Keynesian economics reached the pinnacle of its influence on policy in the early 1960s, when John F. Kennedy appointed prominent Keynesians such as Walter Heller and James Tobin as his economic advisers and promoted a big tax cut in 1963 on Keynesian grounds. On December 31, 1965, John Maynard Keynes appeared on the cover of Time magazine even though he had been dead for more than 20 years.
But a counterrevolution was already underway. Centered at the University of Chicago, its leaders were Milton Friedman and Robert Mundell, both of whom would receive the Nobel Prize in economics. Their work complemented each other's. Friedman was mainly interested in domestic monetary policy and microeconomics; Mundell was primarily concerned with international economics and fiscal policy. Between the two of them they pretty well covered the economic waterfront.
One weakness of Keynesian theory in the 1960s was that its adherents had forgotten that fiscal policy was a means to an end—the end being the injection of additional liquidity into the economy to counter a deflation. But 30 years after Keynes put forward his theory, his followers had come to believe that an expansive fiscal policy was sufficient to stimulate growth on its own. The essential connection to monetary policy was largely lost among the Keynesians. The popularity of the Phillips curve as an explanation for inflation reinforced the deemphasis of monetary policy.
To Keynesians, the key economic variable was the federal budget deficit. It didn't really matter very much whether a deficit arose due to lower revenues or higher spending, they thought. A deficit resulting from a tax cut was just as good as a deficit caused by an increase in spending. Another consequence of this exclusive focus on the deficit was that economists paid little attention to the structure of taxation or spending. As noted in the previous chapter, spending on national defense was just as stimulative as spending on social welfare in the Keynesian model.
But when inflation and unemployment began rising simultaneously in the 1970s, the Keynesians had little to offer except wage and price controls, which failed drastically when instituted by Richard Nixon in 1971. It took a refocus on monetary policy, which Friedman was highly successful in doing, to get at the root of the inflation problem. But monetarism had little to say about growth. That component was added by Mundell, who argued for tax rate reductions to accompany tight money. By increasing the incentive to produce, lower rates would simultaneously reduce unemployment and inflation.
By the late 1970s Republican politicians fully endorsed the idea of tight money to combat inflation and tax cuts to raise growth and reduce unemployment, which came to be called supply-side economics. Still under the influence of Keynesian economics, Democrats strenuously resisted both prongs of this emerging economic program. They saw tight money as ineffective against inflation and highly detrimental to growth because it would raise interest rates, while tax rate reductions would aggravate inflation and be only modestly stimulative to growth. Having tight money and tax cuts at the same time was like putting one's foot on a car's brake and gas pedal at the same time, most Keynesians thought. As economist William Nordhaus of Yale put it in 1981, “I know of no economic theory that says you can suppress money growth to fight inflation and simultaneously use fiscal policy to spur growth.”1
But lacking a coherent cure for the twin problems of slow growth and inflation, dubbed “stagflation,” the Keynesians lost the argument by default. The monetarists and supply-siders slowly gained influence and a powerful advocate in Ronald Reagan. Upon taking office in 1981, he encouraged the Federal Reserve to maintain a tight money policy even in the face of the severe recession that began late that year.2 Reagan was also able to overcome powerful resistance to an across-the-board tax rate reduction and got it enacted into law in August 1981.
Although many analysts continue to insist that an alternative policy mix might have worked better, the rapid decline of inflation and unemployment in the 1980s was widely viewed as vindication of conservative economics, just as the end of the Great Depression during World War II was taken as confirmation of Keynesian economics.
The pinnacle of Keynesian economics—the moment when it reached its highest point of influence before beginning a precipitous fall—can be dated precisely. It came on January 6, 1971. On that day Richard Nixon was taping an interview with Howard K. Smith of ABC News. In an offhand comment after the end of the interview, Nixon mentioned in passing that he was “now a Keynesian in economics.” Smith was struck by the statement because Republicans had historically been unrelentingly hostile to Keynesian economics and deficit spending. To Smith, Nixon's conversion to Keynesianism was equivalent to a Christian saying, “All things considered, I think [Islam founder] Mohammad was right.”3
Ironically, just one week before Nixon's interview, economist Harry Johnson had given the prestigious Richard T. Ely Lecture to the American Economic Association in which he warned that the Keynesians' inability to deal with the growing problem of inflation was making it vulnerable to a counterrevolution from monetarism.4 The monetarists, led principally by Friedman, argued that inflation was almost purely a monetary problem having little, if anything, to do with budget deficits, unemployment, or economic rigidities, as the Keynesians supposed.
In his memoirs, Friedman says he was a conventional Keynesian until after World War II. It's not clear at what point monetary policy became his central interest, but certainly there was a long tradition of emphasis on that subject at the University of Chicago dating back at least to the work of J. Laurence Laughlin, who had been deeply involved in creation of the Federal Reserve System. Being somewhat isolated from the hotbeds of Keynesian thinking in Cambridge, Massachusetts, and Washington, D.C., undoubtedly made it easier to chart a different path. But at the same time, there was less resistance to some of the basic concepts of Keynesian economics in Chicago than is generally believed.5
After publication of Friedman's 1962 book, Capitalism and Freedom, he became closely aligned with the Republican Party and was an adviser to Barry Goldwater's campaign in 1964.6 Friedman soon became the dominant figure in developing the Republican Party's economic policy for the next three decades. Publication of his Monetary History of the United States in 1963 cemented Friedman's reputation as the leading monetary economist in the country. He was often called upon to testify before congressional committees on the subject and Friedman was frequently quoted in the national media.
Like Keynes, to whom he is often compared, Friedman was no ivory tower academic. After publication of the Monetary History, his work increasingly turned away from scientific pursuits to more popular venues, including a column in Newsweek magazine and an extremely popular television series called “Free to Choose” in 1980.
