THE CASE FOR A PROSPERITY DIVIDEND
by Eileen Appelbaum and Richard B. Freeman
The economic boom of the 1990s has re-established the United States as the leading world economy and has brought the benefits of economic progress to virtually all citizens. It has increased employment, raised the real wages of workers across the board, including low-skill and minority workers, enabled the country to reduce welfare rolls, contributed to the fall in crime, and brought the reality of opportunity to all. It is imperative for the well-being of the American people that this economic boom be maintained.
But in early 2001 there are troubling signs that the boom is running out of steam just as the new Bush Administration and Congress begin their task of governing the country. The meltdown of many dotcom firms, signs of bankruptcies in the old economy, slower job growth, layoffs, and a rise in unemployment have shaken consumer confidence and dimmed the prospects for sustaining the boom. State sales tax revenues are declining precipitously, threatening education initiatives and other state programs. In the eyes of Federal Reserve Board Chairman Alan Greenspan, “downside risks predominate,” so that the economy is susceptible to a recession.
The economy needs a stimulus to give it a boost and to get it back on track. In any case, a stimulus will do no harm: except for the volatile energy and food sectors, prices are flat or falling, and inflation is dormant. And with the economy slowing, consumer spending can increase without causing interest rates to rise. Moreover, the country can afford a stimulus. At this writing, the 2001 surplus is projected to reach $281 billion, and the on-budget surplus (which excludes Social Security and Medicare trust fund revenue) is expected to reach $125 billion; in the meantime, the national debt has fallen below 35% of gross domestic product. Put another way, the growing federal government surplus is drawing money from the economy just when the nation needs a burst of spending to avoid a downturn.
Interest rate cuts are not enough
To avoid a serious recession the Federal Reserve has twice lowered interest rates. But reduced interest rates by themselves can do little to prevent a recession since they take so long – eight months to a year – to affect spending. There are a number of reasons why this is so. With regard to consumer spending, credit card interest rates respond sluggishly to changes in the prime rate (Calem and Mester 1995). Also, investment in housing has lost much of its traditional sensitivity to interest rates as a result of financial market deregulation and innovations in mortgage finance such as adjustable rate mortgages. Instead of interest rates, changes in consumer confidence have emerged as a more important factor affecting residential investment (Dudley, McKelvey, Youngdahl, and Hatzius 1999).
As for business investment, it is more responsive to internal cash flow and expected sales than to interest rates. This is especially true of firms that are growing at a moderate or rapid rate. In a slowing economy, when cash flow may be constrained, access to finance plays a more important role in firms’ investment decisions (Fazzarri et al. 1988; Gertzler and Hubbard 1989). Interest rate policy can help avoid a credit crunch and increase access to finance, but the effect will not be felt immediately. Furthermore, the major area of investment today, information technology (IT), is suffering from sector-specific problems that may make the expansive effect of interest rates weaker than usual. Thus, if the economy falls into a major recession, interest rate cuts are unlikely to get it out of trouble, even if Mr. Greenspan times them ideally and gives them his inimitable blessing.
Relying solely on interest rates to avoid recession also runs into another problem. Perhaps the greatest weakness of the U.S. boom has been the rise of consumer and corporate indebtedness to record levels (Maki 2000). The high consumer debt ratio raises the danger that a downturn now may have greater adverse effects on spending than earlier downturns because consumers have run out of credit and savings to fund consumption spending during periods of unemployment (Godley 2000, among others, has made this argument). Reductions in interest rates that induce people and firms to go further into debt may have adverse effects on economic stability, if not immediately then in the near future, possibly making the next downturn more severe.1
Cuts in interest rates are indeed part of any solution to potential economic slowdown, but they are unlikely by themselves to be sufficient medicine to cure the economy.
The administration’s tax plan is not an effective stimulus
The Bush Administration has offered a tax plan that will substantially lower the taxes of the wealthiest Americans over the next decade but have only a modest impact on the tax burden of less-advantaged citizens. The president argues that his proposed tax cut will stimulate the economy, but, as William Gale of the Brookings Institution observed (2001, Executive Summary), “the President’s tax proposal is incredibly poorly designed for that purpose.”
The primary reason that Bush’s proposed tax cut fails as a stimulus is that its implementation is delayed. It would not go into effect until next year, and then the cuts are phased in slowly – to the point where it would take another five years before they are big enough to be felt by the average household. Clearly, then, the tax cuts will not arrive in time to avert a recession.
