On June 18, 2007 an article titled "What If Social Security Were Completely Scrapped?" was posted on the Ludwig von Mises Institute web site. It was written by D.W. MacKenzie who teaches economics at SUNY Plattsburgh. On June 24th, I posted a short response to that article on Mr. MacKenzie's blog which referenced a longer response at this link. Mr. MacKenzie then posted a response to that on his blog and this is a reply to that response.
In the first four paragraphs of Mr. MacKenzie's response, he addresses my point that the studies he cites includes Social Security benefits. Following is the fourth paragraph:
The general point here is that people are NOT irrational about retirement. For people to ignore social security as a part of their retirement income would indicate irrationality, but that is not the case. If SS vanished, people would save more privately and private capital accumulation would increase. Social security DOES NOT imrove the well being of retirees, because people end up saving less in the presence of this program, and this lower saving DOES depress long term economic growth. So the fact that these studies include SS as a part of peoples retirement plan is not inconsistent with my article, but is rather essential to it. I did not make this all clear, but it was a short article.
In fact, I agree that people would generally save more privately if Social Security did not exist. I was simply addressing Mr. MacKenzie's statement that "Several recent studies indicate that most people make adequate provisions for their retirements, and some supposedly save too much". It is not clear from this statement whether he is saying that the studies indicate that most people save enough for their retirements without Social Security (in which case it could be immediately scrapped) or that they save enough given the current reality which includes Social Security. Mr. MacKenzie admits this lack of clarity in the last sentence.
While I agree that people would generally save more, it's not clear that all people would. One of the studies that he cites states that "combined Social Security and DB [Defined Benefit] wealth account for about 30 percent of comprehensive wealth, on average, and about 60 percent among the least-wealthy households". It's not clear that these "least-wealthy households" would or even could save enough to counter the loss of Social Security.
In fact, it should be remembered that additional savings by themselves would not closely replace the loss of Social Security. If Social Security simply paid a lump sum on retirement, this could possibly be the case. But the current program pays retirees a monthly check for the rest of their lives and provides survivor and disability benefits. This could only be closely duplicated by annuities and/or insurance.
In any event, Mr. MacKenzie goes on to address the tables and graphs of the OASDI Internal Rate of Return that I have posted at this link in the next three paragraphs:
As for Mr Davis's claim that I am wrong about the redistributive affects of SS, I think the numbers are quite clear. I suppose that some people have lived long enough to get a 9% rate of return on social security, but are these outliers or average rates of return?. Given my family history I expect that I will do much better than any social security index indicates. It is important to remember that social security is a transfer program, and as such it cannot generate a positive rate of return for everyone across generations. for some to get a positive rate of return on SS means that others will, sooner or later, get a negative return. The figures I have seen indicate on average people fall into the 2.5 to minus 2.5 range, with older generations doing better and younger generations doing worse.
If social security is not a transfer program, but involves real investment, then I want to see proof. I understand that the social security surplus is invested in federal bonds, but this is a shell game. Social security receives interest from those who pay income tax. So people on the SS payroll tax are paying in more to SS, so that they can earn interest into SS through the income taxes that they pay to finanance federal debt. IS that real investment?
If SS is not a transfer program, then I want to know how it involves real investment. If SS can generate real positive return for some, WITHOUT imposing negative returns on others, then I would be fascinated to find out how such alchemy works. Otherwise, there must be something wrong with the numbers that Mr Davis cites. His graphs make it seem that most people get a positive return, and many get a high return, while only a few get around zero or a slightly negative return. How can a transfer program do this? The numbers I cite pose no such problems. From what I can see these numbers are his own, or come from the AARP. The AARP is obviously biased. I, on the otherhand, have a personal interest in preserving this program- I have a high life expectancy (due to family history) and relatively low ability to save (on a college professors income).
At the bottom of the aforementioned page, I have clearly noted that numbers come from a Social Security Administration publication, not the AARP. In any case, the "alchemy" by which a transfer program can "generate real positive return for some, WITHOUT imposing negative returns on others" can be explained in the following simplified example.
