Keep From All Thoughtful Men Read online

  The Statistical Revolution

  The statistical revolution is the story of two men, Nobel laureate Simon Kuznets and economist Robert Nathan. Although these men and their activities remained closely connected, in order to delineate the narrative their story must be told in two parts. The first outlines Kuznets’ pioneering work in the field of national income accounting. The second examines how Nathan applied Kuznets’ work to determine the feasibility of the Victory Program. Interestingly, Nathan was a former student of Kuznets and later was employed by the latter to help work on national accounts for the Commerce Department in the 1930s.43 Still later, on the eve of war, Nathan moved to the Office of Production Management (OPM) to build a statistical abstract of the country’s ability to support an expanded war effort.44 When Nathan later took over the Bureau of Planning and Statistics, a division of OPM, he immediately rehired his former boss and teacher, Kuznets, as his deputy, to help make sense of the masses of economic data assembled to manage war production activities.45

  Simon Kuznets was the recipient of the third Nobel Prize in economics and was a pivotal figure in the transformation of economics from a speculative and ideologically driven discipline into an empirically based science.46 Born in Pinsk, Russia, in 1901, he and his family emigrated to the United States in 1922. He entered Columbia University and earned his B.A. in 1923, his M.A. in 1924, and his Ph.D. in 1926. In 1927 he went to work for the private National Bureau of Economic Research (NBER) where he remained, with brief interruptions for government service, until 1961.47

  From 1932 to 1934 he served in the Department of Commerce, where he constructed the first official estimates of U.S. national income and laid the basis for the department’s national income section. His work during this period established the statistical base that later allowed Nathan to engage in a major battle with the Joint Chiefs of Staff (JCS) and bring about a dramatic change in U.S. strategy for fighting the war, particularly the war in Europe. It is almost impossible to overestimate the effect of Kuznets’ work on the planning and conduct of America’s wartime mobilization. While the United States would still have undertaken the production miracle that swamped the Axis, it would likely have taken years longer and been accomplished only after a great waste of resources.

  To see why and how this was so one must understand what national accounts are, how they came into being, and what they were designed to accomplish. National income and product accounts are the comprehensive set of accounts measuring the total value of final goods and services (often called GDP) produced by the U.S. economy and the total incomes earned in producing that output. This integrated set of accounts and similar detailed sets of regional and industrial accounts allow for the comprehensive analysis of the impact of alternative policy actions, or external events, on the economy, as well as on detailed components of final demand, income, and regions of the country.48 Though the work behind these national accounts often appears arcane and filled with drudgery, the final products have become instrumental in setting policies that keep a modern society functioning.

  The concept of national income goes back to Sir William Petty and Gregory King, who developed estimates of English national income during the seventeenth century. Later economists, including Adam Smith, debated the concept of national accounts and how economists could measure them. Economists were not, however, able to create practical methods to determine or influence government policy until the early twentieth century. For centuries, measurement and application had lagged behind economic theory due to the absence of detailed accurate data on which to build estimates.49

  Prior to World War I, published national income estimates were the work of individuals.50 Most of these attempts were not rigorous and primarily concerned the collection of data on the burden of the public debt or on various taxation schemes. It took the exigencies of a world war to prompt the first American attempts at national product estimates for senior government officials to analyze and use. Adolph C. Miller, an economist with the Federal Reserve Bank of New York, prepared estimates to evaluate the “surplus over necessary consumption and maintenance of capital that could be devoted to the war effort.”51 Moreover, during the war the NBER undertook its first project to study national income—its size, year-to-year variation, and distribution. The purpose of these estimates was to provide an impartial and trustworthy source for use as a basis for correcting social and political problems.52

  It was not until 1926, however, that the first official government estimates, from the Federal Trade Commission, were presented. These estimates, prepared by Francis Walker, showed the value of national product produced by U.S. industry for the years 1918–23. Unfortunately, after completion of this first report the government halted all follow-up work due to a lack of funding. Washington bureaucrats realized the need for such reports only several years later as the economy sank into the Great Depression and the government began to feel the keen lack of a suitable statement on national accounts that politicians could use to guide their efforts to rehabilitate the sinking economy. The NBER (a private organization) tried to provide the required data for policy makers, but it failed to update its earlier estimates and it took many months to calculate new ones. Congress became fully aware of the extent of the problem when in 1931 it called government and private experts to testify on the current economic crisis. Congress was astounded to discover that no expert could provide any national account figures more recent than 1929.53

  At the time, the Depression had been worsening for two years. Nearly a quarter of U.S. workers were in the lines of the unemployed, while many of those who still had jobs were working on a part-time basis. Asset values had plummeted, the banking system was collapsing, deflation was reversing the gears of the economy, and sales were insufficient to maintain businesses. Farm income, on which one-fourth of the population depended, had fallen by half. Neither the public nor elected officials understood the workings of an economy that appeared to be perpetuating the crisis, nor did they know quantitatively its scale and scope. The most up-to-date estimates of national income—that is, economy-wide income—were for 1929, a boom year, that had been marred only by the October stock market “crash,” after which the economic slide began.