Even if Congress makes the tax cut retroactive, the initial tax savings will be far too small to have an effect on the $10 trillion U.S. economy. The first-year tax cut is only $20 billion, and even if households spent the entire amount, the resulting rise in consumer spending would be just 0.3%. Furthermore, the first-year tax cuts would be tiny for typical families. A recent study by the accounting firm Deloitte and Touche, analyzing the effects of the Bush proposal on families that pay income tax, found that a single wage earner with no children earning $62,000 would get just $160 in tax savings in the first year of the Bush plan. A two-earner family with no children earning $38,000 would get $144. A two-earner family with two children earning $70,000 would get $410 (see Table 1 for details). These tax cuts are too small to make much of a difference – either to the families that receive them or to the economy.
And the Bush tax cuts leave lots of people behind. According to researchers at the Center on Budget and Policy Priorities (cited in Gale 2001), the 50 million families with incomes below $26,000 would see their after-tax incomes rise by trivial amounts. Even the expanded child tax credit in the president’s proposal provides little help to these families with the expenses of raising their kids. And about one-third of all children live in families that won’t qualify for the credit at all.
Many hard-working families pay more in payroll taxes than they do in income taxes. For the families in Table 1, payroll taxes amount to $3,000-5,000 a year. But President Bush proposes cutting only income t
axes. As a result, many families will see little change in the total amount of federal taxes they pay – even after the Bush tax plan is fully phased in. Citizens for Tax Justice analyzed the tax plan and found that households in the lowest 20% of the income distribution would receive an average tax cut of just $42; those in the second 20% would get just $187. The president has been photographed with families identified as middle class who will receive $1,600, and perhaps these families will. But 90% of families will receive less than $1,600 from Bush’s tax cut when it is fully phased in, and 27% will get no tax cut at all (cited in Gale 2001).
The benefits of the Bush tax cut go overwhelmingly to high-income households. Under his plan, the richest 20% of households will receive 71% of the tax cuts, and the cut for the richest 1% of households will average more than $39,000. But, as recent research by the Federal Reserve Board shows, high-income households tend to save at much higher rates than other households (Dynan, Skinner, and Zeldes 2000). Thus, tilting the tax cuts to the wealthy is not an effective way to stimulate the economy. It is the middle class and working families who are financially strapped and whose jobs and financial security are threatened by a slowdown. These are the households that can be counted on to spend any tax cut they receive-and it is this spending that can revitalize the economy.
Thus, since President Bush’s proposed tax reductions will take time to appear in people’s pockets and be spent, and because they are targeted to the well-to-do who are less likely to spend than are middle-income or poorer citizens, these potential cuts are unlikely to affect spending behavior with sufficient speed to cure the economy.
Ten years out, however, the tax cuts in the Bush plan really mount up. The cumulative tax cuts in the president’s proposal are large – estimated by the administration at $1.6 trillion and by other tax experts at more than $2 trillion. They risk undoing the successful efforts of the previous Bush and Clinton Administrations’ reduction in federal deficits. And they threaten the nation’s ability to undertake new initiatives and spend on priorities such as a prescription drug plan under Medicare, education, or access to health insurance that Americans say they want. Even worse, when the tax cuts for the wealthy really kick in it may be necessary for the government to make deep cuts in existing programs that people value. Tax cuts this large tie the country’s hands and leave little flexibility for dealing with future exigencies.
Finally, given the likely battle over the type of tax reform and its benefits, the government will find it hard to fine-tune taxes to address possible sudden changes in the economic situation. Appropriately designed, permanent tax cuts can make sense on equity and efficiency grounds and may be a way to augment the disposable income of all citizens. But they are not well-suited to prevent recession in the short term.
What, then, might the U.S. do to try to maintain the longest economic boom in its history?
A prosperity dividend can boost the economy
There is a sensible addition to the country’s anti-recession arsenal that would provide a greater short-run “booster shot” for the economy in its current state than interest rate cuts or a permanent tax cut, while avoiding some of the problems inherent in those anti-recession strategies: the federal government should pay an immediate prosperity dividend to all citizens from the $125 billion budget surplus expected this year.
