Macroeconomics and Empirical Analysis

Jevons Theory Of Sunspot

Empirical work in microeconomics is difficult, but in macroeconomics it is even more difficult because everything tends to be interrelated. One of the earliest contributions was made by W S. Jevons.

William Stanley Jevons's Sunspot Theory, Jevons Theory Of Sunspot

W S. Jevons (1835-1882) was one of the pioneers in mathematical techniques and utility theory. For that work he was highly lauded. Although Jevons is now best known for his microeconomic contributions to neoclassical theory, it is his empirical attempts at measuring macroeconomic relationships that are best known in the history of econometrics. His was one of the early attempts at formal macroeconomic empirical work. Whereas his work in microeconomics was praised, his macroeconomic statistical work on economic cycles was not well received by the profession. In fact, it was often ridiculed.

Jevons was interested in discovering the cause of trade or business cycles that led to fluctuations in prices. Because cyclical behavior did not seem related to individual utility maximization behavior, he thought there must be a cause in nature—some natural phenomenon that led to fluctuations. Preliminary investi­gation led him to believe that the cause of fluctuations in economic activity probably had something to do with the weather. He focused on sunspots (fluctuations in the activity of the sun) as the probable cause.

Jevons's specific hypothesis was that sunspot cycles occur with a periodicity of 11.1 years and these cycles lead to weather cycles that, in turn, lead to business cycles. To test his theory, Jevons looked at agricultural data, about fluctuations in the harvests, that were available from the thirteenth and fourteenth centuries. He then tried to relate those harvest fluctuations to nineteenth-century estimates of sunspot activity that suggested the 11.1-year periodicity. Making the assump­tion that the length of the sunspot cycle had not changed, he compared the two by laying out his data on a grid representing eleven years, eyeballing his data. He noticed a relatively good "fit"; the cycles seemed to match. He then looked at cycles in commercial credit during the nineteenth century and discovered an average cycle of 10.8 years. He concluded that the likely cause of the business cycle was sunspots.

Jevons's sunspot theory was not followed by economists in the nineteenth century, most of whom found it quite bizarre. It deserves mention primarily because it involves an attempt to use statistics to develop and test a macro-economic theory and thereby establishes Jevons as a pioneer of econometric methods.

Moore's Contributions to Macroeconometrics Models

The trade cycle has been an enduring economic phenomenon, and it is not surprising that Moore's contributions to estimating the demand for crops were supplemented by a macroeconomic contribution. In macroeconometrics, Moore's main contribution was to provide both a theory of the trade cycle and an attempt to measure it statistically. In analyzing the trade cycle, Moore's dynamic demand curve had even more justification. Whereas a static theory of demand existed, no similar theory of the trade cycle existed. Moore argued that a priori and ceteris paribus reasoning would be useless in explaining such fluctuations.

Like Jevons, Moore chose weather cycles as the exogenous cause of economic fluctuations. He integrated this view with his finding of an upward-sloping demand for pig iron as an explanation of the trade cycle. His five-part argument went as follows:

1. Rain increases and crop yields rise.
2. Balance of trade rises.
3. Demand, price, and the quantity of producers' goods rise.
4. Employment rises, so demand for crops rises.
5. General prices rise.

When rain decreased, the process would be reversed. Moore used statistical analysis to support this argument.
Moore's initial analysis was criticized by Philip Wright, who in a 1915 article adjusted the measure of rainfall to what was relevant for growing rather than the total yearly amount of rainfall. Wright showed that the statistical relationship breaks down. Wright's argument effectively undermined Moore's statistical analysis, which led Moore to expand his coverage to more countries. In his expanded analysis, Moore found a persistence of eight-year cycles. In 1923 he wrote a second, more careful book on the subject in which he used weather as only one of many economic and social causes.

The broadening of the theory to include multiple causes found sympathetic support, but his deeper analysis of weather cycles, in which he emphasized that Venus comes between the earth and the sun in eight-year intervals, led to the designation of his theory as "the Venus Theory of trade cycles." Moore's particular approach to trade cycles was not pursued by others, but it provided some basis for later econometric work and set the stage for the institutional economists' analysis.

Wesley Clair Mitchell: Heterodox Empiricist

One of the early institutionalists, Wesley C. Mitchell, differed significantly from orthodox neoclassical economists on issues of empirical work in macroeconom­ics. Moore's work provides a useful focal point from which to consider Mitchell's approach to empirical work, which developed during the first half of the twentieth century and was the initial approach of the National Bureau of Economic Research (NBER). One of the reasons that Mitchell's approach gained favor was that there were problems with Moore's more formal statistical approach.

Mitchell's view of the appropriate relationship between theory and factual analysis was laid out in his early work on business cycles:

One seeking to understand the recurrent ebb and flow of economic activity characteristic of the present day finds these numerous explanations [of business cycles] both suggestive and perplexing. All are plausible, but which is valid? None necessarily excludes all the others, but which is the most important? Each may account for certain phenomena; does any one account for all the phenomena? Or can these rival explanations be combined in such a fashion as to make a consistent theory which is wholly adequate?

There is slight hope of getting answers to these questions by a logical process of proving and criticizing the theories. For whatever merits of ingenuity and consistency they may possess, these theories have slight value except as they give keener insights into the phenomena of business cycles. It is by study of the facts which they purport to interpret that the theories must be tested.

