Teaching the New Economics

At MIT, Samuelson taught the ideas he was working on in two courses, Business Cycles and Public Finance. When he took over the first of these courses, he changed its description to reflect the new approach to economics, offering “A statistical, historical, and theoretical examination of the determi­nants of income, production, and employment.

However, although Samuelson praised Keynes for providing a logical system, he struggled to explain the interdependence of its various parts. Samuelson tried a complicated “four quadrant” diagram but it contained some very awkward constructions and was cumbersome to use. One of the criticisms that he made repeatedly was that both the Keynesian and the clas­sical theories neglected forms of wealth other than money. The demand for money in the Keynesian system should depend on these other assets. At one point he suggested that total wealth could be estimated by calculating the present value of income, excluding wages and salaries (the sum of money that would, if invested at the market rate of interest, yield that part of national income not accounted for by wages and salaries).^k Such criticisms appear remarkably prescient, given the role that wealth effects were to play in mac­roeconomics in the late 1940s and 1950s.

Theory dominated Samuelson's course, though data were never far away—hardly surprising, given that he was simultaneously commuting to Washington to discuss issues that he was also discussing in class.^l For exam­ple, citing data that he was using at the NRPB on saving by households in different income brackets, Samuelson paid close attention to the distribu­tion of income, explaining Pareto's law of distribution and using the Lorenz curve and the Gini coefficient to measure inequality.^m The distribution of

j. Ringo, the student whose notes are the source here, noted “PAS. real contribution of General Theory is that it is a logical, interdependent system and takes account of the

problems of ineffective demand.”

k.

He wrote “W =

mY

i

= present value of [income], excluding wages and salaries

(m = roughly 33%).” (“Wealth” has been replaced with “income,” presumably being a note-taking error by Ringo.) The demand function for money should therefore be mY ]

i /

l. Students were required to write term papers on the aftermath of a single war, perhaps the Civil War or the Napoleonic War, a topic that directly paralleled the work he was doing on the First World War.

m. The Lorenz curve is constructed by ranking income recipients from poorest to richest, and then plotting the proportion of income earned against the proportion of the population. If there is perfect equality, the bottom 10 percent of the population will have 10 percent of income, the bottom 50 percent will have 50 percent, and so on. The result will be a straight line at 45 degrees to the horizontal axis. If there is inequality, it will trace out a income, which changed over the cycle, mattered because of its effects on saving. However, because the marginal propensity to consume did not vary very much across households, the effects of redistribution on consumption would be small: it had to be justified on social grounds, not purely because of its effects on saving.

In the course of explaining how saving and investment determined income, he made a number of significant remarks. Samuelson noted that full employment was not a definite point, and he used the term “hyster­esis effect” to denote movements that changed a schedule in an irreversible way. Whereas in print he might have to use generally accepted terminology, in his course he explained his own preferences more forcefully. The student noted: “Samuelson doesn't like ‘ex ante' and ‘ex post' ” and “Doesn't like ‘planned' vs. ‘Unplanned'.... Confusion of schedules, expectations, iden­tity.” The crucial point Samuelson was making was that what mattered was whether investment was maintainable. Ex post, saving and investment would be equal, but investment would not be maintainable except in equilibrium. To avoid having to “drag in a little angel in order to make the process instan­taneous,” it was better to talk in terms of “observable” versus “virtual” mag­nitudes. If this account is not completely clear, it may reflect the fact that, though Samuelson understood the arguments, he had not settled on a par­ticular way to express it.

Even national accounting concepts were unsettled, for economists work­ing in government agencies were then developing different frameworks for the national accounts.ⁿ This no doubt explains why, certainly to modern readers, his discussion appears to lack clarity. What is perhaps most notice­able is the absence of the notion of value added (he simply treated the private sector as a single unit) and the absence of any reference to Leontief's input­output analysis that was later used to clarify the way the national accounts were organized.

When Samuelson turned to government spending, he introduced the balanced-budget multiplier that he and Bill Salant had recently worked out

curve below that line.

n. Samuelson mentioned Kuznets and Milton Gilbert, but he could also have mentioned Robert Nathan or, in Britain, James Meade and Richard Stone. but not yet published. Again, his discussion reflected the extent to which such issues were discussed but had not yet stabilized. There was still no agreement on whether the multiplier should be applied to government expenditure or to the government deficit, and exactly what should be included in the multiplier. He illustrated this by citing the claim by Federal Reserve Chairman Marriner Eccles that the “way to balance budget was to spend freely, raise income, and then collect taxes out of the higher income.” Eccles had, Samuelson explained, forgotten to include the marginal propen­sity to tax in the multiplier: his conclusion might be right, but his logic was wrong.

