate the traditional distinction between college-educated and other workers?</h3>
<p>For many, however, the line between the economic status of college- and noncollege-educated workers is not that clear and has been blurred by rising inequality and labor market discrimination. Over the last 10 years, the college wage premium has declined. In other words, the wage boost from a bachelor’s degree is lower than it used to be, and workers in many jobs that were once considered white collar now experience relatively modest pay and limited control over the pace and content of work (Fast and Gould 2026).</p>
<p>Research also shows that Black college graduates are more likely than white college graduates to work in jobs that don’t require a college degree, and even after controlling for college major, graduates of color, foreign-born graduates, and women of all races are much more likely to work alongside nongraduates (Williams and Wilson 2019; Attewell and Witteveen 2023). In addition to these forms of underemployment, racial and gender pay disparities among college graduates result in lower pay for women and Black and Hispanic men relative to white men (Quadlin, VanHuevelen, and Ahearn 2023; Wilson 2016a; Mora and Dávila 2018). Facing limited returns on their college education, debt-burdened bachelor’s degree holders may even find themselves in a more precarious position than nongraduate co-workers who are free of student debt.</p>
<h3>Which topics will this report cover?</h3>
<p>Taking these factors into consideration, this report updates prior EPI research on the changing demographics of the working class and presents a more expansive definition of this group that includes labor force participants in the bottom 75% of the family income distribution (Wilson 2016b). This report presents the working class as more racially diverse than ever before but also includes workers with and without college degrees employed in a range of blue-, pink<a href="#_note1" class="footnote-id-ref" data-note_number='1' id="_ref1">1</a>- and white-collar jobs that afford different standards of living across the United States. This analysis supports the idea that policies that strengthen labor standards and protect civil and workers’ rights will improve the lives of working people across traditional class lines defined by education or occupation.</p>
<h2>How are we defining the working class in this report?</h2>
<p>This report considers two definitions of the working class. Consistent with previous EPI research, the first definition is based on the share of 18–64-year-olds in the civilian labor force with less than a bachelor’s degree. Education captures much of what people typically associate with class, and in the labor market, a bachelor’s degree is commonly interpreted by employers as a signal of skill acquisition required for many white-collar occupations. Since educational categories have been more consistently coded in the monthly Current Population Survey (CPS) than other commonly used indicators of working-class status, like occupation or income, we use this definition in Part I of the report when comparing working-class demographic and economic outcomes over time.</p>
<p>The second definition, used in Part II of our analysis, is based on the share of 18–64-year-olds in the civilian labor force with family income at or below the 75th percentile. In our sample, this corresponds to individuals with less than $150,000 in family income.<a href="#_note2" class="footnote-id-ref" data-note_number='2' id="_ref2">2</a> The income-based definition captures people in similar economic circumstances, regardless of education or occupation. Implicit in this income-based definition of working class is the assumption that a job loss could result in significant economic hardship for most families below this threshold. We apply this more expansive definition to our more detailed analysis of the working class for disaggregated groups of workers by race, ethnicity, gender, nativity, geography, education, and occupation using pooled 2024–2025 CPS data (Part II).</p>
<p>Though slightly different, there is considerable overlap between these definitions: 82.9% of those who qualify as working class under the education-based definition are also defined as working class using the income-based definition.</p>
<h2>Part I: Changes in working-class demographic characteristics and economic conditions since 1979</h2>
<h3>Shifting demographic characteristics of the working class</h3>
<p>The demographic composition of the working class has changed dramatically over the last five decades, as shifting trends in U.S. immigration and labor force participation among different groups have contributed to a more diverse workforce. For comparability over time, this part of our analysis relies on the education-based definition of working class: 18–64-year-olds in the civilian labor force with less than a bachelor’s degree. Since labor force participation fluctuates across different points in the business cycle, and to varying degrees for different groups of workers, we focus on demographic comparisons across peak years in the business cycle when employment and labor force participation are highest. Those years are 1979, 1989, 2000, 2007, and 2019. The current time period covers 2024 and 2025, which combines the year with the highest rate of prime age employment since the 2019 business cycle peak (2024) and the most recent full year of data (2025) (EPI 2026b).</p>
<p><strong>Figure A </strong>clearly shows that the working class has grown more racially and ethnically diverse over the last 46 years. In 1979, the working class was over 80% white, but after 2000, the combined shares of Black, Hispanic, Asian American and Pacific Islander (AAPI), and American Indian and Alaska Native (AIAN or Native American) workers increased rapidly. While white workers remain the largest single racial group, workers of color collectively accounted for nearly half of the working class in 2024–2025. Currently, the working class is 52% white, 26.8% Hispanic, 15.1% Black, 4.7% AAPI, and 1.4% Native American.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Figure-A"></a><div class="figure chart-325442 figure-screenshot figure-theme-none" data-chartid="325442" data-anchor="Figure-A"><div class="figLabel">Figure A</div><img decoding="async" src="https://files.epi.org/charts/img/325442-35936-email.png" width="608" alt="Figure A" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>The changing race and gender composition of the working class are displayed separately for women and men in <strong>Figure B </strong>so that changes for each group are easier to see over time, but the combined percentages of men and women in each year sum to 100%. Overall, men have maintained a small majority over women in the working class (between 53% and 57%) since 1979. While there hasn’t been a huge shift in the overall gender demographics of the working class, an increase in the share of women between 1979 and 1989 reflects rising women’s labor force participation at a time when men were leaving the labor force (EPI 2026d). Women decline as a share of the working class after 2000 as the share of women with a bachelor’s degree or higher rose faster and surpassed the share of men with college degrees (Census 2022).</p>
<p>The intersection of race and gender in Figure B brings these joint demographic shifts in the working class into even sharper focus. For example, in 1979, white men alone accounted for nearly half (46.3%) of the working class. By 2024–2025, white men were less than a third (29.6%) of the working class. Over the same time, the share of working-class women of color more than doubled from 8.4% in 1979 to 21.2% in 2024–2025, with the largest increase occurring among Hispanic women (from 2.2% in 1979 to 11.1% in 2024–2025).</p>


<!-- BEGINNING OF FIGURE -->

<a name="Figure-B"></a><div class="figure chart-326126 figure-screenshot figure-theme-none chart-has-feature--two-column-chart-group-with-separator" data-chartid="326126" data-anchor="Figure-B"><div class="figLabel">Figure B</div><img decoding="async" src="https://files.epi.org/charts/img/326126-36098-email.png" width="608" alt="Figure B" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>An increase in immigration also contributed to changes in the racial and ethnic composition of the working class since 1979 (Kramer and Passel 2025). While the immigrant population has grown increasingly diverse over the last 46 years, between 2010 and 2024, the 10 largest U.S. immigrant groups have come from Latin American or Asian countries. This has contributed to the growth in working-class Hispanic and Asian workers shown in Figures A and B (MPI 2025). <strong>Figure C</strong> further shows that the share of the working class that is foreign-born has grown from 13.8% in 2000 (the earliest year these data are available in the CPS) to 19.8% in 2024–2025 (a 43.5% increase). Currently, 11.6% of the working class are non-U.S. citizens.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Figure-C"></a><div class="figure chart-325475 figure-screenshot figure-theme-none" data-chartid="325475" data-anchor="Figure-C"><div class="figLabel">Figure C</div><img decoding="async" src="https://files.epi.org/charts/img/325475-35938-email.png" width="608" alt="Figure C" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<h3>Changing economic conditions for the working class</h3>
<p>One of the most prominent narratives regarding the working class has been how people without a bachelor’s degree have been stripped of opportunities to earn a decent living and left behind in the modern economy. All too often, this narrative centers the white working class, and white men in particular, to the exclusion of all others. This section presents a more inclusive picture of changing economic conditions among the working class that depicts both shared class struggles and racial inequities within this group.</p>
<p><strong>Figure D</strong> shows that employment among prime-age (25–54-year-olds) adults without a bachelor’s degree peaked in 2000 but has since fallen for most groups. However, even at the high point in 2000, there were considerable gaps between racial and ethnic groups. As many as 80.9% of prime-age white adults with less than a bachelor’s degree were employed in 2000, compared with just 68.7% of similarly educated Native Americans in that age cohort. While employment among prime-age working-class white workers has fallen 4.0 percentage points since 2000, they have maintained the highest rates of employment among the working class, even as the gaps between white, Hispanic, and AAPI workers narrowed significantly after 2007. Over the decades, employment rates among prime-age working-class Black and Native American workers have been significantly lower than their white counterparts.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Figure-D"></a><div class="figure chart-326700 figure-screenshot figure-theme-none" data-chartid="326700" data-anchor="Figure-D"><div class="figLabel">Figure D</div><img decoding="async" src="https://files.epi.org/charts/img/326700-36012-email.png" width="608" alt="Figure D" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>In addition to a shared peak in employment from the strong economy of the late-1990s–2000, the working class has also shared the experience of inconsistent, often tepid, wage growth since 1979 because of policy choices that suppressed wages for the vast majority of workers (Gould and Wething 2026). <strong>Figure E </strong>shows annualized rates of inflation-adjusted median hourly wage growth for each group between business cycle peaks. This chart notably shows that all racial and ethnic groups suffered real wage losses (adjusted for inflation) during the 1980s, as globalization and concerted efforts to weaken unions stripped many in the working class of good-paying union jobs and bargaining power.</p>
<p>After decades of slower wage growth, median hourly wages grew fastest for all groups between 2019 and 2025, with workers of color experiencing faster wage growth than white workers—from 1.2% a year for Black workers to 1.7% a year for Native Americans. This is consistent with research that shows between 2019 and 2023, the lowest-wage workers—a group that includes many working-class people of color—had the strongest wage growth (Gould and deCourcy 2024).</p>


<!-- BEGINNING OF FIGURE -->

<a name="Figure-E"></a><div class="figure chart-326685 figure-screenshot figure-theme-none" data-chartid="326685" data-anchor="Figure-E"><div class="figLabel">Figure E</div><img decoding="async" src="https://files.epi.org/charts/img/326685-36010-email.png" width="608" alt="Figure E" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->




<!-- BEGINNING OF FIGURE -->

<a name="Figure-F"></a><div class="figure chart-326689 figure-screenshot figure-theme-none" data-chartid="326689" data-anchor="Figure-F"><div class="figLabel">Figure F</div><img decoding="async" src="https://files.epi.org/charts/img/326689-36011-email.png" width="608" alt="Figure F" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>Unfortunately, for most groups, the much-needed wage boost in recent years was not large enough to recover the prior four decades of lost ground, relative to state median wages. The wage boost failed to significantly reduce the share of working-class people whose hourly wage fell below their state median, as racial pay disparities within the working class persist. <strong>Figure F </strong>shows that between 1979 and 2019, the share of working-class Black and white workers whose wages were below the median hourly wage in their state increased by 11.1 and 6.4 percentage points, respectively. While wages among the Hispanic working class grew faster than other groups between 2000–2007 and 2007–2019, Figure F also shows they remained most likely to have wages below their state’s median in 2024–2025 (75.1%), followed closely by Black workers (74.9%). Even after receiving a stronger pandemic recovery wage boost, the median hourly wages of Black ($19.59) and Hispanic workers ($19.99) with less than a bachelor’s degree were significantly lower than those of similarly educated white workers ($22.65), as <strong>Table 1</strong> shows.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Table-1"></a><div class="figure chart-327450 figure-screenshot figure-theme-none" data-chartid="327450" data-anchor="Table-1"><div class="figLabel">Table 1</div><img decoding="async" src="https://files.epi.org/charts/img/327450-36055-email.png" width="608" alt="Table 1" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>Over the same time period that the hourly pay of more working-class wage earners dropped below the state median wage, a declining share of those workers were union members or covered by a union contract (see <strong>Figure G</strong>). The decline is especially stark among Black workers for whom union coverage in 2024–2025 (12.2%) was less than a third of what it was in 1979 (36.9%).</p>


<!-- BEGINNING OF FIGURE -->

<a name="Figure-G"></a><div class="figure chart-326704 figure-screenshot figure-theme-none" data-chartid="326704" data-anchor="Figure-G"><div class="figLabel">Figure G</div><img decoding="async" src="https://files.epi.org/charts/img/326704-36013-email.png" width="608" alt="Figure G" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<h2>Part II: Current demographic and economic characteristics of the working class</h2>
<p>The remainder of this report is based on the more expansive definition of working class that includes all 18–64-year-old labor force participants in the bottom 75% of family income, regardless of educational level or occupation. This portion of the analysis takes a more detailed look at the current demographic and labor market characteristics of the working class in 2024–2025.</p>
<p><strong>Figure H</strong> shows the share of each group, by race, ethnicity, or nativity, that meets this more expansive definition of working class. As shown in Figure H, over 60% of Black and Hispanic adults between the ages of 18 and 64 meet this definition of working class, compared with more than half of Native American and white adults and 43.4% of AAPI adults.<a href="#_note3" class="footnote-id-ref" data-note_number='3' id="_ref3">3</a> Similarly, a larger share of foreign-born non-U.S. citizens is working class—62.1% compared with about 54% of U.S. citizens, both native-born and naturalized.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Figure-H"></a><div class="figure chart-325518 figure-screenshot figure-theme-none" data-chartid="325518" data-anchor="Figure-H"><div class="figLabel">Figure H</div><img decoding="async" src="https://files.epi.org/charts/img/325518-35943-email.png" width="608" alt="Figure H" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p><strong>Figure I</strong> presents the educational distribution of the working class. Even under this broader definition that is unrestricted by educational level, the majority of those in the working class (68.5%) have less than a bachelor’s degree, with the largest group being those with only a high school diploma (32%).</p>


<!-- BEGINNING OF FIGURE -->

<a name="Figure-I"></a><div class="figure chart-325492 figure-screenshot figure-theme-none" data-chartid="325492" data-anchor="Figure-I"><div class="figLabel">Figure I</div><img decoding="async" src="https://files.epi.org/charts/img/325492-36095-email.png" width="608" alt="Figure I" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>Defined in part by labor force participation, a large majority (69.2%) of those in the working class are between 25- and 54-years-old, or prime age. With large shares of younger people attending college and not working, the working class also includes more people ages 55–64 (17.1%) than under age 25 (13.7%) (see <strong>Figure J</strong>) (Gould et al. 2025).</p>


<!-- BEGINNING OF FIGURE -->

<a name="Figure-J"></a><div class="figure chart-325488 figure-screenshot figure-theme-none" data-chartid="325488" data-anchor="Figure-J"><div class="figLabel">Figure J</div><img decoding="async" src="https://files.epi.org/charts/img/325488-35939-email.png" width="608" alt="Figure J" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>In addition to labor force participation, family income at or below the 75th percentile (less than $150,000 per year) is used to identify members of the working class under this more expansive definition. However, the relative precarity of families at the same income level can vary, based on family size and the number of household earners.</p>
<p><strong>Figure K</strong> summarizes basic family structures within the working class, based on marital status and the presence of minor children (under age 18). It shows that almost half of those in the working class are unmarried and childless (46.0%), while 30.8% have at least one child under age 18. The majority of working-class parents are married.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Figure-K"></a><div class="figure chart-325495 figure-screenshot figure-theme-none" data-chartid="325495" data-anchor="Figure-K"><div class="figLabel">Figure K</div><img decoding="async" src="https://files.epi.org/charts/img/325495-35941-email.png" width="608" alt="Figure K" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<h3>Working-class family income</h3>
<p>Although we are using a single-income threshold to define working class, <strong>Figure L</strong> reveals racial differences in family income among the working class that replicate differences in the overall population. For example, white and AAPI families are more likely to be at the upper end of the income distribution. One-third of each group has a family income of $100,000–$149,999 per year, compared with one-fourth of working-class AIAN families and one-fifth of working-class Black and Hispanic families. At the same time, all working-class families of color are more likely to be in the lower end of the income distribution than white families. More than a third of Hispanic (38.4%), Black (38.0%), and AIAN (35.8%) families have an income of less than $50,000 per year, compared with 25.2% of working-class AAPI families and 22.1% of white families. Most working-class Black and Hispanic families have income between $50,000 and $75,000 per year.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Figure-L"></a><div class="figure chart-325612 figure-screenshot figure-theme-none" data-chartid="325612" data-anchor="Figure-L"><div class="figLabel">Figure L</div><img decoding="async" src="https://files.epi.org/charts/img/325612-35953-email.png" width="608" alt="Figure L" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>The number of workers contributing to family income also varies by race and ethnicity. The number of family earners is both a function of family structure and employment status. Consistent with the fact that over half of our working-class sample is not married (55%, as shown in Figure K), <strong>Figure M</strong> shows that most families include a single earner. However, lower rates of employment among Black and Native American families further contribute to their being most likely to be single-earner families (70.8% and 69.3%, respectively). White families (27.8%) are most likely to have two earners, while Hispanic (9.9%) and Asian (7.5%) families are more likely than other groups to include more than two earners.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Figure-M"></a><div class="figure chart-327433 figure-screenshot figure-theme-none" data-chartid="327433" data-anchor="Figure-M"><div class="figLabel">Figure M</div><img decoding="async" src="https://files.epi.org/charts/img/327433-36053-email.png" width="608" alt="Figure M" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<h3>Working-class jobs</h3>
<p>Another consideration in defining what it means to be working class is the type of jobs people have. Separating professional or managerial roles from more traditionally working-class jobs closely aligned with the typical “blue-collar” definition sorts most of the occupations associated with a low degree of control and autonomy at work from those in which workers have more control over the pace and content of their work. As professionals, however, teachers and nurses would not be considered working class under a traditional blue-collar occupation definition of working class, despite teachers’ low pay relative to their years of education and nurses’ lack of control over the pace and content of their work (Allegretto 2026; Stimpfel et al. 2025).</p>
<p><strong>Figure</strong> <strong>N</strong> shows that overall, those employed in nonprofessional and nonmanagement occupations are much more likely to be working class. Between 70% and 87% of people employed in jobs other than professional or management occupations have a family income at or below the 75th percentile, compared with less than 60% of professionals and managers. This further supports the idea that even under our broader definition, working class is more often a label that applies to those in traditional blue-collar jobs.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Figure-N"></a><div class="figure chart-325530 figure-screenshot figure-theme-none" data-chartid="325530" data-anchor="Figure-N"><div class="figLabel">Figure N</div><img decoding="async" src="https://files.epi.org/charts/img/325530-35945-email.png" width="608" alt="Figure N" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>Occupation also captures much of the variation associated with education, given that education is used as a credential for access to different kinds of employment. <strong>Figure O</strong> shows the distribution of major occupation groups within the working class by level of education. Again, even under this broader definition of working class, education is still a major determinant of the types of jobs available to those in the working class. More than two-thirds of the jobs held by members of the working class with a bachelor’s degree or higher are in two major categories—professional and related occupations (43.9%), and management, business, and financial occupations (23.3%)—but the income threshold largely excludes the more highly paid college graduates in these occupations.<a href="#_note4" class="footnote-id-ref" data-note_number='4' id="_ref4">4</a> While the most common jobs among those with less than a bachelor’s degree are service occupations (24.4%), at that level of education, workers are more evenly split between office and administrative occupations (11.8%), transportation and material moving occupations (11.7%), professional and related occupations (10.2%), and management, business, and financial occupations (10.0%).</p>


<!-- BEGINNING OF FIGURE -->

<a name="Figure-O"></a><div class="figure chart-325503 figure-screenshot figure-theme-none" data-chartid="325503" data-anchor="Figure-O"><div class="figLabel">Figure O</div><img decoding="async" src="https://files.epi.org/charts/img/325503-35942-email.png" width="608" alt="Figure O" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>One limitation of using a strictly occupational definition is that the professional and management categories also include jobs that have many of the qualities traditionally associated with membership in the working class: modest pay and limited workplace autonomy, among other things. There are over 220 detailed occupations within the major occupation categories for managers and professionals, but just 22 of those occupations, depicted in <strong>Figure P</strong>, account for half of all working-class managers or professionals. Among those 22 occupations, nine are traditionally women-dominated, or pink-collar, jobs related to teaching and child care (depicted in orange) or nursing (depicted in green) that account for 22.4% of working-class managers and professionals. Seven of the 22 occupations are management positions (depicted in blue) that account for 17.8% of working-class managers and professionals. Besides the generic “all other managers” category, the most common working-class management positions are food service, general and operations, construction, and financial managers. The remainder of the occupations represented in Figure P includes other types of professional occupations (depicted in purple) such as accountants and auditors, social and human resource workers, and technology-related occupations.<br />


<!-- BEGINNING OF FIGURE -->

<a name="Figure-P"></a><div class="figure chart-326197 figure-screenshot figure-theme-none" data-chartid="326197" data-anchor="Figure-P"><div class="figLabel">Figure P</div><img decoding="async" src="https://files.epi.org/charts/img/326197-35987-email.png" width="608" alt="Figure P" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->

