January 26, 2004

Rudolph G. Penner
The Urban Institute
2100 M. Street, NW
Washington, DC 20037

Dear Rudy,

I was glad to see the issue brief that you did for the Center for Retirement Research on the impact of more rapid growth on the solvency of Social Security (“Can Faster Growth Save Social Security?”). I think this issue deserves more attention than it has received. Unfortunately, I can’t say that I completely agree with your analysis.

Based on the overwhelming consensus in the economics profession, I think that the prospects for more rapid growth than predicted by the Social Security trustees are far better than you indicate. I also view the implications of the projected shortfalls as less serious than you – given that voters in the past have always supported tax increases of this size (and larger) when they were needed to sustain the program.

I will address both of these issues, but I first want to raise a more fundamental point. Your analysis minimized the impact that more rapid growth could have on the solvency of Social Security. While I don’t think it is plausible that the economy will grow fast enough to eliminate the shortfall over the seventy-five year planning horizon, I do think that even modestly more optimistic projections have a large impact, because they change the nature of the choices facing people.

Specifically, even a very slight increase in productivity growth above the projected rate, would have a large impact on wage growth and therefore after-tax income, the main determinant of living standards. If the annual rate of productivity growth ends up being just 0.1 percentage point more rapid than is currently projected by the trustees, this would cause wages to be more than 5.0 percentage points higher after fifty years. (This is in addition to the Trustees baseline, which already projects a 100 percent increase in real hourly compensation over this period.) This 5.0 percentage point difference would be as least as large as the tax increases that would be needed to keep the program solvent. In other words, if our concern is the living standards of future workers, even a modest increase in productivity growth is more than enough than to fully offset the negative impact of higher Social Security taxes.

Based on hundreds of talks that I have given, - and conversations that I have had over the last decade, I feel confident in saying that this fact -- that future generations of workers will enjoy higher after-tax wages, under almost any circumstances -- is not widely known. I doubt that even most of the reporters in the elite media (e.g. New York Times, Washington Post, National Public Radio, etc.) are aware of this fact – even though there is no dispute on this point within the economics profession.

In short, my interest in the likelihood that the rate of growth will exceed the Trustees’ projections is only partially its direct impact on the solvency of Social Security. I am at least as interested in the implications that more rapid growth will have for future living standards, and therefore the ability of future generations of workers to pay higher tax rates

The job of economists is first and foremost to make the public aware of the choices it faces, so that the public can make these choices in an intelligent manner. Despite the dozens of books and thousands of articles written on Social Security, hardly anyone is aware of the basic features of the trade-offs that will confront society in coming decades.

This is an extraordinary failing of the economics profession.

Getting to the specific claims on growth – first, the piece focuses exclusively on productivity growth. While this is certainly the more important part of the equation, there are important issues on labor force growth that are not considered. Specifically, the rate of immigration is likely to be faster than the 900,000 annual rate assumed by the trustees, especially in the context of the extreme labor shortage that is projected to result from the retirement of the baby boom generation.[1]

I realize that the projections, as a baseline, assume no change in current law, but it is worth noting that immigration averaged 1.3 million annually in the nineties.[2] Presumably, immigration will be at least as large in future decades, assuming that the labor shortages are as bad as currently projected. Of course, if immigration laws are relaxed in response to labor shortages, then the rate of immigration will be even higher.

Immigration does have a mixed effect on the program. Immigrants don’t only pay taxes, they will eventually receive benefits as well. However, if immigration occurs at a more rapid rate in the next three decades than is currently projected, it will further push out the date at which the program first faces a funding shortfall. With productivity and wages rising further in the additional years of solvency provided by increased immigration, workers will be better able to bear any tax increases they may eventually face.

The discussion of productivity notes that the trustees assume an average rate of productivity growth of just 1.6 percent, but then minimizes the likelihood of faster productivity growth, commenting that the Congressional Budget Office calculated that economy wide productivity growth averaged just 1.73 percent over the last fifty years. The paper does not provide a citation for this claim, however, it is worth noting that the Bureau of Labor Statistics index of productivity growth in the non-farm business sector grew at an average pace of 2.31 percent from 1947 through the third quarter of 2003. Assuming a gap of 0.2 percentage points between growth in the non-farm business sector and the rate of productivity growth in the economy as a whole, this implies a 2.1 percent annual rate of productivity growth over this period.[3]

Of course, productivity has proceeded at a more rapid pace over the last eight years. Since the third quarter of 1995, productivity growth in the non-farm business sector has averaged 3.17 percent. Last year, I attended a session at the American Economic Association that included many of the leading experts on productivity growth, such as Dale Jorgenson, Robert Gordon, and Martin Baily. The consensus within this session was that productivity growth in the non-farm business sector would average between 2.5 and 3.0 percent for the foreseeable future.

