Using an Autoregressive Model to Predict the Price-to-Earnings Ratio and Develop an Investment Strategy

In a previous post, we highlighted an article that showed how useful accounting numbers are. In this post, we will present a concrete example of an application of accounting numbers in portfolio management.

Reference [1] showed that the Price-to-Earnings ratio is a mean-reverting process, and it can be accurately estimated by AR(1), an econometric model. Earnings, on the other hand, follow a trend process and can be modeled by a first-order integrated process with a constant factor that captures relative earnings growth.

I propose a model of expected returns by decomposing stock price into two constituents: earnings and price-to-earnings (PE) ratio. The PE ratio is a mean-reverting process that can be forecasted for short to medium horizons. On the other hand, earnings follow a trend process and can be characterized by a first-order integrated process. Growth in earnings is responsible for the growth of stock prices, but the cyclical nature of the PE process imparts time-series predictability to prices.

After calibrating the models, the author constructed a long-only investment strategy that switches between stocks and bonds.

Using the expected return model, I propose a portfolio switching strategy where investor switches between stocks and bonds based on current expected returns on market-wide stock index and risk-free government bond yields. My strategy provides better risk-adjusted returns by avoiding equity investments in periods of low expected returns. The strategy is suitable for retail investors because it only involves market-wide stock index and bond yield. Moreover, my strategy does not require short-selling of stocks which can be difficult for retail investors.

In short, accounting numbers can be used in econometric models in order to develop an asset allocation strategy that offers better risk-adjusted returns.

The article showed that Earnings and Price-to-Earnings ratio have practical and useful applications. In our opinion, revenue and revenue growth are also important accounting numbers that can be used effectively in portfolio management.

References

[1]  N. Vidhani, Return Predictability using Price-to-Earnings Ratio, 2021,  https://ssrn.com/abstract=3910641

Article Source Here: Using an Autoregressive Model to Predict the Price-to-Earnings Ratio and Develop an Investment Strategy