<br />
</p>
<p>The types of jobs held by members of the working class can also vary by race, ethnicity, and gender, following broader patterns of occupational segregation. <strong>Figure Q</strong> shows how the six most common occupations for working-class men and women are distributed by race, ethnicity, and gender. The distribution across all 10 major occupational categories can be found in <strong>Appendix Table 1. </strong>&nbsp;</p>
<p>Figure Q shows that working-class women’s employment is more concentrated than working-class men’s, with more than 80% of women of any race or ethnicity employed in just five job categories. More than half of all working-class women are employed in just two types of jobs: service, and professional and related occupations. As shown in Figure P, this is consistent with the fact that a large number of working-class professional jobs include female-dominated occupations like teachers, child care workers, and nurses. Service occupations are most common among Black (28.3%), Hispanic (32.7%), and Native American (27.6%) women, while professional occupations are most common among white and AAPI women. Across all racial and ethnic groups, similar shares of working-class women are employed in office and administrative occupations.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Figure-Q"></a><div class="figure chart-325567 figure-screenshot figure-theme-none" data-chartid="325567" data-anchor="Figure-Q"><div class="figLabel">Figure Q</div><img decoding="async" src="https://files.epi.org/charts/img/325567-35948-email.png" width="608" alt="Figure Q" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>By comparison, except for AAPI men, fewer than 80% of working-class men are employed across six job categories. Among men, however, there is more variation in the types of jobs by race or ethnicity. For example, the most common jobs held by Black working-class men are transportation (20.8%) and service (20.7%) occupations, while Hispanic working-class men are most concentrated in construction (21.7%) and service (18.8%) occupations. The most common occupations for working-class AAPI and white men are professional and related, and management, but like AAPI women, AAPI men are largely concentrated in professional and related occupations (31.5%).</p>
<h3>Job characteristics for working-class women and men by race and ethnicity</h3>
<p><strong>Table 2 </strong>further details how the employment characteristics of working-class men and women vary by race and ethnicity, underscoring differences in the work intensity required to sustain working-class family incomes. Among working-class women, Black women are most likely to work full time (83.4%) and to be represented by a union (11.6%). Although working-class Black women are more likely to have full-time jobs than their counterparts, they are still more likely to hold multiple jobs than most other groups of women. Native American women are most likely to hold multiple jobs (8.4%) and to be self-employed (5.7%) but are less likely than Black women to be employed in full-time jobs.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Table-2"></a><div class="figure chart-325608 figure-screenshot figure-theme-none" data-chartid="325608" data-anchor="Table-2"><div class="figLabel">Table 2</div><img decoding="async" src="https://files.epi.org/charts/img/325608-35952-email.png" width="608" alt="Table 2" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>Working-class men are more likely than women to work full time and therefore, less likely to hold multiple jobs. Working-class men are also more likely to be self-employed than working-class women. Among working-class men, AAPI, Black, and Native American men are least likely to be employed full time; however, Black men hold an advantage over other groups of men in union representation (13.0%) and multiple job holdings (5.9%). Hispanic working-class men are most likely to be self-employed (8.1%).</p>
<p>The lower half of Table 2 provides estimated median hourly wages by race, ethnicity, and gender in 2024–2025 as well as the share of workers in each group whose wage falls below the median hourly wage in their state. These estimates differ from those presented in Part I because they are disaggregated by gender and are based on the more expansive income-based definition of working class, which includes some college graduates. Table 2 further shows significant race and gender pay gaps within the working class. For example, larger shares of women than men in each race or ethnic group earn wages lower than their state’s median, including as many as two-thirds of Hispanic (68.3%) and Native American (67.8%) women.</p>
<h3>Working class across the country</h3>
<p>Differences in local labor market conditions, costs of living, and structural inequities across and within state boundaries introduce many more factors to consider in understanding what it means to be working class in different parts of the country. While we don’t attempt to reconcile all those things in this report, we do explore what share of people in each state fits our family income-based definition of working class and how that applies to people with different types of occupations and educational backgrounds.&nbsp;</p>
<p>On average, a larger share of those in nonmetropolitan areas, or in the Midwestern or Southern regions of the country fall within this definition of working class (see <strong>Table 3</strong>), largely because incomes and the cost of living are lower. For example, according to EPI’s Family Budget Calculator, a family of two adults and two children in Boston, Massachusetts (the state with the lowest working-class share), would require over $66,000 more for a modest but adequate lifestyle than they would in the Midwestern city of Sioux Falls, South Dakota, (the state with the highest working-class share) (EPI 2026c). In fact, 26 metropolitan areas require a family income greater than $150,000 for a two adult and two child household to achieve a moderate lifestyle, of which all but one are located in the Northeast and the West.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Table-3"></a><div class="figure chart-325559 figure-screenshot figure-theme-none" data-chartid="325559" data-anchor="Table-3"><div class="figLabel">Table 3</div><img decoding="async" src="https://files.epi.org/charts/img/325559-35946-email.png" width="608" alt="Table 3" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>The map in <strong>Figure R</strong> shows what share of each state’s 18–64-year-old population is working class, including by occupation and education. Since we are using a nationwide family income threshold to identify individuals in the working class, the map also reports the state median hourly wage, as well as working-class median hourly wages by broad occupation and educational category to help contextualize how the working-class family income limit translates to individual workers’ wages in each state.</p>
<p>The map shows that there is considerable variation in working-class shares across states and within regions. While the Midwest and South in aggregate have the highest working-class population shares, in the West, there is notable variance in state working-class population shares. For example, in Idaho and Wyoming, more than 62% of the population is working class compared with states like California and Washington where less than half of the population fit our working-class definition.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Figure-R"></a><div class="figure chart-326089 figure-screenshot figure-theme-none" data-chartid="326089" data-anchor="Figure-R"><div class="figLabel">Figure R</div><img decoding="async" src="https://files.epi.org/charts/img/326089-35979-email.png" width="608" alt="Figure R" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>In each of the five states with the highest working-class population shares—South Dakota (66.7%), Iowa (64.9%), Wyoming (64.7%), Kansas (63.9%), and Nebraska (62.7%)—the state median hourly wage is below $25.00. In the five states (including D.C.) with the lowest working-class population shares—Massachusetts (42.9%); Washington, D.C. (42.9%); New Jersey (44.9%); Maryland (47.8%); and Washington state (47.9%)–the state median hourly wage is over $29.00. Across all 50 states and D.C., the median hourly wage for members of the working class with less than bachelor’s degree ranges from $18.61 in Louisiana to $23.78 in Alaska and New Hampshire.</p>
<p>Figure R also illuminates how filtering the working class by occupation or education can surface different realities faced by workers across the country. Several states register lower-than-expected overall working-class population shares, given their reputation for being low-wage, low-income states. Even after filtering the map by people without a bachelor’s degree, in four states—Mississippi, West Virginia, New Mexico, and Louisiana, in descending order—fewer than 60% are working class. Those four states also have some of the lowest rates of labor force participation in the country, eliminating larger shares of their population from our working-class count. Since the Great Recession, the South has had the lowest labor force participation rate compared with other regions of the country, and people with less than a bachelor’s degree have much lower labor force participation rates than those with a bachelor’s or advanced degree (Childers 2024a; EPI 2026d).</p>
<p>On the other hand, when we use nonprofessional and nonmanagerial (all other) occupations as the filter, it narrows the sample to a group of people who are in the labor force, and working-class shares rise to over 80% in those four states. Mississippi, the state with the lowest median hourly wage and a working-class population share of just 59.8%, has the nation’s highest share of working class among nonprofessional and nonmanagerial workers (89.7%).</p>
<p>This look into state working-class population shares sheds light on some of the conditions that produce unique labor dynamics across regions and states. The long-standing anti-worker policies of the Southern economic development model not only shape the configuration of the working class in the South, including characteristically low wages, but low labor force participation also creates a sub-working class (Childers 2023; Childers 2024b). It also demonstrates the need for strong and comprehensive labor standards and workers’ rights protections nationwide to reduce such large interstate disparities.</p>
<h2>Conclusion</h2>
<p>While there is no single, universally agreed-upon definition of working class, this report proposes that based on similarities in family income and labor force participation, membership may reasonably extend to more college graduates and traditional “white-collar workers” than previously considered. The decline in the college-wage premium, the teacher pay gap, and growing racial and gender pay gaps among college graduates have each contributed to this trend over the last 10–20 years. At the same time, more than 40 years of policy decisions that suppressed wages and led to a sharp decline in union density have undercut traditional opportunities for noncollege graduates to make a decent living, as increasing shares of those in the working class have an hourly wage that falls below their state’s median. In comparing what it means to be working class across different states, we find that the Southern economic development model not only contributes to lower working-class wages in some Southern states but also creates a sub-working class through low labor force participation (Childers 2023).</p>
<p>Finally, this analysis also shows that the working class has long outgrown the old image narrowly represented by white tradesmen in hard hats. Rather, shifting trends in U.S. immigration and labor force participation over the last five decades have created a more racially diverse working class than ever before. In the face of the growing diversity and interconnectedness of the working class, the racist and anti-immigrant policies being pushed by the current administration stand to alienate a larger share of this group. Elected leaders seeking to win support from today’s working class should prioritize issues and policies that improve the lives of workers across race, ethnicity, immigration status, education, and occupation. Previous EPI research demonstrates that some effective strategies for accomplishing this include tripling union density, ending mass deportations, curbing employer misclassification of workers as independent contractors, and setting the federal minimum wage at two-thirds of the national median wage (Bivens et al. 2026; EPI 2026e; Cid-Martinez et al. 2026; Zipperer 2026).</p>
<h2>Appendix</h2>


<!-- BEGINNING OF FIGURE -->

<a name="Appendix-Table-1"></a><div class="figure chart-325563 figure-screenshot figure-theme-none" data-chartid="325563" data-anchor="Appendix-Table-1"><div class="figLabel">Appendix Table 1</div><img decoding="async" src="https://files.epi.org/charts/img/325563-35947-email.png" width="608" alt="Appendix Table 1" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<h2>Notes</h2>
<p data-note_number='1'><a href="#_ref1" class="footnote-id-foot" id="_note1">1. </a> The term pink-collar is commonly used to refer to jobs that are predominantly done by women and pay less than more male-dominated occupations. Examples include many clerical, administrative, and service jobs as well as teaching, nursing and care work.</p>
<p data-note_number='2'><a href="#_ref2" class="footnote-id-foot" id="_note2">2. </a> Family income is reported in bands in the monthly CPS, with $150,000 as the lower limit for the highest income band, and 73.9% of the 18–64-year-old population have a family income below that level.</p>
<p data-note_number='3'><a href="#_ref3" class="footnote-id-foot" id="_note3">3. </a> Though not depicted in the report, the shares of each group that met our education-based definition of working class were AIAN (79.1%), Hispanic (78.8%), Black (71.6%), white (57.8%), and AAPI (39.4%). Shares based on nativity were foreign-born, not a U.S. citizen (68.5%); native-born (63.0%); and foreign-born, naturalized U.S. citizen (54.6%).</p>
<p data-note_number='4'><a href="#_ref4" class="footnote-id-foot" id="_note4">4. </a> We estimate that 36.4% of working-class managers and professionals in our sample are early career or under age 35.</p>
<h2>References</h2>
<p>Abraham, Leanne, Emily Cochrane, and Jeff Adelson. 2026. “<a href="https://www.nytimes.com/interactive/2026/06/17/us/elections/redistricting-maps-black-voters-republicans.html">See How New Republican Maps Split Up Black Voters in 3 Southern States.</a>” <em>New York Times, </em>June 17, 2026.</p>
<p>Allegretto, Sylvia. 2026. <a href="https://www.epi.org/publication/the-teacher-pay-penalty-remained-high-in-2025-teachers-and-students-nationwide-are-still-paying-the-price-of-decades-of-neglect/"><em>The Teacher Pay Penalty Remained High in 2025</em>: <em>Teachers and Students Nationwide Are Still Paying the Price of Decades of Neglect</em></a>. Economic Policy Institute, August 2026.</p>
<p>Attewell, Paul, and Dirk Witteveen. 2023. “Occupational Marginalization, Underemployment, and Earnings Inequality Among College Graduates.” <em>Research in Social Stratification and Mobility </em>83, 100761. <a href="https://doi.org/10.1016/j.rssm.2023.100761">https://doi.org/10.1016/j.rssm.2023.100761</a>.</p>
<p>Bivens, Josh, Heidi Shierholz, Margaret Poydock, Ben Zipperer, Celine McNicholas, and Jennifer Sherer. 2026. <a href="https://www.epi.org/publication/the-case-for-tripling-union-membership-how-rebuilding-union-power-would-strengthen-workers-the-economy-and-our-democracy/"><em>The Case for Tripling Union Membership: How Rebuilding Union Power Would Strengthen Workers, the Economy, and Our Democracy</em></a><em>. </em>Economic Policy Institute, July 2026.</p>
<p>Childers, Chandra. 2023. <a href="https://www.epi.org/publication/rooted-in-racism/"><em>Rooted in Racism and Economic Exploitation: The Failed Southern Economic Development Model</em></a><em>.</em> Economic Policy Institute, October 2023.</p>
<p>Childers, Chandra. 2024a. <a href="https://www.epi.org/publication/rooted-racism-part2/"><em>Breaking Down the South’s Economic Underperformance: Rooted in Racism and Economic Exploitation, Part Two</em></a><em>. </em>Economic Policy Institute, June 2024.</p>
<p>Childers, Chandra. 2024b. <a href="https://www.epi.org/publication/rooted-racism-part3/#epi-toc-2"><em>Southern Policymakers Leave Workers with Lower Wages and a Fraying Safety Net: Rooted in Racism and Economic Exploitation: Part Three</em></a><em>. </em>Economic Policy Institute, July 2024.</p>
<p>Cid-Martinez, Ismael, Nina Mast, Margaret Poydock, and Valerie Wilson. 2026. <a href="https://www.epi.org/publication/misclassifying-workers-as-independent-contractors-is-costly-for-workers-and-social-insurance-systems/"><em>Misclassifying Workers as Independent Contractors Is Costly for Workers and Social Insurance Systems</em></a><em>.</em> Economic Policy Institute, April 2026.</p>
<p>Economic Policy Institute (EPI). 2026a. Current Population Survey Extracts, Version 2026.8.13, <a href="https://microdata.epi.org">https://microdata.epi.org</a>.</p>
<p>Economic Policy Institute (EPI). 2026b. “<a href="https://data.epi.org/labor_force/labor_force_emp/line/year/national/percent_emp/age_group?timeStart=2018-01-01&amp;timeEnd=2025-01-01&amp;dateString=2019-01-01&amp;highlightedLines=age_25_54">Employment by Demographics—Employment-to-Population Ratio</a>,” <em>State of Working America Data Library. </em>Accessed October 20, 2026.</p>
<p>Economic Policy Institute (EPI). 2026c. “<a href="https://www.epi.org/resources/budget/">Family Budget Calculator</a>” (web page). Last modified March 2026.</p>
<p>Economic Policy Institute (EPI). 2026d. “<a href="https://data.epi.org/labor_force/labor_force_lf/line/year/national/percent_lf/gender?timeStart=1979-01-01&amp;timeEnd=2025-01-01&amp;dateString=2000-01-01&amp;highlightedLines=gender_male&amp;highlightedLines=gender_female">Labor Force Participation—Labor Force Participation Rate</a>,”<em> State of Working America Data Library. </em>Accessed October 20, 2026.</p>
<p>Economic Policy Institute (EPI). 2026e. “<a href="https://www.epi.org/event/mass-deportations-an-anti-worker-agenda-that-harms-everyone/">Mass Deportations—an Anti-Worker Agenda That Harms Everyone</a>.” Webinar panel discussion, October 15, 2026.</p>
<p>Fast, Joe, and Elise Gould. 2026. “<a href="https://www.epi.org/blog/class-of-2026-young-college-graduates-face-a-weaker-labor-market-but-a-more-mixed-picture-than-the-headlines-suggest/">Young College Graduates Face a Weaker Labor Market—but a More Mixed Picture Than the Headlines Suggest</a>,” <em>Working Economics Blog </em>(Economic Policy Institute), May 7, 2026.</p>
<p>Gould, Elise, and Katherine deCourcy. 2024. <a href="https://www.epi.org/publication/swa-wages-2023/"><em>Fastest Wage Growth over the Last Four Years Among Historically Disadvantaged Groups: Low-Wage Workers’ Wages Surged After Decades of Slow Growth</em></a> . Economic Policy Institute, March 2024.</p>
<p>Gould, Elise, Sarah Jane Glynn, Hilary Wething, and Josh Bivens. 2025<a href="https://www.epi.org/publication/good-news-and-bad-news-about-u-s-labor-force-participation-many-headwinds-from-the-2010s-are-gone-but-were-not-investing-enough-in-the-future/"><em>.</em><em> Good News and Bad News About U.S. Labor Force Participation: Many Headwinds from the 2010s Are Gone, but We’re Not Investing Enough in the Future</em></a>. Economic Policy Institute, September 2025.</p>
<p>Gould, Elise, and Hilary Wething. 2026<em>. </em><a href="https://www.epi.org/publication/wage-suppression-in-10-charts/"><em>Wage Suppression in 10 Charts</em></a><em>. </em>Economic Policy Institute, September 2026.</p>
<p>Kramer, Stephanie, and Jeffrey S. Passel. 2025. “<a href="https://www.pewresearch.org/short-reads/2025/08/21/key-findings-about-us-immigrants/">What the Data Says About Immigrants in the U.S.</a>” <em>Pew Research Center</em>, August 21, 2025.</p>
<p>Migration Policy Institute (MPI). 2025. “<a href="https://www.migrationpolicy.org/data-tool/largest-us-immigrant-groups">Top Ten Largest U.S. Immigrant Groups, 1960–2024</a>” (web page). Accessed October 20, 2026.</p>
<p>Mora, Marie T., and Alberto Dávila. 2018. <a href="https://www.epi.org/publication/the-hispanic-white-wage-gap-has-remained-wide-and-relatively-steady-examining-hispanic-white-gaps-in-wages-unemployment-labor-force-participation-and-education-by-gender-immigrant/"><em>The Hispanic–White Wage Gap Has Remained Wide and Relatively Steady: Examining Hispanic–White Gaps in Wages, Unemployment, Labor Force Participation, and Education by Gender, Immigration Status, and Other Subpopulations</em></a><em>. </em>Economic Policy Institute, July 2018.</p>
<p>Quadlin, Natasha, Ton VanHeuvelen, and Caitlin E. Ahearn. 2023. “<a href="https://pubmed.ncbi.nlm.nih.gov/37061326/">Higher Education and High-Wage Gender Inequality.</a>” <em>Social Science Research</em> 112, 102873. DOI: <a href="https://doi.org/10.1016/j.ssresearch.2023.102873" target="_blank" rel="noopener">10.1016/j.ssresearch.2023.102873</a>.</p>
<p>Rotondaro, Vinnie. 2026. “<a href="https://newrepublic.com/article/213121/democratic-socialists-america-problem-middle-class-socialism">The Problem with Middle-Class Socialism</a>.” <em>New Republic,</em> July 23, 2026.</p>
<p>Scott, Robert E., Valerie Wilson, Jori Kandra, and Daniel Perez. 2022. <a href="https://www.epi.org/publication/botched-policy-responses-to-globalization/"><em>Botched Policy Responses to Globalization Have Decimated Manufacturing Employment with Often Overlooked Costs for Black, Brown, and Other Workers of Color: Investing in Infrastructure and Rebalancing Trade Can Create Good Jobs for All</em></a>. Economic Policy Institute, January 2022.</p>
<p>Stimpfel, Amy Witkoski, Kathryn Leep-Lazar, Maile Mercer, and Kathleen DeMarco. 2025. “‘Scheduling Is Everything’: A Qualitative Descriptive Study of Job and Schedule Satisfaction of Staff Nurses and Nurse Managers.” <em>Western Journal of Nursing Research</em> 47, no. 10: 912–923. <a href="https://doi.org/10.1177/01939459251330280">https://doi.org/10.1177/01939459251330280</a>.&nbsp;</p>
<p>United States Census Bureau (Census). 2022. “<a href="https://www.census.gov/content/dam/Census/library/visualizations/time-series/demo/fig_6.png">Figure 6: Percent of Population 25 Years and Over, and 25 to 29 Years Old, with Bachelor&#8217;s Degree or Higher by Sex: 1947–2021</a>,” [chart image], <em>CPS Historical Time Series Visualizations. </em>Published February 24, 2022.</p>
<p>Williams, Jhacova, and Valerie Wilson. 2019. <a href="https://www.epi.org/publication/labor-day-2019-racial-disparities-in-employment/"><em>Black Workers Endure Persistent Racial Disparities in Employment Outcomes</em></a><em>. </em>Economic Policy Institute, August 2019.</p>
<p>Wilson, Valerie. 2016a. <a href="https://www.epi.org/publication/african-americans-are-paid-less-than-whites-at-every-education-level/"><em>African Americans Are Paid Less Than Whites at Every Education Level</em></a> (economic snapshot). Economic Policy Institute, October 2016.</p>
<p>Wilson, Valerie. 2016b. <a href="https://www.epi.org/publication/the-changing-demographics-of-americas-working-class/"><em>People of Color Will Be a Majority of the American Working Class in 2032: What This Means for the Effort to Grow Wages and Reduce Inequality</em></a>. Economic Policy Institute, June 2016.</p>
<p>Zipperer, Ben. 2026. <a href="https://www.epi.org/publication/setting-high-standards-for-a-federal-minimum-wage-raising-the-wage-to-two-thirds-of-the-national-median-wage-would-lift-pay-for-nearly-40-million-workers/"><em>Setting High Standards for a Federal Minimum Wage: Raising the Wage to Two-Thirds of the National Median Wage Would Lift Pay for Nearly 40 Million Workers</em></a><em>.</em> Economic Policy Institute, May 2026.</p>
<p>&nbsp;</p>
]]></content:encoded>
											