The core of Friedman's technical work in the late 1960s and early 1970s was refutation of the Phillips curve as neither an adequate explanation for inflation nor an effective policy response to it. Inflation, he argued, had almost nothing to do with unemployment or anything other than excessive money growth by the Federal Reserve. To the extent that there was any trade-off between inflation and unemployment, it was purely temporary. Over time inflation actually raised unemployment by creating malinvestment and economic misallocations.7
It's probably fair to say that the manifest failure of the Keynesians to deal with the growing problem of inflation had more to do with Friedman's ultimate success than the power of his ideas. In a sense, he won by default. The last gasp of the Keynesians was enactment of a surtax in 1968 to sop up excess purchasing power. Monetarists like Friedman argued that this wouldn't work because it didn't get at the monetary roots of inflation. Indeed, because monetary policy remained loose, they predicted that inflation would worsen. When it did, Friedman gained enormous credibility.8
Ironically, Nixon's embrace of Keynesian economics seems to have been a factor causing some liberal economists to reexamine their support for it. In 1973 economist Melville J. Ulmer was forced to admit that “the economic strategy of the Nixon administration is more or less pure Keynesian … the same as that endorsed by the Democratic Party.” Therefore, the failure of Keynesian economics to achieve full employment without inflation could not be blamed on an unwillingness of politicians to fully implement it. This must mean that Keynesian economics was inherently flawed, Ulmer concluded.9
By 1976 Friedman's ideas were so well accepted that he was awarded the Nobel Prize in economics. In his presidential address to the American Economic Association that year, economist Franco Modigliani, a prominent Keynesian who would win the Nobel Prize in 1985, waved the white flag of surrender. “We are all monetarists,” he declared.10
Although monetarism explained inflation, it didn't have much to say about stimulating growth. Adding a growth component to monetarism was the principal contribution of Mundell. He believed that inflation reduced growth mainly through its interaction with the tax system. Workers receiving pay increases were pushed up into higher tax brackets, even though they might not have received any real increase in purchasing power; corporations found that depreciation allowances based on historical cost were inadequate to replace old equipment while taxes were assessed on illusory inventory profits; and investors found that much of their capital gains simply represented inflation but were taxed as if they were real.11 For example, in 1977 investors paid taxes on $ 5.7 billion of nominal gains on sales of corporate stock, but when adjusted for inflation this gain was actually a $ 3.5 billion loss.12
A two-pronged strategy was therefore needed to end stagflation, Mundell said. First, money growth needed to be sharply tightened, preferably by targeting the price of gold—when the price of gold rose, it would be a signal to the Fed to tighten. Thus, Mundell's views on monetary policy dovetailed with Friedman's, although Friedman thought that any linkage to gold was archaic and inefficient.13 Second, tax rates needed to be cut in order to restore incentives and increase the demand for money. Although Friedman had no problem with cutting taxes, this was never a focus of his academic research. Mundell first spelled out this tight-money-and-tax-cuts policy mix in a 1971 paper published by Princeton's International Finance Section:
Monetary expansion stimulates nominal money demand for goods, but, without rigidities or illusions to bite on, it does not lead to real expansion. But growth of real output raises real money demand and thus abets the absorption of real monetary expansion into the economy without inflation. Tax reduction increases employment and growth and this raises the demand for money and hence enables the Federal Reserve to supply additional real money balances to the economy without causing sagging interest rates associated with conditions of loose money. Monetary acceleration is inflationary, but tax reduction is expansionary when there is unemployment.14
At this time, Arthur Laffer was teaching at the University of Chicago business school, where he came to know Mundell's work. Their mutual interest in international monetary issues brought them together and Laffer absorbed Mundell's ideas. In 1974 Laffer organized a conference sponsored by the American Enterprise Institute on the problem of worldwide inflation.15 This was the first opportunity many Washington policymakers had to hear about what came to be called supply-side economics.
Although most of the AEI conference dealt with monetary issues, Mundell reiterated the point made in his 1971 paper about the vital role of tax cuts in the fight against inflation. This was important because of the widespread view that budget deficits caused inflation. Indeed, few now recall that when Representative Jack Kemp first began pushing for big tax cuts in the late 1970s, the principal attack against him was such a policy would be massively inflationary because it would increase the deficit.16
Jude Wanniski, then an editorial writer for the Wall Street Journal, attended the AEI conference and was fascinated by the discussion. He immediately wrote an article for the Journal explaining the Laffer-Mundell world view. This article is probably what led to Mundell's invitation to a White House conference on December 19, 1974, at which he argued for a tight monetary policy to fight inflation and a big tax cut to stimulate growth.17
That same month, Wanniski set up a meeting with Donald Rumsfeld, White House chief of staff, and his deputy, Dick Cheney, at which Laffer drew on a napkin his famous curve for the first time. Unfortunately, Gerald Ford and his economic advisers rejected a tax rate reduction, opting instead for a one-shot tax rebate.18
Wanniski expanded his Mundell-Laffer analysis in a 1975 article for The Public Interest, a quasi-academic journal edited by Irving Kristol. Interestingly, Wanniski's discussion of the tax side of their work appeared only in a footnote. This was the first published statement of Laffer's famous curve, which shows that tax rate cuts could theoretically increase government revenue. As Wanniski wrote:
Taxes should be cut and government spending maintained through deficit financing only when a special condition exists, a condition Mundell and Laffer say exists now. “There are always two tax rates that produce the same dollar revenues,” says Laffer. “For example, when taxes are zero, revenues are zero. When taxes are 100 percent, there is no production, and revenues are also zero. In between these extremes there is one rate that maximizes government revenues.” Any higher tax rate reduces total output and the tax base, and becomes counterproductive even for producing revenues. U.S. marginal tax rates are now, they argue, in this unproductive range and the economy is being “choked, asphyxiated by taxes,” says Mundell. Tax rates have been put up inadvertently by the impact of inflation on all the progressivity of the tax structure. If the tax rate were below the rate that maximizes revenues, tax cuts would reduce tax revenues at full employment. But a multiplier effect operates if the economy is at less than full employment, and the tax cut then raises output and the tax base, besides making the economy more efficient. Even if a bigger deficit emerges, sufficient tax revenues will be recovered to pay the interest on the government bonds issued to finance the deficit. Thus, future taxes would not have to be raised and there would be no subtraction from future output. Tax cuts, therefore, actually can provide a means for servicing the public debt.19
Wanniski later wrote the first book about supply-side economics, The Way the World Works (1978). He also wrote many unsigned editorials for the Wall Street Journal spelling out various aspects of supply-side theory and helped supply-siders such as Laffer and Paul Craig Roberts get published in the Journal, thereby raising their visibility and stature as economic commentators.20 Indeed, when Wanniski left the Journal in 1978 to found a private consulting firm called Polyconomics, where I later went to work, Roberts was hired by the Journal to replace him on the editorial page staff.
