This piece originally appeared on the Huffington Post.

 The recent defeat of its effort to unionize workers at Volkwagen’s Chattanooga, Tennessee factory was a crushing blow to the United Auto Workers, and a setback to the embattled U.S. labor movement, which could have used the morale boost of a high-profile victory.

It was also a big loss for the vast majority of Americans who must work for a living, whether they are union members or not. Without a large robust unionized sector, there is little hope that the relentless spread of low-wage work, job insecurity and economic inequality will be reversed.

Labor unions were key to the post-World War II social contract under which the benefits of economic growth were broadly shared. Collective bargaining agreements set industry-wide standards for wages and working conditions, which put pressure on nonunion firms to keep up or face union organizing drives. Politically, unions were the most important force supporting Social Security, Medicare, unemployment compensation, overtime, job safety, progressive taxes and full employment policies that promoted prosperity far beyond their membership.

Read more

This is a story about misdirection, how authors contort their analysis to answer a question no one is asking but make it seem as if they are answering an important current question, such as ‘why has income inequality increased?’. The consequence is to be grossly misleading or, worse, to present conclusions that are directly opposite of what one’s data show.

The paper in question is titled “US income inequality and assortative marriages” and written by Jeremy Greenwood, Nezih Guner, Georgi Kocharkov, and Cezar Santos for VoxEU.org. The research relates to an increase in positive assortative mating: “how likely a person is to marry someone of similar educational background. Since education is an important determinant of income, these patterns of matching might have an impact on the economy’s distribution of income.” Basically, if higher income men are now more likely to marry higher income women then household income inequality will grow.

The authors conclude that “rising assortative mating together with increasing labour-force participation by married women [emphasis added by me] are important in order to account for the determinants of growth in household income inequality in the US.” So, right out of the gate, a key influence not trumpeted in the headline (rising labour-force participation by married women) is introduced. But we’ll stick with the findings on assortative mating for now. The authors compare assortative mating in 2005 to that of 1960. The selection of the dates for comparison, 1960 and 2005, determines their story and they choose a misleading one. They show their key finding in the very first graph, presented below, which uses Kendall’s tau statistic to measure the relationship between husband’s and wife’s educational levels. The higher the tau statistic ‘the higher is the degree of positive assortative mating.”

Read more

I was glad to see the United Auto Workers (UAW) file objections with the National Labor Relations Board (NLRB) over the nasty campaign by anti-union Tennessee politicians to affect the results of the union election at Volkswagen last week. It would be so enlightening for the NLRB to question Sen. Corker (R-Tenn.) under oath about his alleged conversations with the “real decision makers” at VW, the supposed source of his threat/promise that voting in the UAW would doom the VW plant’s hopes for expansion. Was Corker lying, or were VW executives breaking their neutrality agreement with the UAW and using Corker to help defeat the union? If he was VW’s secret agent, the election should be set aside.

Read more

In a Wall Street Journal op-ed last month, Sylvester Schieber and Andrew Biggs said that census data failed to capture much of the income Americans derive from 401(k) and IRA plans. Though no one denies that the data don’t include some distributions from retirement accounts, the extent of the under-reporting is under dispute, with Biggs and Schieber claiming census data ignore 60% or more of the money that seniors receive.

In an earlier blog post, I noted that Schieber and Biggs may be basing this attention-grabbing claim on Internal Revenue Service measures that count rollovers from one retirement account to another as “income.” At an American Enterprise Institute event today, Paul Van de Water of the Center on Budget and Policy Priorities, who first brought the rollover issue to my attention, asked Schieber whether he based his claims on IRS measures of total or taxable income from retirement plans. Whereas taxable income measures exclude Roth IRA distributions, total income measures include these distributions, but also rollovers. As Federal Reserve and IRS researchers have explained, rollovers are included in some IRS income measures even if taxes aren’t owed on the amounts (and whether or not households report the transactions) because financial service providers report them to the IRS.

