Great Depression

Review: Robert Margo on Ben Bernanke

Ben S. Bernanke, Essays on the Great Depression. Princeton: Princeton University Press, 2000. vii + 310 pp. $35.00 (cloth), ISBN 0-691-01698-4.

Reviewed for EH.NET by Robert A. Margo, Department of Economics, Vanderbilt University.
Published by EH.Net, July 2000.

For many years, economic research on the origins and persistence of the Great Depression bore a striking resemblance to historical research on the causes of the American Civil War. Both sides to the respective debates talked past one another, and little, if any, intellectual progress was made. Beginning in the 1980s, the situation began to change, at least in case of the 1930s, because of a simple methodological innovation. That innovation was to look beyond time series aggregates for a single country, either at dis-aggregated evidence within countries or at aggregate outcomes across countries. Both types of evidence are examined in Ben Bernanke's new book, a collection of (mostly) previously published articles. (Currently, Bernanke is the Howard Harrison and Gabrielle Snyder Beck Professor of Economics, Professor of Economics and Public Affairs, and Chair of the Economics Department, at Princeton University.)

Essays on the Great Depression is divided into three parts, made up of nine substantive chapters in total, plus an index. Following a very brief preface, Chapter 1 ("The Macroeconomics of the Great Depression") presents the basic themes of the book. Using a cross-country panel data set, Bernanke argues that monetary factors, along with the Gold Standard, should be according primary responsibility for the Depression. Adherence to the Gold Standard resulted in monetary "meltdown" -- the declines in the money stock, accordingly, were non-neutral, because of negative effects of financial crisis and sticky nominal wages on output growth. Countries that remained on the Gold Standard did far worse than countries that abandoned it, and a case can be made that staying on the Gold Standard was more or less an exogenous event - that is, a "natural experiment."

Part Two is made up of three chapters. Chapter 2 ("Nonmonetary Effects of the Financial Crisis in the Propagation of the Great Depression") is justly famous. The paper argues that the US financial crises of the early 1930s had (negative) effects on the real economy, essentially by driving up what Bernanke calls the "cost of credit intermediation." One of two chapters in the book to rely on aggregate time series data for a single country, Bernanke shows (p. 58) that including deposits of failed banks and liabilities of failed business in an otherwise standard time-series regression helps to (negatively) predict output growth. Chapter 3 ("The Gold Standard, Deflation, and Financial Crisis: An International Comparison," written with Harold James) reports a similar finding using the international panel, while Chapter 4 ("Deflation and Monetary Contraction in the Great Depression: An Analysis by Simple Ratios," written with Ilian Mihov) uses very simple methods (an identity linking the price level to the nominal money supply, the monetary base, international reserves, and the quantity and price of gold reserves) to decompose the sources of world-wide deflation before and after 1931.

Part Three, on labor markets, is made up of five chapters. Chapter 5 ("The Cyclical Behavior of Industrial Labor Markets: A Comparison of the Prewar and Postwar Eras," written with James Powell) introduces the second major data set examined in the book, pre-war monthly data on output, employment, hours, and average hourly earnings for eight US manufacturing industries. Among its many findings is the striking observation (p. 175) that prewar employers used shorter hours to reduce labor during a business cycle downturn, while postwar employers have relied on layoffs. Chapter 6 ("Employment, Hours, and Earnings in the Depression: An Analyses of Eight Manufacturing Industries"), perhaps the most influential of the lot, argues that nominal average hourly earnings did not decline as much as one might expect during the downturn (that is, nominal wages were sticky) because average weekly hours fell as well, and workers were unwilling to accept drastic declines in the former given the declines in the latter. Chapter 7 ("Unemployment, Inflation, and Wages in the American Depression: Are There Lessons for Europe?" written with Martin Parkinson), the shortest in the book at eight pages, examines whether the experience of the US in the 1930s can shed light on certain features of recent European macroeconomic behavior; namely, the persistence of high unemployment, and the apparent lack of any influence of high unemployment on either the rate of inflation or real wage growth. Bernanke concludes that, while there are some useful lessons, "there are also large enough differences to make inferences about policy treacherous" (p. 253). Chapter 8 ("Procyclical Labor Productivity and Competing Theories of the Business Cycle: Some Evidence from Interwar Manufacturing Industries," also written with Martin Parkinson) uses the US industry data to investigate "short run increasing returns to labor" (SRIRL) during the interwar period. The key finding is that the degree of SRIRL appears to have been similar in magnitude to the postwar period, which Bernanke claims is "troubling for the technology shocks explanation of procyclical productivity (and thus for the real business cycle hypothesis)." Chapter 9 ("Nominal Wage Stickiness and Aggregate Supply in the Great Depression" written with Kevin Carey) considers several econometric refinements to Eichengreen and Sach's well-known 1985 Journal of Economic History article on the gold standard and the Great Depression. It concludes that the refinements do little to alter Eichengreen and Sach's findings, and concurs with Eichengreen and Sachs that nominal wage stickiness was an important mechanism for propagating monetary declines in the early 1930s.

