We discussed in the last video that investors can choose investments by considering and optimising the following trade-off: investment reward minus risk aversion, times investment risk. Today, we see how to precisely measure our preference metric. The first part about investment reward is simple enough: in almost all cases we will measure investment reward by the expected rate of return E(r). The expected return of a financial asset return r is the probability-weighted average of its possible values. For example if there are n possible values of returns so r1 with probability p1, r2 with probability p2, up to rn with probability pn; then the expected return is given by the following equation. So you have, expected return is equal to the product of the probability and the return value, summed over all n possible values. The expected return is a probability concept. Expected returns are unknown; for example we do not know what is the expected return on a stock like Amazon or Microsoft. In reality we therefore use a statistical estimator to estimate these expected returns. The most often used statistical estimator of the expected return is the sample average return. If you have a sample of T returns, then the sample average return is computed as the sum over all T returns divided by T. Here the hat symbol over E(r) indicates that we obtain an estimator of E(r), not the true unknown value. For example you can compute the average monthly return of Amazon stock since it was issued, to estimate the expected return over the next month. Note that the sample average return is not the only statistical estimator of the expected return. You may think for example, that recent returns are more representative of future returns, and consequently use the average return on the most recent 5 years, instead of using all historical returns. But what about investment risk? We start with the simplest measure of risk - the variance. While it is not a perfect measure of risk, we begin with variance because it happens to be the most widely used measure. The variance is a measure of dispersion of returns around expected return. As risk-averse investors, we prefer investments with lower variance, because we are more certain of the return we will realise on our investment. The variance is also a probability-weighted average, like the expected return. However, in this case, it is not a probability-weighted average of all returns, but rather we take the average of squared deviations of returns from their expected value. As we can use a sample average return to estimate the expected return, we can use a sample variance of returns to estimate the true unknown variance. Here we divide by T-1, in order to obtain the unbiased estimator. We also often refer to this type of risk as the volatility, which is the square root of variance. The volatility of annual returns usually is a number such as 5-10% for bonds, and 15-25% for stocks in developed markets. Now that we have a precise measure of investment reward and risk, we can express our basic preference metric as the expected return minus half of our risk aversion, times the variance of returns. As will become clear in module 2, we scale here risk aversion by a factor of 1/2, just for mathematical convenience. Note that we have not discussed how to measure risk aversion. From experiments in psychology, we know that typical values range from 2 to 10. There are consultancy firms that specialise in measuring your risk aversion. A simpler approach is to look at portfolios obtained with different levels of risk aversion, and determine with which you are more comfortable. So let's come back to our stock vs. bond example from the last video. This table contains the average return and volatility computed over the whole sample, for both investments. Stocks have outperformed bonds over the 1971 to 2015 period, but they also experienced higher volatility. How can we then determine which one is preferable on a standalone basis? This new row shows the preference metric computed with a risk aversion of 5. An investment in the S&P 500, corresponds to a preference equivalent return of 5.29%, whereas bonds correspond to a preference equivalent return of 5.94%. By "preference equivalent return", I mean that that the preference metric would be the same for the stock investment than for an investment with no risk that pays a return of 5.29% We can use a preference equivalent return in order to compare investments that have different average return and volatility. Naturally, we prefer investments with the highest preference equivalent return. Therefore, an investor with a risk aversion of 5, would have been better off leaving all his savings in bonds during this period. Because bonds'preference equivalent return at 5.94% is higher than the one for stocks. But hold on! There are no one-size-fits-all solution. Consider an investor who is willing to take on more risk, in order to potentially reap higher returns. The last row here, shows a preference equivalent return when you have a risk aversion of 2. In this case it is 8.77% for stocks, and only 6.59% for bonds, which indicates this less risk-averse investor would have been better off choosing stocks as an investment. But even if the preference metric is harder for one versus the other, is it still good enough? We usually compute returns in excess of the risk-free rate Rf, which is obtained by investing in a safe asset, for example a short-term bond from a developed country's government. We consider it as the ultimate benchmark, because we can always park our money in the safe asset and achieve this rate of return with no risk. Clearly, an asset that has a preference metric of 5%, is not as attractive when the risk-free rate is 8%, than when it is 2%. So note that the preference metric of the risk-free security is its rate of return Rf. Over the 1971 to 2015 period, the average risk-free rate of return was 5.21%, meaning that both our risk aversion of 2 and risk aversion of 5 investors would have preferred to invest either in the stock market or their long-term bond, instead of the risk-free asset. So let's wrap up with one question. If your risk aversion is 2, do you prefer an investment that has an expected return of 8% and a volatility of 15%, or an investment with an expected return of 9% and a volatility of 20%? Here, be careful not to use these numbers in percentage, but rather in their decimal form. The possible answers are: a) I prefer the first investment b) I prefer the second investment, or c) I am indifferent between the 2 investments. The answer is a. You prefer the first investment, which has a preference equivalent return of 0.8 minus your risk aversion of 2, divided by 2, times the volatility of 0.15, squared in order to get the variance. Which is equal to 5.75%, hereas the preference equivalent return of the second investment is only equal to 5%. So we've discussed in this module the different types of investors and what kind of preference they have. We have also examined how to choose between different investments. In the next lesson we dive deeper into financial markets in which thousands of investments are available. The choice will then become selecting the appropriate mix of these investments.