For over a century, researchers have scrutinized the relationship between stock prices and corporate earnings. Notably, economist Robert Shiller has amassed extensive data indicating a consistent correlation between these two financial metrics across long periods. This article delves into the nature of this relationship and examines the utility of shifts in correlation for predicting future market returns.
Particularly relevant for investors, understanding this connection can provide a clearer framework for anticipating stock market trends. The analysis focuses on data from the S&P Composite, stretching from 1871 to December, allowing for a comprehensive examination of how earnings influence price movements over time.
The long-term correlation between earnings and prices
In investigating the relationship between earnings and stock prices, two primary objectives emerge. First, the data offers an intuitive explanation for market behavior over extended periods, specifically those exceeding a decade. Such a timeframe is essential for retirement planning and asset allocation strategies.
Second, by tracking the correlations between these variables, we can assess whether fluctuations in correlation serve as reliable indicators for future market performance. The analysis seeks to determine if historically low correlation periods lead to stronger or weaker stock performance in subsequent years.
Historical correlation data
The analysis utilizes monthly averages of earnings-per-share and stock prices from the S&P Composite to derive correlation statistics. The data, which encompasses over 150 years, consistently demonstrates high correlation coefficients across various timeframes.
For instance, the overall correlation from 1871 to stands at a remarkable 0.977. When narrowing the focus to the last 100 years, the correlation remains strong at 0.974, even after the introduction of investor protections post-1940, where it was recorded at 0.973. This indicates that despite regulatory changes, the relationship between earnings and stock prices has remained stable.
Fluctuations in correlation and their implications
While the long-term correlation appears robust, it is essential to recognize that these correlations are not static; they exhibit variability over shorter windows, such as five, ten, and twenty years. For example, during the tumultuous early 20th century, marked by two world wars and the Great Depression, the correlation dipped to 0.6 at its lowest.
As we examine shorter timeframes, the data reveals even greater fluctuations. The rolling ten-year correlations have been observed to fall below zero during significant historical events, including the end of both world wars and the inflation crisis of the late 1970s.
Assessing predictive power
To explore whether changes in earnings-price correlation could serve as predictive tools for future returns, regressions were performed. The relationship between annualized returns and correlation levels yielded varied results. Despite a strong historical linkage, no significant predictive relationship emerged for shorter time horizons.
For example, the rolling twenty-year periods showed a correlation coefficient of only 0.24, while the ten-year periods dropped to a mere 0.06. In contrast, the longest rolling fifty-year window produced a correlation of 0.53, suggesting that only over extended periods does the correlation possess some predictive potential.
Conclusion: Insights for investors
The findings from this comprehensive analysis underscore the importance of understanding the long-term relationship between earnings and stock prices. While earnings are a critical factor in explaining market behavior over time, they do not necessarily provide the necessary insights for market timing or predicting short-term returns.
Investors should be cautious in interpreting fluctuations in correlation as signals for buying or selling. Instead, recognizing the inherent strength of the long-term relationship between earnings and stock prices can offer a more grounded perspective on market dynamics.