Wanniski was responsible for converting Kristol and Bob Bartley, editorial page editor of the Wall Street Journal, to supply-side economics. By his own admission, Kristol knew almost nothing about economics at the time. But he saw the political potential of supplyside economics to refocus the Republican Party's economic thinking away from stability and toward growth. Upon his retirement, Bartley cited his support for supply-side economics as among his greatest accomplishments.21
By the mid-1970s, inflation had created an economic crisis throughout the West. Even those on the political left were forced to admit that Keynesian economics offered no way out from the twin problems of rising prices and slow growth. Whatever was done to cure one problem would only make the other worse. The idea that inflation caused unemployment, rather than curing it as the Phillips curve posited, became widespread both among economists and policymakers. In July 1976 Johannes Witteveen, managing director of the International Monetary Fund, spoke for many when he said, “There seems to have developed of late such an unusual sensitivity to inflation that an acceleration in the rate of price increase in all likelihood would have significant adverse effects on demand, production and employment.”22
Soon after, British Prime Minister James Callaghan of the Labor Party eloquently expressed his frustration with the lack of options presented by Keynesian economic thinking. In a September 28, 1976, speech to a party conference, Callahan said:
We used to think that you could just spend your way out of a recession and increase employment by cutting taxes and raising government spending. I tell you, in all candor, that that option no longer exists and that it only worked on each occasion since the war by injecting bigger doses of inflation into the economy, followed by higher levels of unemployment as the next step.23
A few months later, German Chancellor Helmut Schmidt of the Social Democratic Party declared that “the time for Keynesian ideas is past because the problem of the world today is inflation.” By the end of 1977, the popular media had proclaimed the death of Keynes.24
Increasingly, academic economists followed suit. A 1978 article in the Journal of Economic Literature, a publication of the American Economic Association that tries to reflect the consensus view of the profession, said, “There appears to be no long-run trade-off between inflation and unemployment.”25 This marked the death of the Phillips curve among professional economists.
It also marked the death of Keynesian economics. “By about 1980, it was hard to find an American academic macroeconomist under the age of 40 who professed to be a Keynesian,” economist Alan Blinder of Princeton later wrote.26
Writing in The Public Interest in 1981, Martin Feldstein of Harvard said that while Keynesian ideas may have worked in the 1930s, “it has become very clear that those ideas were not appropriate for the U.S. economy of the 1960s and 1970s when they achieved their greatest acceptance and influence.”27
Speaking to the annual meeting of the American Economic Association that same year, Edmund Phelps of Columbia, winner of the Nobel Prize in 2006, said that while Keynesian economics might one day be reborn, “for now, the old Keynesian notion of fiscal stimulus is so beset by doubts that fiscal policy, if not truly incapacitated, is in a deactivated state.”28
In the midst of these developments, an important legislative fight took place that highlighted both the weakness of the Keynesian approach to fiscal policy and the failure of standard revenue-estimating methods to account for the supply-side effects of tax changes. It involved a cut in the long-term capital gains tax rate.
The roots of the capital gains controversy dated back to 1969, when the maximum long-term tax rate on capital gains was increased to 35 percent from 25 percent. On April 13, 1978, Representative William Steiger, Republican of Wisconsin, introduced legislation to return the top rate to 25 percent. The ensuing debate was extremely important to the development of supply-side economics for two reasons. First, a number of highly respected economists such as Feldstein argued vigorously that a cut in the capital gains tax rate would almost immediately recoup the static revenue loss through a combination of unlocking and increased investment. Second, the obvious expansionary potential of a capital gains tax cut clearly illustrated the supply-side argument that tax rate cuts could be stimulative by increasing incentives without any Keynesian impact on disposable income. Thus the 1978 capital gains debate was seminal in the development and acceptance of supply-side theories by mainstream economists.29
Of course, the idea that cutting the capital gains tax might not reduce federal revenue was not a new one. It had been long recognized that capital gains are a unique form of income because taxpayers have the freedom to decide when and whether to realize gains for tax purposes. When rates are high they encourage a lock-in effect that reduces revenues as investors hold on to potentially taxable gains. Therefore, a rate reduction might result in a rapid unlocking of past gains that could cause revenue to rise quickly.30
Feldstein was outspoken in his conviction that the Steiger bill would increase federal revenue almost immediately and testified before Congress to that effect. The basis for his testimony was a National Bureau of Economic Research paper that circulated on Capitol Hill in June 1978. Other prominent economists forecasting higher revenues from the Steiger bill included Gary Ciminero of Merrill Lynch Economics and Michael Evans of Chase Econometrics. Although the Steiger bill was officially scored as a revenue-loser, after it was enacted into law in 1978 the Joint Committee on Taxation (JCT) conceded that the bill would probably raise revenue, which was confirmed by subsequent research.31
The capital gains proposal set off a fierce debate over how to estimate the revenue effects of changes in tax policy that was critical to advancing supply-side economics and undermining Keynesian economics. The focus of the debate was on the mathematical models used by economists to forecast the economy. These models were often used to evaluate public policies and almost universally had Keynesian underpinnings. This tended to bias public policy in favor of Keynesian policies long after they were generally discredited.32
This was especially a problem after enactment of the Budget Act of 1974 because no tax bill could even be considered in Congress unless there was provision for it in the annual budget resolution. This meant that Congress needed to know in advance the precise revenue loss expected from a tax cut before it could be voted upon. Although revenue estimates had been done for many years previously, they had never been an essential requirement for consideration of tax legislation.