Suppose that a population consists of eighty people, all born on January 1st and consisting of exactly one person of each age from zero to 79. Now suppose that all of these people live "average lives" whereby they begin working for an average wage on their 20th birthday, retire on their 60th birthday, and die on their 80th birthday. Furthermore, suppose that a baby is born each January 1st such that the population continues to have 80 members, evenly distributed in age. Finally suppose that they have a sustainable transfer program by which each working person pays 15 percent of their income into a pool which is split up evenly between the retirees. Since there will always be 2 workers per retiree (40 workers and 20 retirees), each retiree will receive 30 percent of the average wage.
Now, if wages never change then the amount that a person will pay into the transfer program over their lifetime will be
(Wage per year) * (40 years) * (15 percent) = (Wage per year) * 6
and the amount that they will receive in retirement will be
(Wage per year) * (20 years) * (30 percent) = (Wage per year) * 6
Hence, this would be the case that Mr. MacKenzie seems to expect of a transfer program, where the total return is zero. However, suppose that there was a constant rate of inflation of 3 percent per year and that wages also increased at that rate. Then all of the wages per year in the second "benefit formula" would be greater than all of the wages per year in the first "contribution formula", meaning that each retiree would get a positive return. In fact, since wages rose at the exact same rate as inflation, one could say that they remained constant in real, inflation-adjusted dollars. Hence, if the unit of measurement is real dollars then, once again, the contributions and benefits would be equal and each retiree would receive a "real return" of zero.
Now, suppose that wages rise faster than inflation. According to numbers from the numbers from the U.S. Census Bureau, this has been the case since at least 1955. Then, even if the unit of measurement is real dollars, all of the wages in the second benefit formula would be greater than the wages in the first contribution formula. This would mean that all retirees would be receiving a positive real return.
Of course, this does not prove that current retirees are receiving a positive real return under the actual system. But it does show that a sustainable transfer program can provide positive real returns to its participants. In a sense, one could look at this as a sustainable pyramid scheme, driven by increasing real wages which, in turn, are driven by increases in technology and productivity. Of course, Social Security has also benefited from some unsustainable "pyramid schemes" such as the increase in the Social Security tax rate (it rose from 1 percent in 1940 to 6.2 percent today). However, increases in technology and other sources of prosperity do provide a sustainable source of positive returns that can be tapped into.
Regarding the effect of life expectancy on the rate of return, Mr. MacKenzie continues:
The proposition that SS transfers from poor to wealthy, based on life expectancies, has been around for a long time. I first heard it is grad school, from a competent econometrician. I have not perfomed calculations on this personally, but have relied on others. Mr Davis claims that SS is structered to counteract these affects. I am not presently in a position to dispute his numbers on SS payments, but even if his numbers are accurate, they do not address the life expectancy issue (at least I did not see any way that SS payments address this), and this point does not address the capital accumulation issue. The bottom line is that SS penalizes people with short life expecancies, which is due partially to life style choices, but largely to family genetics and personal circumstances (i.e. poverty).
The proposition that Social Security transfers from poor to wealthy, based on life expectancies, may have been around for a long time but I have never seen a serious study that purports to give evidence of it. As I mentioned, the rates of return that I posted came from a Social Security Administration publication. This publication did say the following regarding the this question:
For each sex, family grouping, and year-of-birth cohort the internal rates of return decrease as earnings increase. This is because the benefit formula is weighted toward beneficiaries with lower earnings. The advantage for lower earners is partially offset by their lower life expectancy.
An economic letter from the Federal Reserve Bank of San Francisco gives the approximate rates of return by income in the following excerpt:
By design, Social Security redistributes income by means of a progressive benefit formula. Low-paid workers are awarded a greater fraction of their pre-retirement income than high-paid workers. This progressive benefit structure is more than enough to offset the regressive nature of the OASI payroll tax whereby income above a specified maximum level (currently $72,600) is not subject to the tax. Caldwell, et al. (1998) compute internal rates of return on contributions for several categories of OASI participants (see Figure 1, Panels A-D). The authors find that, under current OASI rules, today's lowest paid workers (those in the bottom 20% of the income distribution based on lifetime earnings) can expect internal rates of return between 4% and 5% after adjusting for inflation (Panel A). Today's middle income workers can expect real rates of return between 1% and 2%. Today's highest paid workers can expect real rates of return below 1% and may even face negative rates of return if born after 1975.
If Mr. MacKenzie can provide a link to a credible study that purports to show that Social Security transfers from poor to wealthy or that the return for average people fall into the 2.5 to minus 2.5 range, I'll be more than happy to look at it. Until then, I'll just have to stick with the sources that I have.