  Realizing that it was trying to direct economic policy without the necessary tools, in June 1932 Congress passed a law ordering “That the Secretary of Commerce report estimates of the total national income of the United States for each of the calendar years 1929, 1930, 1931, including estimates of the portions of the national income originating from agriculture, manufacturing, mining, transportation, and other gainful industries.”54

  The person originally placed in charge of this project, Frederick Dewhurst, soon quit, and the Commerce Department borrowed Simon Kuznets from the NBER to take over. Kuznets had already been working on a new set of estimates for the bureau, where he had been busily remaking the nation’s basic economic models.55 To help him, the Department of Commerce provided Kuznets with a small staff of assistants, among whom his former student, Robert Nathan, was the leading light.

  Almost a year to the day after assuming control over the project, Kuznets provided Congress with the Commerce Department’s first complete set of estimates. These estimates indicated that between 1929 and 1932 national income had dropped by more than 50 percent. Incomes in manufacturing had dropped by 70 percent, while incomes in construction had dropped by more than 80 percent. Government was the only industry that had grown over the period, although the federal government remained relatively small—federal tax receipts claimed only 3 percent of GDP in 1932.56 The finished product, National Income, 1929–32, was printed as a Senate document in 1933.57 The report—261 pages of tables and explanation—was the Senate’s first bestseller, and the initial run of 4,500 copies immediately sold out at twenty cents a copy.

  The report also found that between 1929 and 1932 national income had fallen by more than 50 percent and national income paid out had fallen by 40 percent, while business savings had b
ecome negative in 1930 and remained negative through 1935 as businesses drew down their financial reserves or borrowed to remain in operation, when fixed costs and wages exceeded revenues. In terms of the new statistics, this meant that national income paid out exceeded national income produced.58 The United States was, in economic terms, a losing proposition.

  Soon after completion of the estimates, Kuznets returned to the NBER and Nathan took over the ongoing work on future economic estimates. By May 1938 Nathan was presenting estimates not only of total income, but also of income per capita broken down by types of payment and individual states. In addition to business taking an interest for the first time (to help in marketing efforts), these new, more-discrete estimates became the basis for targeted federal relief efforts in each of the states. From this point, the government published these estimates annually, and these statistical methods remained basically unchanged until the mid-1950s.59

  Economic historians credit Kuznets with transforming the field of national income accounting by bringing a far greater precision to the task than economists had ever achieved before. He accomplished this by rooting his estimates firmly in classical economic theory and by solving numerous problems related to using imperfect raw data to create the theoretical conception of “national income.”60 In later publications, the Department of Commerce heralded the creation of Kuznets’ national income estimates as one of the great inventions of the twentieth century.61 Nobel laureate Paul Samuelson said of them, “That for the first time the President, Congress, and the Federal Reserve had a methodology to judge whether the economy was contracting or expanding, and whether it needed a boost or should be reined in a bit.” Without measures of economic aggregates like GDP, Samuelson believed policy makers were adrift in a sea of unorganized data. The GDP and related data acted like beacons that helped policy makers steer the economy toward key economic objectives.62

  During the 1930s, national income statements became regularly issued products of the Department of Commerce and were generally accepted as the broadest readings on U.S. economic conditions.63 The public followed them closely, while the Roosevelt administration and Congress used them to plan and evaluate fiscal policy. By the time Commerce first provided GNP figures in 1942, national income statements had already become the most cited U.S. macroeconomic statistics. The Roosevelt administration realized that these new measures provided an authoritative means of describing the dire economic conditions that its proposed New Deal programs would address, thus it included them in the marketing of these programs to the public.

  For example, within two weeks of the release of the first national income report, Secretary of Commerce Daniel C. Roper was already citing the more than 50 percent drop in national income between 1929 and 1932 in speeches around the country explaining New Deal programs.64 Although few if any policy makers thought about it at the time, it did not take a great logical leap to see that, if the government could use these statistics to drive policies for the Depression-era economy, then they would prove an equally valuable tool in managing the economy and military production during a major conflict.

