The equity risk premium measures the extra return demanded by investors for choosing a risky asset over a riskless investment. First, a risk free asset is an investment that the investor knows the expected return with certainty. When the investor does not know what the expected return is, the investment is not risk free, and the investor as a result demands a higher return. The equity risk premium is a measure of this average risk in equities.
The risk premium is based on two basics ideas: risk aversion of the investor and the riskiness of the average risk investment. The risk aversion of investors is how willing the investors are to take risk. For example, when the economy is doing well, investors would be more willing to invest in risky assets. The riskiness of the average risk investment is dependent on what investors think the average riskiness is. As a result, the risk premium changes over time. Since the risk premium is different for each individual, the equity risk premium should be a weighted average of all investors based on how much they are investing into the market.
The risk premium can be measured in one of three ways: (1) major investors can be surveyed for what they think the risk premium is, (2) the risk premium can be measured based on historical data, and (3) the premium can be implied from current market data. However, though the survey method would give the most accurate number, it is not feasible to do so. As a result, we are limited to calculating the risk premium either from historical data or from current market data.
The historical risk premium is the most commonly used method in risk and return models. While the idea of using a historical risk premium is accepted, the actual method of calculating the premium is still very varied. For example, the estimated historical risk premium has ranged from 4% to 12% depending on the analyst. There are three main reasons for this huge spread: the time period used, the risk free rate used, and the type of averaging (geometric, arithmetic) used.
The time period can be very variable as investors can calculate the risk premium using data from the last ten, twenty, thirty, fifty, or more, years of data. One main reason why no “correct” time period exists for calculating the premium is due to the conflict between theory and application. As more years of data are used, the standard error of estimation becomes smaller, as seen below:
| Estimation Period |
Standard Error |
| 5 Years |
8.94% |
| 10 Years |
6.32% |
| 25 Years |
4.00% |
| 50 Years |
2.83% |
Thus, in order to achieve an acceptable standard error, the estimation period needs to be fairly large. However, the problem with such a large estimation period is that the current market is sure to be very different from what it was fifty years ago. Furthermore, the risk aversion of investors is very likely to have changed over the years as well; as a result, historical data going back to 1926 will not be an accurate representation of current events and expectations.
The risk free rate used can change the risk premium as well depending on what was used in the calculations. If investors choose a short maturity treasury bill, the risk premium will be larger; similarly, the risk premium will be smaller if a longer maturity is used. Typically, the choice of the risk free rate will be consistent with the valuation mythology, or the ten year bond rate.
In addition, when dealing with historical premiums, the arithmetic and geometric average can affect the premium. The arithmetic average is a simple average of the annual returns whereas the geometric average measures compound change. Both methodologies can be used as strong arguments can be made for either method.
Given all the variables in calculating the historical risk premium, however, a much better way to calculate the risk premium is to use the implied equity risk premium, which assumes that the current market is correctly priced. Based on the dividend discount model, the required return on equity can be calculated, and when the risk free rate is subtracted out, the risk premium is calculated.
| Value = |
Expected Dividends Next Period
|
|
(Required Return on Equity - Expected Growth Rate in Dividends)
|
Based on the above formula, three of the four variables are known. The value is the current value of the market (the S&P 500); the expected dividend and expected growth can both be estimated. For example, if the S&P 500 is at 900, the expected dividend yield for the next period is 3%, and the expected growth in dividends is 6%. Plugging in the numbers and solving the unknown gives you 9%. This 9% is the required return on the equity; when you subtracted out the risk free rate of about 6%, we can estimate a 3% risk premium.
This method has several advantages over using the historical risk premium. First, it is very simple to calculate - it can be calculated using information found on the front page the Wall Street Journal. Furthermore, there are fewer things that can go wrong when calculating this risk premium. The data is also very current, and does not require a large amount of historical data, data that is typically not available when looking at emerging markets.
The most accurate and realistic measure of the equity risk premium can be found by using the implied risk premium method. Using any other premium will bring a market basis into the valuation. Currently (as of February 21, 2007) the implied equity premium is 4.05%, while the most commonly accepted historical premium is 4.91%. In order to achieve a risk premium of 4.91%, the S&P 500 would need to be approximately 17% lower than its current level - such a gross deviation from the correct number will no doubt to cause serious problems in investment decision.