Revenue estimates for Congress are done by the JCT, although Treasury estimates were often sufficient for its needs.33 These estimates were usually made by accountants for one or two years out. But by the 1970s, economists had replaced the accountants and computers started to be used in lieu of adding machines. Moreover, the budget act generally required five-year revenue estimates, which increased the need for economic forecasts on which to base the estimates.
These forecasts were prepared by the Congressional Budget Office and generally did not incorporate any macroeconomic effects from tax changes. The JCT revenue estimate was largely a function of the CBO baseline forecast. Nor did the JCT incorporate behavioral effects in its revenue estimates. The effect of holding economic variables constant regardless of the magnitude of a proposed tax change sometimes led to absurd results. For example, Senator Bob Packwood, Republican of Oregon and former chairman of the Senate Finance Committee, once asked the JCT how much revenue would be raised by a 100 percent tax rate on all incomes over $ 200,000. The JCT reported that it would raise $ 204 billion in 1990, $ 232 billion in 1991, $ 263 billion in 1992, and $ 299 billion in 1993.34 Of course, the true revenue yield would have been zero since no one would have realized any taxable income over $ 200,000 if they couldn't keep any of it.
Supply-siders argued that if a tax cut raised GDP, then the base of taxation would be larger, thus recouping some of the revenue lost from the rate cut. Conversely, a tax increase might reduce GDP, thereby reducing the revenue potential. In other words, a 10 percent tax rate cut might only reduce revenues by 6 or 7 percent, and a 10 percent tax rate increase might only raise revenues by 6 or 7 percent.
CBO's forecasts relied on commercial econometric models, such as those of Data Resources, Inc. (DRI), Wharton Econometric Forecasting Associates (WEFA), and Chase Econometrics. Largely based on Keynesian assumptions, these models tended to make the budgetary cost of tax cuts high relative to equivalent government spending programs and thereby biased the legislative process in favor of temporary tax cuts designed to stimulate consumption and against supply-side tax cuts such as marginal tax rate reductions.
An example of how this worked in practice can be seen in one of the CBO's earliest studies, which looked at various fiscal options for reducing unemployment. Because of the Keynesian underpinnings used to evaluate the alternatives—spending drives growth in the Keynesian model while saving is a drag on it—increased government spending appeared preferable to tax cuts. Direct spending created more jobs per $ 1 billion increase in the deficit than tax cuts because all of the former was assumed to be spent while some of the latter was saved. In political terms, therefore, supply-side tax cuts were at a disadvantage compared with Keynesian-style public service jobs programs. Permanent tax rate reductions were also deemed more costly and less effective than temporary tax rebates.35
Paul Craig Roberts was the first supply-sider to recognize that the breakdown of the Keynesian system and institution of the new congressional budget process created an opportunity to promote the supply-side approach to fiscal policy.36 Given the constraints of the econometric models and the budget process, it suddenly became very important to be able to show that certain types of tax cuts did not increase the deficit as much as direct spending programs.
Traditional tax-writers, such as Senator Russell Long, Democrat of Louisiana and Senate Finance Committee chairman, now found themselves at a disadvantage relative to the appropriators. The CBO gave spending programs the benefit of a Keynesian multiplier in calculating their economic effects. But tax cuts were calculated by the JCT on a static basis, as if they had no economic impact on growth or incentives.
Long was no supply-sider, but he had been around a long time and seen a lot of tax changes and their effects at close hand. His experience told him that there was something to the supply-side argument that tax cuts would not lose as much revenue as static forecasts said they would. During a 1977 hearing, Long had this to say:
Revenue estimates have a way of being very, very far off base because of the failure to anticipate everything that happens. … Now, when we put the Investment Tax Credit on, we estimated that we were going to lose about $ 5 billion. … Instead of losing money, revenues went up in corporate income tax collections. Then we thought it was overheating the economy. We repealed it. We thought that the government would take in more money. But instead of making $ 5 billion, we lost $ 5 billion. Then, after a while, we thought we made a mistake, so we put it back on again. Instead of losing us money, it made us money. Then, after a while, we repealed it again and it did just exactly the opposite from what it was estimated to do again by about the same amount. It seems to me, if we take all factors into account, we wind up with the conclusion that taking the Investment Tax Credit alone and looking at it by itself, it is not costing us any money. Because the impression I gain from it is that it stimulates the economy to the extent, and brings about additional investment to the extent, that it makes us money rather than loses us money.37
Long commissioned Michael Evans, an experienced econometric modeler, to build a supply-side model for the Senate Finance Committee. By the time Evans finished his work, however, control of the committee had shifted to the Republicans and Senator Bob Dole of Kansas became its chairman. Bob Lighthizer, the new chief of staff of the Finance Committee, told me personally that he had no interest in the project because it originated on the Democrats' watch. When I asked Evans himself what happened to the model, he wrote me to say that he made one copy, sent it to the Finance Committee in fulfillment of his contract, and had no other copies.38
Economists associated with the rational expectations school played a part as well in undermining the foundations of Keynesian economics. From the point of view of supply-siders, a key element of their critique related to econometric models. They argued that people learn from policy changes and thus change their behavior accordingly. People may react to a policy one way the first time and differently the second time. Interestingly, Keynes basically agreed with this idea.39
It was in this atmosphere that Kemp and Senator William Roth, Republican of Delaware, introduced the legislation that defined supply-side economics. It grew out of Kemp's desire to duplicate the Kennedy tax cut by having a pure, across-the-board individual income tax rate reduction, without the many corporate provisions that had been the central features of his earlier tax efforts.