	</item>
		<item>
		<title>AI and the economy: Your questions, answered</title>
		<link>https://www.epi.org/publication/ai-and-the-economy-your-questions-answered/</link>
		<pubDate>Mon, 16 Nov 2026 13:00:45 +0000</pubDate>
		<dc:creator><![CDATA[Josh Bivens]]></dc:creator>
		<guid isPermaLink="false">https://www.epi.org/?post_type=publication&#038;p=327528</guid>
					<description><![CDATA[This FAQ addresses the potential effect of artificial intelligence (AI) on the U.S. economy, particularly on labor markets. Many recent concerns about advances in AI—for example, its implications for national security, cybersecurity, or media disinformation—are largely outside our areas of expertise.]]></description>
										<content:encoded><![CDATA[<p><strong><span class="dropped">T</span></strong>his FAQ addresses the potential effect of artificial intelligence (AI) on the U.S. economy, particularly on labor markets. Many recent concerns about advances in AI—for example, its implications for national security, cybersecurity, or media disinformation—are largely outside our areas of expertise. But we&nbsp;do&nbsp;have considerable expertise regarding AI’s potential effect on labor markets and the broader impacts of technological change on the economy.</p>
<p>Further, many of AI’s potential effects on the economy are largely distinct from its consequences in these other areas, making the economic effects amenable to separate, granular analysis. For example, if poorly regulated and recklessly deployed AI systems lead to disruptive cybersecurity incidents, this could obviously have some economic effects (imagine a large bank suffering a cyberattack and depositors’ funds being frozen or inaccessible for a time). But regulatory failures and irresponsible management of AI systems that lead to cyberattacks do not really signal anything about the economic potential for AI tools to replace human labor at large scale or about how a rapid evaporation of AI investment or stock market gains would ripple through the economy. Nor does the right policy to mitigate possible cybersecurity or disinformation problems exacerbated by AI have much at all to do with the optimal policy to ensure working families’ prosperity and economic security in an age of AI.</p>
<p>Across all realms of AI policy, much attention has been paid to regulations that are specifically tailored to AI. This might be appropriate in some fields. However, we suggest that not enough attention has been paid to simply applying protections, norms, and institutions that predate AI to its uses. For example, tort law already makes it possible to hold companies liable for harms their products create. While the application of these existing legal remedies to AI hacking incidents could potentially be clarified, it doesn’t seem to require a huge new field of law or legislation to make this possible. Further, we should significantly <a href="https://www.epi.org/publication/federal-ai-legislation/">strengthen</a> the application and enforcement of existing protections for workers. For example, employers should be held liable when they discriminate against workers, rather than being able to evade accountability because the discriminatory impact stemmed from their use of an algorithm or other automated system.</p>
<p>This notion that existing best practices, policies, and ideas are likely to be the best solutions to the problems AI may cause holds particularly true when it comes to its economic effects. The single largest concern expressed in this realm is AI’s potential to permanently displace very large amounts of human labor in a short period of time, causing economy-wide joblessness and wage declines. This FAQ will examine the economic evidence surrounding many of these concerns, but one key conclusion to state here is that even if such a scenario seemed plausible, it is highly unlikely that policies tailored specifically to AI tools’ potential uses in workplaces could ever keep up with, let alone proactively block, the ways that millions of private-sector firms are adopting these tools. What would reduce the most harm in such scenarios is implementing ideas already firmly on the policy agenda for many: keep the macroeconomy well managed and unemployment rates very low; provide effective protection for workers’ rights to form unions and bargain in the workplace over issues like how technology is implemented; offer a more protective social safety net that makes spells of joblessness less devastating for workers experiencing them; and implement progressive taxes—particularly higher tax rates on income derived from capital.</p>
<p>It is understandable that people are anxious about what AI could do to workers’ jobs and wages in coming years; the U.S. labor market has been rigged by policy for decades to make most workers insecure and keep their wages suppressed. AI deployed into this rigged labor market, as is happening currently, is unlikely to provide much benefit to most workers and could inject more churn and uncertainty. But making the policy changes noted above and deploying AI in a high-road economy where policy forces labor markets to perform better could allay many of these worries. In fact, if power were rebalanced in our economy to ensure workers are protected and empowered, AI might deliver real tangible benefits to most workers in the form of faster wage growth.</p>
<p>To be clear, arguing that broad, pre-existing ideas for how to unrig labor markets and generate shared prosperity are the best response to the economic concerns of AI does not mean we’re not taking AI “seriously” or recognizing its potential harms across many other realms. Instead, our recommendation stems from granular analysis of how AI might affect the economy and labor markets.</p>
<a name='1'></a>
<div class="shadow-box immi-faq-shadowbox">
<h2>How should we think about AI&#8217;s likely effects on jobs and wages?</h2>
<div class="callout-text">
<p>The effect of AI on the wages and employment of typical U.S. workers will, like most technological advances, hinge almost entirely on the broader economic, social, and political context. Crucially, the U.S. labor market has been largely rigged by policy to suppress wages and opportunities for most workers. In this context, there is little upside for U.S. workers from widespread deployment of AI tools. There are also widespread fears that AI might threaten jobs and wages on a large scale. But were the labor market substantially unrigged, AI tools could well raise the living standards of working people. Nothing inherent to AI lets us forecast its potential effects on U.S. workers, for good or bad. This will be determined by policy.</p>
</div>
<div class="epi-togglable-container  "><div><a href="#" class="epi-togglable-link toggler" data-close-text="Close" data-open-text="Read more">Read more</a></div><div class="epi-togglable-target togglee" style="display:none;">
<p>Being relatively agnostic about the direct effect of technology and AI on labor markets does not imply that we think labor markets have been working well for U.S. workers. On the contrary, unemployment has been too high and wage growth too low for <em>most</em> years in recent decades. But the roots of labor market dysfunction—both past and future—have very little to do with technological changes. Instead, this dysfunction is driven by the concerted policy push to exacerbate the extreme imbalance of power between typical workers and the corporate managers and capital owners who hire and supervise them.</p>
<p>It is important to get the facts and analysis right on the questions of why labor markets have not delivered enough jobs or acceptable wage growth, and what the real threats are to decent jobs in the future. The current fixation on AI—and the resulting pressures to respond with policies that wrap tightly around AI-specific parameters—distracts from the more fundamental problem of imbalanced power in labor markets, pulling policy in less useful directions.</p>
</div></div>
</div>
<div class="pdf-page-break "></div>
<a name='2'></a>
<div class="shadow-box immi-faq-shadowbox">
<h2>What’s AI’s most immediate threat to the U.S. economy?</h2>
<div class="callout-text">
<p>AI-related spending accounts for roughly half of U.S. growth in recent quarters—a narrow and fragile base for economic growth. That spending comes from two sources: (1) tangible investment in data centers, hardware, and research, and (2) consumption financed by the AI-driven stock bubble. Both spending flows could reverse quickly if expectations change about the future profitability of AI tools. If this reversal happens, the economy may well enter recession unless policymakers react nimbly and smartly.</p>
</div>
<div class="epi-togglable-container  "><div><a href="#" class="epi-togglable-link toggler" data-close-text="Close" data-open-text="Read more">Read more</a></div><div class="epi-togglable-target togglee" style="display:none;">
<p>There are two main spending flows associated with the race to develop and sell AI tools for mass-market use. The first is tangible business investment—what we’ll call capital expenditures, or <em>capex</em>. Many firms are hoping to be in the first wave of companies whose AI products are adopted, so they are currently investing heavily in data centers, servers, racks, and research. This wave of AI-related spending is driving demand for land, hardware and software components, construction materials, and the labor needed to build data centers, develop the models, and market them.</p>
<p>The other spending flow comes from the sharp rise in the market valuation of AI-related companies, which has driven a stock market boom—or, likely, bubble—in recent years. Part of households’ annual consumption spending comes from wealth (not just income). According to <a href="https://economics.mit.edu/sites/default/files/publications/jep.14.2.99.pdf">general estimates</a>, $1 in increased stock market wealth leads to around 3 cents in additional consumption spending. To the degree AI is driving above-normal valuations of the stock market relative to current fundamentals, the extra consumption spending out of this “excess” wealth can be attributed to AI.</p>
<p>To measure the AI effect on capex, macroeconomic observers have looked at the categories of business investment where the AI build-out is concentrated—data centers, information processing equipment, software, and R&amp;D—and calculated how much these categories are contributing to GDP growth over and above historical norms. This excess contribution has recently reached roughly 0.7% of GDP. In an economy where growth runs at roughly 2%, this would constitute roughly a third of all <em>increases</em> in GDP. That said, some portion of this capex spending—likely a third to a half—probably stems from increased imports, which are then subtracted out of the contribution to domestic growth. Very roughly, we estimate that the AI capex build-out has likely contributed 0.5% to GDP growth in the past year.</p>
<p>

<!-- BEGINNING OF FIGURE -->

<a name="Figure-A"></a><div class="figure chart-325901 figure-screenshot figure-theme-none" data-chartid="325901" data-anchor="Figure-A"><div class="figLabel">Figure A</div><img decoding="async" src="https://files.epi.org/charts/img/325901-36060-email.png" width="608" alt="Figure A" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->

</p>
<p>For the consumption effect associated with AI, we look at today’s ratio of price to earnings (P/E ratio) in the overall stock market (the S&amp;P 500 specifically). This measures how expensive it is to buy a dollar’s worth of earnings. Today, it is more expensive than in any period in history except the very late 1990s–2000, an unambiguous bubble that was also driven by wild enthusiasm about tech and saw a very large correction. Today’s P/E ratio is close to 40. A more normal ratio—though still high in historical terms—would be 30. If the stock market fell from a P/E ratio of 40 to one of 30 in a relatively short period of time, that would constitute a 25% decline in household wealth in corporate equities. That decline in wealth would imply a consumption reduction of about $200 billion, or roughly 0.7% of GDP.</p>
<p>

<!-- BEGINNING OF FIGURE -->

<a name="Figure-B"></a><div class="figure chart-326134 figure-screenshot figure-theme-none" data-chartid="326134" data-anchor="Figure-B"><div class="figLabel">Figure B</div><img decoding="async" src="https://files.epi.org/charts/img/326134-36061-email.png" width="608" alt="Figure B" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->

</p>
<p>Combined, the contributions of these two AI-related spending flows to GDP growth likely exceed 1% of GDP. Since the first quarter of 2025, growth <a href="https://fred.stlouisfed.org/graph/?g=1XYEd">has averaged roughly 2%</a>, which implies that over half of all growth in the past year was being propelled by AI-related spending.</p>
<p>But this AI-related spending is a narrow and fragile base for overall growth. A genuinely healthy economy should be able to generate enough demand to keep unemployment low and stable without a stock market bubble. And both the stock market valuations and the determination to keep spending heavily on data centers and other capex hinge on the belief among companies and investors that AI companies’ eventual profits will justify them.</p>
<p>If the potential profitability of this investment in AI ever appears lower than expected, this could lead to a sharp deflation of the current stock market bubble. A drop in the market value of AI companies would almost surely also convince at least a few of the big capex spenders that their investments will not pay off, and they will likely cut back. All this means that the AI contribution to GDP growth could evaporate quickly if just one or two dominoes fall.</p>
<p>This would leave a large hole in aggregate demand (spending by households, businesses, and governments) that would cause a recession unless policymakers reacted nimbly to fill it. This could largely be accomplished with <a href="https://www.epi.org/publication/next-recession-bivens/">standard recession-fighting monetary and fiscal policy</a>. But there would still be pain. And if political obstacles prevented the quick deployment of standard recession-fighting policy, the recession could linger and cause lasting damage.</p>
</div></div>
</div>
<div class="pdf-page-break "></div>
<a name='3'></a>
<div class="shadow-box immi-faq-shadowbox">
<h2>Is the current fixation on AI the first time a technology-driven stock market bubble has threatened the broader economy?</h2>
<div class="callout-text">
<p>Today’s AI boom bears striking parallels to the internet build-out and stock market bubble of the late 1990s. That episode clearly led to a stock market bubble, and the resulting deflation of that bubble led to a very stubborn recession. But the late 1990s boom also gave the U.S. labor market four to five very strong years before the bubble burst—something the current AI frenzy has largely failed to do.</p>
</div>
<div class="epi-togglable-container  "><div><a href="#" class="epi-togglable-link toggler" data-close-text="Close" data-open-text="Read more">Read more</a></div><div class="epi-togglable-target togglee" style="display:none;">
<p>The parallels between today’s AI frenzy and the late 1990s build-out of the internet are striking. The late 1990s <a href="https://www.epi.org/blog/how-ai-spending-is-impacting-the-u-s-economy/">boost to GDP from tech-related capex</a> was even larger than today’s, and by a significant amount. The late 1990s stock market bubble (measured by P/E ratios) was the only one in history larger than today’s AI bubble, and it was even more <a href="https://cepr.org/voxeu/columns/unpacking-us-tech-valuations-agnostic-assessment">concentrated</a> in tech-related stocks than today’s is.</p>
<p>Taken together, the productivity boost delivered by the internet boom from the mid-1990s through the mid-2000s was very large—roughly in line with the more optimistic projections of how much AI could boost productivity growth in the decades ahead. By 2005, the economy was <a href="https://www.nber.org/system/files/chapters/c11074/c11074.pdf">about 15% larger</a> than projecting forward from the early 1990s would have suggested.</p>
<p>In short, the introduction of the internet was a huge economic story. We don’t think of it that way for the same reason that fish don’t notice water: It’s now so fundamental to daily life that it no longer registers. But it was a mammoth change with measurable, obvious economic effects.</p>
<p>One important difference, however, is that the internet boom lifted the labor market broadly. The late 1990s saw large reductions in unemployment relative to the previous five (or 10) years, and inflation-adjusted wages grew strongly up and down the wage distribution <a href="https://www.epi.org/blog/focus-on-the-boom-not-the-slump-the-feds-new-policy-framework-needs-to-stop-cutting-recoveries-short-epi-macroeconomics-newsletter/">for the first time in a generation</a>.</p>
<p>The AI boom has not benefited the broader labor market in the same way. Since late 2022, when ChatGPT was first introduced, the U.S. labor market <a href="https://www.epi.org/blog/low-wage-workers-faced-worsening-affordability-in-2025/">has clearly softened</a>, and wage growth has decelerated, most sharply for the bottom half of the wage distribution. The AI boom began with lower unemployment and tighter labor markets than the internet boom did, so it is striking that it has not been associated with a stronger—or even an equally strong—labor market.</p>
<p>Part of the reason is policy—not AI-specific policy but broader policy decisions. Many <a href="https://www.epi.org/publication/the-trump-administrations-macroeconomic-agenda-harms-affordability-and-raises-inequality/">policy choices over this year slowed growth</a> in the large parts of the economy outside the AI boom. For example, in normal times, much of the Trump administration’s agenda in 2025 would have been seen as very bad for business investment. Its tariff policy was extreme and chaotic; its mass deportation policy threatened to weigh heavily on both the labor supply and the customer base of very specific industries; the large deficit increases locked in by the 2025 Republican tax and spending megabill (even with the large spending cuts) seemed likely to keep upward pressure on interest rates; and the administration’s historically unprecedented meddling with the Fed’s policy independence also put upward pressure on rates and increased uncertainty.</p>
<p>AI-related investment has been strong in recent years even in the face of these poor broader policy choices simply because this happened to be the historical moment when the technology matured enough for firms to race to deploy it.</p>
<p>But non-AI-related investment—over half of all business investment—has clearly been hurt by the policy trajectory of the past year and a half. This non-AI investment decelerated through 2025 and then turned firmly negative by year-end. In short, the policy headwinds for investment really do seem to be binding on all areas of the economy not being swept along by the AI build-out.</p>
<p>

<!-- BEGINNING OF FIGURE -->

<a name="Figure-C"></a><div class="figure chart-325965 figure-screenshot figure-theme-none" data-chartid="325965" data-anchor="Figure-C"><div class="figLabel">Figure C</div><img decoding="async" src="https://files.epi.org/charts/img/325965-36062-email.png" width="608" alt="Figure C" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->

</p>
</div></div>
</div>
<div class="pdf-page-break "></div>
<a name='4'></a>
<div class="shadow-box immi-faq-shadowbox">
<h2>Is AI inevitably bad for the jobs and wages of typical workers over the longer run?</h2>
<div class="callout-text">
<p>There’s no obvious reason to think so. Economists’ best summary measure of how rapidly technological change is moving forward is the pace of <em>productivity growth</em>—growth in the average output and income created in an hour of work. Throughout history, periods of fast productivity growth (that is, periods of rapid technological advance) have been <em>good</em>, not bad, for the wages and jobs of typical workers. What has overwhelmingly determined how workers fare is <em>policy</em>, not the pace or nature of technological change, and there is no sign yet that AI will break that pattern. Current policy is notably bad for workers’ jobs and wages and should be fixed, but that poor policy environment predates AI by decades.</p>
</div>
<div class="epi-togglable-container  "><div><a href="#" class="epi-togglable-link toggler" data-close-text="Close" data-open-text="Read more">Read more</a></div><div class="epi-togglable-target togglee" style="display:none;">
<p>U.S. labor markets have performed poorly for most workers for decades. But the root of this poor performance has largely <em>not</em> been technological change that displaced workers or made their skills obsolete. Instead, the main drivers of dysfunctional labor market performance for most U.S. workers have been intentional policy changes meant to tilt the balance of bargaining power in labor markets away from typical workers and toward the owners and managers of corporate business. These policy changes tore down institutions (like the minimum wage and unions) that had supported workers’ leverage and bargaining power in earlier decades and introduced new institutions (like <a href="https://www.epi.org/publication/noncompete-agreements/">noncompete agreements</a>) that further eroded workers’ ability to win pay increases that matched overall economic growth.</p>
<p><a href="https://www.epi.org/productivity-pay-gap/">Too slow wage growth for the bottom 80% of workers</a>—and the rise in inequality that came with it—began in the late 1970s and accelerated through the 1980s and 1990s. The early conventional wisdom about this inequality, especially among economists, was that labor markets are generally competitive and efficient, so inequality must stem from deep market forces shifting the relative demand and supply for different types of workers—things like technology and the pace of educational attainment. According to this view, policy institutions like the minimum wage and unions could affect only the small sliver of workers they touched directly, and any such effect introduced its own costly distortions. If one believed this, then the 30-year period of equal and fast wage growth after World War II could not have been driven by distortionary institutions like unions and a robust minimum wage, and the attack on unions and minimum wages could not be a convincing explanation of why inequality increased after 1979.</p>
<p>A tide of <a href="https://jhr.uwpress.org/content/57/s/s1">cutting-edge labor market research</a> has overturned most of this conventional wisdom and shown that while technology—including AI—is a prime determinant of how fast the overall economy grows, how that growth is <em>distributed</em> as jobs and wages is mostly a function of <a href="https://www.epi.org/unequalpower/home/">policies and institutions</a>. Like any technology, AI will help set the pace of growth, but whether typical workers share in it or whether today’s inequality widens even further will be decided by the policies and institutions that govern workers’ wages and labor market power.</p>
<p>The political and economic implications of this shift in understanding can hardly be overstated. If inequality were simply the product of rapid technological changes interacting with competitive and efficient markets, then it would be regrettable but essentially nobody’s fault, and little could be done about it. That is a very convenient story for the powerful. But it does not fit the historical record or economic evidence. Instead, across history, rapid technological progress has generally been <em>good</em> for workers, not bad: Periods of fast productivity growth have been periods of lower unemployment and slower growing inequality.</p>
<p>

<!-- BEGINNING OF FIGURE -->

<a name="Figure-D"></a><div class="figure chart-325933 figure-screenshot figure-theme-none" data-chartid="325933" data-anchor="Figure-D"><div class="figLabel">Figure D</div><img decoding="async" src="https://files.epi.org/charts/img/325933-36063-email.png" width="608" alt="Figure D" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->

</p>
<p>Whether U.S. workers thrive in an era of AI remains mostly a policy question. The decisions that will determine it have to do with the level of labor market power available to typical workers through policies like the level of minimum wages, the determination of macroeconomic policymakers to sustain full employment, the effectiveness of labor law in allowing workers to unionize and bargain collectively, and how robust the U.S. welfare state is.</p>
</div></div>
</div>
<div class="pdf-page-break "></div>
<a name='5'></a>
<div class="shadow-box immi-faq-shadowbox">
<h2>Won’t AI automate many jobs away?</h2>
<div class="callout-text">
<p>AI will likely destroy some <em>specific jobs</em> through automation—but it is also highly likely to create jobs more broadly over time by boosting incomes for those in the non-automated sectors. How much any particular group of workers suffers when displaced from their jobs will be determined by policy decisions. Throughout U.S. history, we have generally been terrible at managing these transitions in worker-friendly ways. It would obviously be better if we changed that and put in place the policy supports workers need to get through them.</p>
</div>
<div class="epi-togglable-container  "><div><a href="#" class="epi-togglable-link toggler" data-close-text="Close" data-open-text="Read more">Read more</a></div><div class="epi-togglable-target togglee" style="display:none;">
<p>Rapid technological change concentrated in any one sector can reduce employment <em>in that sector</em>. So perhaps AI really will, say, reduce the number of software coding jobs.</p>
<p>But if this technological change allows the same output to be produced with fewer hours of work, it raises incomes across the economy. Sticking with software coders, if the price of software design stays the same as its productivity rises, incomes within the sector will rise, and the workers and capital owners within the sector will have higher incomes they can use to buy <em>other</em> sectors’ goods and services. More likely, competition pushes the price of software design down, and those who purchase software design services (including purchasing indirectly as consumers buying goods and services from other businesses) will see their real (inflation-adjusted) incomes rise.</p>
<p>This increased demand stemming from these productivity gains means that job losses within the AI-using sector can be counterbalanced (or more than counterbalanced) by expanding employment in other sectors <em>as long as aggregate demand (economy-wide spending by households, businesses, and governments) is maintained.</em></p>
<p><a href="https://www.brookings.edu/articles/is-automation-labor-displacing-productivity-growth-employment-and-the-labor-share/">Autor and Salomons (2018)</a> analyzed these cross-cutting effects of technological changes within particular economic sectors. They found that the <em>own-effect</em> of a positive productivity shock—its effect on jobs within the sector where it happens—is indeed negative: Fewer hours of work are needed to produce the sector’s output, and rising demand for that now less expensive output doesn’t fully make up the difference. However, the <em>cross-effect</em>—the boost that sector’s productivity growth gives to employment <em>in other sectors</em>—is strongly positive and generally large enough to outweigh the negative own-effect for total employment.</p>
<p>This offset is not always automatic. Market mechanisms could fail, at least in part, to channel higher incomes driven by technological advance into higher demand elsewhere in the economy. If the initial large income gains in the AI-using sector flow to already wealthy households, for example, these households are unlikely to <a href="https://www.epi.org/publication/inequalitys-drag-on-aggregate-demand/">spend rather than save all of these gains</a>. That is why policymakers should keep aggregate demand high enough to keep unemployment low. They have <a href="https://www.epi.org/publication/next-recession-bivens/">ample tools to do it</a>, from lower interest rates set by the Federal Reserve to redistributive fiscal policy that funds more generous public benefits out of progressive taxes.</p>
<p>In reality, however, market mechanisms likely will do much of the work in terms of maintaining aggregate demand in the face of productivity increases. Recessions and damagingly high increases in unemployment rates are overwhelmingly <em>not</em> caused by bursts of productivity growth. Instead, as the economy’s productivity potential grows year after year, those gains tend to translate relatively smoothly into growth in demand. When demand does falter, it is usually because of an outside shock, like an asset bubble popping or central bankers deliberately slowing the economy to forestall inflation.</p>
</div></div>
</div>
<div class="pdf-page-break "></div>
<a name='6'></a>
<div class="shadow-box immi-faq-shadowbox">
<h2>If AI can both create and destroy jobs, does this mean policymakers can totally ignore its effects?</h2>
<div class="callout-text">
<p>No. If anything, the opposite is true: Policymakers should be doing far more to protect working families from the hardship and uncertainty of earning a living in labor markets that are volatile and offer little protection. If AI-induced technological change turns out to be large and causes significant job displacement, it will hit a U.S. labor market with few worker protections and create real pockets of concentrated harm—as every other economic change of the past several decades has. But the main driver of this harm is the weakness of U.S. labor market protections, not the specifics of the technology.</p>
</div>
<div class="epi-togglable-container  "><div><a href="#" class="epi-togglable-link toggler" data-close-text="Close" data-open-text="Read more">Read more</a></div><div class="epi-togglable-target togglee" style="display:none;">
<p>To understand the scale, start with how much churn the labor market already absorbs. About 1.5 million workers involuntarily <a href="https://fred.stlouisfed.org/series/JTSLDL">lose their jobs <em>every single month</em></a> in the U.S. economy—about 18 million jobs lost involuntarily each year. If AI raised the growth of productivity by 0.5 percentage points over the next decade <em>and</em> there were no countervailing job creation from increased incomes, this would very roughly translate into another 700,000 jobs each year being shed, with workers in those jobs needing to find alternative employment. Compared with the 18 million separations the economy already churns through, that is a real but manageable increase.</p>
<p>Even so, many individual workers may suffer from sector-specific job displacements due to AI, even in years when overall unemployment is low. The harm these displacements cause is real and significant: The labor market is not frictionless, and it can be a painful stretch of time before a displaced worker finds comparable work elsewhere. Some workers (particularly older workers) may never find a specific job as good as the one they lost.</p>
<p>But, again, most of this suffering is <em>not an AI-specific problem</em>. The <em>vast majority</em> of painful, forced job separations are not about AI currently and likely won’t be in the future. <em>Workers displaced for reasons not due to AI deserve protection too.</em> The solution is not a search for novel AI-specific policies but the broad, labor market-wide policies that help in every case: stronger social insurance, more widespread collective bargaining, and sustained high-pressure labor markets. These policies are all highly desirable regardless of the pace of technological change.</p>
<p>It is the absence of these broad-based labor market and social insurance policies that makes losing a <em>specific job</em> so damaging for some U.S. workers. Because the guaranteed baseline of job quality and social protection is so low for U.S. workers, a job that does provide some of those things is rare, and workers know it will be hard to replace. This imparts a strong sense of <em>rational</em> loss aversion for workers who happen to hold a high-quality job.</p>
<p>One reason specific jobs are occasionally highly valued in the U.S. labor market is that they come bundled with nonwage compensation—like health and retirement benefits, which are not universally available. Making those benefits universal through more protective social insurance would make losing any particular job far less damaging. Another key social insurance system—unemployment insurance (UI)—is too stingy in the U.S., imposing large income losses while workers search for alternative employment after a layoff. Making UI more generous would directly reduce the pain of a layoff.</p>
<p>Another reason some specific jobs can be highly valued in the U.S. economy is that they are unionized. Union jobs should be much more common than they currently are, but decades of employer hostility and policy indifference have made large-scale organizing in newly created jobs incredibly challenging. The sectors that remain unionized do so largely because of a historical legacy that saw their unions formed decades ago, so a worker separated from one today has little chance of finding another union job elsewhere. <em>Anything</em> that destroys today’s union jobs, technological progress included, will leave their former holders worse off. If policy changed to make it more likely for newly created jobs to be unionized, much of the damage from loss of union jobs spurred by technological churn could be ameliorated.</p>
<p>Additionally, the U.S. economy has spent much of the past four decades <a href="https://www.epi.org/blog/focus-on-the-boom-not-the-slump-the-feds-new-policy-framework-needs-to-stop-cutting-recoveries-short-epi-macroeconomics-newsletter/">with excessively high unemployment</a>, which radically increases the cost of losing a job. When unemployment is low and vacancies are high, a laid-off worker can find a new job quickly, putting employers under constant pressure to keep job quality high enough to retain and attract new employees.</p>
<p>If U.S. policymakers built a more protective social insurance system, restored the effective right to organize unions, and maintained high-pressure labor markets with low unemployment, then a large part of the damage done by <em>any</em> job displacements—including any driven by AI—would disappear.</p>
</div></div>
</div>
<div class="pdf-page-break "></div>
<a name='7'></a>
<div class="shadow-box immi-faq-shadowbox">
<h2>Will AI lead to mass unemployment across the U.S. economy?</h2>
<div class="callout-text">
<p>Almost surely not. The historical evidence is clear that business cycles with faster productivity growth see <em>lower</em> unemployment, not higher. And the ease with which AI or any capital investment can substitute for human labor at scale is often overstated.</p>
</div>
<div class="epi-togglable-container  "><div><a href="#" class="epi-togglable-link toggler" data-close-text="Close" data-open-text="Read more">Read more</a></div><div class="epi-togglable-target togglee" style="display:none;">
<p>Throughout modern history in advanced economies, mass unemployment has never been caused by technological advances that accelerate productivity growth. Unemployment crises are instead caused by a shortfall in aggregate demand (economy-wide spending by households, businesses, and governments). Within a given business cycle, faster productivity growth is accompanied by lower, not higher, unemployment.</p>
<p>

