Profitable timing of the stock market with the senior loan officer survey

https://doi.org/10.1016/j.frl.2023.103733Get rights and content

Highlights

  • The loan standards question is shown to be predictive of subsequent quarterly stock returns.

  • A simple alpha and risk model is developed to time investments in the stock market.

  • Out-of-sample, market timing the S&P 500 with the loan standards signal outperformed indexing by a large margin.

Abstract

The loan standards question in the Federal Reserve's quarterly Senior Loan Officer Survey is shown to be predictive of quarterly stock returns a month or two after its release. This is an apparent violation of semi-strong form stock market efficiency. Out-of-sample, we use this signal and develop a simple risk and alpha model to market time the S&P 500. It outperformed the S&P 500 with a Sharpe (1966) ratio of 1.9 versus 0.34 for passive investment.

Introduction

In this paper, we show that a question on the Federal Reserve's Senior Loan Officer Survey is a negative and significant predictor of quarterly stock returns. We show how a simple market timing strategy with this survey question would let a portfolio manager initiate trades a couple months after the survey release date that would have a Sharpe (1966) ratio over five times that of a passive investment in the S&P 500.

Lown et al. (2000), Lown and Morgan (2006) and Cunningham (2006) found that “Net Percentage of Domestic Banks Tightening Standards for Commercial and Industrial Loans to Large and Middle-Market Firms,” was predictive of GDP growth in the United States and was negative and significant. That is tightening of loan standards was associated with lower future GDP growth.

Bassett et al. (2014) finds that rising loan standards were predictive of falls in lending, rising credit spreads, lower GDP, and cuts to the Fed Funds rate. Lown and Morgan (2002) find rising loan standards are associated with a drop in lending and higher interest rates. Intuitively, higher credit standards should raise the cost of debt for bank-dependent borrowers. That should translate into less capital investment and to ultimately lower GDP. The Fed eventually responds to shrinking money creation and lower output with cuts to the Fed funds rate.

Predicting GDP growth with this survey question is useful, but its ability to predict future stock prices flies in the face of the efficient market hypothesis and our notions of rational expectations. The present paper adds further support to the finding that the “loan standards” survey question predicts stock returns a few months after the survey results are publicly released. That is a violation of the semi-strong version of the efficient markets hypothesis as defined by Fama (1970). It appears that the stock market has not been taking into account this piece of publicly available information, and we show how investors could time the market to earn over three times the returns out-of-sample with one-third the realized standard deviation of returns.

Chava et al. (2015) found that Senior Loan Officer survey response predicted future stock returns in the first twenty-four years of that survey question's release. We show it was still predictive after over thirty-two years. In contrast to Chava et al. (2015), we show how profitable a market timing strategy based on a simple risk model would be compared to indexing the S&P 500. Park (2011) finds that, as with the U.S. stock market, the loan standards question from a Bank of Canada survey was a negative and significant predictor of Canadian stock market. The loan standards question may also predict banking sector stock returns, according to Sohn and Park (2016).

Section snippets

Data

In April 1990, the Senior Loan Officer Opinion Survey on Bank Lending Practices added the following survey item referred to as, “Net Percentage of Domestic Banks Tightening Standards for Commercial and Industrial Loans to Large and Middle-Market Firms, Percent, Quarterly, Not Seasonally Adjusted.” It is often called the “loan standards” question. A positive reading means that credit conditions are tightening for large and middle market firms. A negative reading means credit conditions are

Regression results

We estimated the following Ordinary Least Squares (OLS) regression:RQ,t=α+βSLOt1+ε

RQ, t is the quarterly return to the S&P 500. α is an intercept coefficient to be estimated, β is the slope coefficient to be estimated, and ε is a random, mean zero error term. SLOt-1 is the loan standards survey question in the month or two prior to the start of the quarter.

We tested six different time periods, the full sample of 128 quarters, the first sixty quarters or fifteen years, the second 60 quarters,

Market timing

To calculate the economic significance of the Senior Loan Officer survey data on predicting lagging quarterly stock returns, we use the first fifteen years of data from model 2 of Table 1 to create an alpha model to drive our stock market buying decisions each quarter.

A risk model is required to create actual stock allocations in the out-of-sample period from Q3 2005 to Q1 2022. We want to build a portfolio of equal variance to the S&P 500, as a whole, in expectation. That is our risk model.

Conclusion

The loan standards question in the Fed's Senior Loan Officer survey has been shown to be a significant predictor of the S&P 500′s quarterly stock returns. An asset manager could put on stock trades a month or two after the survey is released and earn abnormal returns. Using a strategy where the investor tries to mimic the volatility of the stock market, we show that this hypothetical loan standards market-timing portfolio would have outperformed the S&P 500, out-of-sample, by over three to one

Funding

None.

Declaration of Competing Interest

None.

References (15)

There are more references available in the full text version of this article.

Cited by (0)

View full text