The Great Depression remains a watershed moment in American economic history. For that reason, it has rightly attracted significant interest from the

Transcription

1 The Great Depression remains a watershed moment in American economic history. For that reason, it has rightly attracted significant interest from the economics profession. Not surprisingly then, a plethora of theories have been developed to understand what caused it, what made it so bad, and what made it go on for so long. One aspect of the Depression that both remains understudied and that still resonates today is the role of political uncertainty in driving economic outcomes. Robert Lucas, in an interview with the Wall Street Journal, recently opined that the most important but hardest to measure [factor retarding recovery from the Depression was] FDR s demonization of business and the uncertainty it engendered. Schumpeter (1942) actually emphasized a similar point just a few years after the Depression. Huey Long was a contemporary of the former president who was more radical, more populist, and more hostile to business than FDR was. This paper studies the effect of political uncertainty on economic outcomes with a case study of Louisiana under Huey Long. There is a literature on the role of economic uncertainty, such as uncertainty over demand or technology, in the Great Depression. Romer (1990) ties the uncertainty caused by the 1929 crash stock market to the decline in consumer durables purchases in late 1929 and Federer and Zalewski (1994) argue that interest-rate uncertainty was a channel through which banking crises and the collapse of the international gold standard negatively impacted the real economy. Recent work by Mathy (2011) has reemphasized the importance of uncertainty shocks using VAR analysis and a plausibly calibrated DSGE model. For post-war business cycles, Baker et al. (2011) have found large negative effects of policy uncertainty engendering economic uncertainty while other papers such as Born and Pfeifer (2011) have struggled to find anything at all. All of the aforementioned papers on postwar business cycles and the Great Depression (save Baker et al. (2011)) have taken changes in economic uncertainty as some sort of shock originating outside the model. This comes at the cost of difficult question regarding identification. Are changes in economic uncertainty really a cause or effect of changes in the macroeconomy? Changes in political uncertainty due to the legislative process and elections provide one natural mechanism for creating changes in the level of economic uncertainty. In fact, Higgs (1997) argues that the second New Deal induced uncertainty in the private sector over whether property rights would be respected, thus inhibiting private investment. While suggestive, the Higgs study lacks any quantitative measure for political uncertainty. More generally, the Depression provides a rich environment in which to locate these changes in political uncertainty. It saw the rise of a number of populists who advocated radical changes in economic policy. In many cases, they combined these proposals with sharp criticism of business. Many of the most strident populists like Father Coughlin never actually gained political office and thus never had a chance to effect policy making, directly. One of those populists who actually attained an office of some significance was Huey Long, who rose to power in Louisiana in 1928 and continued to run the state even after being elected U.S. Senator in Beginning in his first campaign for the Louisiana Railroad Commission, Long campaigned on a populist platform, telling crowds that his opponent was a tool of big business (Hair, 1991). The Railroad Commission was 1

2 responsible for the regulation of oil and gas pipelines in addition to that of all public transportation in the state. Previous commissioners tended to do little and to acquiesce to the wishes of the businesses they regulated (Hair, 1991). Long, however, used his post to launch populist attacks on Standard Oil, Louisiana s largest oil company. He called the company an octopus and highway bandit (qtd. in Hair 1991, p. 92), and demanded the governor work with the legislature to declare pipeline companies public utilities. When the governor refused, Long call him the criminal who disgraces the gubernatorial chair (qtd. in Hair 1991, p. 92). After losing his first attempt at the governorship, Long finally won in He dubbed the wealthy parasites, and as a U.S. senator he proposed a national Share the Wealth program that would have imposed a 100 percent tax on incomes over $1 million and wealth over $5 million. This became the basis of the first bill he introduced in the U.S. senate, which was proposal to cap annual incomes at $1,000,000 (Hair, 1991). Louisiana business was perhaps equally frightened by the antidemocratic actions Long used to secure power in the state. As governor, he fired political opponents and forced state employees to contribute money to his political machine (White, 2006). In 1934, Long reduced the power of local governments, particularly that of New Orleans (White, 2006). He passed a tax on newspapers, which he called a tax on lying. He also forced through a tax on Standard Oil that had nearly led him to be impeached in Later Long strengthened the Bureau of Criminal Identification, an agency empowered to make warrantless arrests throughout Louisiana (Hair, 1991). In reaction to Long s consolidation of power, in early 1935 a coalition including a former Louisiana governor and the mayor of New Orleans organized a paramilitary organization, the Square Deal Association, to oppose Long (Hair, 1991). They took over a courthouse in Baton Rouge in January 1935, leading Long to call out the National Guard and impose martial law, under which Baton Rouge newspapers were prohibited from criticizing the state government (White, 2006). Hair (1991) compares the state at this time to a banana republic with a personality cult around Long. By focusing just on Louisiana, which one can think of as a small open economy with a fixed exchange rater relative to the rest of the United States, we are able to minimize questions of reverse causality since what happens to Louisiana s economy has minor effects on the aggregate economy. Furthermore, it allows us to estimate the effects using a diff-in-diff setup taking Mississippi as the control group. The vast majority of previous studies we are aware of on the effects of political uncertainty have used time series variation to attempt to identify the effect with Baker and Bloom (2012) the one exception. The first pressing question is how to measure the political uncertainty generated by Long s rhetoric and legislative maneuverings. We develop two measures that draw on Louisiana-specific data. First, like modern studies such as Bloom (2009) as well as Romer (1990) and Mathy (2011) for the Depression, we use the volatility of stock prices as a proxy for uncertainty. Not only does this measure have the virtue of being straightforward to calculate, there are theoretical reasons to believe that stock price volatility reflects political uncertainty (Pastor and Veronesi, 2012). The stock volatility measure is constructed using data from the New Orleans Stock Exchange, whose listings are derived from digitized records of the Times-Picayune. A New Orleans stock index 2