Kemp was already drawing parallels to his early tax proposals—such as the business-oriented Jobs Creation Act—with John F. Kennedy's legislation when I joined his staff in 1976. Norman Ture was a Kemp adviser and an important link to the Kennedy experience because he had worked for Wilbur Mills, who chaired the House Ways and Means Committee during the time of the Kennedy tax cut.
In August 1976 Kemp received data from the Congressional Research Service on the estimated revenue loss from the Kennedy tax cut.40 By comparing these revenue loss figures with actual revenue collections from the 1960s, Kemp concluded that the Kennedy tax cut increased federal revenue. His point was more of an assertion than hard evidence since he had no data on what aggregate revenues were expected to be in the absence of the Kennedy tax cut. Interestingly, however, Walter Heller, who chaired Kennedy's Council of Economic Advisers, soon made the case for him. In testimony before the Joint Economic Committee on February 7, 1977, he was asked by Senator Jacob Javits, Republican of New York, to comment on Kemp's anal y sis of the CRS memo. I was in the hearing room when Heller responded:
What happened to the tax cut in 1965 is difficult to pin down, but insofar as we are able to isolate it, it did seem to have a tremendously stimulative effect, a multiplied effect on the economy. It was the major factor that led to our running a $ 3 billion surplus by the middle of 1965, before escalation in Vietnam struck us. It was a $ 12 billion tax cut, which would be about $ 33 or $ 34 billion in today's terms. And within one year the revenues into the Federal Treasury were already above what they had been before the tax cut. … Did it pay for itself in increased revenues? I think the evidence is very strong that it did.41
Heller was later embarrassed to have provided the supply-siders with the proof they lacked and tried to take it back.42 But as a witness to the event, I had no reason to think he was not stating a sincere belief. Indeed, a review of statements by Kennedy, his advisers and his supporters at the time clearly indicates they expected the tax cut would in fact raise federal revenue. As Kennedy said in his Economic Club of New York speech on December 14, 1962, “It is a paradoxical truth that tax rates are too high today and revenues are too low, and the soundest way to raise the revenues in the long run is to cut the rates now.”
During floor debate on September 24, 1963, Wilbur Mills, manager of the Kennedy tax cut in the House of Representatives, said, “There is no doubt in my mind that this tax reduction bill, in and of itself, can bring about an increase in the gross national product of approximately $ 50 billion in the next few years [$ 350 billion today]. If it does, these lower rates of taxation will bring in at least $ 12 billion in additional revenue [$ 84 billion today].”43
Contemporary analyses by the CEA and economists Arthur Okun and Lawrence Klein show that Mills was definitely in the ballpark with his estimate.44 Once the impact of the Kennedy tax cut became a political issue in the late 1970s, further analyses were undertaken. DRI and WEFA were contracted to study the impact of the Kennedy tax cut.45 After reviewing these studies, the CBO drew the following conclusion:
The effect of the 1964 tax cut on the federal deficit has been a matter of controversy. … The direct effect of the tax cut was to reduce revenues by some $ 12 billion (annual rate) after the initial buildup. The increase in output and later in prices produced by the tax cut, according to the models, recaptured $ 3 to $ 9 billion of this revenue at the end of two years. The result was a net increase in the federal deficit of only about 25 to 75 percent of the full $ 12 billion.46
Thus, while the Kennedy tax cut may not have paid for itself immediately, there is overwhelming evidence that the federal government did not lose nearly as much revenue as expected, owing to the expansionary effect of the tax cut on the economy.47
The real importance of the feedback argument had to do with how much additional federal borrowing was necessitated by tax cuts. If they led to higher interest rates, it would be plausible to argue that this could offset much of the beneficial impact of tax cuts on incentives. In this respect, it is also important to know whether a tax cut increased private saving. If saving expands, it is reasonable to include this in revenue feedback estimates.
In a 1981 study for the Federal Reserve Bank of San Francisco, economist Paul Evans concluded that the Kennedy tax cut actually raised saving by more than the total amount of the tax cut. That is, households saved more than 100 percent of the tax cut. Therefore, the Kennedy tax cut could not put upward pressure on interest rates due to an increase in the federal budget deficit. Supply-siders often made this point with regard to the Reagan tax cut when questions were raised about its impact on federal borrowing and interest rates.48 Thus Ronald Reagan was really not too far off when he asserted in his October 1981 news conference that the Kennedy tax cut paid for itself.