The budget surplus is largely the result of the economic boom. It is America’s return on 10 years of good economic performance, to which everyone in America has contributed. To fight off a potential recession, the government should declare a dividend and send a tax-free check to every permanent resident of the United States – man, woman, and child. This one-time payment would give a quick lift to the economy by spurring consumer spending.
A prosperity dividend that transfers the $125 billion surplus from the government to the private sector could be as large as $440 per person. This amount, equal to 1.8% of consumer spending or 1.2 % of GDP, is large enough to stimulate the country’s $10 trillion economy substantially.
Studies of the reaction of household spending to tax refunds or changes in tax rates suggest that a substantial part of the prosperity dividend would be spent soon after it is received (Parker 1999; Souleles 1999). Evidence shows that consumption spending is sensitive to changes in current income (Hall and Mishkin 1982; Carroll and Summers 1991) and that consumer spending changes with a household’s cash flow (Hatsopoulos, Krugman, and Poterba 1989).2
To be sure, inflows that have a sunset provision do not get factored into permanent income. Rather, they are windfall gains. At one time, economists concerned with the theory of consumption argued over whether windfall gains – transitory income – had any effect on consumption spending. From the point of view of economic stimulus, much of this argument is beside the point, for the debate focused over a narrow definition of consumption spending that excluded purchases of durable goods. And most economic studies find that one-time payments have a substantial impact on consumer spending (Bodkin 1959; Bird and Bodkin 1965; Lee 1975), although this impact is usually smaller than a comparable change in permanent income.3 This has been true for the tax surcharge of 1968 (Okun 1977, 1971); the tax rebate of 1975 (Poterba 1988); income tax refunds in general (Souleles 1999); and first-time changes in Social Security taxes (Parker 1999). A study of the Alaska dividend, through which the state distributes the economic rewards of its ownership of land and mineral resources, shows that households increase their spending in response to this anticipated annual income receipt, though they space expenditures from the anticipated dividend fairly evenly over the year (Hsieh 2000).
It is possible that some consumers will use the prosperity dividend to reduce their debt or increase their financial assets. However, a recent analysis of household consumption and income tax refunds (Souleles 1999) finds that households facing liquidity constraints – younger households, households with high debts and few liquid assets, and households with low income – quickly increase their spending on food and nondurable goods. The wealthiest households quickly increase their spending on durable goods and may later use part of their refund check on summer trips. The main finding with respect to the marginal propensity to consume out of tax refunds is that, across all households, “almost two-thirds of every extra dollar of refund is spent within a quarter” (Souleles 1999, 955). There is significant spending on cars as well as on other durable goods in the period immediately following the refund.
Even if consumers spend 70% of their prosperity dividend checks in a short time period – about what they spend out of income tax refunds (Souleles 1999) – consumer spending would increase by 1.2%, giving an immediate lift to GDP of 0.9%, a substantial stimulus this year.4
The $440-per-person figure is a reasonably sized number to give the U.S. economy a “booster shot.” But the number is not etched in stone. The precise amount needed to secure the continuing boom will differ depending on how substantial the slowdown appears, how responsive the economy turns out to be to the current round of interest rate cuts, and
other factors that affect the extent of the potential downturn. The amount can be readily matched to the perceived need. Recently, the AFL-CIO leadership suggested a $400 prosperity dividend; other groups might prefer larger or smaller numbers, but there is a clear range that makes economic sense. The $2,000 or so that Alaska gives to its citizens from its Permanent Fund would be far too much, while a dividend of $85 or so – equivalent to the tax savings in the first year of the Bush Administration’s tax plan – would be too little. The key point is that a prosperity dividend will have an immediate impact on consumers who have run out of consumer credit and will thus boost consumer confidence and spending just when such a boost is needed.
The prosperity dividend is a new “out of the box” solution, but it is one well-suited for dealing with the unprecedented U.S. economic situation of a large surplus and a slowing economy. The prosperity dividend is workable, and it will affect spending faster than will lowering interest rates or legislating a tax cut. Evidence from past tax rebates and related spending programs indicate that it will stimulate spending. And by giving equal amounts to all persons, the prosperity dividend will spur spending more than an otherwise comparable proportionate tax rebate that gives the bulk of the rebate to the wealthiest Americans. Finally, because it is a once-and-for-all payment that does not tie the government’s hands for the future, a prosperity dividend can complement any particular longer-term program of lowering tax rates, reducing the national debt, shoring up Social Security, providing a prescription drug plan, or investing in education.