<!-- BEGINNING OF FIGURE -->

<a name="Figure-E"></a><div class="figure chart-325935 figure-screenshot figure-theme-none" data-chartid="325935" data-anchor="Figure-E"><div class="figLabel">Figure E</div><img decoding="async" src="https://files.epi.org/charts/img/325935-36069-email.png" width="608" alt="Figure E" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->

</p>
<p>It is worth stepping back from this historical evidence to think through <em>why</em> technological advance has never led to mass unemployment. The rest of this answer walks through some of that logic. We offer these illustrations to take the current fears of mass unemployment seriously, not to suggest that any of these scenarios are likely or have characterized the real economic history of the U.S.</p>
<p>One important economic reason why AI or any technological change is unlikely to cause mass unemployment is simply that technology is not free. Firms will replace labor with capital (including AI tools) only when the costs of doing so are less than the savings. But while shedding labor saves firms money, investing in new capital <em>costs</em> money.</p>
<p>And if the demand for AI tools rises rapidly as businesses across the economy try to replace large swathes of their workforce, the price of these tools will rise, and eventually (or even quickly) the price of investing in AI tools will exceed the benefits to firms.</p>
<p>Because so much of what is called “technological change” in recent years has occurred in online spaces that seem unconstrained by capital costs, it is easy to underestimate how costly AI tools will likely be if a broad swathe of U.S. employers try to buy them at scale to economize on labor costs. But AI tools require vast amounts of capital, labor, and energy to build and maintain (see <strong>Figure A</strong> about the scale of investments being made to build up AI capacity and enter the race to try to sell AI tools across the U.S. economy).</p>
<p>Further, if AI opened up historically large-scale possibilities for the replacement of human labor in the future, this would radically raise economy-wide productivity <em>and</em> reduce the <em>relative</em> cost of labor, as unemployment reduced workers’ leverage to demand wage growth that keeps pace with productivity. This reduction in the relative cost of labor would provide some businesses a large incentive to figure out how to profitably employ some of the now-jobless workers—who, again, are now quite a bit cheaper relative to other inputs into production (i.e., AI tools).</p>
<p>In short, mass unemployment due to technological change is highly unlikely to be an <em>equilibrium</em> in the economy—a place the economy remains at for long periods—unless policymakers completely fail at keeping aggregate demand (economy-wide spending by households, businesses, and governments) stable. Have policymakers failed in the past at this fundamental part of their jobs, allowing excess unemployment to linger far longer than necessary? Definitely. But that had nothing to do with AI, and the toleration of mass unemployment that might occur due to an AI-driven acceleration of productivity would be a failure of these policymakers, not an inevitability of the technology.</p>
<p>Also, this scenario—where AI generates extremely rapid productivity growth and the <em>relative</em> price of labor must fall to keep unemployment in check in the face of this AI-biased technological change—does not imply <em>absolute</em> reductions in wages. Instead, it only implies wages fall further behind productivity growth. But if productivity growth is supercharged by AI (and that’s the entire root of this scenario), workers’ wages and living standards could still grow in absolute terms, even more rapidly than they have in the recent past. Only wages <em>relative</em> to capital costs have to fall. To be clear, this isn’t a particularly happy scenario: Rising inequality is objectively bad, even if all incomes see some absolute gains. But it’s not a purely pessimistic scenario either.</p>
<p>Finally, much of the discourse that seems to assume a very easy replacement of human labor with AI tools underappreciates the complexity of most jobs in the economy. <em>Jobs</em> are a bundle of specific <em>tasks</em> that are often hard to clearly describe (let alone fully replace). AI will replace many tasks, but this will often just allow workers to focus more time and attention on the other tasks their jobs entail. Many of the tasks of modern jobs are extremely hard to define in binary terms that would allow easy automation. This difficulty in crisply describing many job tasks often leads to sneering about the alleged worthlessness of “<a href="https://www.programmablemutter.com/p/ai-isnt-management-try-explaining">middle managers</a>” or “<a href="https://substack.jurgenappelo.com/p/how-ai-will-achieve-peak-bullshit-work">bullshit jobs</a>” (to borrow anthropologist David Graeber’s term), but it stems from the fact that we have an amazingly complex economy built on ever-finer divisions of labor and information across workers and functions. Just because any single person has <a href="https://messyjobs.ai/">almost no idea what the key tasks</a> are for most of the workforce <em>doesn’t mean these tasks are unimportant or easy to replace</em>. It just means the economy is complex.</p>
</div></div>
</div>
<div class="pdf-page-break "></div>
<a name='8'></a>
<div class="shadow-box immi-faq-shadowbox">
<h2>How much might AI boost productivity growth—and has the U.S. labor market ever absorbed that fast a productivity acceleration before?</h2>
<div class="callout-text">
<p>Most credible estimates project that AI could boost productivity growth over the next decade by an average of 0.5–1.5 percentage points per year. Even at the high end of 1.5 percentage points per year, that implies an acceleration of productivity growth that the U.S. economy has easily absorbed in the past with no labor market distress. In fact, periods of productivity acceleration have been good for U.S. workers; it is productivity slowdowns that have coincided with bad outcomes.</p>
</div>
<div class="epi-togglable-container  "><div><a href="#" class="epi-togglable-link toggler" data-close-text="Close" data-open-text="Read more">Read more</a></div><div class="epi-togglable-target togglee" style="display:none;">
<p>The effect AI will have in boosting productivity growth is still genuinely unclear. There are clear reasons to think it could run faster for a stretch as AI is incorporated into firms’ workflow, but as with all technological innovations, the pace and the ultimate depth of that penetration remain highly uncertain.</p>
<p>As of the first quarter of 2026, there is very little reason to think AI has contributed anything measurable to U.S. productivity growth. Since the last business cycle peak (the last quarter of 2019), productivity growth has averaged 1.96%. Looking only at the period since the third quarter of 2022 (when ChatGPT was introduced to the public), productivity growth has averaged 2.08%—a small acceleration in a very noisy data series. The pre-2022 numbers also include the large productivity declines stemming from the supply chain bottlenecks caused by the pandemic and the Russian invasion of Ukraine in the early 2020s. <strong>Figure F</strong> shows the simple level of productivity over time, along with an extrapolated trend for recent years that shows what the level of productivity would be today if the 2017–2019 productivity trend held after the rollout of ChatGPT. Two things stand out: the large drop in the level of productivity as the early 2020s supply chain snarls formed, and the fact that as of the first quarter of 2026, the level of productivity is only about 0.2% higher than what would have been extrapolated from it following its immediate pre-Covid trend.</p>
<p>Further, even this small edge over projection is entirely driven by the first quarters of 2023—the gap between actual productivity and trend productivity has actually declined in the past two years. None of this looks like evidence of a large technological innovation that is diffusing rapidly across ever-larger shares of the U.S. economy.</p>
<p>

<!-- BEGINNING OF FIGURE -->

<a name="Figure-F"></a><div class="figure chart-326014 figure-screenshot figure-theme-none" data-chartid="326014" data-anchor="Figure-F"><div class="figLabel">Figure F</div><img decoding="async" src="https://files.epi.org/charts/img/326014-36070-email.png" width="608" alt="Figure F" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->

</p>
<p>Finally, even the quite optimistic projections of how much AI might add to U.S. productivity growth in the coming decade do not imply a pace of productivity growth the U.S. labor market has never successfully absorbed before. <strong>Figure G</strong> shows a range of projections of how much AI could raise productivity growth over the next decade and compares the implied acceleration with past episodes of rising and falling productivity growth in the U.S. economy. We intentionally choose historical periods that have seen sharp changes in productivity growth to compare with projections of AI-driven productivity growth.</p>
<p>

<!-- BEGINNING OF FIGURE -->

<a name="Figure-G"></a><div class="figure chart-326010 figure-screenshot figure-theme-none" data-chartid="326010" data-anchor="Figure-G"><div class="figLabel">Figure G</div><img decoding="async" src="https://files.epi.org/charts/img/326010-36071-email.png" width="608" alt="Figure G" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->

</p>
<p>U.S. history has clearly seen sharp changes in productivity growth roughly as large as the more optimistic projections of AI’s effect on productivity. Importantly, the historical period that saw a rapid acceleration of productivity growth—the late 1990s and early 2000s—was an unusually strong period in the U.S. labor market, with lower unemployment and faster, broader wage growth than had been seen in a generation. The beginning of wage suppression, in contrast, followed the sharp productivity slowdown of 1973–1979. The other sharp productivity slowdown, from 2007 to 2019, was also a long period of extreme dysfunction in U.S. labor markets, with extended periods of excess unemployment and many years of near-stagnant wage growth.</p>
<p>In short, the actual history of the U.S. economy and credible projections of AI’s effect on productivity do not warrant extreme concern about AI-caused labor displacement.</p>
</div></div>
</div>
<div class="pdf-page-break "></div>
]]></content:encoded>
											