Work on the legislation began in early 1977. It was my job to figure out exactly what it meant to “duplicate” the Kennedy tax cut, given that the rate structure had changed dramatically in the years since 1964, when the Kennedy tax cut was rammed through Congress by Lyndon Johnson. Working together with Bruce Thompson from Roth's office, Pete Davis of the JCT, Norman Ture, and others, we eventually decided to reduce the top statutory rate from 70 percent to 50 percent and the bottom rate from 14 percent to 10 percent. We felt that this was roughly comparable to Kennedy's reduction in the top rate from 91 percent to 70 percent and the bottom rate from 20 percent to 14 percent.
It took about a year before the Kemp-Roth proposal began to get attention. Interestingly, the idea that it would stimulate growth enough to pay back some of the static revenue loss was not especially controversial at first. Indeed, Bert Lance, Jimmy Carter's Office of Management and Budget director, had testified shortly before KempRoth was introduced to that effect:
My personal observation is that as you go through the process of permanent tax reduction, that there is an awfully good argument to be made for the fact that the revenues of the government actually increase at a given time. I think it has been proven in previous circumstances. I have no problem in following that sort of thing.49
Treasury Secretary Michael Blumenthal was also sympathetic to the idea that lowering tax rates would raise revenue: “The simple-minded notion underlying all of this is that if it works I would hope there would be a bigger pie and higher levels of activity producing more revenue.”50
When the CBO reviewed the Kemp-Roth bill, it estimated that feedback effects would recoup between 14 percent and 19 percent of the static revenue loss the first year, rising to between 26 percent and 38 percent in the fourth. This is consistent with what the supplysiders themselves thought would happen. Contrary to popular belief—including Ronald Reagan's—they never thought there would be no revenue loss at all. Clearly, there would be large revenue losses in the short-run. But the supply-siders thought the net revenue loss would be much less than the static estimates predicted.51
Although supply-siders certainly thought there would be increases in economic growth, investment, and labor supply from marginal tax rate cuts, this was not by any means the only way they expected revenues to be recouped. They anticipated many changes in behavior that would have the effect of increasing taxable income. Among the most important areas where expansion of the tax base was anticipated was from shrinkage of the so-called underground economy. I vividly remember reading economist Peter Gutmann's path-breaking 1977 article, which estimated the underground economy equaled about 10 percent of recorded gross national product. Although not motivated solely by tax evasion, high tax rates unquestionably contributed heavily to its growth. Hence, tax rate reduction would cause some underground economic activity to move above ground, so to speak, and become taxable.52
Supply-siders further anticipated that workers would alter their compensation so as to increase the taxable portion of their wages. In particular, tax-free fringe benefits would be less attractive relative to cash wages. There is considerable evidence that rising tax rates in the 1970s were behind much of the growth in fringe benefits such as health insurance.53
Investors were also expected to alter their portfolios in ways that would raise taxable income. For example, with lower rates, the value of tax losses and other deductions, such as for Individual Retirement Accounts, would no longer be as valuable. Home ownership, with its many tax advantages, would no longer be as appealing relative to renting. And taxable dividends and interest would become more attractive compared with more lightly taxed capital gains and tax-free municipal bonds. All of these factors were expected to have a powerful effect on raising taxable income even in the absence of any growth effects from lower tax rates.54
Supply-siders believed that spending would fall automatically to some extent if expansionary tax cuts were enacted. They saw much government spending for such things as unemployment compensation and welfare as costs of slow growth that would become much smaller as employment rose.55 Hence, the supply-siders believed that the impact on the deficit from something like Kemp-Roth should account for both the revenue reflows and the automatic reduction in spending for cyclical spending programs. That is why Laffer always emphasized the impact of Kemp-Roth on the deficit and not just on revenues. In his first formal statement on the legislation in 1978, he said, “Kemp-Roth would partially redress the counterproductive structure of current tax rates leading to a substantial increase in output, and may well, in the course of a very few years, reduce the size of total government deficits from what they otherwise would have been.”56
Laffer also frequently talked about the effect of revenue reflows on all levels of government.57 Because state and local governments tended to run budget surpluses in the aggregate, higher revenues in that sector would add to national saving. In 1978 congressional testimony Laffer observed:
As I look at the Roth-Kemp bill, it cuts tax rates across the board over three years by approximately 30 percent. … On the federal level, there is quite a reasonable chance that within a very short period of time, a year or two or three, that not only will the cut in taxes cause more work output and employment, but the incomes, profits, and taxes, because of the expansion of the tax base, would actually increase. It is very clear to me that a cut in these tax rates, along the lines you suggested … would increase state and local revenues substantially. There is no ambiguity there. Any increase in incomes, productivity and production will increase state and local revenues substantially. If you take the government as a whole, it is likely that more revenues will increase.58
Laffer pulled all these points together in a 1979 academic paper in which he included higher federal revenues, higher state and local government revenues, lower government spending, and higher private saving in the reflows expected from a tax rate reduction. “The relevant question,” he wrote,
is not whether revenues actually rise or not but whether a change in tax rates is “self-financing.” Therefore, one should focus not only on the specific receipts for which the rates have been changed but also on other receipts, on spending, and on savings. Other receipts must rise if a rate is reduced. The expansion of activity will elicit a greater base upon which all other unchanged rates will obtain greater revenue. Government spending at all levels will fall because of lowered unemployment, reduced poverty, and thus less welfare. Likewise, government employees will require less in real wages because with lower tax rates the same real wages will yield greater after-tax wages, and so on. Finally, a cut in tax rates will yield greater savings in order to finance any deficit. Using a broader interpretation these tax rates and revenue positions should refer basically to the self-financing nature of tax rate changes.59
Laffer never made a precise estimate of the economic or revenue impact of the Kemp-Roth bill or the Reagan tax cut in 1981. The closest he ever came to saying that the Reagan tax cut would pay for itself was in a 1981 academic paper:
It is reasonable to conclude that each of the proposed 10 percent reductions in tax rates would, in terms of overall revenues, be self-financing in less than two years. Thereafter, each installment would provide a positive contribution to overall tax receipts. By the third year of the tax reduction program, it is likely that net revenue gains from the plan's first installment would offset completely the revenue reductions attributable to the final 10 percent tax rate cut. It should be noted that a significant portion of these revenues would accrue to state and local governments, relieving much if not all of the fiscal stress evident in these governmental units as well.60