So, which measures are Schieber and Biggs using—total or taxable income from retirement plans? Schieber wasn’t able to answer, deferring to a coauthor, Billie Jean Miller, who wasn’t present. This should raise eyebrows, since Schieber has been making the same point for years and Social Security Administration researcher John Woods raised the rollover issue back in 1996.

Without claiming any familiarity with IRS data, it appears to me that Schieber and Miller are using measures of total income from these accounts, at least with respect to pension and annuity income. In Exhibit 5 of their Journal of Retirement article, for example, Schieber and Miller cite an IRS measure of pension and annuity income of $812 billion for 2008. According to summary statistics published by the IRS, total pension and annuity income was $845 billion in 2008, whereas taxable pension and annuity income was $506 billion (all figures have been rounded). Though Schieber and Miller’s figure is somewhat smaller than the IRS measure of total pension and annuity income, the difference could be due to revisions or differences between the internal IRS data and micro-data made available for public use.

Over the last six years, the labor force participation rate dropped by several percentage points. There is a debate over how much of that drop is a direct result of the lack of job opportunities in the Great Recession and its aftermath (changes that are generally labeled cyclical), and how much is instead a result of long-run trends, such as baby boomers beginning to retire (changes that are generally labeled structural). A recent blog post in the Wall Street Journal said that among Federal Reserve officials, the view that much of the decline is structural is gaining traction. If true, that’s a problem. My read of the data shows that most of the decline is cyclical, so if the Fed believes it’s structural, it means they believe there’s less slack in the economy than there really is.

Part of the misunderstanding is that there are two components of structural change. First, there are population shifts. Age groups that tend to have lower labor force participation rates are now a larger share of the population (think retiring boomers). These are called “compositional” shifts. Accounting for purely compositional changes by gender and age, more than 40 percent of the decline in the labor force participation rate over the last six years can be accounted for. Many people doing a quick analysis on this topic tend to stop there.

However, the other component of structural change is made up of long-run trends in labor force participation within age/gender groups. The labor force participation rate among people under age 25 has been declining since the 1980s, in part due to increasing college and university enrollment. The continuation of that long-run trend accounts for an additional structural decline in the overall labor force participation rate over the last six years. The projected trend in labor force participation rate of workers age 25-54 was virtually flat, so that trend did not meaningfully contribute to structural changes over the last six years. The trend labor force participation of workers age 55+, however, was expected to rise significantly over this period, particularly for women, as cohorts with much stronger labor force attachment throughout their 20’s, 30’s, and 40’s than the cohorts that preceded them began aging into this age bracket. In other words, that’s a structural change that should have substantially contributed to an increase in labor force participation over this period.

Putting these factors together—the compositional changes between gender/age groups and the structural trends within gender/age groups—the result is that only around a quarter of the total drop in labor force participation over the last six years is structural. This means that around 75 percent of it is cyclical. In other words, there are now around 5.8 million workers who are not in the labor force but who would be if job opportunities were strong. If these workers were in the labor market looking for work, the unemployment rate right now would be 10.0 percent instead of 6.6 percent. That is a lot of additional slack in the labor market.

It has been pointed out that it is likely that at least some of the workers who are out of the labor force due to cyclical factors are people who gave up and decided to retire early. Given that retirees are less likely to reenter the labor force when job opportunities improve, improving economic conditions may not draw these workers back in. This means that labeling them as being out of the labor force due to cyclical factors may not be very useful. However, it is important to note that there are large participation gaps for workers age 54 or younger, who are unlikely to be early retirees. In fact, more than 70 percent of the 5.8 million missing workers are under age 55. These missing workers under age 55—4.2 million of them—are therefore unlikely to be deterred from entering or re-entering the labor force when job opportunities strengthen.

While much of the reaction to the most recent CBO Budget and Economic Outlook (released earlier this month) focused on the labor market impacts of the Affordable Care Act, it’s important to note that this report actually contained a multitude of interesting findings and updated projections. Among the most important is CBO’s revised projection of the costs of federal health care programs—Medicare, Medicaid, ACA subsidies, and some smaller programs as well—over the next decade. For the fourth year in a row, CBO revised these cost projections downward. The figure below shows CBO’s projections of these costs in the decade following each Budget and Economic Outlook published since 2011.