The prosperity dividend is workable
As the Alaska experience has shown, distributing a dividend check from the government to the people from the cumulated budget surplus is administratively feasible. In 2000, the state sent a check for nearly $2,000 to every man, woman, and child who was a permanent resident and had lived in the state for all of 1999. The Alaska dividend is administratively simple, and a federal prosperity dividend could easily be modeled on Alaska’s Permanent Fund Dividend.
To qualify for the prosperity dividend, a resident of the U.S. would have to be able to answer yes to all of the following statements.
- I was a resident of the U.S. during all of calendar year 2000;
- On the date I applied for the prosperity dividend, I had the intent to remain a resident of the U.S. indefinitely;
- I have not claimed residency in any other country at any time since December 31, 1999;
- If absent from the U.S. for more than 180 days, I was absent on an allowable absence (attending college, serving in the U.S. Armed Forces, receiving medical treatment, employed on a vessel of the U.S. Merchant Marine, required by employer, etc.);
- I was physically present in the U.S. for at least 72 consecutive hours at some time during 1999 or 2000.
The application form would be short. It would require name, address, telephone number, date of birth, Social Security number, signature, the names / addresses / telephone numbers of two adult residents of the U.S. who can verify the applicant’s residency, and a statement that the applicant understands that a deliberate misrepresentation or reckless disregard of facts on the application may result in a civil penalty and a fine of up to $5,000.
If the program dropped the requirement of permanent residency and instead sent the prosperity dividend to anyone currently living in the U.S. who worked here last year or resided here last year, then a simpler procedure could be established for those who paid either payroll or income taxes. For households that paid income taxes, the check could reflect the number of people in the household. Those who paid or filed neither payroll nor income tax would have to submit an application. Those who paid only payroll taxes would have to submit applications for non-working household members.
The prosperity dividend can be timed
Government efforts to stimulate an economy through spending programs like infrastructure investment can fail because the government is unable to time the stimulus correctly to deal with the problem. Spending programs like this require detailed plans, the gathering of human and physical resources, and so on. A program announced in a given year may spend much of its money two or three years later. Over the long run, some government efforts to stimulate the economy can fail because consumers may recognize that these efforts won’t necessarily raise the output of the economy and thus must be paid for in the future, and consumers will adjust their decisions accordingly (Barro 1989). As a result, many economists are leery of making massive changes in taxes or spending to stimulate an economy.
These problems would not arise with the prosperity dividend. The dividend will quickly transfer money from the government to citizens, many of whom risk facing serious liquidity constraints on their ability to spend. It is designed to raise spending and confidence in this critical period when the economy teeters on recession, not to alter long-run fundamentals. As a one-shot, temporary anti-recession tool the dividend would expire once the risk of recession subsided.
The prosperity dividend will increase domestic production
An immediate boost to the U.S. economy requires spending on domestic production, not on goods and services from other countries. It is plausible that the prosperity dividend (or any other scheme that puts money in the hands of consumers) will fail to stimulate the economy because consumers may spend the money largely on foreign imports. Imports of goods are about 12% of GDP, while imports of services are about 3% (Mann 1999), and so it is certainly possible that part of the stimulus will simply leak out of the U.S. economy. Studies of the relationship between GDP growth and growth in imports have found that increases in GDP typically increase spending on imports proportionately more than spending on domestic goods (see Mann 1999, Chapter 8, for a review of these estimates). So, even though most of the stimulus will be spent on domestic goods and services, the effect will be reduced by spending on imports.5
However, the nature of the prosperity dividend suggests that spending on imports is likely to be less of a problem with this type of stimulus than it would be with other types of increases in household income. Younger and poorer households – who will receive their proportionate share from the prosperity dividend – will increase their consumption of food and their expenditures on services in the quarter after they receive the check (Souleles 1999) compared with the richest households, who will spend it on consumer durables. Demand for cars and other durable goods will increase, though some of it will be satisfied with imports. Nevertheless, there will be a significant positive impact on hard-hit parts of the manufacturing economy, such as autos and steel, where thousands of workers have already lost their jobs. The fact that manufacturing has shed a large number of jobs in the last four months implies that the U.S. has the capacity to respond to the sudden shock of a $100-125 billion increase in consumer spending, i.e., firms have simply to reverse their previous layoff decisions.