	</item>
		<item>
		<title>EPI comment on DHS elimination of the 60-day grace period for workers in the E-1, E-2, E-3, H-1B, H-1B1, L-1, O-1, and TN visa programs</title>
		<link>https://www.epi.org/publication/epi-comment-on-dhs-elimination-of-the-60-day-grace-period-for-workers-in-the-e-1-e-2-e-3-h-1b-h-1b1-l-1-o-1-and-tn-visa-programs/</link>
		<pubDate>Tue, 10 Nov 2026 19:32:04 +0000</pubDate>
		<dc:creator><![CDATA[Daniel Costa]]></dc:creator>
		<guid isPermaLink="false">https://www.epi.org/?post_type=publication&#038;p=328334</guid>
					<description><![CDATA[Submitted via Office of Policy and United States Citizenship and Immigration Service Department of Homeland Security 5900 Capital Gateway Camp Springs, MD Re: U.S.]]></description>
										<content:encoded><![CDATA[<p><a href="https://www.federalregister.gov/documents/2026/09/11/2026-18631/eliminating-the-discretionary-60-day-grace-period"><em>Submitted via FederalRegister.gov</em></a></p>
<p>Office of Policy and Strategy<br />
United States Citizenship and Immigration Service (USCIS)<br />
Department of Homeland Security (DHS)<br />
5900 Capital Gateway Drive<br />
Camp Springs, MD 20746&nbsp;</p>
<p><strong>Re: U.S. Department of Homeland Security, </strong><a href="https://www.federalregister.gov/documents/2026/09/11/2026-18631/eliminating-the-discretionary-60-day-grace-period"><strong><em>Eliminating the Discretionary 60-Day Grace Period</em></strong></a><strong>, Notice of Proposed Rulemaking, </strong><strong>CIS No. 2858-26, DHS Docket No. USCIS-2026-0364,</strong><strong> RIN 1615-AD22, </strong><strong>91 Fed. Reg. 57807 (September 11, 2026).</strong></p>
<p>The Economic Policy Institute (EPI) is a nonprofit, nonpartisan think tank established in 1986 to include the needs of low- and middle-income workers in economic policy discussions. EPI conducts research and analysis on the economic status of working America, proposes public policies that protect and improve the economic conditions of low- and middle-income workers—regardless of immigration status—and assesses policies with respect to how well they further those goals. EPI submits these comments on the Department of Homeland Security’s (DHS) Notice of Proposed Rulemaking (NPRM) regarding the elimination of the discretionary 60-day grace period for certain nonimmigrants in the E-1, E-2, E-3, H-1B, H-1B1, L-1, O-1 or TN classifications. EPI has researched, written, and commented extensively on the U.S. system for labor migration, including in particular, the H-1B and L-1 programs, and other temporary work visa programs and green cards. EPI has published extensively on issues related to H-1B wage levels and employer usage and abuse of H-1B and other visa programs.<a href="#_note1" class="footnote-id-ref" data-note_number='1' id="_ref1">1</a></p>
<p><strong>EPI opposes the proposed rule</strong> and believes it arbitrary and capricious, and will potentially harm over a million workers in nonimmigrants statuses, as well as U.S.-born workers who are similarly employed, and have significant negative economic impacts across numerous industries that have not been considered or calculated by DHS. The ultimate impact will be to grant employers with even more power to exploit H-1B and other visa programs for university-educated workers with impunity and to degrade standards generally for workers in professional occupations. Rather than eliminating the 60-day grace period established at 8 CFR § 214.1(l)(2), DHS should extend the grace period to 120 days.</p>
<h2>Over one million workers will be impacted by the proposed rule, many of whom are in the green card backlog.</h2>
<p>DHS does not attempt to estimate the total population of workers currently employed in the impacted visa classifications of E-1, E-2, E-3, H-1B, H-1B1, L-1, O-1 and TN. The Office of Homeland Security Statistics could likely have provided an estimate for the NPRM. The lack of an estimate prevents the public from understanding the size of the total stock of workers for whom the NPRM would strip of the ability to leave an abusive or lawbreaking employer, forcing them to risk their ability to reside and work in the United States, or to even have enough time to gather their belongings before they are considered out of status and at risk of detention and deportation.</p>
<p>We have provided a rough estimate shown below in <strong>Table 1 </strong>on the total stock of workers employed in the impacted classifications as of fiscal year (FY) 2019, which is just over one million. (a more recent year was not chosen because of data limitations in multiple classifications.) The vast majority are in the H-1B program with over 583,0000 workers employed, as well as the L-1 program with 200,000, and E-2 and TN where over 100,000 are employed in each.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Table-1"></a><div class="figure chart-328259 figure-screenshot figure-theme-none" data-chartid="328259" data-anchor="Table-1"><div class="figLabel">Table 1</div><img decoding="async" src="https://files.epi.org/charts/img/328259-36113-email.png" width="608" alt="Table 1" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>Further, a significant share of workers in this population are on the path to lawful permanent residence (i.e., getting a green card), meaning their employer has submitted an I-140 petition to adjust their status. But because of annual numerical limits and country caps on green cards, there are significant wait times that can often last decades for applicants. In total, there are an estimated 1.2 million people in the employment-based green card backlog.<a href="#_note2" class="footnote-id-ref" data-note_number='2' id="_ref2">2</a> While there is no official estimate of the shares in the different visa or status classifications, data on permanent labor certifications—the first step towards getting a green card for most H-1B workers—show that the vast majority are likely to be in H-1B status. In 2023, 64.3% of permanent labor certifications were for H-1B workers and 4.8% were for L-1 workers, accounting for a total of 69.1% of labor certifications in just those two classifications.<a href="#_note3" class="footnote-id-ref" data-note_number='3' id="_ref3">3</a></p>
<h2>U.S. temporary work visa programs lack adequate regulation, worker protections, and labor standards enforcement.</h2>
<p>Through numerous reports, commentary, and testimony over many years, EPI has shown how temporary work visa programs—including the categories implicated by the proposed rule that allow employers to recruit and hire mostly university-educated workers from abroad—represent a flawed employment model that restricts workers’ rights by giving employers control over their workers’ immigration status.<a href="#_note4" class="footnote-id-ref" data-note_number='4' id="_ref4">4</a> The H-1B program’s inadequate wage rules and worker protections, and lack of enforcement from the U.S. Department of Labor (DOL), and the fact that the E-1, E-2, E-3, L-1, O-1 and TN programs have zero wage or worker protections, are directly contributing to the undercutting of U.S. labor standards.<a href="#_note5" class="footnote-id-ref" data-note_number='5' id="_ref5">5</a> The programs lack adequate regulations and enforcement to protect workplace rights and fail to provide workers with a pathway to becoming lawful permanent residents or U.S. citizens, leaving workers indentured for years and even decades. Given this reality, DHS should put in place measures—like grace periods, visa portability, and deferred action for workers with bona fide workplace complaints—that can act as safety valves for workers employed with nonimmigrant visas that need to leave an abusive and/or illegal employment situation. Instead, DHS is making conditions much worse for this cohort of workers by removing their freedom to walk away from an employer and seek employment elsewhere. It is exacerbating the already disastrous situation it has caused for workers when it rescinded the process established in 2023 that allowed workers the opportunity to apply for immigration protections when assisting a qualifying labor agency in conducting an employer investigation.</p>
<h2>The ability to leave an abusive employment situation is a fundamental worker right and DHS is proposing to eliminate it, which will degrade wages and working conditions.</h2>
<p>The vast majority of workers in the United States can leave an abusive employment situation—for example when they are not being paid the wages promised or when conditions are illegally hazardous—and seek alternative employment without a penalty. They are also able to report labor and employment law violations to the relevant state and local labor standards enforcement agencies without fear or any immediate penalty. This freedom to leave an employer acts as a check on the power of employers, so that they do not offer conditions that are unbearably and illegally poor. In short, workers who can change jobs or leave a lawbreaking employer without fear can bargain more effectively for fair and decent wages and seek career advancement opportunities they are qualified for. While this may seem like common sense, there is research to support it. For example, Mukhopadhyay and Oxborrow found that the wage gain associated with moving from an H-1B visa to an employment-based green card was $11,860 per year.<a href="#_note6" class="footnote-id-ref" data-note_number='6' id="_ref6">6</a></p>
<p>Statutes and regulations in temporary work visa programs, however, put migrant workers in a nonimmigrant status into a much more precarious and dangerous situation. Such workers do not have equal rights: Most visa programs do not permit migrant workers to easily leave a job or permit them to change jobs, making it difficult for them to leave an abusive employment situation. In essence, these visa programs operate in practice to create a labor market monopsony for employers—awarding employers greater leverage over their workers<a href="#_note7" class="footnote-id-ref" data-note_number='7' id="_ref7">7</a>—and growing research has shown that even modest amounts of employer monopsony power are utterly corrosive to workers’ ability to bargain for better wages.<a href="#_note8" class="footnote-id-ref" data-note_number='8' id="_ref8">8</a> In the limited visa programs where workers may have the ability to change jobs, many find it difficult in practice, because the U.S. government has not created a sensible and holistic system allowing them to find new opportunities in a reasonable amount of time.<a href="#_note9" class="footnote-id-ref" data-note_number='9' id="_ref9">9</a></p>
<p>In addition, many workers have invested thousands of dollars to recruiters to obtain their temporary job, often taking on significant amounts of debt to do so, whether it’s in the H-2A, H-2B, or H-1B programs, or for employment-based (EB) green cards. If leaving a job means sacrificing a valid immigration status and the opportunity to continue working in the United States and to earn back the investments made, then workers are much more likely to opt to remain in an abusive employment situation, which degrades wages and working conditions—not just for the migrant workers themselves, but also for U.S. workers who are similarly employed.</p>
<p>Migrant workers in the impacted classifications will be far less likely to file complaints when employers violate the law, if it means risking a retaliatory termination that would result in the worker’s immediate loss of status and investments and triggering the need to depart the United States. If an employer unlawfully retaliates against a worker, there would be insufficient time to file a complaint with the relevant government agency, whether seeking reinstatement or back pay for wage and hour violations—and if the worker is a union member, there would be insufficient time to file a grievance through any established mechanisms. In fact, there would be no time to do so, since the worker must depart immediately after their termination.</p>
<p>As we’ve stated on numerous occasions and in multiple publications, workers in H-1B and other nonimmigrant statuses at a disadvantage in the workplace vis-à-vis their U.S. counterparts because they have significantly less job mobility based on their temporary and precarious immigration status. But with the 60-day grace period created in its 2016 regulation, DHS took a major step towards leveling the playing field between employers and workers in the E-1, E-2, E-3, H-1B, H-1B1, L-1, O-1, and TN classifications. While nevertheless insufficient to remedy the known flaws in work visa programs, the grace period was supported broadly by unions and worker advocacy groups. By proposing to eliminate one of the few worker protections in the targeted visa programs—impacting over one million workers—DHS is taking a major step backwards when it comes to protecting labor standards for U.S. workers and migrant workers. Overnight, one million workers will become indentured to their employers. Those one million workers are employed in fields as diverse as information technology, artificial intelligence, health care and medicine, engineering, and finance. U.S. workers in those fields will also be impacted negatively by their colleagues lacking workplace rights.</p>
<h2>Many of the workers impacted by the rule have resided in the United States for over a decade and have spouses employed in the United States and U.S. citizen children.</h2>
<p>Many of the workers impacted by the rule have resided in the United States for over a decade and have U.S. citizen children, and have made long term economic investments, such as purchasing homes, and are well integrated into their local communities. Cutting off the ability for these workers to find alternative employment and in many cases to continue on the path to a green card if they are terminated or laid off at work, or because their employer has engaged in unethical or illegal behavior, and forcing them to immediately depart the country, is not justified by any of the arguments or evidence presented by DHS in the NPRM.</p>
<p>Although no official count is available, there are hundreds of thousands of spouses of H-1B, L-1, and O-1 principal nonimmigrants who are likely employed and will also be impacted by elimination of the 60-day grace period. They will lose their employment and their employers will abruptly lose employees who will be forced to immediately depart the United States, disrupting the economy and labor market. Tens to hundreds of thousands of children—some share of whom are in a nonimmigrants status, and some share of whom are U.S. citizens—will also be impacted. Those children will have their education severely disrupted and interrupted, as parents will have no time to formally withdraw them from school and get their records in order, or to find another school in the country of origin.</p>
<p>The significant equities accrued, especially by long-term nonimmigrant workers, and the interest and well-being of the families who rely on them, significantly outweighs the administrative burdens which DHS overstates and primarily cites as their justification for eliminating the 60-day grace period.</p>
<h2>DHS has not calculated the economic impact of stripping over one million workers of an essential component of their workplace rights.</h2>
<p>DHS has failed to consider or estimate the significant costs that will be incurred not just by the more than one million nonimmigrant workers themselves, but also U.S. citizens, lawful permanent residents, and other workers with employment authorization documents due to the increased restriction on worker mobility in these visa programs. In fact, various economic studies have shown that limiting worker mobility has a direct negative effect on the wages of all workers in a labor market, even those not subject to those limitations.<a href="#_note10" class="footnote-id-ref" data-note_number='10' id="_ref10">10</a> Eliminating the 60-day grace period will ultimately result in one million workers being indentured to their employers in numerous critical industries.</p>
<h2>Conclusion: DHS should rescind the proposed rule and instead increase the grace period to 120 days.</h2>
<p>As discussed above, a grace period of only 60 days is insufficient to permit workers in the E-1, E-2, E-3, H-1B, H-1B1, L-1, O-1 and TN classifications—who have been recruited from abroad and hired to participate and contribute to the U.S. labor market—to get their affairs in order, which may require selling a home and pulling children out of school. It is irrational, cruel, and arbitrary, and capricious to require such workers to depart the United States immediately, solely due to the whims of U.S. employers. Such workers should be permitted an adequate amount of time to find alternative employment in the United States, as well as to get their affairs in order if they choose to depart the country. Their spouses who are legally employed should also be permitted to benefit from an adequate grace period, which will benefit not just them, but the U.S. employers who hire them, who are provided with an opportunity to seek a replacement with a grace period, rather than having an employee who is required to depart the country immediately when their spouse’s employment is terminated.</p>
<p>DHS should rescind the NPRM, but as noted above, the 60-day grace period status quo, and the portability provisions in the implicated classifications, are inadequate to protect workers and labor standards. Instead, DHS should issue a new proposed rule to increase the grace period to 120 days. If deemed necessary, DHS could determine a shorter grace period for nonimmigrant workers who have resided in the United States for shorter periods. For example, a worker employed in nonimmigrant status for over three years could be entitled to a 120 day grace period, while a worker employed for fewer than three years could be entitled to a shorter period of at least 90 days.</p>
<p>Regards,</p>
<p>Daniel Costa, Esq.<br />
Director of Immigration Law and Policy Research<br />
Economic Policy Institute</p>
<hr>
<h3>Endnotes</h3>
<p data-note_number='1'><a href="#_ref1" class="footnote-id-foot" id="_note1">1. </a> See for example, Daniel Costa and Ron Hira, <a href="https://www.epi.org/publication/h-1b-visas-and-prevailing-wage-levels/"><em>H-1B visas and prevailing wage levels: A majority of H-1B employers—including major U.S. tech firms—use the program to pay migrant workers well below market wages</em></a>, Economic Policy Institute, May 4, 2020; Ron Hira and Daniel Costa, <a href="https://www.epi.org/publication/new-evidence-widespread-wage-theft-in-the-h-1b-program/"><em>New evidence of widespread wage theft in the H-1B visa program: Corporate document reveals how tech firms ignore the law and systematically rob migrant workers</em></a>, Economic Policy Institute, December 9, 2021.</p>
<p data-note_number='2'><a href="#_ref2" class="footnote-id-foot" id="_note2">2. </a> National Foundation for American Policy, <a href="https://nfap.com/research/new-nfap-policy-brief-employment-based-immigration-backlogs-and-waiting-times/"><em>Employment-Based Immigration Backlogs and Waiting Times</em></a>, NFAP Policy Brief, August 2026.</p>
<p data-note_number='3'><a href="#_ref3" class="footnote-id-foot" id="_note3">3. </a> Office of Foreign Labor Certification, “<a href="https://www.dol.gov/sites/dolgov/files/ETA/oflc/pdfs/PERM_Selected_Statistics_FY2023_Q4.pdf">Permanent Employment Program – Selected Statistics, Fiscal Year (FY) 2023</a>,” Employment and Training Administration, U.S. Department of Labor.</p>
<p data-note_number='4'><a href="#_ref4" class="footnote-id-foot" id="_note4">4. </a> For more background on temporary work visas, their flaws, and proposed reforms, see Daniel Costa,&nbsp;<a href="https://www.epi.org/publication/temporary-work-visa-reform/"><em>Temporary Work Visa Programs and the Need for Reform: A Briefing on Program Frameworks, Policy Issues and Fixes, and the Impact of COVID-19</em></a>, Economic Policy Institute, February 2021.</p>
<p data-note_number='5'><a href="#_ref5" class="footnote-id-foot" id="_note5">5. </a> See for example, Daniel Costa and Ron Hira, <a href="https://www.epi.org/publication/h-1b-visas-and-prevailing-wage-levels/"><em>H-1B visas and prevailing wage levels: A majority of H-1B employers—including major U.S. tech firms—use the program to pay migrant workers well below market wages</em></a>, Economic Policy Institute, May 4, 2020; Ron Hira and Daniel Costa, <a href="https://www.epi.org/publication/new-evidence-widespread-wage-theft-in-the-h-1b-program/"><em>New evidence of widespread wage theft in the H-1B visa program: Corporate document reveals how tech firms ignore the law and systematically rob migrant workers</em></a>, Economic Policy Institute, December 9, 2021; Lauren Kaori Gurley, “<a href="https://www.vice.com/en/article/jgmpvb/analysis-claims-migrant-tech-workers-have-been-underpaid-by-tens-of-millions">Analysis Claims Migrant Tech Workers Have Been Underpaid by Tens of Millions</a>,” Vice News, December 9, 2021; Centro de los Derechos del Migrante, <a href="https://cdmigrante.org/wp-content/uploads/2018/01/Coerced-under-NAFTA_-Abuses-of-Migrant-Workers-in-TN-Visa-Program.pdf"><em>Coerced under NAFTA: Abuses of Migrant Workers in the TN Visa Program and Recommendations for Reform</em></a>, 2017; George Avalos, “<a href="https://www.mercurynews.com/2014/10/22/workers-paid-1-21-an-hour-to-install-fremont-tech-companys-computers/">Workers Paid $1.21 an Hour to Install Fremont Tech Company’s Computers</a>.”&nbsp;<em>The Mercury News</em>, October 22, 2014; Farah Stockman, “<a href="https://www.bostonglobe.com/editorials/2013/06/11/your-child-teacher-victim-human-trafficking/dQz2fYPwg6Xkgt1aV6HaiL/story.html">Teacher Trafficking: The Strange Saga of Filipino Workers, American Schools, and H-1B Visas</a>,”&nbsp;<em>Boston Globe</em>, June 12, 2013; Maria Perez, “American farms recruit Mexican veterinarians for jobs as animal scientists — but the real work is milking cows and cleaning pens for low pay,” <em>Milwaukee Journal Sentinel</em>, December 4, 2019 (updated Feb. 18, 2020).</p>
<p data-note_number='6'><a href="#_ref6" class="footnote-id-foot" id="_note6">6. </a> Sankar Mukhopadhyay and David Oxborrow,&nbsp;“<a href="https://link.springer.com/article/10.1007/s13524-011-0079-3">The Value of an Employment-Based Green Card</a>,”&nbsp;<em>Demography</em>&nbsp;49: 219–237, February 2012.</p>
<p data-note_number='7'><a href="#_ref7" class="footnote-id-foot" id="_note7">7. </a> Bivens and Shierholz broadly define “monopsony power” as “the leverage enjoyed by employers to set their workers’ pay.” See Josh Bivens and Heidi Shierholz,&nbsp;<a href="https://www.epi.org/publication/what-labor-market-changes-have-generated-inequality-and-wage-suppression-employer-power-is-significant-but-largely-constant-whereas-workers-power-has-been-eroded-by-policy-actions/"><em>What Labor Market Changes Have Generated Inequality and Wage Suppression?: Employer Power Is Significant but Largely Constant, Whereas Workers’ Power Has Been Eroded by Policy Actions</em></a>, Economic Policy Institute, Dec. 12, 2018.</p>
<p data-note_number='8'><a href="#_ref8" class="footnote-id-foot" id="_note8">8. </a> Eric M. Gibbons et al., “<a href="http://ftp.iza.org/dp12096.pdf">Monopsony Power and Guest Worker Programs</a>,” IZA Institute of Labor Economics, Discussion Paper no. 12096, January 2019.</p>
<p data-note_number='9'><a href="#_ref9" class="footnote-id-foot" id="_note9">9. </a> For a discussion about H-1B and TN visa portability in practice, see Daniel Costa, “Is portability a panacea? Changing employers in US temporary migration programmes,” chapter 8 in Christiane Kuptsch and Fabiola Mieres (eds.),&nbsp;<a href="https://www.ilo.org/resource/news/new-ilo-publication-explores-path-social-justice-migrant-workers"><em>Temporary labour migration: Towards social justice?</em></a>&nbsp;edited volume, International Labour Organization, February 2025.</p>
<p data-note_number='10'><a href="#_ref10" class="footnote-id-foot" id="_note10">10. </a> See for example, the economic literature on noncompete clauses, including, Evan Starr, “<a href="https://pubs.aeaweb.org/doi/pdfplus/10.1257/jep.20251457">The Economics of Noncompete Clauses</a>,” <em>Journal of Economic Perspectives</em> 40, no. 1 (2026): 139-166;, Matthew Johnson, Kurt Lavetti, and Michael Lipsitz, “<a href="https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3455381">The Labor Market Effects of Legal Restrictions on Worker Mobility</a>,” available at SSRN, June 6, 2020.&nbsp;</p>
]]></content:encoded>
											
	</item>
		<item>
		<title>The economics of deficits, debt, and interest rates: Part 2: Structural deficits today are too large</title>
		<link>https://www.epi.org/blog/the-simple-economics-of-deficits-and-debt-structural-deficits-today-are-too-large/</link>
		<pubDate>Mon, 02 Nov 2026 13:00:06 +0000</pubDate>
		<dc:creator><![CDATA[Josh Bivens]]></dc:creator>
		<guid isPermaLink="false">https://www.epi.org/?post_type=blog&#038;p=327691</guid>
					<description><![CDATA[This is part two of a series on the economics of debt, deficits, and today&#8217;s high interest rates—and how policymakers should respond.]]></description>
										<content:encoded><![CDATA[<div class="box">
<p><em>This is part two of a series on the economics of debt, deficits, and today&#8217;s high interest rates—and how policymakers should respond. Read <a href="https://www.epi.org/blog/deficits-debt-and-interest-rates-some-questions-answered/">part one here.</a>&nbsp;</em></p>
</div>
<p>The effects of deficits and debt on the broader economy are largely straightforward—with one big exception: They differ radically depending on whether the economy is operating near its productive potential or is instead operating well below it (which occurs most notably during recessions).<span id="more-327691"></span></p>
<p>The main economic effects of deficits in these two states of the world are essentially as follows:</p>
<ul>
<li>Large deficits during times when the economy is operating near its productive potential (or during times of <em>full employment</em>) put upward pressure on either interest rates (more likely) or inflation (less likely).
<ul>
<li>The higher interest rates in turn can “crowd out” productive private-sector capital investments. If this happens chronically, this can shrink the size of the capital stock U.S. workers have available to them—with the capital stock being factories and other buildings, equipment, software, and research and development efforts. This smaller capital stock in turn reduces productivity, or the income generated in an average hour of work in the U.S. economy, hence shrinking the growth of potential living standards.</li>
</ul>
</li>
<li>Large deficits when the economy is operating well below potential actually stabilize the economy and aid its return to potential.
<ul>
<li>Economies start operating below full employment when aggregate demand—spending by households, businesses, and governments—falls short of what is needed to fully employ productive resources. By increasing spending and lowering taxes (hence boosting private-sector disposable income that can be spent), higher deficits boost aggregate demand and push the economy closer to potential.</li>
<li>Because private-sector investment responds to income growth as well as interest rates, deficits run during times of less than full employment can actually “crowd in” productive private-sector investment.</li>
</ul>
</li>
</ul>
<h4><strong>The one noncontroversial effect of budget deficits: They raise aggregate demand</strong></h4>
<p>If there is one noncontroversial statement about budget deficits, it is that <em>they raise the economy’s aggregate demand</em>. Whether this increase in aggregate demand is good or bad depends on the broader economy.</p>
<h5><strong>Deficits when private spending is strong</strong></h5>
<p>When private-sector sources of demand are strong—consumers are buying a lot, businesses are demanding lots of new capital goods to undertake production, and imports and exports are roughly balanced—all available productive resources (like workers) are fully employed and the economy does not need the public sector to inject extra demand into circulation with larger deficits; additional increments to demand will not bring forth more output. Instead, additional demand sent into a full employment economy will bid up <em>prices </em>rather than cause more output to be produced.</p>
<p>If the Federal Reserve is determined to not let strong demand bid up prices and push inflation above its preferred 2% target, it will respond to this excess of demand by raising its short-term policy rates to try to tamp down private sources of demand. Higher interest rates <a href="https://economics.mit.edu/sites/default/files/publications/Semiparametric%20Estimates%20of%20Monetary%20Policy-V2.pdf">are generally effective in reducing demand</a> through lower consumption, business investment, and larger trade deficits.</p>
<p>Besides the Fed’s response, higher deficits in a full employment economy can also put upward pressure on interest rates through financial crowding out. At full employment, the supply of savings available to be loaned to households or businesses looking to make long-lived investments is largely fixed. Federal deficits <em>by definition</em> mean that the government has to go into financial markets, where supplies of savings and demands for investment meet, and increase <em>its own</em> demand for a share of these savings. Deficits are financed by selling government bonds to households and businesses—that is, by convincing savers to lend their money to the government for a period of time. The extra demand for borrowing meets the fixed supply of savings and pushes up the price of loans, or interest rates.</p>
<p>Because most capital investment by businesses (building new factories and buildings, purchasing new equipment and software, and paying salaries of workers engaged in research and development) is financed by borrowing, higher interest rates stemming from large budget deficits in economies that are already strong will slow capital investment over time. As the price of something rises—in this case, the price of borrowing to finance new investment—less of it will be demanded.</p>
<p>This reduction in capital investment slows long-run economic growth. A key contributor to growth in productivity—the amount of income generated in an average hour of work in the economy—is how much capital workers have available to them to do their jobs. Construction workers today are far more productive than 50 years ago because they’re more likely to use mechanized earth movers rather than shovels for excavation.</p>
<p>If deficits crowd out capital investment, this can slow productivity growth. Productivity growth is the ceiling on how fast average living standards can rise in an economy. So when large deficits during full employment crowd out of investment, this can damage economic growth. A key point in this analysis is that it is higher interest rates that are the transmission mechanism for this damage: If interest rates do not rise, there is no damage being done.</p>
<h5><strong>Deficits when private spending is weak</strong></h5>
<p>When private spending is too low to fully employ all available productive resources, the boost to aggregate demand from large deficits actually helps to employ these resources and move the economy closer to its productive potential.</p>
<p>Some increases in deficits that push the economy closer to its productive potential during downturns <a href="https://www.cbo.gov/publication/62568">happen automatically</a>. Spending rises on safety net programs like unemployment insurance, and this supports private consumption. Tax collections fall as private incomes fall, buffering disposable incomes from the full fall in pre-tax incomes.</p>
<p>Some increases in deficits that push the economy back to its potential come from discretionary legislation passed by Congress in direct response to the downturn—relief packages like the <a href="https://obamawhitehouse.archives.gov/sites/default/files/cea_8th_arra_report_final_draft.pdf">American Recovery and Reinvestment Act (ARRA) of 2009</a> and the <a href="https://bfi.uchicago.edu/insight/research-summary/cares-impact-on-welfare/">CARES Act of 2020</a>.</p>
<p>Weak private spending places sharp downward pressure on inflation and interest rates as it generates idle resources (like rising unemployment). Given this, a boost to demand from larger deficits tends to <em>not</em> manifest as faster inflation or higher interest rates. A boost to demand when there are idle resources mostly causes output, not prices, to rise. And the higher output and income generated by higher deficits <em>create</em> savings that keep interest rates from rising in response to greater demand for investment.</p>
<p>Because this scenario sees no rise in interest rates, and because investment responds positively to output growth, larger deficits that boost economic activity <a href="https://www.brookings.edu/wp-content/uploads/2012/03/2012a_delong.pdf">almost certainly “crowd in” private investment</a>, leading to more investment and higher productivity than would have occurred had deficits not been allowed to rise.</p>
<h4><strong>Excessive structural deficits are hurting the U.S. economy today</strong></h4>
<p>The upshot is that deficits can hurt if they’re too large when the economy is already near full employment, and that the data signature for this damage is rising interest rates. For decades, large budget deficits have not been accompanied by higher interest rates, making it very hard to argue that the economy’s long-run growth prospects were being damaged by crowded out private investment.</p>
<p>Over the past year, however, interest rates have risen noticeably. In fact, as of September 2026, 10-year Treasury rates have reached their highest levels since the early 2000s (though still far below highs reached between the mid-1970s and mid-1990s when—among other things—inflation expectations were much higher). Because inflation is a bit higher now than in the early 2000s, real (inflation-adjusted) rates have not risen quite as fast, but they are always noticeably higher than the average of the past 20 years. Between this and the fact that unemployment is likely not too far above full employment at the moment and the fact that inflation remains above the Fed’s 2% target level, it seems clear that too large U.S. deficits are putting unwelcome pressure on the larger economy.</p>
<p><iframe style="overflow: hidden; width: 540px; height: 433px;" src="https://fred.stlouisfed.org/graph/graph-landing.php?g=20dZ0&amp;width=540&amp;height=383" frameborder="0" scrolling="no"></iframe></p>
<p>Too large U.S. deficits are not the only driver of recent interest rate increases. Another contributor is <a href="https://ec.europa.eu/eurostat/web/products-euro-indicators/w/2-21102026-cp">larger deficits being run by many European countries</a>. Many of these countries have <a href="https://www.consilium.europa.eu/en/press/press-releases/2025/07/08/council-activates-flexibility-in-eu-fiscal-rules-for-15-member-states-to-increase-defence-spending/">loosened fiscal rules to spend more on the military</a> in the face of the Russian invasion of Ukraine. The Japanese economy is also stronger than it has been in decades, and <a href="https://www.nytimes.com/2026/09/17/business/japan-interest-rates.html">interest rates there are rising</a> for the first time in decades. Finally, there has been a huge surge in <a href="https://www.epi.org/blog/how-ai-spending-is-impacting-the-u-s-economy/">private investment demand</a> driven by the race between AI companies to dominate the market.</p>
<p>On the one hand, it seems hard to worry that higher interest rates are crowding out private-sector investment if it is precisely strong private investment contributing to higher rates. But the investment in AI is clearly not interest-sensitive (at least so far). It is driven by exogenous events and factors: Now just happens to be the time that the technology has reached tipping points where scalability and profitability seem possible, hence sparking a race between private companies to be the first to capitalize on this. Additionally, investment in sectors in the U.S. not being pulled along by the AI boom is <a href="https://www.epi.org/blog/how-ai-spending-is-impacting-the-u-s-economy/">notably weak in 2026</a>, posting negative growth. If U.S. fiscal deficits were lower and this led to lower interest rates, this would help <a href="https://www.epi.org/publication/the-trump-administrations-macroeconomic-agenda-harms-affordability-and-raises-inequality/">investment in non-AI sectors</a>.</p>
<p>Noting that too large structural deficits are likely slowing U.S. growth today does not map directly onto policy solutions. It’s not the case that the damage from too large deficits right now is so overwhelming that anything that reduces deficits should be done in response. In a later post, I will walk through the considerations about how policymakers should optimally respond to too high structural deficits. They definitely should do <em>something—</em>but it’s not the case that they should just do <em>anything</em>.</p>
]]></content:encoded>
											