While health care costs remain the fastest growing portion of the federal budget, and are still projected by CBO to grow significantly faster than the overall economy over the next decade, the downward revisions of the past three years are quite significant. Put simply, since 2011, CBO’s projection of what the level of federal health spending will be in 2021 has dropped by $183 billion, or about 10.4 percent.

To put this in perspective, when lawmakers passed the counterproductive, indiscriminate sequestration spending cuts as part of 2011’s Budget Control Act (BCA), they were looking at projections of federal health spending over the following decade that the latest estimates indicate were $900 billion too high. The $1.2 trillion in “sequestration” cuts over the decade, with their damaging effect to public investment and to the still-incomplete economic recovery, look even more unnecessary in this light.

Genuinely informed budget wonks know that the BCA cuts were particularly perverse because they took an ax to the portion of the budget— discretionary spending—that is not projected to grow in coming years, relative to the economy. To the extent that long-run budget projections highlight a need for restraining spending, this is entirely driven by the rapid rise in health care costs—both public and private. And these health care costs have rapidly decelerated in recent years. This deceleration began even before the provisions of the Affordable Care Act (ACA) meant to restrain cost-growth went into effect. Health care costs are so important in driving long-run budget trends that if the cost-growth slowdown of the past five years continues, there will be no long-run budget deficit problem. Lawmakers who actually cared about long-run budget deficits, not to mention living standards of typical Americans, would reverse the damaging cuts to discretionary spending and instead continue to press for efficiencies that further slow health care cost growth.

Mark Price and Estelle Sommeiller’s new paper traces the trajectory of top incomes in American states and regions from 1917 through 2011. Mapping this data across the continental United States and over the last century suggests both important similarities across states, and some key differences.

On the map below, the states tip from green to red when the top 10 percent’s share of income exceeds one-third. In the early years, inequality (as measured by the high income share of top earners), is starkest in the Northeast. This inequality is generalized by the impact of depression and war in the 1930s and 1940s, but once the policy innovations of the New Deal (collective bargaining, retirement security, labor standards, and financial regulation) take hold, inequality eases: by the middle 1950s, only New York and the Deep South are still colored red.

The pattern across the last generation is just as telling. Of the sixteen states to top this threshold in 1972, the only ones outside the South were the tri-state home of big finance—New York, New Jersey, and Connecticut. As we scroll forward from there, the states in which the top 10 percent claim less than a third of total income gradually diminish, disappearing entirely by 1989. By 2011, the top 10 percent are claiming almost 60 percent of income in New York and Connecticut, and over 40 percent in all but three states (Iowa, Nebraska, and South Dakota).

2014 is shaping up to be a much better year for Social Security than any in the recent past. People seem to finally be recognizing that Social Security is the one leg of the retirement stool that’s working well and providing genuine, much-needed retirement security. And they’re realizing that given this, kicking away at it doesn’t make a lot of sense.

Today provides another welcome bit of news in this regard, with the Obama administration announcing that the cut to the Social Security cost of living adjustment (COLA) that was in their fiscal 2014 budget proposal will not show up again this year.

This is good news. The oft-repeated claim that using the “chained” consumer price index for urban consumers (C-CPI-U) was simply a technical improvement to the Social Security COLA was always flat-wrong. Given that Social Security is the bedrock pillar of retirement income security for so many Americans, paring benefits back is a terrible idea.

Politically, this announcement amounts to yet another confirmation that the pursuit of a budget “Grand Bargain” is dead. Even the best versions of this bargain—near-term stimulus to boost the recovery combined with a mix of tax increases and cuts to Social Security, Medicare, and Medicaid to reduce projected long-run budget deficits—were pretty bad (after all, why would boosting recovery and lowering unemployment be something that only one party cared about and would have to sacrifice something else for?). But given that Republicans in Congress refused to accept stimulus (and instead have demanded, and gotten, extreme anti-stimulus in recent years) or any tax increases, this left only cuts to the social insurance programs that have provided the majority of income growth for low- and moderate-income households over the past generation. Even more bizarrely, these same Republicans refused to actually support any specific cut to these programs and would even attack the president for offering them.