But the perspective of the investigation would be disturbed if we set out to test each theory in turn by collecting evidence to confirm or to refute it. For the point of interest is not the validity of any writer's views, but clear comprehension of the facts. To observe, analyze, and systematize the phenomena of prosperity, crisis, and depression is the chief task. And there is better prospect of rendering service if we attack this task directly, than if we take the round about way of considering the phenomena with reference to the theories.

This plan of attacking the facts directly by no means precludes free use of the results achieved by others. On the contrary, their conclusions suggest certain facts to be looked for, certain analyses to be made, certain arrangements to be tried. Indeed, the whole investigation would be crude and superficial if we did not seek help from all quarters. But the help wanted is help in making a fresh examination into the facts.

Mitchell's approach was pragmatic; it did not see a significant role for actually testing theories but instead saw theories as a backdrop useful in interpreting empirical observation. Consistent with this view, Mitchell saw economics not as a science but as an art to aid in policy formation. Ultimately, for Mitchell there was no unchanging theory that could be specified in a neat model; the economy was far too complicated and was undergoing continual structural change. Given such complicated changes, creating general theories was game-playing; the only acceptable theory was educated common sense, and the economy could be understood only through careful integration of common sense and statistical analysis.

Although data were, in a formal scientific sense, inappropriate to test theories, data were appropriate to test various hypotheses about the behavior of cycles. In a later book, Measuring Business Cycles (Burns and Mitchell, 1946), Mitchell tested Schumpeter's hypothesis about the relationship among different cycles and rejected it. He also tested his own hypothesis that there was a long-term secular change in cyclical behavior, which he and his co-author also rejected. They did find changes, but they were irregular and random. Thus, they were able to "test" hypotheses informally through a combination of formal statistical tests—such as correlation tests and F tests of significance—and judgment based on knowledge of institutions and of the data. Whereas in the scientific approach the formal tests determine the validity or falsity of a theory, for Mitchell such tests were merely aids to common sense and subjective judgment. Mitchell's approach to data and economic empirical analysis was used by the mainstream macro-economists in the United States during the 1930s.

Measurement and Data Collection

Some data, such as the price of coal, can be simply collected and used. Often data that fit theoretical constructs must be constructed. Quantifiable concepts must be determined, and then data must be collected. That work is often difficult and demanding. Let's consider some examples.

Economists use a concept of the general level of prices, but there is no such measure of all prices in the economy. Since the 1940s, increases in the general level of prices (inflation) have received a good deal of attention. Before inflation could be measured, a considerable amount of work had to be done on the construction of measures of the general level of prices. In The Making of Index Numbers (1922), Irving Fisher (1867-1947) examined some of the problems encountered in constructing index numbers designed to measure prices and economic activity. It was hoped that by eliminating intermediate goods and appropriately weighting final goods, one could construct an index that measured changes in the general level of prices. With such a measure it would be possible to add greater precision to the concept of inflation and to test hypotheses about the relationships between changes in the money supply and changes in prices. It should be noted that the money supply does not exist as simple data to be collected and analyzed: measures of the money supply must be constructed. There are many economists who have spent much of their lives working in the areas of measurement and data collection.

Other economists have made major contributions in the area of national income accounting. Keynesian theory cried out for measures of national income, consumption, spending, savings, and investment spending. These macro-theo­retical concepts required the resolution of tremendously difficult conceptual issues before quantitative work could be done in data collection. Sir Richard Stone (1913-1991) and James Meade (1907-1995), both Nobel Prize winners in economics, developed national income accounts for Britain that fit the Keynesian theoretical mold.

In the United States, national income accounting was studied by Simon Kuznets (1901-1985), a Nobel Prize winner, who wrote his Ph.D. dissertation under Wesley Mitchell and went on to work under Mitchell at the National Bureau of Economic Research. Kuznets's major contributions were the construc­tion of measures of national income for the United States and the use of statistics to measure and compare growth patterns in different countries. The national income accounts Kuznets helped to create were an important ingredient in the post-Keynesian macroeconometric models.

Wassily Leontief, also a Nobel Prize winner in economics, played a role in organizing the collection of data; he designed input-output analysis, a practical planning tool for handling interrelationships in the economy. Leontief has been strongly critical of modern mainstream economic model building, which is devoid of empirical content. He advocates concentrating on the practical appli­cation of economics and working with data rather than building sophisticated mathematical models. His paper, "Theoretical Assumptions and Nonobserved Facts" (1971), is one of the best criticisms of ivory-tower model building one can find.

Two other economists who made contributions in data collection were Abram Bergson (1914- ) and Alexander Gerschenkron (1904-1978). Bergson was an able theorist who wrote a classical paper in welfare economics at the age of twenty-four, while he was a graduate student. He became the dean of American Sovietologists and did seminal work in the measurement of economic activity in the Soviet Union. Before the breakup of the Soviet Union, it was often alleged that Soviet planners used measures of their economy's activity that were gener­ated in the United States, because they were more reliable than their own statistics. Bergson was instrumental in building the Russian Center at Harvard University into a major research center for the study of Soviet society.

Gerschenkron was born in Russia, like Kuznets, but received his economic training in Vienna during the 1920s. He was a colleague of Kuznets and Bergson at Harvard. Although the volume of his written work is small, he was a professor's professor who had a command of many languages and published critical essays on Pasternak and on Nabokov's translation of Pushkin's Eugene Onegin. Gerschenkron did important work on the measurement of growth, particularly in the Soviet Union, and was able to show how the selection of the base year used for an index of industrial production influenced the growth rates shown by that index. His research disclosed that Soviet growth was not as rapid as Soviet planners indicated, because their measurements were biased.