From The Panic Of 1873 To The Crisis Of 2007: Bank Credit, Business Cycles, And Financial Markets

From The Panic Of 1873 To The Crisis Of 2007: Bank Credit, Business Cycles, And Financial Markets

Mendoza College of Business, University of Notre Dame

January 11, 2016

Abstract:

What is the relationship between bank credit, macroeconomic risk, and financial markets? Despite its fundamental importance, this question is among the most controversial issues in financial economics. In this paper, I use an extensive new monthly data set spanning 140 years that allows us to unmask even the shortest or mildest variations in bank credit in the U.S. and study its relationship with macroeconomic risk and financial markets. I find that a decrease in bank credit is associated with an increase in macroeconomic risk and consequently higher expected returns in financial (equity) markets. I also examine how relationship between bank credit, macroeconomic risk, and financial markets has evolved over time in response to major changes in policy regimes (such as the foundation of the Federal Reserve). Finally, I am also able to distinguish empirically between two competing channels proposed in the literature through which bank credit interacts with macroeconomic risk and financial markets. I find that the empirical evidence is primarily consistent with the “credit risk channel” rather than the “bank lending channel”. My results imply that time-variation in macroeconomic risk, and not market frictions, (such as agency costs, capital constraints, behavioral biases) may be relatively more important for understanding the relationship between bank credit, economic activity, and financial markets.

From The Panic Of 1873 To The Crisis Of 2007: Bank Credit, Business Cycles, And Financial Markets – Introduction

Understanding the relationship between bank credit, macroeconomic risk (i.e. variation in economic activity), and financial markets is one of the most important issues in financial economics. As one of the main source of funds, fluctuations in bank credit may affect aggregate consumption and investment, and may have significant real effects. In fact, the relationship between bank credit and macroeconomic risk is presumably at the root of widespread government policies concerning bank regulation and supervision. Clearly, the design of these public policies depends on our understanding of this relationship. Despite the fundamental nature of this issue, significant controversy in the literature still remains about how, if at all, does bank credit relates with macroeconomic risk and financial markets.

On one hand, an extensive theoretical literature implies that banks are a veil that can be ignored because bank credit has no real effects on macroeconomic risk or financial markets. This “irrelevance view”, most associated with ideas of Modigliani and Miller (1958) among others, suggests that real economic decisions are independent of the source of funds. As such, it appears that fluctuations in bank credit should not be central to our understanding of economic and financial outcomes.

On the other hand, economists since at least Mill (1965) have suggested that a decrease in bank credit is associated with an increase in macroeconomic risk. The extant literature has proposed two competing hypothesis to explain this negative correlation. The “credit risk channel” (of Bernanke and Gertler (1995) and Kiyotaki and Moore (1997), among others) suggests that effects primarily flow from economic conditions to bank lending. In contrast, the “bank lending channel” (of Rajan (1994), Thakor (1996), and Holmstrom and Tirole (1997), among others) suggests that it is problems at banks that subsequently spill over and increase risk for the entire economy. The jury is still out as to which of these competing channels is relatively more important.

In the aftermath of the financial crisis of 2008 – 2009, yet another view to emerge is that bank credit increases macroeconomic risk (i.e. a positive correlation). Recent work in this spirit includes Schularick and Taylor (2012), Baron and Xiong (2015), and Jorda, Schularick, and Taylor (2013), among others. The implicit assumption in these papers is that “excessive” bank credit fuels speculative over-investments, and that such inefficient investment increases risk for future macroeconomic performance.

One approach to resolve the spectrum of opposing viewpoints expressed in this literature may be to more accurately and consistently measure bank credit over a large number of business cycles and more precisely correlate variations in bank credit with variations in economic activity, and financial markets.

Thus far, the ability of researchers to do this has been limited by the availability of bank credit data at a sufficiently high sampling interval – the extensive literature on this topic primarily utilizes annual or quarterly data. As Watson (1994) points out annual data is far from ideal for such analysis, since it can mask short or mild contractions or expansions. In this paper, I use an extensive new monthly data set spanning 140 years that allows us to unmask even the shortest or mildest variations in bank credit and economic activity in the U.S., and study its relationship with macroeconomic risk and financial markets over a large number of business cycles.

This data set, which comprises hand-collected data on bank credit for the U.S. from historical financial records, provides several additional unique advantages. First, monthly data allows us to correlate bank credit with actual variation in economic activity (as measured by say the index of industrial production). Thus, my analysis does not rely on definitions of economic activity that could be dated or arbitrary.2 Second, this data offers the highest frequency look at the interaction between the U.S. banking system, macroeconomic activity, and financial markets over the longest sample available. This is especially useful because of the way financial markets respond to new information – with quarterly or annual data on bank credit, researchers run the risk of missing all the interesting variation in financial markets (i.e. equity returns) in response to changes in bank credit. In addition, this data also allows us to examine how the relationship between bank credit, macroeconomic risk, and financial markets has evolved over time in response to major changes in policy regimes, such as the foundation of the Federal Reserve System. Third, my focus on U.S. data alone ensures that all banks (and firms) are subject to uniform institutional structures, legal frameworks, and regulations.3 Fourth, since the U.S. is one of the few markets for which data are available for a broad cross-section of traded firms over a long sample, my data allows me to investigate whether the relationship between bank credit and financial markets varies in the cross-section. Finally, the monthly data on bank credit compliments the economic data currently available to researchers and may prove to be a valuable resource for future time series analysis by scholars interested in this area. The data used in this paper also offer one of the few monthly time series of consistent quality that are available both during the prewar and the postwar period.