The Hidden Cost of Volatility Drag

A fundamental concept in finance has long been debated among investors: the relationship between stock market volatility and excess return predictability. Many studies have shown that excess returns are, to some extent, predictable – but how do we unlock this predictive power?

The binary dependent dynamic probit model offers a promising solution. By analyzing the signs of excess returns, researchers can forecast direction with greater accuracy than traditional time series models. This is particularly useful for identifying potential economic downturns, as a binary recession indicator has proven to be an effective predictor.

A key insight lies in the fact that some assets exhibit significant volatility and higher-order conditional moments. These factors have statistically significant explanatory power in predicting excess returns, with average trading strategy returns often outperforming buy-and-hold or ARMAX models. For instance, a study by Cristoffersen et al. (2006) found that the return is decomposed into a sign component and absolute value component, which are modeled separately before the joint forecast is constructed.

Why Most Investors Miss This Pattern

One reason why most investors overlook this pattern lies in the complexity of financial markets. The noise in observed returns can be overwhelming, making it challenging to discern patterns. Additionally, the time series nature of excess returns means that traditional statistical summary statistics may not accurately capture the underlying dynamics. As a result, directional predictability remains a neglected area of research.

A 10-Year Backtest Reveals...

A recent study by Leung et al. (2011) employed a dynamic error correction probit model to forecast excess stock returns for an extended period. The results showed that the recession forecast obtained from the model appears to be the most useful predictive variable, with accurate forecasts in-sample and statistically significant relationships between the forecasted value and excess return.

What the Data Actually Shows

The data does indeed support this claim. A review of empirical studies has demonstrated that excess returns are, to some extent, predictable – but the direction of excess returns is generally more informative than their overall level or mean. Furthermore, the recession forecast constructed for a binary recession indicator has been shown to be an effective explanatory variable in predicting excess returns.

Three Scenarios to Consider

For investors seeking to capitalize on this predictive power, three scenarios are worth considering:

1. Buy-and-hold strategy: By employing a buy-and-hold approach with accurate forecasts, investors may see significant gains over the long term. 2. Risk-averse portfolios: Investors willing to take on more risk may benefit from using excess returns as an input for their portfolio optimization models. 3. Asset allocation decisions: By incorporating excess return forecasts into asset allocation decisions between stock and risk-free interest rate investments, investors can potentially optimize their portfolios.

As the data continues to evolve, researchers will be able to refine and extend these findings. While there is still much work to be done, this study has provided a valuable framework for understanding how excess returns are predictably linked to market volatility.