	</item>
		<item>
		<title>The working class is more diverse than ever before</title>
		<link>https://www.epi.org/publication/the-working-class-is-more-diverse-than-ever-before/</link>
		<pubDate>Fri, 30 Oct 2026 13:40:07 +0000</pubDate>
		<dc:creator><![CDATA[Valerie Wilson]]></dc:creator>
		<guid isPermaLink="false">https://www.epi.org/?post_type=publication&#038;p=327793</guid>
					<description><![CDATA[Regardless of the definition you use, today’s working class is diverse in terms of race, ethnicity, and gender.]]></description>
										<content:encoded><![CDATA[<p>The working class is a concept that carries tremendous political weight. But there is no official definition of what it is, and popular views of a mostly white working class are out of date.</p>
<p>Over the past 50 years, increases in women’s participation in the labor force and immigration have resulted in more women, Hispanics, and Asian Americans in the workforce. <strong>New data below from a forthcoming <a href="https://www.epi.org/publication/todays-working-class-is-more-diverse-than-ever-a-new-picture-emerges-as-demographic-change-and-economic-inequality-challenge-old-perceptions/">EPI report</a></strong> show that the working class is more diverse than ever before—no matter how you define it.</p>
<h4><strong>When defined by education, the working class is nearly half workers of color </strong></h4>
<p>The traditional definition of the working class comprises workers with less than a bachelor’s degree. Under this definition, in 1979, the working class was over 80% white. Today, workers of color collectively account for nearly <em>half</em> of the working class.</p>
<h5><strong>Working class by race and ethnicity</strong></h5>
<ul style="list-style-type: circle;">
<li>52% white</li>
<li>26.8% Hispanic</li>
<li>15.1% Black</li>
<li>4.7% AAPI</li>
<li>1.4% Native American</li>
</ul>
<h5><strong>Working class by gender</strong></h5>
<ul style="list-style-type: circle;">
<li>56.4% men</li>
<li>43.6% women</li>
</ul>
<h4><strong>When defined by income, the working class is similarly diverse</strong></h4>
<p>The working class could reasonably include more college graduates and traditional “white-collar workers” than were previously considered. That’s because the wage boost from getting a college degree is lower than it used to be, and underemployment—along with racial and gender pay disparities—can mean the economic status of graduates of color and women is more like some nongraduates.</p>
<p>By this broader measure, the working class can be considered the bottom 75% of U.S. households, regardless of education level or job type. This works out to a family income of below $150,000 per year.</p>
<p>Under this definition, the working class remains similarly diverse in terms of race and ethnicity. While men are still a majority of the working class, a larger share of women fit within this definition of working class, given lower pay in jobs predominantly held by college-educated women.</p>
<h5><strong>Working class by race and ethnicity</strong></h5>
<ul style="list-style-type: circle;">
<li>53.1% white</li>
<li>24% Hispanic</li>
<li>15.2% Black</li>
<li>6.5% AAPI</li>
<li>1.2% Native American</li>
</ul>
<h5><strong>Working class by gender</strong></h5>
<ul style="list-style-type: circle;">
<li>52.5% men</li>
<li>47.5% women</li>
</ul>
<p>However you define it, today’s working class is diverse by race, ethnicity, and gender. Improving economic conditions for today’s working class requires policies that meet the needs of <em>all</em> these workers. &nbsp;</p>
<p><em><strong>Source:</strong> Author&#8217;s analysis of pooled 2024–2025 Current Population Survey basic monthly microdata from EPI Microdata Extracts, Version 2026.8.13, </em><em><a href="https://microdata.epi.org/">https://microdata.epi.org</a>.</em></p>
<p>&nbsp;</p>
<p>&nbsp;</p>
]]></content:encoded>
											
	</item>
		<item>
		<title>Voters in Colorado and Florida face major tax questions on the November ballot</title>
		<link>https://www.epi.org/blog/voters-in-colorado-and-florida-face-major-tax-questions-on-the-november-ballot/</link>
		<pubDate>Fri, 23 Oct 2026 14:37:28 +0000</pubDate>
		<dc:creator><![CDATA[David Cooper]]></dc:creator>
		<guid isPermaLink="false">https://www.epi.org/?post_type=blog&#038;p=327377</guid>
					<description><![CDATA[Key Colorado Amendment 87 would help fix a structural funding problem that has undermined the state’s public services for 35 years.]]></description>
										<content:encoded><![CDATA[<div class="quick-card">
<h4><strong>Key Takeaways:</strong></h4>
<ul>
<li>Colorado Amendment 87 would help fix a structural funding problem that has undermined the state’s public services for 35 years. The measure would modestly raise taxes on the top 3% of taxpayers while generating $2 billion in revenue.</li>
<li>Florida Amendment 3 would devastate public services across the state. It would cut annual local government revenue by an estimated $12 billion, equivalent to 17% of statewide local government compensation. This translates to 123,000 jobs at risk of being cut from local government payrolls across the state.</li>
</ul>
</div>
<p>Lots of media attention around the upcoming midterm elections is understandably focused on which party will win control of Congress, as well as the many state legislative chambers and governorships that are up for grabs. However, there are also over 150 state ballot measures going before voters next month, some of which will be tremendously consequential in shaping state economies for years to come. Two such proposals warrant greater attention: Amendment 87 in Colorado and Amendment 3 in Florida. Both would be transformative for their respective states—the former in positive ways, the latter in catastrophic ones.<span id="more-327377"></span></p>
<h4>Colorado Amendment 87 would make the state tax system fairer while boosting revenues for essential public services</h4>
<p>On November 3rd, Colorado voters will vote on <a href="https://leg.colorado.gov/ballots/graduated-income-tax-2/files/blue-book">Amendment 87</a>, a measure that would replace the state’s constitutionally mandated flat tax—which requires a teacher’s aide making $40,000 a year to pay the same tax rate as a billionaire tech CEO—with a progressive, graduated income tax. The new tax system would maintain current tax rates on households and businesses earning $100,000–$500,000 a year, lower taxes for those making less than $100,000 a year, and raise taxes modestly on households and businesses making more than $500,000 a year. All told, these changes <a href="https://bellpolicy.org/graduated-income-tax-faq/">would mean a modest tax increase on the richest 3%</a> of Colorado taxpayers, with everyone else in the state paying the same or less than before.</p>
<p>In a state and country shaped by tremendous income inequality, even modest changes to tax rates on the highest earners—like those being considered in Colorado—can yield substantial new revenues. Analysts estimate Amendment 87 <a href="https://bellpolicy.org/graduated-income-tax-faq/#how-much-revenue-could-a-graduated-income-tax-generate">would generate over $2 billion in new funds</a> to support public education, health, and child care in the state. (The state’s general fund gross revenues for FY25–26 <a href="https://content.leg.colorado.gov/sites/default/files/FY26-27ApprReptA.pdf">were projected to be about $17 billion</a>, so a $2 billion increase is sizable.)</p>
<p>In many ways, the Colorado measure is not very different from other high-earner taxes passed in Maine, Washington, Rhode Island, and Hawaii this past year. Massachusetts’ “fair share” tax, adopted by voters in 2022, has been a huge success, exceeding revenue projections and <a href="https://massbudget.org/2024/10/07/two-years-fsa/">funding significant investments</a> in public transit, free school meals, free community college, and early childhood education.</p>
<p>Amendment 87 is notable because it would begin to fix the state’s uniquely broken funding system. In 1992, Colorado voters adopted a “Taxpayer’s Bill of Rights” (TABOR) amendment that caps state revenue even when the state economy grows. Colorado is the only state that has such a measure, and for over 30 years, it has kept the state’s public services from keeping up with a growing and aging population. <a href="https://www.epi.org/publication/the-teacher-pay-penalty-remained-high-in-2025-teachers-and-students-nationwide-are-still-paying-the-price-of-decades-of-neglect/#epi-toc-5">Low teacher salaries</a> have driven high teacher turnover in Colorado public schools, many of which have <a href="https://coloradofiscal.org/tabor-the-colorado-con-2/">increasingly had to shift to four-day school weeks</a>. The state workforce has <a href="https://progov21.org/Download/Document/YX1532">similarly struggled with turnover</a>, and federal funding cuts to Medicaid, among other programs, exacerbate existing strains on programs serving <a href="https://bellpolicy.org/the-state-of-aging-in-colorado-2025/">one of the fastest aging populations in the country</a><a name="_Hlt243574891"></a>.</p>
<p>Unsurprisingly, opponents are making typical arguments that the measure would cause high earners and small businesses to flee the state en masse—arguments that <a href="https://www.asanet.org/wp-content/uploads/attach/journals/jun16asrfeature.pdf">empirical research</a> has <a href="https://www.cbpp.org/blog/states-should-forge-ahead-with-new-revenues-in-face-of-misleading-tax-migration-claims">thoroughly debunked</a>. Colorado’s Governor Jared Polis, <a href="https://www.propublica.org/article/how-these-ultrawealthy-politicians-avoided-paying-taxes">himself a wealthy tech entrepreneur</a>, has claimed the measure would hurt small businesses, though independent analysis from the Institute on Taxation and Economic Policy (ITEP) found that the corporate tax changes would fall overwhelmingly on large multinational corporations with shareholders located across the globe. ITEP estimates that 90% of the corporate impact would fall on people who aren’t Coloradans.</p>
<p>Amendment 87 would go a long way toward fixing a nearly 35-year-old structural funding problem that has hamstrung public services in Colorado while making the state tax system fairer. If voters there see through the fearmongering, it could help motivate lawmakers and advocates in other states to pursue similar tax changes critical to bolstering vital public services, combating inequality, and improving outcomes for everyday working families.</p>
<h4>Florida Amendment 3 would lead to substantial job losses and public service cuts</h4>
<p>Florida <a href="https://ballotpedia.org/Florida_Amendment_3,_Homestead_Tax_Exemptions,_Property_Assessments,_and_Spending_Restrictions_Amendment_(2026)">Amendment 3</a>, a constitutional amendment placed on the ballot by the state legislature at the request of Governor Ron DeSantis, would fundamentally undermine Florida local governments’ ability to raise revenue and fund public services. The measure would dramatically increase the state nonschool property tax exemption on homesteads from the current $25,000 to $250,000 in 2028, indexing it to inflation thereafter; lower the 10% assessment growth cap on all other properties to 5%; and require the legislature to establish procedures allowing local governments to raise homestead exemptions even further—opening the door for potential full elimination of nonschool property taxes. The state Office of Economic and Demographic Research estimates that the amendment would <a href="https://www.floridapolicy.org/posts/amendment-3-local-funding-cuts-calculator">cut local tax revenue by nearly $14 billion in the first two years alone</a> and structurally reduce local tax revenues by nearly $12 billion annually thereafter.</p>
<p>Funding losses of this magnitude would be devastating for public services <a href="https://amendment3.floridapolicyinstitute.org/">in every community across the state</a>. Keep in mind that when analysts describe funding cuts and reductions in “public services,” what they’re often really describing is people—who work on behalf of their neighbors and communities—losing their jobs.</p>
<p>To make Amendment 3’s potential damage concrete, we can compare the estimated permanent funding losses with current local government compensation and employment data in the state to get a rough sense of the potential shock to local government employment.</p>
<p><iframe id="datawrapper-chart-fh74x" style="width: 0; min-width: 100% !important; border: none;" title="Amendment 3 funding losses represent large shares of local public jobs" src="https://datawrapper.dwcdn.net/fh74x/8/" height="798" frameborder="0" scrolling="no" aria-label="Choropleth map" data-external='1'></iframe><script type="text/javascript">(function(){function e(){window.addEventListener(`message`,function(e){if(e.data[`datawrapper-height`]!==void 0){var t=document.querySelectorAll(`iframe`);for(var n in e.data[`datawrapper-height`])for(var r=0,i;i=t[r];r++)if(i.contentWindow===e.source){var a=e.data[`datawrapper-height`][n]+`px`;i.style.height=a}}})}e()})();</script></p>
<p>The map above shows the projected permanent revenue losses that would result from Amendment 3 for every Florida county. It also shows what these losses represent as a share of total compensation of local government employees in each county. (<strong>Table 1 </strong>below contains the same data for all counties and the state overall.) The data show that the projected funding losses are equal to roughly 17% of total local government compensation statewide, rising as high as 39% of compensation in Flagler County.</p>
<p>The map and table also report local government employment as of 2025, which statewide totaled almost 720,000 workers. <strong>If local governments were forced to cut staffing in direct proportion to their funding losses, it would mean job losses for almost 125,000 local public-sector workers across the state.</strong> In Flagler County, a 39% loss in public-sector employment translates to over 1,500 people losing their jobs. Notably, Flagler county—a relatively small Atlantic-Ocean-adjacent county north of Daytona Beach—had fewer than 29,000 jobs total in 2025, meaning that a loss of 1,500 jobs would be over 5% of total employment (public and private) in the county.</p>
<p>Of course, not every county and municipality will absorb these funding losses entirely through staffing cuts; most will likely adopt a mix of staffing cuts, reductions in services, fewer supply purchases, canceled contracts with private vendors, and possibly sales of public assets. Each of these would have ripple effects on local households and businesses. If a county, for instance, is forced to stop contracting with a local lawn care or building services company, those formerly contracted local businesses could also have to cut staff.</p>
<p>Local governments will also likely try to increase revenues from other sources, such as service fees and tax increases on entities not eligible for the amendment’s exemptions. For instance, Amendment 3 does not provide an expanded tax exemption for non-homestead properties—meaning that commercial property owners will likely see their property taxes rise, which <a href="https://flhousingactionlab.substack.com/p/new-analysis-estimates-amendment">they could very likely pass on to renters</a> and tenant businesses.</p>
<p>This would shift more of the local tax bill away from homeowners and onto renters, who are typically less wealthy. With no income tax and a large portion of state and local revenues generated by sales and excise taxes, Florida already has <a href="https://itep.org/whopays/florida-who-pays-7th-edition/">the most regressive tax system in the country</a>. The lowest-income taxpayers in the state pay effective tax rates nearly five times higher than the top 1% of taxpayers.</p>
<p>If voters adopt Amendment 3 next month, the consequences for communities across the state will be dire. Researchers at the Florida Policy Institute have described in detail how Amendment 3 would gut resources for <a href="https://www.floridapolicy.org/posts/fact-sheet-potential-impact-of-amendment-3-on-floridas-public-hospital-districts">public hospitals</a>, <a href="https://www.floridapolicy.org/posts/local-services-for-florida-children-and-parents-would-see-deep-funding-cuts-under-amendment-3">child and family services</a>, and <a href="https://www.floridapolicy.org/posts/a-one-two-punch-fema-cuts-and-amendment-3s-property-tax-reductions-could-both-hit-florida-at-once">emergency and disaster response services</a>, among other things. In the short run, especially as the labor market is <a href="https://www.epi.org/blog/weak-september-jobs-report-shows-hiring-and-wage-growth-slow/">looking increasingly unsteady</a>, enacting massive funding cuts to local governments is incredibly foolish just for the direct impacts it will have on jobs and spending in local economies. In the long run, the harm to the welfare of Florida residents from a severely diminished public sector could be devastating.</p>
<p>Together, these measures show how much is at stake on state ballots this November. Colorado could begin to fix a decades-old structural funding problem, while Florida may create a new one of its own—one that would strip billions from local governments and the services they provide.</p>


<!-- BEGINNING OF FIGURE -->

<a name="Table-1"></a><div class="figure chart-327340 figure-screenshot figure-theme-none" data-chartid="327340" data-anchor="Table-1"><div class="figLabel">Table 1</div><img decoding="async" src="https://files.epi.org/charts/img/327340-36049-email.png" width="608" alt="Table 1" class="fig-image-from-url rsImg"><div class="fig-features donotprint"></div></div><!-- /.figure -->