In a nutshell, cuts to Social Security, including the adoption of the C-CPI-U to cut the COLA, are bad policy and bad politics. This makes dropping it from the president’s latest budget a wise move.

So, efforts to cut Social Security are in retreat. Even more hopefully, the case for expanding it has gone from something only liberal bloggers would dare write about to something that a handful of U.S. Senators have started discussing.

Yes, the public debate on Social Security looks to be getting a lot smarter these days, which is awfully welcome.

Tuesday’s CBO report on the effects of increasing the minimum wage has generated a lot of discussion. While some of the CBO’s findings are consistent with our own analysis, we have some serious disagreements. Here’s our take on the report, particularly CBO’s estimates on employment and income (we focus on their estimates of the effects of increasing the minimum wage to $10.10 by 2016).

The report finds that 16.5 million workers who make below $10.10 would get a raise, and an additional 8 million workers who make slightly above $10.10 would also likely get a bump (since employers like to preserve internal wage ladders). This is right in line with our estimates of the likely impact.

They found that the increase in the minimum wage would benefit mostly adults who need the earnings from their minimum wage job to make ends meet: less than 12 percent of the people who would get a raise are under age 20 and more than 70 percent of the total earnings would go to workers in families whose income is less than three times the poverty threshold. For context, in 2013, three times the poverty threshold for a family of three was around $55,700. This too is right in line with our analysis.

CBO also found that 900,000 people would be lifted out of poverty. We agree that raising the minimum will lift a significant number of people out of poverty, and if anything, CBO’s estimate here seems conservative. CBO is a bit vague on how they came to their conclusion about the effect on poverty levels, but from what we can tell, it seems that they looked at current income levels, expected poverty levels in 2016, simulated how peoples’ incomes would change following the minimum wage hike, and estimated the change in the number of people in poverty. This is a perfectly reasonable approach; however, there’s a good body of research that has looked at the real-world experience of how minimum wage hikes have affected poverty levels. A recent paper by Arin Dube looks specifically at this question and estimates that in the past, for every 10 percent increase in the minimum wage, we’ve seen a 2.4 percent decrease in the number of people in poverty. This implies that increasing the minimum wage to $10.10 could reduce the number of people in poverty by as much as 4.5 million.

Read more

In The Increasingly Unequal States of America: Income Inequality by State, Mark Price and Estelle Sommeiller develop estimates for top income shares, from 1917 through 2011, for American states and regions. The national version of this story is now quite familiar: the iconic Piketty and Saez curve that starts high in the early years of the twentieth century (with the top 10 percent claiming about half of all income and the top 1 percent claiming about one-fifth), drops with the political innovations of the New Deal, and then climbs again—returning, by 2012, to its early-century heights—as those innovations are dismantled. So what does the state and regional breakdown tell us?

The full results are plotted below. For each year, the states (the light blue dots) are plotted by top income shares. The middle of the shaded box marks the median state (half are lower on the chosen measure, half are higher), and then the states are broken into quartiles: the middle quartiles (marked by the boxes) surround the median, the outer quartiles run along the lines between the boxes and the outer tick marks.

A few patterns stand out. First, the general sweep of the graph echoes the national story. The arc of inequality from Gilded Age to New Deal and back again is experienced across every state, not just in a few of them. Second, the policy innovations that dampened inequality—collective bargaining, retirement and unemployment security, labor standards, financial regulation, progressive taxation—also narrowed the variation across states. In the middle years of the last century strong federal policies trumped (or overcame) the economic and political differences among the states.  And third, the erosion of those policies, beginning in the 1970s, saw both inequality and the variation across states widen again.