3 will be constructed as a market capitalization weighted index of all stocks listed on the Louisiana Stock Exchange. Stocks that are also listed on the New York Stock Exchange or are not specific to Louisiana will be excluded from the index. Stock volatility will be calculated as the standard deviation of the log daily return of this Louisiana stock index. Volatility in this measure should be closely related to Louisiana specific uncertainty. The second measure follows that of Alexopoulos and Cohen (2009), who create an uncertainty index that counts the number of articles in the New York Times that mention the word uncertainty. Baker et al. (2011) construct a similar newspaper index for policy uncertainty by restricting attention to articles with both economic policy and uncertainty in the article itself. Our newspaper measure is constructed using information from the Times-Picayune, which is the paper of record in Louisiana. However, we adjust the terminology employed to reflect changes in usage over time. To take one example, we use commerce rather than economic though other idiosyncratic substitutions will be required as well such as the Kingfish for Huey Long. We first validate both the stock volatility and newspaper index measures by linking them to particular political events in Louisiana at the time. For example, in September 1935, Huey Long s unexpectedly assassinated throwing into turmoil his whole political machine provides an excellent occurrence to test for the relationship. We pair these uncertainty measures with an outcome variable derived from detailed plantlevel data on monthly employment counts from the Census of Manufactures. Besides data on manufacturing plants in Louisiana, we employ data on plants in Mississippi, previously collected in Ziebarth (2013). The Mississippi plants provide a natural control group with no polarizing figure governing that state at the time. In addition, the manufacturing sector of Mississippi is quite similar to that of Louisiana at this time. While there exist other state-level employment datasets at this time (Wallis, 1989) that would allow for this treatment-control setup, these are only at an annual frequency, which severely limits the power to identify any effect. Furthermore, using the Census data allows for us to consider an even more stringent test focusing on plants along the border of the two states following the approach in Holmes (1998). It also allows for us to examine heterogeneity in the effects across plants of different sizes and industry. We have results for one particular event so far: the assassination of Huey Long in Across a variety of specifications, we find a null effect of Long s assassination suggesting limited scope for the role of political uncertainty. This includes specifications ranging from using all of the data to focusing on plants near the Mississippi-Louisiana border. To emphasize, it is quite surprising that there is no effect given narrative and financial market evidence suggest that the business community, in general, reacted positively. For example, Louisiana bond prices rose sharply. What these overall results hide is the heterogeneity in response not only across industries but across plant sizes. In fact, we find large negative effects for the largest plants upon Long s death. There are also particular industries such as manufactured ice that appear to be positively affected by Long s death. In the end, more work is necessary to separate out the effects due to an increase in uncertainty versus changes in the expected policy. 3

4 In the end, we think this study can contribute in important ways to both the modern and historical literature on the role of political uncertainty. From an historical perspective, if the effects of even Huey Long s term in power are nil, it would appear to cast doubt on the quantitative relevance of political uncertainty. By focusing on this small, open economy relative to the national economy, we provide an upper bound on the effect since it is relatively easy for businesses to move operations to an adjacent state. We also would argue a further reason for taking these estimates as an upper bound is the fact that FDR s policies were much more moderate relative to Huey Long. For the modern literature, the paper offers an important contribution by studying the effects in a diff-in-diff framework. References Alexopoulos, M. and J. Cohen (2009). Uncertain times, uncertain measures. Unpublished, University of Toronto. Baker, S., N. Bloom, and S. Davis (2011). Measuring economic policy uncertainty. Unpublished, Stanford University. Baker, S. R. and N. Bloom (2012). Does uncertainty reduce growth? Using disasters as natural experiments. Unpublished, Stanford University. Bloom, N. (2009). The impact of uncertainty shocks. Econometrica 77, Born, B. and J. Pfeifer (2011). Policy risk and the business cycle. Bonn Economics Discussion Papers. Federer, J. and D. Zalewski (1994). Uncertainty as a propagating force in the Great Depression. Journal of Economic History 54, Hair, W. (1991). The Kingfish and his realm. Louisiana State University Press. Higgs, R. (1997). Regime uncertainty: Why the Great Depression lasted so long and why prosperity resumed after the war. Independent Review 1, Holmes, T. J. (1998). The effects of state policies on the location of industry: Evidence from state borders. Journal of Political Economy 106, Mathy, G. (2011). Uncertainty shocks and the Great Depression. Unpublished, UC-Davis. Pastor, L. and P. Veronesi (2012). Uncertainty about government policy and stock prices. Journal of Finance 64, Romer, C. (1990). The Great Crash and the onset of the Great Depression. Quarterly Journal of Economics 105, Schumpeter, J. A. (1942). Capitalism, socialism and democracy. Harper and Row. Wallis, J. J. (1989). Employment in the Great Depression: New data and hypotheses. Explorations in Economic History 26, White, R. (2006). Kingfish: The reign of Huey P. Long. Random House. 4

5 Ziebarth, N. L. (2013). Identifying the effects of bank failures from a natural experiment in Mississippi during the Great Depression. AEJ: Macroeconomics 5,