<!-- END OF FIGURE -->


<p>&nbsp;</p>
]]></content:encoded>
											
	</item>
		<item>
		<title>A pro-worker AI policy playbook: Strengthening unemployment insurance and building a better safety net</title>
		<link>https://www.epi.org/publication/a-pro-worker-ai-policy-playbook-strengthening-unemployment-insurance-and-building-a-better-safety-net/</link>
		<pubDate>Wed, 21 Oct 2026 09:00:20 +0000</pubDate>
		<dc:creator><![CDATA[Mary Beech]]></dc:creator>
		<guid isPermaLink="false">https://www.epi.org/?post_type=publication&#038;p=326852</guid>
					<description><![CDATA[Policymakers can act now to reduce the harm to workers of losing a job for any reason, by strengthening the safety net to support them until they find another.
&#160;]]></description>
										<content:encoded><![CDATA[<p><img decoding="async" src="https://files.epi.org/uploads/EPI-Workshop_WeBuild-Cobrand-2.png" width="430"></p>
<div class="quick-card">
<h4><strong><span style="font-size: 18px; font-family: 'Harriet Display', serif; color: #1372ae;">Key takeaways</span></strong></h4>
<ul>
<li><span style="font-size: 14px;">Policy choices will determine whether AI helps or harms workers.</span></li>
<li><span style="font-size: 14px;">Given that, today’s priority should be to strengthen the safety net for workers to meet any scenario and to acknowledge how AI is already affecting workers’ rights, protections, and livelihoods.</span></li>
<li><span style="font-size: 14px;">To strengthen the unemployment insurance system that provides funds to workers who have lost their jobs, we should do the following:&nbsp;</span></li>
</ul>
<ul>
<li style="list-style-type: none;">
<ul style="margin-left: 20px !important; padding-left: 20px !important;">
<li style="list-style-type: circle !important; border-left: none !important;">increase the amounts and duration of unemployment benefits&nbsp;&nbsp;</li>
</ul>
</li>
</ul>
<ul>
<li style="list-style-type: none;">
<ul style="margin-left: 20px !important; padding-left: 20px !important;">
<li style="list-style-type: circle !important; border-left: none !important;">expand and standardize eligibility to receive benefits</li>
</ul>
</li>
</ul>
<ul>
<li style="list-style-type: none;">
<ul style="margin-left: 20px !important; padding-left: 20px !important;">
<li style="list-style-type: circle !important; border-left: none !important;">align employer incentives with worker access to benefits</li>
</ul>
</li>
</ul>
<ul>
<li style="list-style-type: none;">
<ul style="margin-left: 20px !important; padding-left: 20px !important;">
<li style="list-style-type: circle !important; border-left: none !important;">center delivery on worker needs and power</li>
</ul>
</li>
</ul>
<ul>
<li style="list-style-type: none;">
<ul style="margin-left: 20px !important; padding-left: 20px !important;">
<li style="list-style-type: circle !important; border-left: none !important;">build stronger data infrastructure to support workers while they are searching for new jobs</li>
</ul>
</li>
<li><span style="font-size: 14px;">Establishing a stand-alone unemployment program solely focused on AI raises important considerations. These include program design, stakeholder roles, eligibility criteria challenges, the resources needed to be effective, and vulnerability to reauthorization hurdles in Congress.</span></li>
</ul>
</div>
<p>&nbsp;</p>
<p><i>This is the first brief in a series exploring comprehensive federal policy recommendations around AI and workers.</i></p>
<div class="pdf-page-break "></div>
<h2>Introduction</h2>
<p>Amid the seemingly endless predictions about AI, one point is essential to remember: How this plays out is not predetermined. Policy choices determine whether workers are helped or harmed by technological change. Even without AI, many workers already face harmful disruptions. Every year, 20 million workers are laid off in the U.S. economy<a href="#_note1" class="footnote-id-ref" data-note_number='1' id="_ref1">1</a>—due to employer mismanagement, once-profitable products becoming less so, offshoring, automation, and other factors. For individual workers, losing a job fo<em>r any reason</em> and for <em>any period of time</em> is traumatic and damaging in large part because we lack adequate social insurance programs to smooth those transitions.</p>
<p>As lawmakers debate how to best regulate AI deployment in our workplaces, today’s priority should be to strengthen the safety net for workers in any scenario and to acknowledge how AI is already affecting workers’ rights, protections, and livelihoods. Decades of underinvestment, coupled with attacks on unions and other labor supports, have left workers without an adequate safety net.</p>
<p>We can improve our current systems to better provide livable benefits for workers who lose their jobs because of large-scale economic shocks that are not their fault and outside their control. We can also learn from other periods of workforce disruption—the COVID-19 pandemic, trade-related shocks, and the 2008 recession—about which policy choices were effective and which fell short.</p>
<p>When workers lose their jobs due to factors outside their control, they are considered “displaced” or “dislocated.” If they meet certain requirements, these workers are eligible for unemployment insurance (UI). To strengthen outcomes for workers, this brief will provide recommendations to:</p>
<ul>
<li>increase UI benefit amount and duration</li>
<li>expand and standardize eligibility to receive benefits</li>
<li>align employer incentives with worker access to benefits</li>
<li>center worker needs and power</li>
<li>strengthen data infrastructure to support workers</li>
</ul>
<p>These recommendations should be considered amid conversations focused on creating a stand-alone unemployment program geared toward AI-driven job loss. These discussions should weigh considerations and potential pitfalls, including:</p>
<ul>
<li>a blank slate for program design</li>
<li>relevant roles for key stakeholders</li>
<li>eligibility criteria challenges</li>
<li>the allocation of resources, time, and awareness to be effective</li>
<li>reauthorization vulnerability</li>
</ul>
<p>If policymakers are serious about easing the impact of AI-related job transitions for workers, the priority solution should be strengthening already existing social insurance programs to reduce the economic consequences of losing any job and to provide adequate support to workers while they seek other employment.</p>
<p>Changes to the current system or establishing a new program will require robust resources, including time, money, and expertise. This investment can help all workers, regardless of their risk of AI-based displacement. With a strong safety net, workers who lose their jobs can get the assistance they need, including the opportunity to find work that is well suited to their skills and experiences, and businesses are more inclined to improve their practices because they know workers aren&#8217;t trapped in their current jobs by economic necessity.<a href="#_note2" class="footnote-id-ref" data-note_number='2' id="_ref2">2</a> This moment presents an opportunity to strengthen the labor market for everyone and to ensure workers have the safety net they deserve.</p>
<h2>How to strengthen our current system</h2>
<p>Any effective approach to fixing the UI system must be comprehensive and should apply regardless of whether a worker is displaced due to AI. Numerous experts have weighed in on the best approach—including the National Academy of Social Insurance (NASI) Unemployment Insurance Taskforce,<a href="#_note3" class="footnote-id-ref" data-note_number='3' id="_ref3">3</a> the National Employment Law Project (NELP),<a href="#_note4" class="footnote-id-ref" data-note_number='4' id="_ref4">4</a>,<a href="#_note5" class="footnote-id-ref" data-note_number='5' id="_ref5">5</a> and The Century Foundation,<a href="#_note6" class="footnote-id-ref" data-note_number='6' id="_ref6">6</a> among others. This brief draws from those recommendations to strengthen outcomes for workers. Other important issues, such as modernizing UI system technology and infrastructure, fraud detection and prevention, and the use of AI in benefit determinations and delivery, also shape worker outcomes but fall outside this brief&#8217;s scope.</p>
<h3>Increase benefit amount and duration</h3>
<p>First and foremost, UI should be designed to ensure that workers can afford housing, food, and other essentials during their period of joblessness. The percentage of wages that UI replaces for workers has declined over time and has not kept pace with inflation.<a href="#_note7" class="footnote-id-ref" data-note_number='7' id="_ref7">7</a> In 2025, national average UI benefits replaced well under half of national average wages.<a href="#_note8" class="footnote-id-ref" data-note_number='8' id="_ref8">8</a> This gap leaves workers across the income spectrum unable to maintain their standard of living and can push those whose earnings were close to the minimum wage into poverty.</p>
<h4>Ensure benefit amounts are consistent across states</h4>
<p>Benefit amounts vary across states. In Arizona and Florida, UI benefits replace less than 15% of average weekly wages.<a href="#_note9" class="footnote-id-ref" data-note_number='9' id="_ref9">9</a> In some states, the weekly benefit amount is just $222 per week.<a href="#_note10" class="footnote-id-ref" data-note_number='10' id="_ref10">10</a> While state variation in benefit amounts hurts all workers in stingy states, it weighs especially heavily on Black and Hispanic workers.</p>
<p>One recent study finds that differences in state rules cause the wage replacement rate for Black workers to be meaningfully lower than the rate for white workers, and that the discrepancies harm overall well-being.<a href="#_note11" class="footnote-id-ref" data-note_number='11' id="_ref11">11</a> These disparities compound. Even among those likely eligible, Black workers are less likely than white workers to receive UI benefits—a gap explained primarily by structural factors, not a difference in eligibility, that has held steady for three decades.<a href="#_note12" class="footnote-id-ref" data-note_number='12' id="_ref12">12</a></p>
<p>Wage replacement rates should be consistent across states. A recent federal bill has proposed changes to increase the replacement rate to 75% and provide additional weekly funding for dependents.<a href="#_note13" class="footnote-id-ref" data-note_number='13' id="_ref13">13</a> Given growing wage inequality, the rate also should vary by income level. In one proposal, the lowest earners would receive benefits that represent 85% of their wages.<a href="#_note14" class="footnote-id-ref" data-note_number='14' id="_ref14">14</a></p>
<h4>Extend the duration of benefits across states</h4>
<p>The duration of benefits also varies greatly from state to state. Until 15 years ago, eligible workers across the country had access to 26 weeks of benefits, a generally agreed-upon standard.<a href="#_note15" class="footnote-id-ref" data-note_number='15' id="_ref15">15</a> After the Great Recession however, federal law barred states from cutting benefit amounts, so many states instead cut the duration of unemployment benefits to stabilize their trust funds and lower tax rates for employers.<a href="#_note16" class="footnote-id-ref" data-note_number='16' id="_ref16">16</a> Some states now only provide 12 weeks of benefits, a short bridge to find a new job well-matched to workers’ skills and experiences.<a href="#_note17" class="footnote-id-ref" data-note_number='17' id="_ref17">17</a> A minimum of at least 26 weeks of benefits should be a uniform requirement across states.</p>
<p>Workers cannot currently depend on the Extended Benefits program, which was intended to provide support during economic downturns by providing additional weeks of benefits during periods of high and rising unemployment. The triggers for Extended Benefits have not been effective, leaving workers to wonder if they are available. In their absence, workers must rely on temporary programs funded by Congress, which depend on political priorities.<a href="#_note18" class="footnote-id-ref" data-note_number='18' id="_ref18">18</a> More reliable triggers, higher wage replacement levels, and a longer duration of benefits would make the Extended Benefits program more effective.<a href="#_note19" class="footnote-id-ref" data-note_number='19' id="_ref19">19</a></p>
<h3>Expand and standardize eligibility for benefits</h3>
<p>Only certain workers are eligible for UI, leaving others stranded without benefits if they lose their jobs. Eligibility criteria and the base period used to calculate benefits are set by state law, rather than federal statute. This state-by-state variation can make it difficult for workers to know whether they qualify for assistance and plan accordingly.</p>
<p>Generally, the group of workers ineligible for benefits includes independent contractors, job seekers who have not been recently employed, and those who have exhausted their benefits. Some employers have incorrectly classified many workers as independent contractors, and this misclassification prevents those workers from requesting access to the UI benefits to which they would otherwise be legally entitled.<a href="#_note20" class="footnote-id-ref" data-note_number='20' id="_ref20">20</a></p>
<p>UI should cover a broader range of workers, including part-time workers, independent contractors, job seekers who have not been recently employed, and seasonal and temporary workers. During the COVID-19 pandemic, Congress tasked states with implementing new UI programs for gig and platform workers who fell outside of the usual verification system, creating significant challenges.<a href="#_note21" class="footnote-id-ref" data-note_number='21' id="_ref21">21</a> Barring changes to account for the unique situation of these workers, these same issues will arise again.<a href="#_note22" class="footnote-id-ref" data-note_number='22' id="_ref22">22</a> By incorporating these workers into the UI system now, instead of waiting for a crisis, future challenges can be mitigated.</p>
<p>Policymakers should also consider stronger presumptive eligibility standards, under which workers are presumed eligible for benefits and paid promptly, with the burden placed on employers or the state to prove otherwise. Lowering the barriers for workers to receive benefits could meaningfully increase uptake among workers who currently go without benefits.</p>
<h3>Align employer incentives with worker access to benefits</h3>
<p>All eligible workers should be able to access UI benefits if they want them. State UI programs are generally financed through employer experience rating: Employers pay more taxes when they lay off more workers who are eligible for UI.<a href="#_note23" class="footnote-id-ref" data-note_number='23' id="_ref23">23</a> Therefore, the system’s setup incentivizes employers to limit their tax liability by restricting eligibility for the program in different ways. This can include not hiring certain workers as employees, not proactively informing workers about their eligibility, attempting to reclassify workers as independent contractors, and other forms of suppressing UI program access.</p>
<p>A Bureau of Labor Statistics survey found that only about a quarter of workers who were unemployed in 2022 applied for benefits.<a href="#_note24" class="footnote-id-ref" data-note_number='24' id="_ref24">24</a> Further, a recent survey found that 19% of recently unemployed workers encountered actions from an employer that were meant to deter them from applying, in addition to a quarter of all UI filers who said that their employer contested their benefits.<a href="#_note25" class="footnote-id-ref" data-note_number='25' id="_ref25">25</a> Low-income workers more frequently reported employer contestations , and workers who reported contestations of their claims by employers were less likely to receive benefits and to experience greater material and psychological burdens during the filing process.<a href="#_note26" class="footnote-id-ref" data-note_number='26' id="_ref26">26</a> There is a real opportunity to better align employer incentives to support workers in accessing UI benefits.</p>
<p>Presumptive eligibility, discussed above, would address some of these access barriers by shifting the burden of proof away from workers, but it would not resolve the underlying incentive employers have to minimize their tax liability. There have been many proposals for how to address this issue: broadening the tax base, payroll contributions from employees to UI, or increasing state taxes to fund programs.<a href="#_note27" class="footnote-id-ref" data-note_number='27' id="_ref27">27</a> Most basically, employers could pay taxes based on the total number of workers they lay off, not just those eligible for UI.<a href="#_note28" class="footnote-id-ref" data-note_number='28' id="_ref28">28</a> These questions are tied to broader considerations about how to finance a more effective program.</p>
<h3>Center on worker needs and power</h3>
<p>The UI system should be structured around ensuring that workers have the support they need as they seek reemployment. This requires a real investment in active labor market policy, including communicating with workers and unions about their needs in order to align training and placement. Congress should allocate funds to revitalize the Employment Service in the U.S. Department of Labor, which once played a key role in supporting reemployment.<a href="#_note29" class="footnote-id-ref" data-note_number='29' id="_ref29">29</a> For example, in a 1986 New Jersey pilot program, job search assistance measurably helped dislocated workers return to work sooner and increased their earnings.<a href="#_note30" class="footnote-id-ref" data-note_number='30' id="_ref30">30</a> A well-resourced and staffed U.S. Employment Service can leverage state labor market information—frequently used by employers—to help job seekers find opportunities aligned with their skill sets.<a href="#_note31" class="footnote-id-ref" data-note_number='31' id="_ref31">31</a></p>
<p>Worker organizations can also play an important role in the delivery of benefits, including facilitating training and connections to job search assistance.<a href="#_note32" class="footnote-id-ref" data-note_number='32' id="_ref32">32</a> In 2022, unemployed people who were members of a union were twice as likely to apply for UI benefits, and among applicants, 30 percentage points more likely to receive those benefits.<a href="#_note33" class="footnote-id-ref" data-note_number='33' id="_ref33">33</a> This sizable gap suggests that when workers have organizational support navigating the UI system, they are far more likely to receive the benefits they are entitled to, reinforcing the case for involving worker organizations directly in UI outreach, filing assistance, and perhaps even delivery. Furthermore, unions routinely negotiate with employers, both in collective bargaining agreements and when bargaining before layoffs, to secure agreements that employers will not contest workers’ applications for UI benefits.</p>
<p>In some Northern European countries, unions are responsible for administering unemployment benefits, and scholars have found that this system leads to stronger unions and worker voice in the economy.<a href="#_note34" class="footnote-id-ref" data-note_number='34' id="_ref34">34</a> Policymakers in the United States should consider how unions and other worker organizations could deliver unemployment benefits to workers.</p>
<h3>Strengthen data infrastructure to improve outcomes</h3>
<p>State workforce agencies collect quarterly wage records from employers to facilitate UI administration to gauge eligibility for benefits as well as benefit levels. States share wage records with the federal Census Bureau, and these data are included in longitudinal data sets that can follow workers over time. These data are used for both research and the administration of other policies (such as facilitating child support payments or tracking the performance of job training programs). Wage records provide an important but limited set of information that varies by state. A growing number of states have expanded their wage records to include additional information such as job titles, total hours worked, and job location.<a href="#_note35" class="footnote-id-ref" data-note_number='35' id="_ref35">35</a></p>
<p>State-level data are a valuable source of information about what is happening to workers, but the lack of standardization across states makes it difficult to gain a clearer picture of employment. To remedy this, lawmakers should create standardized requirements for enhanced wage records to analyze who is losing their job and where workers are finding new opportunities.</p>
<p>The enhanced information coming from states would be especially useful with the advent of AI since no single federal data source can reliably track how AI adoption is reshaping demand for specific occupations in real time or identify when firms are reducing jobs through attrition. Further, workers would benefit from better information about opportunities that align with their skill set.<a href="#_note36" class="footnote-id-ref" data-note_number='36' id="_ref36">36</a> Better integration with job posting data—including the National Labor Exchange—and reemployment systems would help workers navigate the job-search process.</p>
<h2>Considerations for a stand-alone program</h2>
<p>The UI system should be improved regardless of how AI affects employment and job opportunities. Additionally, some policymakers have suggested creating a stand-alone program geared toward AI-driven job loss. There is precedent for this approach. Congress has acted nine times since 1958 to create temporary, additional UI benefits, most recently during the COVID-19 pandemic.<a href="#_note37" class="footnote-id-ref" data-note_number='37' id="_ref37">37</a> There is much to be learned from successes and challenges of these previous programs.</p>
<p>Some of these programs also included benefits or services beyond unemployment insurance for workers. For example, Trade Adjustment Assistance has, at various points, included relocation assistance, training programs, a health coverage tax credit, and wage insurance for older workers who took a lower-paying job.<a href="#_note38" class="footnote-id-ref" data-note_number='38' id="_ref38">38</a> A new stand-alone program could similarly go beyond cash benefits, offering workers support to develop new skills, cover the costs of a transition, or find new opportunities.</p>
<h3>Blank slate for program design</h3>
<p>A new program can address the key structural issues with the current system discussed in the previous section. Any new program should be uniform across states so that benefit levels, duration, and eligibility are consistent for every worker regardless of where they live. A new program also presents the opportunity to build in the other improvements outlined above, including benefit levels set to a livable amount, expanded eligibility, and stronger data infrastructure. A new program could also adopt more fundamental changes to the claims process, such as the presumptive eligibility standards that Michigan moved toward during the COVID-19 pandemic.<a href="#_note39" class="footnote-id-ref" data-note_number='39' id="_ref39">39</a></p>
<p>As discussed in the worker power section, a new program should be explicitly structured around building worker power, for example, by involving worker organizations in program administration or benefit delivery. Further, a new program can focus on effective job placement by leveraging relationships between employers, unions, and other worker organizations.</p>
<p>Finally, a new program should also be designed to avoid UI&#8217;s employer-financing structure, which—as discussed—in most states, disincentivizes employers from making sure workers get the benefits they are due. Left unaddressed, that same structure would allow employers to game or underreport AI-related job loss in a new program.</p>
<h3>Relevant roles for key stakeholders</h3>
<p>A successful program requires effective coordination among stakeholders. As with UI and previous stand-alone programs, states would likely retain a central role in delivering benefits and services, given their existing infrastructure and more direct knowledge of local labor market conditions. Local agencies can provide labor market information to workers, worker organizations, and employers to facilitate job placement. Any program should carefully consider how to balance flexibility for states to tailor the program to meet local needs, while maintaining strong, uniform standards across states.</p>
<p>As discussed, worker organizations and unions can play a central role in program administration. Worker organizations can support workers in accessing relevant benefits and training opportunities, ensuring that worker needs and outcomes are central to programmatic strategy and implementation.</p>
<p>Employers bear significant responsibility for the program&#8217;s success. They are often the first to know a layoff is coming and should proactively connect workers to available services, for example, by supporting on-site services when requested by state or local Rapid Response teams. Employers can proactively share specific information about their talent and skills needs with labor agencies and worker organizations, and hire workers during or after relevant training.</p>
<h3>Eligibility criteria challenges</h3>
<p>Creating clear and consistent eligibility criteria will be a significant challenge in establishing a stand-alone program for workers affected by AI-related job displacement. For instance, a program to compensate Department of Energy workers harmed by nuclear materials has faced significant challenges determining eligibility, including the complexity of reconstructing historical exposure to nuclear materials to establish causation.<a href="#_note40" class="footnote-id-ref" data-note_number='40' id="_ref40">40</a> Challenges in establishing causation have also been an enormous barrier to worker relief through the Trade Adjustment Assistance program.<a href="#_note41" class="footnote-id-ref" data-note_number='41' id="_ref41">41</a></p>
<p>In another example, some pandemic unemployment assistance benefits are still hung up over eligibility disputes years later.<a href="#_note42" class="footnote-id-ref" data-note_number='42' id="_ref42">42</a> And from the start of the pandemic, there have been hotly contested debates about who should receive benefits.<a href="#_note43" class="footnote-id-ref" data-note_number='43' id="_ref43">43</a> Moreover, concerns about fraud can also complicate eligibility determinations. In the wake of fraud during the COVID-19 pandemic, many states adopted stricter verification requirements to confirm eligibility.<a href="#_note44" class="footnote-id-ref" data-note_number='44' id="_ref44">44</a> A new program focused on AI-driven job loss would need to balance similar integrity concerns against the risk of preventing eligible workers from receiving the benefits they are due.<a href="#_note45" class="footnote-id-ref" data-note_number='45' id="_ref45">45</a></p>
