Money and Stock Prices: Market Efficiency and the Lag in Effect of Monetary Policy

The author starts with a review of the literature.  In particular, the "Predictive Monetary Portfolio" explanation theorizes how more money supply causes an imbalance of money in investors' portfolios, causing them to trade money for stocks or other assets, causing a lagged increase in stock prices.  However, the efficient market hypothesis would claim there should be no lag and no ability for an investor to make excess profits using this lag; several studies (i.e., Sprinkel and Homa/Jaffee) have seemed to contradict the EMH in this regard.

The author suggests that prior studies (in particular Cooper 1972) have been faulty in regard to the EMH in that they don't distinguish between whether money supply affects stock prices (which is agreeable with the EMH) or money supply affects stock prices with a lag (which would go against the EMH).  The author proposes a "Non-Predictive" Monetary Portfolio model that explains how current and anticipated future money supply changes affect current stock prices, which would be in line with the EMH.

The author then goes into ways that an investor may earn excess returns that do not necessarily counter the EMH.  For example, he may have a superior predictive model unavailable to other investors using publicly available data, or he may have access to superior data that is costly to obtain.  Also, certain trading rules may appear to provide excess returns, but may in fact have performance calculation flaws (e.g., the ineffective treatment of dividends in short sales, neglect of brokerage and other costs uncharacteristic of a buy-and-hold strategy, only showing returns for the in-sample period and neglecting an out-of-sample period, failure to adjust for risk, use of unattainable prices for the transactions, using data that would be unavailable at the time of trading decision).

The author then proposes a regression to test whether the growth in money supply explains stock returns.  The dependent variable is stock returns, and the independent variables in the equation are the current growth rate in money supply and several lagged growth rates in the money supply.  The thought is that if the lagged variables turn out to be significant, then there is evidence to rebuff the EMH and to conclude that trading rules can be made to exploit the disconnect; however, if the contemporaneous variable is significant and the lagged variables are not, then there is evidence to claim the market is efficient and there is no way to abnormally profit from the discrepancy.  The authors caution the researcher to use money supply data as of or after the publishing data (not the as of date), in order to prevent using data that was unavailable at the time of decision making.

The authors find in their regression, that over the entire sample period of August 1948 - March 1970, the independent variables were determined to be insignificant at a 5% level, and the adjusted R^2 was below 0.04; both of which would support the EMH and the inability to earn excess returns from lagged monetary supply variables.  They then test their non-predictive MP regression and find slightly significant support for contemporaneous explanation between money supply changes and stock returns.  He then uses a regression to anticipate the non-predictive MP model that takes into account contemporaneous and future anticipated money supply levels; he finds the independent variables to be significant at a 1% level and the R^2 to be sufficiently high to err in favor of the EMH.

The author then does a trading rule simulation by taking the regression equation for the 1/47 - 9/56 period, and using it to determine whether to invest in stocks or commercial paper depending on whether the regression's estimate of the stock returns exceed the commercial paper return.  The backtesting shows that when incorporating the publication lag of the monetary data into the equation, the trading rules do not beat a buy-and-hold strategy; however, not incorporating the publication lag produces extremely high returns compared to the buy-and-hold strategy.  This result emphasizes the importance of timing the data appropriately in the model; otherwise, a false interpretation could be construed.  The result is that if someone knew what the money supply figures would be before they are published, they could profit from a strategy using money supply as an independent variable.

The author then goes into exploring prior studies.  First, he bashes Sprinkel's landmark study by explaining how Sprinkel backed into his trading strategy ex post, which would be impossible for an investor to do ex ante.  Palmer's study (which followed Sprinkel's) also gets a good thrashing due to his use of a moving average of stock returns, which would be correlated with the lagged variables causing misleading results.  He then pointed out how Reilly and Lewis found evidence to support the EM, but continued to naively support the Sprinkel study.  He then goes on to rebuff Keran's, Hamburger and Kochen's, Homa and Jaffee's, and Cooper's studies citing biases in choices and use of unavailable data or neglecting to use appropriate tests to backup their conclusions.

Rozeff, M. S. (1974). MONEY AND STOCK PRICES: Market efficiency and the lag in effect of monetary policy. Journal of Financial Economics, 1(3), 245–302.