  To see just how valuable a policy instrument they proved, one can note the long-term effects engendered by these estimates. According to the U.S. Bureau of Economic Analysis, the national accounts contributed to a reduction in the severity of business cycles and a post–World War II era of strong economic growth. Prior to World War II the lows of the business cycle were much more severe and more frequent. There were six severe depressions between 1854 and 1945, each with an average duration of nearly three years. Including recessions as well as depressions, the average downturn between 1854 and 1945 was twenty-one months, with a contraction occurring on average once every four years. During the postwar era the length of the average contraction has been halved to eleven months, with a contraction occurring once every five years.65 The postwar era also stands out as a period of unprecedented growth for the United States. Real GDP per capita and real wealth have almost tripled since 1948. This period of economic prosperity has not only dramatically improved standards of living but also has contributed to large improvements in social conditions, cutting poverty in half, raising living standards, and adding to leisure time. Bank runs, financial panics, and depressions that were recurring problems before World War II became a thing of the past. This success was due in significant part to the timely, comprehensive, and accurate data on the economy provided by the national accounts.66

  In 1935 Nathan began writing a series of annual survey articles presenting the national income statistics for the preceding year, analyzing them in detail. The next year, the Department of Commerce published a statistical compendium, National Income in the United States, 1929–35, which presented revised and extended statistics and explained the concepts behind the report.67 Roosevelt began citing national income statistics in speeches as early as 1935, for example, in his statement of September 1935 on the state of the economy and the federal budget. In April 1938, in his message to the Congress requesting additional spending for the new recovery program to address problems caused by the 1937 recession, the president described economic developments over the period 1929–37 in national income terms.68 And he described the goal for the program in national income terms as well: “We must start again on a long, steady, upward incline in national income.” Starting with the annual budget message to the Congress in January 1939, which presented his fiscal year 1940 budget, the president cited national income statistics as the primary measures of the state of the economy.69 In his 1939 message, he also highlighted the importance of these measures to economic policy by suggesting how different levels of national income would generate different levels of federal tax receipts.70

  The critical element to grasp from the president’s use of national income statistics to support his policies is that he trusted their accuracy and apparently believed that he could use them to plan national economic strategy. This trust and belief was not limited to the president. His advisers, government economists, and any politician who thought of himself as a “New Dealer” shared the president’s faith. This common shared faith in the efficacy of national income statistics would be of critical importance in the final showdown between the economic planners and the military over the feasibility of the Victory Program.

  Gross National Product and World War II

  The Department of Commerce launched GNP statistics to answer a policy question for which the national income accounts undertaken by Kuznets and Nathan were proving inadequate. Could the nation meet the Victory Program requirements? If so, at what costs would it be to the civilian standard of living and price stability? As was the case for national income accounts in 1932, the GNP concept elucidated in 1942 was not new, having been discussed and partially formulated during the 1930s. While economists made progress in developing theoretical and statistical standards for GNP, it took a policy requirement (the requirement for economic information during world war) to push the U.S. government to develop an authoritative, consensus-based statistical measurement. When finished, GNP made up the other side of the national income equation—the production side, to match the income-earned side (approximated by national income). It thus helped provide a more complete picture of the economy.

  In his budget message to Congress in January 1940, four months after Germany had invaded Poland, Roosevelt asked for a modest defense supplemental appropriation for fiscal year 1940 and a likely increase in defense spending for fiscal year 1941. In 1940 defense expenditures had reached more than $1 billion, approximately 14 percent of the budget. In his January 1941 budget message, Roosevelt asked for $25 billion in defense expenditures, 62 percent of the budget, reflecting a “world at war.” In his January 1942 budget message, President Roosevelt asked for $53 billion for defense, 90 percent of the budget, reflecting “a nation at war in a world at war.” It was the computations of the economists creating national income and GNP accounts that provided the president and his advisers with the basis for estima
ting what the United States was capable of supporting in terms of a rapid and massive growth in military spending.

  During the week before the January 1942 budget message and shortly after the attack on Pearl Harbor, the president announced the goal of increasing the share of national income spent on war production from the current 17 percent to 50 percent by 1943. Such an increase in the speed and scale of the mobilization program was beyond experience, and was “A national effort of gigantic magnitude,” according to the president. To accomplish this task, the United States would have to sustain major economic dislocations as it readjusted to a wartime economy.

  The U.S. rearmament program, begun in 1940, had already boosted national income above the 1929 level for the first time, to achieve an almost 25 percent increase before Pearl Harbor. The rise was steep, and by December 1941 national income was almost 40 percent above its level of less than two years earlier. Putting the country on full war footing would boost income even more, but purchases of consumer goods and services, which had boomed in 1941, would find themselves stymied because the country would need to cut back production for civilian purposes to make way for the war’s programs. Rationing, wage and price controls, and other consumption-damping regulations were on the table. Statistics measuring the total amount as well as the composition of goods and services being produced were a base requirement for evaluating the risks of shortages of civilian goods and services and the bidding up of prices, but those statistics were not available in the United States at the beginning of 1942. National income sufficed at that time as an informed measure of the economy’s size, but it was not up to the task of evaluating production constraints and trade-offs because it measured only the income earned in production and not the greater market value of the goods and services produced. In these early days, using current economic statistics for mobilization planning was akin to a corporation bidding on contracts without knowing the capacity of its factories or the financial facilities at its disposal.