More on Martin Feldstein

ProGrowth Liberal just posted interesting information about Martin Feldstein on Social Security. Here is an extract from my new book, The Confiscation of American Prosperity, regarding Feldstein's obsession with the subject.

Feldstein first came to national attention in 1974, the same year that Arthur Laffer produced his famous napkin. Feldstein published a model that "proved" that Social Security caused enormous losses for the US economy. According to Feldstein, Social Security was reducing personal savings by 30 to 50 percent. He estimated that if Social Security had not existed, the stock of plant and equipment in the United States would have been as much as 50 percent larger and total personal income 20 percent greater than the level in 1971 (Feldstein 1974). Since Social Security had only been functioning 24 years at the end of the time period that his data covered, Feldstein's article implies that the present effect of Social Security on total personal income today would be far higher ‑‑ perhaps almost 50 percent since the program has had another 35 years at the time of this writing.

The same Jude Wanniski, who popularized supply side economics, later recalled, "I came across a paper that a fellow at Harvard had written on Social Security, saying it was causing the national saving rate to decrease. And I thought, 'Great .... I've got to publish it'" (Bernasek 2004). In other words, because Feldstein's results were welcome, people of influence rushed to embrace him.

The only problem was that Feldstein's work was seriously flawed. A few weeks before the election of Ronald Reagan at the 1980 annual meeting of the American Economic Association in Denver and after Feldstein had already ascended to the head of the National Bureau of Economic Research, two less famous economists, Selig D. Lesnoy and Dean R. Leimer, reported that they were unable to replicate Feldstein's results (later published as Leimer and Lesnoy 1982). Upon analyzing Feldstein's work, they discovered that his results critically depended upon an elementary programming error. With that error corrected, Feldstein's data no longer had the disastrous effects Feldstein claimed. Instead, his model showed that Social Security could have actually had a positive impact on savings.

In all fairness, errors in economic model building are extremely common. In 1982, the Journal of Money, Credit, and Banking began a project to replicate previously published articles. The results were unsettling to say the least. Sixty‑six percent of the authors were unable or unwilling to supply the materials necessary to rerun the model. The authors who responded did so after an average delay of 217 days. All but one of these articles had problems, including programming errors, such as Feldstein committed (Dewald, Thursby and Anderson 1986). This project was hardly likely to inspire confidence in the scientific rigor of economics.

Feldstein admitted his programming error. Undeterred, he soon rejiggled his model. By adding a few new assumptions, he was able to "prove" once again that Social Security was still destructive. Some years later, in 1996, Feldstein gave his own Richard T. Ely lecture. There, Feldstein regaled his audience with new data demonstrating one more time the harmful effects of Social Security. According to Feldstein, the present value of privatizing Social Security would be an astounding $20 trillion dollars ‑‑ about twice the GDP of the United States (Feldstein 1996, p. 12).

In a 2005 Wall Street Journal opinion piece, disingenuously entitled, "Saving Social Security," Feldstein returned once more to his bête noire. This time he was arguing in support of an unpopular piece of Republican legislation to mix Security and private accounts. Feldstein promised great benefits from this "reform": "A higher national saving rate would finance investment in plant and equipment that raises productivity and produces the extra national income to finance future retiree benefits" (Feldstein 2005b). So, Feldstein would rescue Social Security by gutting it.

Earlier in the year, the American Economic Association had given Feldstein a platform to renew his attack on Social Security in his presidential address. Here Feldstein adopted a new pitch. He protested that the program did too little to redistribute income from the rich to the poor. His argument was that because the rich live longer than the poor, they will have more opportunity to benefit from Social Security (Feldstein 2005a).

Without bothering to contest Feldstein's questionable calculations about the redistributional impact of Social Security, this last attack is especially notable for its unusual rhetorical turn. Not too long ago, the same Professor Feldstein discussed the question of inequality with the New York Times. Feldstein began as if he took the subject seriously, observing, "Why there has been increasing inequality in this country has been one of the big puzzles in our field and has absorbed a lot of intellectual effort." Feldstein's own intellectual effort in this debate left something to be desired. Rather than address the question of inequality seriously, he merely trivialized the question, responding to the reporter: "But if you ask me whether we should worry about the fact that some people on Wall Street and basketball players are making a lot of money, I say no" (Stille 2001).