I have never been a New Economy optimist, and I would be very hesitant about making seventy-five year projections based on an eight-year upturn in productivity growth,[4] however, it seems odd to me to choose a productivity growth projection that is both below the consensus among experts within the field and lower than the average for the period in which reliable data exists. For this reason, a productivity growth assumption of 2.0 percent seems very reasonable.

Increasing the projected growth rate from 1.6 percent to 2.0 percent would reduce the projected shortfall by approximately 0.41 percentage points or approximately 21 percent. It would also have the effect of pushing out the date at which the program will first be short of funds to approximately 2047. At that point in time, real hourly compensation would be projected to be 115 percent higher than it is today. Even if these workers faced a tax rate that was ten percentage points higher than the current tax rate (more than twice the amount needed to sustain Social Security) their after-tax income would still be more than 80 percent higher than it is today.[5]

On this point of taxation, your brief argues that the public will not support higher tax rates as it gets richer. This is not my reading of the evidence. From 1946 to 1959 federal revenue averaged 16.9 percent as a share of GDP. This rose to 18.6 percent in the nineties. Furthermore, there was a large shift in the composition of this revenue, with payroll taxes rising from just 7.9 percent in 1946 of total federal revenue to 40.4 percent in 2003. This suggests to me that people have been willing to pay higher taxes to support programs (Social Security and Medicare) that they value. Note that very few politicians have lost their offices when they supported tax increases for these programs in prior decades. You may have some reason for believing that the future will be different from the past in this respect, but this argument is not presented in the issue brief.[6]

To summarize, I don’t believe that the paper does justice to the importance of economic growth in the Social Security debate. First, recent experience with immigration, and the consensus concerning productivity growth among experts in the field, suggests that the potential for more rapid growth than the trustees assume is much greater than the brief implies. Second, recent history provides solid evidence that the public is willing to pay a higher portion of a growing income to support these programs.

Finally, and most importantly, the brief does not clearly address the impact of higher growth rates on workers’ ability to pay taxes. The public should be aware of the fact that under any plausible scenario, future generations of workers will enjoy much higher after- tax compensation than do current workers, even if they have to pay a higher Social Security tax rate. Under the trustees current projections, after-tax compensation (in today’s dollars) would be more than 80 percent higher in fifty years, than it is today. If the trustees adopted the low end of the current consensus on productivity growth, the projected increase in after-tax compensation over the next fifty years would be in excess of 100 percent.

The fact that most of the public is unaware of these facts – on which virtually all economists agree -- speaks to an incredible failure of the economics profession. Hopefully we will succeed in educating the public on the true situation facing Social Security before any major reform plan is put in place.

                                                Sincerely,

                                                Dean Baker
                       
                        Co-Director

cc: Alicia Munnell 

 

[1] Another issue worth considering in the context of making realistic projections for the program is that the distribution of wage income is likely to change if the country faces a labor shortage. Approximately 40 percent of the projected shortfall can be attributed to the fact that wage income has become increasingly unequal in the two decades since the Greenspan Commission measures were put in place. As a result, the portion of wage income that goes to high-end earners, and is therefore not subject to Social Security taxes, has risen from 10 percent in 1983 to almost 16 percent at present. If a labor shortage put upward pressure on the wages of less skilled workers, this would increase the portion of wage income subject to the tax and improve the financial situation of Social Security.

[2] Bureau of the Census, 2001, Table QT-02, “Profile of Selected Social Characteristics,” 2000 Census.

[3] One of the factors that may have added to the gap between productivity growth in the non-farm business sector and economy-wide productivity in the past was the growth of employment in less productive jobs in the government sector. This cannot be an important factor in projections for the future, since these are baseline projections that by definition assume no change in government policies, and therefore no additional government programs to increase employment.

[4] It is worth noting that the 1991 Technical Panel on Assumptions did construct a productivity projection for the future that gave disproportionate weight to the recent past (1991 Advisory Council on Social Security, pp 22-23).

[5] The paper notes in passing that Social Security costs may prove greater than the Trustees assume, because life expectancy may increase more rapidly than is assumed. It is important to note that this implies a somewhat different problem. Most of the potential impact of the uncertainty about mortality rates is in the more distant future. In this sense, it raises an issue of how much young workers will have to be taxed to pay for their own retirements, not for the retirement of their parents or grandparents. The plausible range of uncertainty of life expectancies of older workers and those already retired is not very large.

[6] It is also worth noting that voters in other industrialized countries have been willing to support far higher tax burdens to sustain their retirement programs. This is also evidence that suggests that voters in the United States would be willing (albeit, not happy) to pay higher taxes if needed to support Social Security.