The multiplier was also topical owing to the problem of America's wartime relationship with its allies and debates about the need to support European countries after the war; it provided evidence to counter the conservative claim that the United States could not afford to support other countries, for additional spending might lead to additional output. Samuelson used multi­plier reasoning to establish whether foreign lending would reduce domestic investment by raising interest rates, and whether it would raise exports. The latter might happen through two mechanisms: the “classical mechanism,”’ whereby the foreign loan altered prices in the two countries; and the “modern mechanism,” which changed spending by changing incomes.

Samuelson also extended the Keynesian model to allow for changes in the capital stock. Here, too, he was grappling with problems for which there was no accepted mode of analysis. Clearly the problem of capital accumulation mattered, but the reason why he discussed it was probably that he wanted to address the argument made by Oskar Lange about the existence of an opti­mum propensity to consume.^o Discussing this required him to talk about capital accumulation, for Lange's arguments raised the question of whether it was appropriate to maximize national income, the capital stock, or the rate of investment.

Samuelson had touched on fiscal policy in his course on business cycles in the first quarter, but went into it in much more detail in the course he taught on public finance during the summer. Public Finance was a new course. It focused on public works spending, deficits, and their effects on aggregate demand and employment. Even though Samuelson covered a lot of theory in the lectures, the reading exposed the students to up-to-date institutional and empirical material, including several NRPB and NBER reports.²⁶ If there was a textbook for the course, it was Deficit Spending and the National Income by

o. See chapter 18 this volume.

Henry Villard (1941).^p It probably approached the topic much as Samuelson would have done. Its coverage—90 pages on the business cycle, no fewer than 160 pages on the multiplier, and 150 pages analyzing “recent public net income-increasing expenditure”—shows clearly the state of discussion: the concept of the multiplier was still one that had not stabilized and needed detailed explanation. One reason for this was that the multiplier was tied up with definitional issues necessary to use it in practice.

Samuelson started the course by noting that government activities had to be appraised on the basis of their direct usefulness and their influence on effec­tive demand. The latter was important because public works were the “only respectable weapon government has which operates in anti-cyclical manner” (he seems not to have said what the non-respectable weapons were). After review­ing discussions of public works in the thirties, Samuelson jumped ahead to the postwar situation and the consequences of demobilizing 10 million men and women. He made the same point as was obsessing him in Washington— namely, that there had been insufficient planning to implement the public spending needed to prevent unemployment. He presented the theory about public works spending in terms of three schools of thought. The first school, associated with Herbert Hoover, was that policy was all about adjusting the timing of public works that would be undertaken regardless.

Samuelson defended Hansen's policies against two accusations that con­servatives often made: that they were socialist and that they would lead to an unsustainable burden of government debt. He contended that, with the exception of public utilities such as the Tennessee Valley Authority and the Boulder Dam—both enormous projects—public works spending had not led to socialism. Though there was a sense in which public works competed with private activities, the effect was tiny. Also, as federal spending rose, state and local government spending, especially relief, fell. Samuelson defended Social

p. Ringo noted that it should be “read ‘copiously.’ ” Security spending, arguing that Roosevelt’s policies did not involve any break with previous policies and that they would not undermine the private enterprise economy. Indeed, there were good reasons why the Social Security Act, passed in 1935, was the most popular New Deal measure. Samuelson countered a range of arguments about why government debt was a problem, ranging from fear that the government would one day default to the idea that high values of debt were inflationary: it was government spending, not government debt, that was inflationary, and it was interest payments, not the debt itself, that was a burden. Britain had been able to bear a high debt burden for hundreds of years, and in any case, much American debt was internal, involving just transfers within the United States. He dismissed the argument, made by Gayer, that government debt had been the cause of the 1937 depression. Though he clearly sided with Hansen on much, he dis­missed as “nonsense” Hansen’s idea that a consumption boom would stop as soon as the stimulus was withdrawn, whereas an investment boom would be self-sustaining.

Despite Samuelson’s convictions that mathematics was important, he used it sparingly.⁴ Samuelson obviously believed that the important issues that students needed to think about were not ones to which answers could be found simply by solving sets of equations.