This dismissal of the question of inequality was not some uncharacteristic, off‑hand remark. In an earlier article, entitled, "Reducing Poverty Not Inequality," Feldstein described the proper approach to an imagined increase in inequality occurring because a small number of affluent people received $1000 each at no cost to the rest of society. For Feldstein, only a "spiteful egalitarian" would not welcome such an improvement in society (Feldstein 1999, p. 34).

Of course, Feldstein and his fellow 'spiteful inegalitarians' have been adamant in their hostility to any redistribution of income toward the less fortunate. Such policies threaten to hinder the magical trickle down upon which all progress supposedly defends. Suddenly, however, when it gave credence to his attack on Social Security, Professor Feldstein refurbished himself as a populist advocate of redistribution of income from the rich to the poor by arguing that Social Security benefited the rich. Professor Feldstein never bothered to explain why the rich are so hostile to this program that benefits them so lavishly.

One might expect such a flurry of conflicting arguments from an unscrupulous salesman who wants to earn his commission from a confused customer, but not from one of the most prominent academic economists in the country. One might suspect that ideology rather than an objective search for the truth is at work.

Feldstein did not limit his political activism to Social Security. For example, he used the Wall Street Journal to publicize his work predicting that Clinton's economic taxes would harm the economy while raising little revenue (Feldstein 1993). Unlike his Social Security work, this article made a specific prediction. Unfortunately for Feldstein, his estimates turned out to be demonstrably false. The economy experienced a sudden burst of prosperity during the rest of the Clinton administration.

Alicia H. Munnell, a former student of Feldstein whom he thanked in the acknowledgements to his original Social Security paper and who later rose to become a member of the President's Council of Economic Advisers and Assistant Secretary of the Treasury for Economic Policy, offered this damning verdict in a Business Week article following the Denver meeting: "I get the feeling that the NBER does adopt a position on an issue ‑‑ explicit or implicit ‑‑ and then they go about generating research to support the position" (Anon. 1980). In light of Feldstein's later work, I see no reason to revise her evaluation.

Even if an economist avoids rudimentary programming errors and questionable procedures in handling the data, problems with economic models still remain. The economy is far too complex to reduce it to a mathematical equation or a computer model, even a very large and sophisticated one. As a result, such models necessarily rely on simplifying assumptions.

Although Feldstein proved nothing with his unrelenting attacks on government programs, he demonstrated how clever economists, armed with sophisticated mathematical and statistical techniques, along with the help of well‑trained graduate assistants, are capable of manipulating models to get whatever results they desire. As economists like to joke, that if you torture the data long enough they will confess. So, although economists such as Feldstein can give their work the appearance of scientific precision, their work must necessarily remain suspect.

For example, Social Security's presumably negative effect on saving was at the core of Feldstein's model, but saving has a contradictory effect on the economy. Some models assume that saving encourages investment, while others assume that saving depresses demand, which, in turn, holds back investment. No matter which assumption about the effect of saving economists choose, they can point to reputable theories and models that support them. Admittedly, as economists marginalized Keynesian theory, the models that show the positive influence of saving have become more common. That shift does not reflect an advance in knowledge, but rather a consequence of the right‑wing offensive.

Also, economists can pick and choose among various time periods and data sets, avoiding combinations that do not confirm what they want to find. While such models ‑‑ including many of the models to which I have referred in this book ‑‑ might suggest new lines of research or raise questions about previously accepted truths, they cannot constitute proof by any means.

So, economists may build their models and pundits or politicians can foist the results of these models on the unsuspecting public as if they were scientific evidence, but they are not grounded in science. For example, almost two decades after the errors in Feldstein's original model had been revealed, conservative ideologists, such as those at the Heritage Foundation, still continue to trumpet his long‑discredited calculation as serious evidence of the damage done by Social Security (see, for example, Mitchell 1998).

I believe that Social Security is one of the most effective government programs ever devised in the United States, but I can neither prove nor disprove that assertion with a computer model. In fact, Feldstein's results might possibly turn out to be correct after all, but nobody can know for certain. Different economists have come up with a wide range of estimates (see Lesnoy and Leimer 1985).

Unfortunately, the public rarely has the opportunity to hear about the full range of economic information. Ideological filters determine who gets hired or tenured in economics departments. Those economists who manage to defy the conventional wisdom face the added barrier of getting their work published in "reputable" journals. Even if such papers manage to find their way into journals, they lack the "megaphone" of powerful agencies, such as the Heritage Foundation, which give wide distribution to long‑discredited material without much fear of being exposed. So, ultimately what the public learns about how the economy works are those results that conform to the desires of the rich and powerful.