<p>It is virtually impossible to systematically link job loss with AI adoption. In many instances, AI’s effects on a worker or industry may be indirect. For example, workers may lose their jobs in part due to AI adoption months or years earlier or because of broader economic disruption caused by AI’s impact on different industries. Even setting aside these implementation challenges, restricting eligibility to AI-driven job loss raises a separate concern: fairness. AI is not the only source of economic disruption. Global trade (including offshoring and outsourcing) and the green energy transition continue to displace workers across industries.</p>
<p>Other workers should be entitled to the same benefits as those directly displaced by AI. Drawing arbitrary boundaries around a single cause of job loss will create inequities, a risk compounded if states must independently interpret what counts as AI-driven job loss, therefore creating inconsistent standards nationwide.</p>
<h3>Resources, time, and awareness to be effective</h3>
<p>Standing up a new program will take significant time and resources. In addition to providing benefits, this includes making meaningful investment in technological systems and program administration, a particular challenge in light of years of underinvestment in modernizing UI systems.<a href="#_note46" class="footnote-id-ref" data-note_number='46' id="_ref46">46</a> Federal and state governments need adequate resources for user experience testing, efficient application determinations, coordination with relevant partners, data collection and analysis, and programmatic evaluation. The challenges of implementing relief programs during COVID-19 are instructive. Slow initial payments, overwhelming caseloads that strained public servants, technological breakdowns, and significant confusion about eligibility requirements demonstrated the need to proactively develop systems that can respond more nimbly in times of crisis.<a href="#_note47" class="footnote-id-ref" data-note_number='47' id="_ref47">47</a></p>
<p>Starting a new program also requires sustained public outreach, as new and stand-alone programs have historically struggled because of a lack of public familiarity. For example, over half of workers did not know they had access to emergency sick leave during COVID-19.<a href="#_note48" class="footnote-id-ref" data-note_number='48' id="_ref48">48</a> In another example, when the Reagan administration made changes to the Trade Adjustment Assistance program, applications plummeted, perhaps because many workers believed the program had been abolished.<a href="#_note49" class="footnote-id-ref" data-note_number='49' id="_ref49">49</a> Further, most workers did not know that other services including training, job counseling, and relocation assistance were available through the program.<a href="#_note50" class="footnote-id-ref" data-note_number='50' id="_ref50">50</a></p>
<p>A stand-alone program could also come at the expense of the existing UI system, which the vast majority of states say is underfunded.<a href="#_note51" class="footnote-id-ref" data-note_number='51' id="_ref51">51</a> From 1999 to 2019, federal funding for UI administration declined by 30% on an inflation-adjusted basis.<a href="#_note52" class="footnote-id-ref" data-note_number='52' id="_ref52">52</a> If policymakers decide to divert resources that would otherwise go to the existing UI system, the system’s further deterioration could harm workers served by that system in both the short and long term.</p>
<h3>Reauthorization vulnerability in Congress</h3>
<p>If an AI-specific UI program requires congressional reauthorization, it will be vulnerable to the difficulty of coordinating bipartisan action in Congress. Several previous programs have been subject to a sunset date, requiring periodic reauthorization to continue accepting new claims. Sunsets and other reauthorization requirements leave workers vulnerable to congressional inaction, especially if the impacts of AI on the workforce play out over an extended time period.</p>
<p>For an example of a program caught up in this type of reauthorization limbo, Trade Adjustment Assistance lapsed multiple times, including most recently in 2022, leaving additional workers without access to the program. Another example is The Emergency Unemployment Compensation program, which was created in 2008 and lapsed five times between 2010 and 2013, before expiring for good at the end of that year.<a href="#_note53" class="footnote-id-ref" data-note_number='53' id="_ref53">53</a> This immediately cut off benefits for 1.3 million workers and for more than 3.6 million workers by the following September.<a href="#_note54" class="footnote-id-ref" data-note_number='54' id="_ref54">54</a></p>
<h2>Financing and the path forward</h2>
<p>The current worker safety net has failed to expand enough to cover the basic realities of today&#8217;s labor market, and inadequate benefit levels, incomplete coverage, and diminished worker representation are persistent problems. Deployment of AI and related technologies in the workplace adds urgency to addressing these issues.</p>
<p>Implementing the recommendations in this issue brief requires a significant investment that should be considered, along with program parameters and design. As AI drives up stock prices and generates wealth for a select few, policymakers are considering how the public should share in these gains. Ideas include increasing taxes for wealthy individuals, closing the gap between labor and capital tax rates, taxing AI use, and taxing data center development, with the proceeds from these new taxes financing programs that redistribute income more broadly across the economy. Considerations for potential taxation approaches will be discussed in a forthcoming brief.</p>
<p>This moment requires big thinking and corresponding action. While AI drives changes to the economy, improvements to the worker safety net are long overdue. We can learn from the past to improve the labor market for everyone.</p>
<h2>Acknowledgments</h2>
<p>Angela Hanks, Alex Hertel-Fernandez, Ben Zipperer, Gillian Gutierrez, Jacob Leibenluft, Janelle Jones, Josh Bivens, Michele Evermore, Patrick Oakford, Raj Nayak, and Stephen Nuñez.</p>
<div class="pdf-page-break "></div>
<h2>Notes</h2>
<p data-note_number='1'><a href="#_ref1" class="footnote-id-foot" id="_note1">1. </a> Bureau of Labor Statistics, “Layoffs and Discharges: Total Nonfarm” [JTSLDL]. Accessed from <a href="https://fred.stlouisfed.org/series/JTSLDL">FRED, Federal Reserve Bank of St. Louis</a>, October 2026.</p>
<p data-note_number='2'><a href="#_ref2" class="footnote-id-foot" id="_note2">2. </a> Patrick Oakford, Josh Bivens, and Celine McNicholas, <a href="https://www.epi.org/publication/federal-ai-legislation/"><em>Federal AI Legislation: An Evaluation of Existing Proposals and a Road Map Forward</em></a>, Economic Policy Institute, September 2024.</p>
<p data-note_number='3'><a href="#_ref3" class="footnote-id-foot" id="_note3">3. </a> National Academy of Social Insurance, Unemployment Insurance Task Force, <a href="https://www.nasi.org/wp-content/uploads/2024/07/Unemployment-Insurance-Task-Force-Final-Report.pdf"><em>Unemployment Insurance Task Force Final Report</em></a>, July 2024.</p>
<p data-note_number='4'><a href="#_ref4" class="footnote-id-foot" id="_note4">4. </a> Amy Traub, <a href="https://www.nelp.org/insights-research/boosting-economic-resilience-seven-ways-states-can-support-more-workers-with-unemployment-insurance/"><em>Boosting Economic Resilience: Seven Ways States Can Support More Workers with Unemployment Insurance</em></a>, National Employment Law Project, July 2025.</p>
<p data-note_number='5'><a href="#_ref5" class="footnote-id-foot" id="_note5">5. </a> Michele Evermore, <a href="https://www.nelp.org/app/uploads/2026/08/Policy-Brief-AI-is-New-The-Challenge-of-Mass-Unemployment-Isnt.pdf"><em>AI Is New. The Challenge of Mass Unemployment Isn&#8217;t</em>.</a> National Employment Law Project, September 2026.</p>
<p data-note_number='6'><a href="#_ref6" class="footnote-id-foot" id="_note6">6. </a> Michele Evermore and David Balducchi, <a href="https://tcf.org/content/commentary/a-roadmap-for-unemployment-insurance-reform/"><em>A Roadmap for Unemployment Insurance Reform</em></a>, The Century Foundation, October 2023.</p>
<p data-note_number='7'><a href="#_ref7" class="footnote-id-foot" id="_note7">7. </a> National Employment Law Project, <a href="https://www.nelp.org/app/uploads/2023/11/Policy-Advocacy-Brief-Benefit-Amounts-11-2023.pdf."><em>Policy Advocacy Brief: Benefit Amounts</em></a>, November 2023.</p>
<p data-note_number='8'><a href="#_ref8" class="footnote-id-foot" id="_note8">8. </a> U.S. Department of Labor, Employment and Training Administration, “<a href="https://oui.doleta.gov/unemploy/ui_replacement_rates.asp">UI Replacement Rates Report</a>,” Benefit Accuracy Measurement data. Public data prepared by the OUI Division of Performance Management, Accessed August 7, 2026.</p>
<p data-note_number='9'><a href="#_ref9" class="footnote-id-foot" id="_note9">9. </a> Megan Rivera, <a href="https://equitablegrowth.org/why-unemployment-insurance-alone-is-not-enough-in-the-ai-era/"><em>Why Unemployment Insurance Alone Is Not Enough in the AI Era</em></a>, Washington Center for Equitable Growth, July 2026.</p>
<p data-note_number='10'><a href="#_ref10" class="footnote-id-foot" id="_note10">10. </a> Julie Su, Andrew Stettner, and Michele Evermore, <a href="https://tcf.org/content/report/our-unemployment-system-needs-modernizing-trump-is-doing-the-opposite/"><em>Our Unemployment System Needs Modernizing</em>.<em> Trump Is Doing the Opposit</em>e.</a> The Century Foundation, September 2025.</p>
<p data-note_number='11'><a href="#_ref11" class="footnote-id-foot" id="_note11">11. </a> Daphné Skandalis, Ioana Marinescu, and Maxim Massenkoff, <a href="https://equitablegrowth.org/working-papers/racial-inequality-in-the-u-s-unemployment-insurance-system/"><em>Racial Inequality in the U.S. Unemployment Insurance System</em></a>, Washington Center for Equitable Growth Working Paper, July 2023.</p>
<p data-note_number='12'><a href="#_ref12" class="footnote-id-foot" id="_note12">12. </a> Elira Kuka and Bryan A. Stuart, <a href="https://www.nber.org/papers/w29595"><em>Racial Inequality in Unemployment Insurance Receipt and Take-Up</em></a>, National Bureau of Economic Research Working Paper no. 29595, December 2021.</p>
<p data-note_number='13'><a href="#_ref13" class="footnote-id-foot" id="_note13">13. </a> Office of U.S. Senator Michael Bennet, “<a href="https://www.bennet.senate.gov/2025/07/16/bennet-wyden-beyer-reintroduce-bicameral-bill-to-overhaul-unemployment-insurance/">Bennet, Wyden, Beyer Reintroduce Bicameral Bill to Overhaul Unemployment Insurance</a>” (press release), July 16, 2025.</p>
<p data-note_number='14'><a href="#_ref14" class="footnote-id-foot" id="_note14">14. </a> Josh Bivens, Melissa Boteach, Rachel Deutsch, Francisco Díez, Rebecca Dixon, Brian Galle, Alix Gould-Werth, Nicole Marquez, Lily Roberts, Heidi Shierholz, William Spriggs, and Andrew Stettner, <a href="https://www.epi.org/publication/unemployment-insurance-reform/"><em>Reforming Unemployment Insurance: Stabilizing a System in Crisis and Laying the Foundation for Equity</em></a>, a joint report of the Center for American Progress, Center for Popular Democracy, Economic Policy Institute, Groundwork Collaborative, National Employment Law Project, National Women&#8217;s Law Center, and Washington Center for Equitable Growth, June 2021.</p>
<p data-note_number='15'><a href="#_ref15" class="footnote-id-foot" id="_note15">15. </a> National Academy of Social Insurance, Unemployment Insurance Task Force, <a href="https://www.nasi.org/wp-content/uploads/2024/07/Unemployment-Insurance-Task-Force-Final-Report.pdf"><em>U</em><em>nemployment Insurance Task Force Final Report</em></a>, July 2024.</p>
<p data-note_number='16'><a href="#_ref16" class="footnote-id-foot" id="_note16">16. </a> National Academy of Social Insurance, Unemployment Insurance Task Force, <a href="https://www.nasi.org/wp-content/uploads/2024/07/Unemployment-Insurance-Task-Force-Final-Report.pdf">U<em>nemployment Insurance Task Force Final Report</em></a>, July 2024.</p>
<p data-note_number='17'><a href="#_ref17" class="footnote-id-foot" id="_note17">17. </a> Julie Su, Andrew Stettner, and Michele Evermore, <a href="https://tcf.org/content/report/our-unemployment-system-needs-modernizing-trump-is-doing-the-opposite/"><em>Our Unemployment System Needs Modernizing</em>.<em> Trump Is Doing the Opposit</em>e.</a> The Century Foundation, September 2025.</p>
<p data-note_number='18'><a href="#_ref18" class="footnote-id-foot" id="_note18">18. </a> Michele Evermore, “<a href="https://tcf.org/content/commentary/theres-a-growing-bipartisan-consensus-to-fix-unemployment-insurance/">There&#8217;s a Growing Bipartisan Consensus to Fix Unemployment Insurance</a>” (commentary), The Century Foundation website, July 17, 2024.</p>
<p data-note_number='19'><a href="#_ref19" class="footnote-id-foot" id="_note19">19. </a> Michele Evermore, “<a href="https://tcf.org/content/commentary/theres-a-growing-bipartisan-consensus-to-fix-unemployment-insurance/">There&#8217;s a Growing Bipartisan Consensus to Fix Unemployment Insurance</a>” (commentary), The Century Foundation website, July 17, 2024.</p>
<p data-note_number='20'><a href="#_ref20" class="footnote-id-foot" id="_note20">20. </a> Ismael Cid-Martinez, Nina Mast, Margaret Poydock, and Valerie Wilson, <a href="https://www.epi.org/publication/misclassifying-workers-as-independent-contractors-is-costly-for-workers-and-social-insurance-systems/"><em>Misclassifying Workers as Independent Contractors Is Costly for Workers and Social Insurance Systems</em></a><em>, </em>Economic Policy Institute, April 2026.</p>
<p data-note_number='21'><a href="#_ref21" class="footnote-id-foot" id="_note21">21. </a> Julie Su, Andrew Stettner, and Michele Evermore, <a href="https://tcf.org/content/report/our-unemployment-system-needs-modernizing-trump-is-doing-the-opposite/"><em>Our Unemployment System Needs Modernizing</em>.<em> Trump Is Doing the Opposit</em>e.</a>&nbsp;The Century Foundation, September 2025.</p>
<p data-note_number='22'><a href="#_ref22" class="footnote-id-foot" id="_note22">22. </a> Adam Grundy and Lucie Alleyne, “<a href="https://www.census.gov/library/stories/2025/07/nes-gig-economy.html">Nonemployer Businesses Are Key Contributors to Gig Activities and Economic Growth Nationally</a>” (blog post), U.S. Census Bureau website, July 1, 2025.</p>
<p data-note_number='23'><a href="#_ref23" class="footnote-id-foot" id="_note23">23. </a> Michele Evermore, “<a href="https://tcf.org/content/commentary/theres-a-growing-bipartisan-consensus-to-fix-unemployment-insurance/">There&#8217;s a Growing Bipartisan Consensus to Fix Unemployment Insurance</a>” (commentary), The Century Foundation website, July 17, 2024</p>
<p data-note_number='24'><a href="#_ref24" class="footnote-id-foot" id="_note24">24. </a> Bureau of Labor Statistics, U.S. Department of Labor, “<a href="https://www.bls.gov/opub/ted/2023/most-unemployed-people-in-2022-did-not-apply-for-unemployment-insurance-benefits.htm">Most Unemployed People in 2022 Did Not Apply for Unemployment Insurance Benefits</a>” (blog post), <em>The Economics Daily (TED)</em>, April 14, 2023.</p>
<p data-note_number='25'><a href="#_ref25" class="footnote-id-foot" id="_note25">25. </a> Amy Traub, Alexander Hertel-Fernandez, and Sanjay Pinto, <a href="https://www.nelp.org/insights-research/the-unemployed-worker-study/"><em>The Unemployed Worker Study</em></a>, National Employment Law Project, April 2025.</p>
<p data-note_number='26'><a href="#_ref26" class="footnote-id-foot" id="_note26">26. </a> Alexander Hertel-Fernandez and Alix Gould-Werth, &#8220;<a href="https://www.journals.uchicago.edu/doi/abs/10.1086/739410">Employer Contestations and Access to Unemployment Insurance: Implications for Benefit Receipt, Material Hardship, and Administrative Burden</a>,&#8221; <em>Social Service Review</em> 100, no. 1 (March 2026): 38–68.</p>
<p data-note_number='27'><a href="#_ref27" class="footnote-id-foot" id="_note27">27. </a> Michele Evermore and David Balducchi, <a href="https://tcf.org/content/commentary/a-roadmap-for-unemployment-insurance-reform/"><em>A Roadmap for Unemployment Insurance Reform</em></a>, The Century Foundation, October 2023.</p>
<p data-note_number='28'><a href="#_ref28" class="footnote-id-foot" id="_note28">28. </a> Josh Bivens, Melissa Boteach, Rachel Deutsch, Francisco Díez, Rebecca Dixon, Brian Galle, Alix Gould-Werth, Nicole Marquez, Lily Roberts, Heidi Shierholz, William Spriggs, and Andrew Stettner, <a href="https://www.epi.org/publication/unemployment-insurance-reform/"><em>Reforming Unemployment Insurance: Stabilizing a System in Crisis and Laying the Foundation for Equity</em></a>, a joint report of the Center for American Progress, Center for Popular Democracy, Economic Policy Institute, Groundwork Collaborative, National Employment Law Project, National Women&#8217;s Law Center, and Washington Center for Equitable Growth, June 2021.</p>
<p data-note_number='29'><a href="#_ref29" class="footnote-id-foot" id="_note29">29. </a> Michele Evermore, <a href="https://www.nelp.org/app/uploads/2026/08/Policy-Brief-AI-is-New-The-Challenge-of-Mass-Unemployment-Isnt.pdf"><em>AI Is New. The Challenge of Mass Unemployment Isn&#8217;t</em>.</a> National Employment Law Project, September 2026.</p>
<p data-note_number='30'><a href="#_ref30" class="footnote-id-foot" id="_note30">30. </a> Stephen A. Wandner, <a href="https://doi.org/10.17848/9781441680389"><em>Solving the Reemployment Puzzle: From Research to Policy</em></a>, W.E. Upjohn Institute for Employment Research, January 2010.</p>
<p data-note_number='31'><a href="#_ref31" class="footnote-id-foot" id="_note31">31. </a> Michele Evermore, <a href="https://www.nelp.org/app/uploads/2026/08/Policy-Brief-AI-is-New-The-Challenge-of-Mass-Unemployment-Isnt.pdf"><em>AI Is New. The Challenge of Mass Unemployment Isn&#8217;t</em>.</a> National Employment Law Project, September 2026.</p>
<p data-note_number='32'><a href="#_ref32" class="footnote-id-foot" id="_note32">32. </a> Alexander Hertel-Fernandez, “<a href="https://peri.umass.edu/publication/solidarity-insurance-reimagining-unemployment-insurance-as-a-platform-for-worker-power/">Solidarity Insurance: Reimagining Unemployment Insurance as a Platform for Worker Power.</a>” (policy brief), Political Economy Research Institute, University of Massachusetts Amherst, July 2026.</p>
<p data-note_number='33'><a href="#_ref33" class="footnote-id-foot" id="_note33">33. </a> Bureau of Labor Statistics, U.S. Department of Labor, “<a href="https://www.bls.gov/news.release/archives/uisup_03292023.htm">Characteristics of Unemployment Insurance Applicants and Benefit Recipients—2022</a>” (economic news release), March 29, 2023.</p>
<p data-note_number='34'><a href="#_ref34" class="footnote-id-foot" id="_note34">34. </a> Bruce Western, <a href="https://press.princeton.edu/books/paperback/9780691010335/between-class-and-market"><em>Between Class and Market: Postwar Unionization in the Capitalist Democracies</em></a> (Princeton, N.J.: Princeton Univ. Press, 1997).&nbsp;</p>
<p data-note_number='35'><a href="#_ref35" class="footnote-id-foot" id="_note35">35. </a> Michele Evermore and Will Raderman, <a href="https://www.niskanencenter.org/wp-content/uploads/2026/01/The-case-for-enhanced-wage-records.pdf"><em>T</em></a><a href="https://www.niskanencenter.org/wp-content/uploads/2026/01/The-case-for-enhanced-wage-records.pdf"><em>he Case for Enhanced Wage Records: Lessons from States</em></a>, National Academy of Social Insurance and The Niskanen Center, January 2026.</p>
<p data-note_number='36'><a href="#_ref36" class="footnote-id-foot" id="_note36">36. </a> Michele Evermore, “<a href="https://tcf.org/content/commentary/theres-a-growing-bipartisan-consensus-to-fix-unemployment-insurance/">There&#8217;s a Growing Bipartisan Consensus to Fix Unemployment Insurance</a>” (commentary), The Century Foundation website, July 17, 2024.</p>
<p data-note_number='37'><a href="#_ref37" class="footnote-id-foot" id="_note37">37. </a> Katelin P. Isaacs and Julie M. Whittaker, “<a href="https://www.congress.gov/crs-product/IF10336">The Fundamentals of Unemployment Compensation</a>,” Congressional Research Service, IF10336, updated January 12, 2026.</p>
<p data-note_number='38'><a href="#_ref38" class="footnote-id-foot" id="_note38">38. </a> Jacob Leibenluft, <a href="https://equitablegrowth.org/research-paper/lessons-from-past-trade-adjustment-policies-to-support-displaced-workers-in-the-era-of-artificial-intelligence/"><em>Lessons from Past Trade Adjustment Policies to Support Displaced Workers in the Era of Artificial Intelligence</em></a><em>,</em> Washington Center for Equitable Growth, March 2026.</p>
<p data-note_number='39'><a href="#_ref39" class="footnote-id-foot" id="_note39">39. </a> Michele Evermore, “<a href="https://www.nelp.org/app/uploads/2020/05/NELP-Michele-Evermore-Testimony-Michigan-UI-Processing-5-21-2020.pdf">Michigan Unemployment Claims Processing</a>,” testimony on behalf of National Employment Law Project to a hearing before the Michigan Legislature, Joint Select Committee on the COVID-19 Pandemic, May 21, 2020.</p>
<p data-note_number='40'><a href="#_ref40" class="footnote-id-foot" id="_note40">40. </a> U.S. Government Accountability Office, <a href="https://www.gao.gov/products/gao-16-74"><em>Energy Employees Compensation: DOL Generally Followed Its Procedures to Process Claims but Could Strengthen Some Internal Controls</em></a>, GAO-16-74, March 2016.</p>
<p data-note_number='41'><a href="#_ref41" class="footnote-id-foot" id="_note41">41. </a> Jacob Leibenluft, <a href="https://equitablegrowth.org/research-paper/lessons-from-past-trade-adjustment-policies-to-support-displaced-workers-in-the-era-of-artificial-intelligence/"><em>Lessons from Past Trade Adjustment Policies to Support Displaced Workers in the Era of Artificial Intelligence</em></a><em>,</em> Washington Center for Equitable Growth, March 2026.</p>
<p data-note_number='42'><a href="#_ref42" class="footnote-id-foot" id="_note42">42. </a> Julie Kay, “<a href="https://www.wcnc.com/article/news/health/coronavirus/north-carolina-unemployment-benefits-still-delayed-five-years-after-pandemic/275-5b6093e8-1e0f-46bc-811a-d446166114a3">‘No Accountability’: Pandemic-Era Unemployment Benefit Delays Persist, 5 Years Later</a>,” <em>WCNC Charlotte</em>, April 3, 2025.</p>
<p data-note_number='43'><a href="#_ref43" class="footnote-id-foot" id="_note43">43. </a> Annelies Goger, Tracy Hadden Loh, and Nicole Bateman, <a href="https://www.brookings.edu/articles/debunking-myths-about-covid-19-reliefs-unemployment-insurance-on-steroids/"><em>Debunking Myths About COVID-19 Relief&#8217;s ‘Unemployment Insurance on Steroids,’</em></a> The Brookings Institution, May 2020.</p>
<p data-note_number='44'><a href="#_ref44" class="footnote-id-foot" id="_note44">44. </a> National Employment Law Project, <a href="https://www.nelp.org/insights-research/id-verification/"><em>ID Verification</em></a>, November 14, 2023.</p>
<p data-note_number='45'><a href="#_ref45" class="footnote-id-foot" id="_note45">45. </a> National Academy of Social Insurance, “<a href="https://waysandmeans.house.gov/wp-content/uploads/2026/03/Michele-Evermore-Testimony.pdf">Testimony of Michele Evermore, Senior Fellow, National Academy of </a><a href="https://waysandmeans.house.gov/wp-content/uploads/2026/03/Michele-Evermore-Testimony.pdf">Social Insurance</a>,” testimony before the U.S. House of Representatives, Committee on Ways and Means, Work and Welfare Subcommittee, March 5, 2026.&nbsp;</p>
<p data-note_number='46'><a href="#_ref46" class="footnote-id-foot" id="_note46">46. </a> Julie Su, Andrew Stettner, and Michele Evermore, <a href="https://tcf.org/content/report/our-unemployment-system-needs-modernizing-trump-is-doing-the-opposite/"><em>Our Unemployment System Needs Modernizing</em>.<em> Trump Is Doing the Opposit</em>e.</a> The Century Foundation, September 2025.</p>
<p data-note_number='47'><a href="#_ref47" class="footnote-id-foot" id="_note47">47. </a> Michele Evermore, <a href="https://www.nelp.org/app/uploads/2026/08/Policy-Brief-AI-is-New-The-Challenge-of-Mass-Unemployment-Isnt.pdf"><em>AI Is New. The Challenge of Mass Unemployment Isn&#8217;t</em>.</a> National Employment Law Project, September 2026.</p>
<p data-note_number='48'><a href="#_ref48" class="footnote-id-foot" id="_note48">48. </a> Emma Jelliffe, Paul Pangburn, Stefan Pichler, and Nicolas R. Ziebarth, “<a href="https://www.pnas.org/doi/full/10.1073/pnas.2107670118">Awareness and Use of (Emergency) Sick Leave: US Employees&#8217; Unaddressed Sick Leave Needs in a Global Pandemic</a>,” <em>Proceedings of the National Academy of Sciences</em> 118, no. 29 (July 20, 2021): e2107670118.</p>
<p data-note_number='49'><a href="#_ref49" class="footnote-id-foot" id="_note49">49. </a> U.S. Congress, Office of Technology Assessment, <a href="https://www.princeton.edu/~ota/disk2/1987/8730/873005.PDF"><em>Trade Adjustment Assistance: New Ideas for an Old Program</em></a>, 1987.</p>
<p data-note_number='50'><a href="#_ref50" class="footnote-id-foot" id="_note50">50. </a> U.S. Congress, Office of Technology Assessment, <a href="https://www.princeton.edu/~ota/disk2/1987/8730/873005.PDF"><em>Trade Adjustment Assistance: New Ideas for an Old Program</em></a>, 1987.</p>
<p data-note_number='51'><a href="#_ref51" class="footnote-id-foot" id="_note51">51. </a> Bipartisan Policy Center, “<a href="https://bipartisanpolicy.org/article/administrative-failures-plague-state-unemployment-insurance-programs/">Administrative Failures Plague State Unemployment Insurance Programs</a>,” July 1, 2020.</p>
<p data-note_number='52'><a href="#_ref52" class="footnote-id-foot" id="_note52">52. </a> Bipartisan Policy Center, “<a href="https://bipartisanpolicy.org/article/administrative-failures-plague-state-unemployment-insurance-programs/">Administrative Failures Plague State Unemployment Insurance Programs</a>,” July 1, 2020.</p>
<p data-note_number='53'><a href="#_ref53" class="footnote-id-foot" id="_note53">53. </a> Katelin P. Isaacs and Julie M. Whittaker, <a href="https://www.congress.gov/crs-product/R42444"><em>Emergency Unemployment Compensation (EUC08): Status of Benefits Prior to Expiration</em></a>, Library of Congress, Congressional Research Service, CRS Product Number R42444, August 11, 2014.</p>
<p data-note_number='54'><a href="#_ref54" class="footnote-id-foot" id="_note54">54. </a> U.S. House of Representatives, Ways and Means Committee Democrats, <a href="https://democrats-waysandmeans.house.gov/hot-topics/unemployment-insurance"><em>Unemployment Insurance</em></a> (fact sheet), 2014.</p>
]]></content:encoded>
											
	</item>
	
</channel>
</rss>
