Regardless of the level of diversification, why is it more rational to invest in a low β portfolio than a high β one?

The issue is that CAPM’s predictions and actual statistics diverge. As the graph shows, low β portfolios generate higher returns than CAPM predicts, and vice versa.

Although high β portfolios generate higher returns, the risk-return ratio is better for low β portfolios.

Why does this happen? What is the reason? I didn’t know either, so I asked ChatGPT:

  1. Market Anomalies: CAPM assumes all investors have the same information and preferences. In reality, market anomalies and behavioral biases can affect asset prices. Factors such as value (Book-to-Market), momentum (panic), and size (Big vs. Small) influence returns, and these are not accounted for in CAPM.
  2. Non-Market Risk: CAPM relies solely on systematic risk. However, individual stocks also have unsystematic risks. Low β stocks might also have better fundamental foundations (there could be some correlation between market and specific risks, so the degree of portfolio diversification will significantly impact the outcomes).
  3. Investor Behavior: Investor sentiment can influence stock prices. High β stocks may experience greater price volatility due to investor sentiment, while low β stocks might remain unnoticed or undervalued. For example, it’s trendy now to invest in ESG companies, which are environmentally conscious, but this doesn’t necessarily mean that tobacco or beer manufacturers yield lower returns.
  4. Changing Market Conditions: Economic or geopolitical events can impact asset returns. These changes may not be adequately reflected in CAPM (the changing range of market fluctuations might affect unanticipated returns—CAPM predicts expectations, not real returns, so the more unexpectedly the market behaves, the more predictions deviate from actual statistics).
  5. Time Horizon: CAPM considers a single investment horizon. In reality, investors might have different time horizons and risk preferences, leading to variations in returns (if we break this statistic into different time segments or add another 30 years of data, we might see a different picture).

P.S. There is a modified version of the CAPM model that takes into account the size and balance sheet value of an organization. Since these two factors are necessary for evaluating local companies, I will write about this separately later.

Photo from the book:

Principles of Corporate Finance, 12th Edition
By Richard Brealey, Stewart Myers and Franklin Allen