These vague statements about relatively fast reflows from acrossthe-board tax rate reductions, however, contrast with more detailed estimates by Norman Ture and Michael Evans that explicitly included supply-side effects. Ture's estimate of Kemp-Roth in 1978 saw substantial revenues losses, net of feedback, even 10 years after enactment, when revenues would still be $ 53 billion (in 1977 dollars) below baseline.61 Evans's figures were very similar, showing a current dollar increase in the deficit in 1987 of $ 61 billion.62
Among the strongest supporters of Kemp-Roth in Congress was Representative David Stockman, Republican of Michigan, who became Reagan's OMB director in 1981 and, famously, broke with him over the problem of deficits.63 Stockman would often speak on the House floor and in committee in favor of passing an across-theboard tax rate reduction and against tax increases for the purpose of reducing budget deficits. For example, on March 1, 1978, he said,
A tax increase to achieve budget balance would be even less appropriate. Such a move would “crowd-out” output just as surely as more pump priming will “crowd-out” investment. Mr. Speaker, these considerations make clear that the time is ripe for implementing the only new fiscal policy idea that has been proposed in decades: A deliberate across-the-board reduction in marginal tax rates for the purpose of reducing the burden of government on the productive sectors of the economy.64
In testimony before the Senate Finance Committee on July 14, 1978, Stockman explicitly refuted charges that Kemp-Roth would lead to larger deficits and inflation:
These charges are based on a total misunderstanding of what Kemp-Roth is all about. We are not merely advocating a simple tax cut, an election-year gimmick. Instead, we view this measure as just one policy step in a whole, new fiscal policy program based on the supply side of the economy; based on the idea of getting more labor, capital, innovation, risk-taking and productivity into the economy by removing government barriers and deterrents, the most important of which, I would suggest to the committee today, is the rapidly rising marginal tax rates that Congressman Kemp has just discussed. … I would like to suggest to the committee that if this proposal that we are making is properly looked at, that these scare stories about these horrendous fiscal results, the deficits, cannot be validated at all. If you understand that we are substituting tax cuts and an incentive, supply-side approach for pump priming and demand stimulus, it can clearly be done with a large surplus produced within less than four years.65
There were many other occasions as well when Stockman defended unilateral tax cuts.66 Having worked with him closely at the time, I never heard the slightest concern from him that the basic supply-side message was not sound. And this was long before there was even the remotest prospect of Reagan being elected or of Stockman becoming OMB director. Remember also that Stockman supported former Texas governor John Connally during the primaries, endorsing Reagan only after Connally dropped out of the presidential race. So I conclude that he was not merely posturing about tax cuts in hopes of getting a cabinet appointment, but expressing a sincere belief.
In 1980 Reagan essentially adopted the Kemp-Roth bill as his principal campaign economic issue. After taking office, one of his first actions was to send a proposal to Congress on February 17, 1981, requesting passage of a tax cut closely modeled on Kemp-Roth. It showed a loss in revenue of $ 53.9 billion the first year, rising to $ 221.7 billion by 1986. No revenue feedback was assumed.67
Interestingly, old-time Keynesian economists were more optimistic about the potential revenue reflows of the Reagan tax cut than was the White House. For example, Richard Musgrave of Harvard, dean of America's public finance economists, testified before the Joint Economic Committee in early 1981 that the Reagan plan would likely recoup 18 percent of the static revenue loss through increased demand and another 30 percent to 35 percent through increased supply.68 Gardner Ackley, the CEA's chairman under Lyndon Johnson, compared Reagan's plan favorably to Kennedy's, saying,
I think the response to the proposed Reagan tax cuts would be similar to that of the Kennedy tax cuts. I think, yes, in a general way, taking the tax cut part by itself, independently of everything else. I think we would find a response of aggregate demand very substantially to a tax cut, and this would tend, as it did following 1963, to stimulate additional production and employment and investment. It would do so again today. The results of that would be beneficial.69
Laffer testified that it would take ten years before the Reagan tax cut paid for itself, a view he said was consistent with what he had said about Kemp-Roth.70 Joseph A. Pechman of the liberal Brookings Institution was largely in agreement with Laffer's assessment. Speaking at the same congressional hearing, Pechman said he was pleased
to hear that Arthur Laffer did not exaggerate some of the things that have been attributed to the supply-side economists. What he told us was that, if you reduce taxes or increase the net return to saving and to labor, there will be an increase in the incentives to work and to save. I think every economist, regardless of his persuasion, would agree with that.71
Again, Stockman spoke forcefully about the need for tax rate reduction to accompany budget control. Because of bracket creep and still-high inflation rates, future revenue projections always tended to show budget balance within reach a few years out. But spending always increased by a greater amount. Hence, tax reduction was essential to holding down the growth of spending, which was the Reagan administration's goal, not budget balance. As Stockman told the Senate Finance Committee at his confirmation hearing as OMB director:
It is my very strongly held belief that if we fail to cut taxes then we have no hope, over the next 3 or 4 years of bringing the budget into balance, and of closing this enormous deficit that we face again this year. Of course, there are those who will show you a paper projection, a computer run, and will try to demonstrate that if we can keep the rate of inflation high and allow the tax rates on businesses and individuals to continue to creep up, we will then automatically, by fiscal year 1983 or 1984, have a balanced budget. But that is pure mythology. That is only a computer projection. That is only a paper exercise that would never come true in the real world. We have had those forecasts made every year for the last 4 or 5, but as we have moved down the path toward the target year, these balanced budgets have seemed to disappear like the morning haze. There are reasons for that. The primary reason is that the tax burden today is so debilitating that it prevents the economy from growing, and without a growing economy we simply cannot hope to achieve a balanced budget.72
Eventually, Stockman broke with Reagan over the problem of budget deficits. But he always conceded that the tax cut was not really the cause of them because all it did was offset tax increases that would have resulted automatically from inflation. In his 1986 book, The Triumph of Politics, Stockman wrote:
The Carter revenue estimates assumed the greatest sustained period of income tax bracket creep in U.S. history. But when you started with an inflation-and bracket-creep-swollen revenue level and trotted it out four or five years into the future, fiscal miracles were easy. … With high inflation, the Reagan program amounted to little more than indexing the fiscal status quo; the Kemp-Roth tax cut simply offset bracket creep.73
A number of analysts pointed out that the Reagan tax cut was not even big enough to fully offset bracket-creep, including the New York Times, which attacked it for this very reason. Subsequent analyses confirmed that the Reagan tax cut did little more than effectively index the tax system, keeping aggregate revenues from rising as a share of GDP.74 Indeed, federal revenues as a share of GDP were actually higher in the decade of the 1980s than they had been in the 1970s and only slightly less than in the 1990s. Revenues averaged 17.93 percent of GDP in the 1970s, 18.25 percent in the 1980s, and 18.56 percent in the 1990s, according to CBO data.