In terms of impacts on the balance of payments, the Fed’s reductions in interest rates should help alleviate problems. Such reductions ought to lower the value of the dollar and increase
foreign spending on U.S. goods and reduce U.S. spending on foreign goods. The prosperity dividend thus will work better as the Federal Reserve reduces interest rates.
A prosperity dividend is fair
Economists often shy away from fairness considerations, but it is difficult to shy away from fairness when the economic well-being of so many Americans is at stake. After-tax income of the top 1% of households nearly doubled between 1977 and 1999, while for the bottom half of households after-tax income fell (Mishel, Bernstein, and Schmitt 2000, 58). Not until the economic boom of the late 1990s did the benefits of economic growth finally reach the majority of American workers. At a minimum, then, fairness would seem to dictate that prudent steps should be taken to maintain the prosperity that has brought gains to all without exacerbating disparities further.
Some have taken another view of fairness and argue that the benefits of the surplus should be proportionate to income taxes paid, so that the wealthiest, who pay the most in taxes, would get the largest share of any surplus distributed to the population. But while some people may not owe much income tax, other taxes fall more heavily on less-affluent households than on the best off. Everyone pays taxes. Everyone who works pays payroll taxes, for example – a tax that falls more lightly on the top 20% of income earners, especially the top 1%, than on the rest of Americans. Even those who don’t work pay excise taxes on gasoline and utilities, and they frequently pay sales taxes or property taxes. And even with a flat per-person payment, a tax-free prosperity dividend would have greater monetary value to the wealthy, who pay a higher marginal tax rate, than to other citizens. A $440 tax-free dividend to Bill Gates would be worth $700 in taxable income to him but just $440 to someone who pays nothing in income taxes. Some may regard this as unfair to the poor, who surely need a larger check from the government than the wealthy.
A stimulus that distributes a prosperity dividend equally to everyone living permanently in the U.S. treats us all equally, yet it would mean the most to families with young children. In the case of a single mother with two young children, the family would receive a total of $1,320 in checks from the government; a married couple with two children would receive $1,760. These payments would raise the incomes of families with low incomes proportionately more than they would raise the income of families with higher incomes. And these low-income families will make the stimulus work by spending all or most of their prosperity dividend checks on goods and services.
Conclusion: Boost the economy with a prosperity dividend now
In 2001 the U.S. economy faces unusual circumstances. The combination of the extended boom, high consumer and business debts, a huge federal surplus, and the danger of a substantial recession have not been experienced before. The prosperity dividend seems well designed for this environment. It is likely to succeed in stimulating the economy, and it meets the test of past experience.
There is no economic or policy reason to tie a critical stabilization package designed to deal with a potential cyclical downturn to a massive redistribution of income to the wealthy via the tax system. A tax package that benefits the wealthiest will be less effective and more risky than a prosperity dividend that treats all citizens equally. The immediate problem of stimulating the economy in the short run needs to be separated from the problem of changing the level of taxes or spending that will affect the long-run state of the economy and the distribution of economic rewards.
A prosperity dividend of $440 tax-free from the surplus will get money into the hands of consumers faster than a tax cut; will increase spending more than a tax cut because the money will go equally to all permanent residents, rather than mostly to high-income taxpayers; and will provide the temporary stimulus the economy needs to maintain full-employment prosperity. Once the threat of recession is past and the economy is growing and producing surpluses, the country can again weigh the alternative uses for this economic bounty.
1. At the same time, interest rate cuts will lower the debt servicing of many existing loans, which should reduce the risk of bankruptcy. [ return to text ]
2. Laibson (1997) has developed a model of consumer behavior that explains why this should be so. His theory suggests that households will spend most of the prosperity dividend. [return to text ]
3. In a study with a small number of observations for Israelis, Kreinen (1961) found a small response of consumption spending to a lump sum restitution of payments from Germany. [ return to text ]
4. Given the high levels of household indebtedness, some might use the money to retire consumer debt, a move that in turn would enable them to spend more in the future. [return to text ]
5. In addition, spending increases could lead to an increase in the interest rate and a rise in the dollar, which would depress exports. However, the weak economy and the Fed’s policy to reduce interest rates makes it unlikely that the dollar will rise. [return to text ]
Barro, Robert. 1989. “The Ricardian Approach to Budget Deficits.” Journal of Economic Perspectives, 3 (Spring): 37-54.
