Gerhard Tintner’s Review of Business Cycles in the United States of America, 1919–1932 (1941)

Gerhard Tintner’s book review assesses J. Tinbergen’s attempt to construct and statistically test a comprehensive model of American business cycles. The reviewed book, published in 1939 as the second volume of Statistical Testing of Business-Cycle Theories, belongs to the League of Nations investigation of cyclical fluctuations and follows a study of investment activity. Tintner’s central judgment combines admiration for the integration of economic theory and statistical analysis with reservations about the reliability of particular methods and conclusions. He establishes the scale of Tinbergen’s project at the outset:

The author has endeavored nothing less than the construction of an economic model which should take into account all the important factors and economic variables influencing business cycles in the United States during the post-war period.

For Tintner, this ambition helps explain both the study’s importance and its imperfections. His review moves from statistical foundations through the model’s component relationships to its implications for policy and competing cycle theories. Throughout, he distinguishes the achievement of constructing an extensive empirical system from the stronger claim that its results securely establish economic causation.

Tinbergen uses multiple-correlation analysis, expressing variables as deviations from a trend designated as “normal.” Relationships are predominantly linear, justified by the possibility of approximating functions locally, while regression coefficients are assumed constant despite acknowledged changes over the period. Tintner notes the attention paid to collinearity but argues that the distinctive problems of time-series analysis receive insufficient treatment. Ordinary significance tests cannot simply be transferred to these data without modification:

It would be necessary at least to test the residuals for randomness and especially to make certain that no bias has been introduced; for instance, there should be no positive or negative correlation between the residuals.

This objection bears directly on the evidential status of the model. Tintner wants scrutiny of the unexplained component before treating estimated relationships as dependable. Tinbergen’s procedure constructs a system with as many equations as variables, introduces lags empirically, and eliminates variables to obtain a dynamic equation with oscillatory solutions. Its usefulness for testing cycle theories therefore depends on the adequacy of both the economic specification and the statistical treatment.

The review describes the model’s successive components: definitional relationships, demand for goods and services, supply and price equations, money and capital markets, and income formation. Nonlinear relationships are often replaced by linear approximations. Demand equations introduce lagged effects; wage rigidity is represented through a lag in an equation relating wages to employment, living costs, productivity, and institutional factors. Stock prices require more complicated relationships because linearity breaks down. Tintner emphasizes the computational scale: the system includes twenty capital variables, twenty-four income-related variables, fourteen price variables, ten physical-quantity variables, and others.

Eliminating all variables except profits yields the final equation used to investigate cyclical movement. Without a stock-exchange boom or hoarding, high profits stimulate consumption and investment, while commodity stocks and speculative gains also matter. Introducing speculation and hoarding changes the movement substantially. Building activity appears to follow a partly independent, longer cycle associated with the life span of houses. The model thus differentiates mechanisms and conditions rather than assigning every fluctuation to a single cause.

Tintner regards the policy deductions as especially ambitious. Compensatory public investment produces greater damping and shorter cycles; consumption stabilization produces still greater damping. Wage rigidity has negligible influence, whereas price stabilization increases fluctuations. These results are explicitly conditional:

All these conclusions are valid only if no stock-exchange boom occurs and are also subject to other qualifications.

The qualification limits their interpretation as general policy prescriptions. Intervention may also work by changing the regression coefficients themselves, particularly those governing consumption. Stocks of consumption goods, share prices, dividends, profits, and speculative variables consequently enter the policy discussion as parts of the model’s transmission mechanisms.

Tinbergen’s concluding appraisal of theories surveyed in Haberler’s Prosperity and Depression is likewise preliminary. Small estimated effects of short-term interest rates, and only moderate effects of long-term rates, support skepticism toward purely monetary explanations such as Hawtrey’s. The acceleration principle does not directly explain short investment fluctuations, although capital-construction periods are important. The results show no systematic lead or lag between consumers’ and producers’ goods production, and little influence from changing production costs. Overinvestment theory encounters difficulties because capital goods were not fully employed and elastic credit could offset saving deficiencies. Harvest irregularities affect prices and consumption but apparently exert little influence on the whole system; consumer demand, especially for agricultural goods, is comparatively stable.

The resulting account presents depressions as readjustments to earlier disproportionalities, with automatic recovery capable of generating another boom. Nevertheless, appropriate policy might prevent or lessen fluctuations. Tintner closes by questioning the price equations and interest-rate conclusions, and by stressing, with reference to Keynes’s criticism of the earlier volume, the insufficiently explicit treatment of expectations. These objections qualify rather than overturn his assessment:

But there is no doubt that Professor Tinbergen’s book will stand out as a very courageous and competent pioneer work which brilliantly combines mathematical economics and statistics for a study of business-cycle phenomena.

The review’s significance lies in this carefully maintained distinction: comprehensive econometric modeling is a major advance, but its theoretical and policy conclusions remain dependent on statistical validity, economic specification, and the conditions under which the model operates.