Overall, Essays on the Great Depression is a mixed bag. Two of the papers -- "Non-Monetary Effects" and "Employment, Hours, and Earnings" -- are certifiable classics, and all the chapters are worth reading (or re-reading, as the case may be). The quality of the prose is several notches better than the usual fare in professional economics journals. The empirical analysis demonstrates a practiced eye for what is important and what is not, attention to historical and institutional detail, and a willingness to explore alternative explanations. Conditional on when they were written, the statistical analyses are state-of-the-art, an order of magnitude beyond what passed for econometric sophistication in economic history journals at the time.

On the other hand, the value-per-dollar ratio is not particularly high. Eight of the nine chapters are essentially copied (the type-setting is new and consistent throughout) from their original sources, and six of these, being from the American Economic Review, Journal of Political Economy, and the like, are easily available (indeed, I bet most economists who considering buying this book will have the originals, or most of them, on their shelves already). The other two chapters were originally published in NBER volumes, hardly less accessible except to the most library-challenged. That leaves the three-page preface, the previously unpublished Chapter 4, and an index as the new material -- not much for the reader's $35.00. I'm not questioning Bernanke's right to reprint his admittedly influential articles, nor Princeton University Press's decision to package them in a handsome volume complete with a striking periodphotograph on the cover and laudatory blurbs from Peter Temin, Barry Eichengreen, and Randall Kroszner on the back cover. However, in the long run, scholarship on the Great Depression, along with readers' pocketbooks, would have been better served had these essays been re-written into real chapters; with a real introduction and conclusion; regressions re-estimated taking into account new data and new econometric techniques; and new textual material interwoven throughout -- in brief, if Bernanke had written a real monograph. Even with a proper introduction and conclusion, such a book, I suspect, would be a good deal shorter than this volume's 301 pages of rather small print -- there is much unnecessary repetition, indeed confusion, across the various chapters.

As a case in point, consider the treatment of wage stickiness throughout. In Chapter 2, using his international panel, Bernanke (p. 31) concludes that "countries in which nominal wages adjusted relatively slowly toward changing price levels experienced the sharpest declines in manufacturing output." Table 9 of Chapter 3 reports first-difference regressions of industrial production (in logs) using evidently the same data set, noting that (p. 101) "only when the PANIC variable [the dummy variable for financial distress] is included does nominal wage growth have the correct (negative) sign ... but it is not statistically significant." In Chapter 6, Bernanke (p. 236) simulates his US industry model assuming "perfect wage adjustment to the cost of living." Remarkably, this assumption "had virtually no effect on the ability to track" employment and hours, suggesting that "the importance of lagged adjustment for explaining observed real wage behavior" during the Depression "may not have had great allocative significance." Then, drawing again on the international evidence, Chapter 9 notes (p. 300) that "the correlation across countries of high nominal wages and low output is interpretable as an allocational effect of sticky wages." Exactly what is going on here? Bernanke clearly views wage stickiness in the 1930s as an economic puzzle, since few of the standard post-WW2 institutional stories (e.g., unions, efficiency wages, and the like) seem to have explanatory power. While I don't disagree with Bernanke on this point, greater familiarity with the economic history literature would have alerted him to the stylized fact that wage stickiness long-predated the 1930s, in the United States and other countries.

While I applaud Bernanke's willingness to move beyond the standard US time series, readers should keep in mind that all of the US data come from conventional, if somewhat neglected sources, as do the international data. Other than occasional asides, not much attention is paid to data quality, and only rarely do readers get any sense of the sort of archival work that might improve matters for future research. Some of the regressions (for example, in Chapter 7) were estimated prior to Christina Romer's revisions to the standard (that is, Lebergott) labor force series, and thus invite re-estimation. Finally, in a book of essays about the Great Depression, one might have expected more attention paid to unemployment, particularly its uneven incidence across the workforce, and the strikingly high levels of long-term unemployment prevailing in the US and elsewhere.

Criticisms aside, Essays on the Great Depression displays one of the great contemporary masters of applied macro-econometrics at work. In the grand scheme of things, I am glad that Ben Bernanke is fascinated by the Great Depression. Hopefully his personal obsession will translate in an expanded audience for the work of economic historians. In any case, his book will more than do as a fine and ready example of why the past continues to have useful macroeconomics.

Robert A. Margo is Professor of Economics and of History at Vanderbilt University, Nashville, TN, and a Research Associate of the National Bureau of Economic Research. He will be on leave during the 2000-01 academic year as Visiting Senior Scholar and Visiting Research Professor of Economics at Bard College in Annandale-on-Hudson, NY. His most recent books are Wages and Labor Markets in the United States, 1820-1860 (University of Chicago Press, 2000) and Women's Work? American Schoolteachers, 1650-1920 (with Joel Perlmann, forthcoming, University of Chicago Press, 2001).

Copyright © 2000 by EH.NET. All rights reserved. This work may be copied for non-profit educational uses if proper credit is given to the author and the list. For other permission, please contact the EH.NET Administrator (administrator@eh.net; Telephone: 513-529-2850; Fax: 513-529-3308).

Posted: 10 July 2000