Bird, Roger C., and Ronald G. Bodkin. 1965. “The National Service Life-Insurance Dividend of 1950 and Consumption: A Further Test of the Strict Permanent Income Hypothesis.” Journal of Political Economy, 73 (October): 499-515.
Blinder, Alan. 1981. “Temporary Income Taxes and Consumer Spending.” Journal of Political Economy, 89 (February): 26-53.
Bodkin, Ronald. 1959. “Windfall Income and Consumption.” American Economic Review, 49 (September): 602-14.
Calem, Paul S., and Loretta Mester. 1995. “Consumer Behavior and the Stickiness of Credit-Card Interest Rates.” American Economic Review, 85 (December): 1327-36.
Carroll, Christopher D., and Lawrence H. Summers. 1991. “Consumption Growth Parallels Income Growth: Some New Evidence.” In B. Douglas Bernheim and John Shoven, eds., National Saving and Economic Performance. Chicago, Ill: University of Chicago Press.
Dudley, Bill, Ed McKelvey, John Youngdahl, and Jan Hatzius. 1999. “Housing Activity Down but Not Out.” Goldman Sachs U.S. Economics Analyst, October 15.
Dynan, Karen E., Jonathan Skinner, and Stephen P. Zeldes. 2000. “Do the Rich Save More?” Working Paper No. 2000-52. Washington, D.C.: Federal Reserve Board.
Fazzari, Steven M., R. Glenn Hubbard, and Bruce C. Petersen. 1988. “Financing Constraints and Corporate Investment.” Brookings Papers on Economic Activity. Washington, D.C.: Brookings Institution.
Gale, William G. 2001. “The Budget Outlook and Tax Policy.” Testimony before the Committee on the Budget, United States Senate, February 13, 2001. Washington, D.C.: Br
Gertler, Mark, and R. Glenn Hubbard. 1989. “Financial Factors in Business Investment.” Financial Market Volatility. Kansas City, Kan.: Federal Reserve Bank of Kansas.
Godley, Wynn. 2000. “What If They Start Saving Again.” London Review of Books, 22(13).
Hall, Robert E., and Frederick Mishkin. 1982. “The Sensitivity of Consumption to Transitory Income: Estimates From Panel Data on Households.” Econometrica, 50: 461-81.
Hatsopoulos, George N., Paul R. Krugman, and James M. Poterba. 1989. “Overconsumption: The Challenge to U.S. Economic Policy.” Working Paper. American Business Conference; cited in Laibson 1997.
Hsieh, Chang Ti. 2000. “Do Consumers React to Anticipated Income Changes? Evidence From the Alaska Permanent Fund.” Princeton, N.J.: Princeton University, unpublished manuscript.
Kreinin, Mordechai. 1961. “Windfall Income and Consumption: Additional Evidence.” American Economic Review, 51 (June): 388-90.
Laibson, David. 1997. “Golden Eggs and Hyperbolic Discounting,” Quarterly Journal of Economics, 112 (May): 443-76.
Lee, Tong Hun. 1975. “More on Windfall Income and Consumption.” Journal of Political Economy, 83 (April): 407-17.
Maki, Dean. 2000. “The Growth of Consumer Credit and the Household Debt Service Burden.” Working Paper. Washington, D.C.: Federal Reserve Board.
Mann, Catherine L. 1999. Is the U.S. Trade Deficit Sustainable? Washington, D.C.: Institute for International Economics.
Okun, Arthur. 1977. “Did the 1968 Surcharge Really Work? A Comment.” American Economic Review, 67 (March): 166-9.
Okun, Arthur. 1971. “The Personal Tax Surcharge and Consumer Demand, 1968-1970.” Brookings Papers on Economic Activity, 1. Washington, D.C.: Brookings Institution, pp. 167-212.
Parker, Jonathan. 1999. “The Reaction of Household Consumption to Predictable Changes in Social Security Taxes.” American Economic Review, 89 (September): 959-73.
Poterba, James. 1988. “Are Consumers Forward Looking? Evidence From Fiscal Experiments.” American Economic Review, 78 (May): 413-8.
Souleles, Nicholas. 1999. “The Response of Household Consumption to Income Tax Refunds.”American Economic Review, 89 (September): 947-58.