With the emergence of large budget deficits after passage of the Reagan tax cut in 1981, the issue of its supply-side effects was answered in the minds of many. The simple cause-and-effect relationship seemed obvious: tax cut enacted, deficits emerged, therefore tax cuts caused deficits.75
Economist Lawrence Lindsey was the first to look at the effect of the Reagan tax cut on revenues after the fact, taking into account the economy's actual performance as opposed to projections based on forecasts and assumptions. In his initial effort, he concluded that on net the tax cut induced reflows of about 25 percent of the static revenue loss through behavioral effects. An estimate by some CBO economists came to a similar conclusion.76 Lindsey's final calculation was that reflows paid for about a third of the direct cost of the tax cut, including both Keynesian demand-side effects and supply-side effects. “So who was right about the effect of tax changes on the economy,” asked Lindsey, “the Keynesians or the supply-siders?”
The answer is both, at least in part. The Keynesians were right in claiming that such a substantial reduction in rates would powerfully boost demand, a point the supply-siders never denied but perhaps underestimated. The demand-side revenue feedbacks and the combined behavioral feedbacks (supply-side and pecuniary) turned out to be roughly equal. On the other hand, the revenue results vindicate the supply-siders' most important claim: The tax cut produced quite large changes in taxpayer behavior. That claim, strongly confirmed by results, ran directly counter to Keynesian theory and most Keynesian predictions. The combined supply-side and pecuniary effects recouped well over one-third of ERTA's estimated direct cost, a very powerful response.77
So if the tax cut was too small to fully offset bracket creep and feedback effects recouped a third of the static revenue loss, then where did the huge budget deficits come from? Obviously, much came from higher spending on defense and other programs, as well a deep recession in 1981 and 1982. But Paul Craig Roberts argues that most of it came, ironically, from the enormously greater success against inflation than anyone thought was possible.78
Remember, the conventional wisdom said that even without additional demand stimulus in the form of a tax cut, it would take many years to get inflation down from its double-digit level in 1980 to the low single digits. Indeed, a simple Okun's Law calculation would have suggested the need for something like another Great Depression to bring inflation down to tolerable levels.79
Since taxes are assessed on nominal incomes, not real incomes, the fall of inflation from 12.5 percent in 1980 to about 4 percent in 1982 and throughout the 1980s simply collapsed the expected tax base. Ironically, the Reagan administration was counting on inflation coming down fairly slowly—even though it was frequently attacked for being far too optimistic on this score—which would increase revenues from bracket creep even as tax rates were cut. The Reagan administration anticipated an increase in the GNP deflator of 36 percent between 1981 and 1986. It actually came in at 21 percent.80 When inflation came down far faster than anyone inside or outside the administration thought possible, revenues inevitably came in far lower than expected as well.
To put this effect into perspective, the Carter administration's last budget forecast 12.6 percent inflation in 1981 and 9.6 percent in 1982. It further estimated that each one percentage point decline of inflation below forecast would reduce revenues by $ 11 billion.81 With actual inflation coming in at 8.9 percent in 1981 and 3.8 percent in 1982, this suggests that lower-than-expected inflation alone increased the deficit by about 50 percent, adding $ 41 billion to the deficit in 1981 ($ 180 billion in today's dollars) and $ 64 billion in 1982 ($ 270 billion today).
It has been argued that budget deficits offset all of the stimulative effect of the 1981 tax cut.82 This extreme view, however, isn't shared by most economists. Even many of Reagan's political opponents concede that bringing down inflation so rapidly at far less economic cost than was previously imaginable was a remarkable accomplishment. Moreover, the rebound of growth and productivity in the 1980s, after the malaise of the 1970s, was at least in part due to the stimulative effect of the tax cut.
In 1989 Nobel Prize–winning economist Paul Samuelson of MIT admitted, “The latter half of the 1980s, historians will recognize, has been an economic success story.” Even Bill Clinton's CEA conceded that the 1981 tax cut had been a major factor in stimulating growth. “It is undeniable that the sharp reduction in taxes in the early 1980s was a strong impetus to economic growth,” it said in the 1994 Economic Report of the President.83