Predictability has become a term of art in Finance. That is to say, when we say it, it doesn't mean what it usually means. If I ask, are returns predictable? I don't mean, can I say for certainty if this stock or the market as a whole will go up. I don't even mean, can I tell good investments from bad ones? Return predictability means, can I use some information today to judge what the expected future return over some horizon will be?
Suppose that the long run average for stock returns is 8%. The question, then, is the expected return on stocks over the next year always 8%. This would be true if stocks followed a random walk, which for a long time was a basic assumption in Finance research. If stocks follow a random walk, then all the ups and downs that prices take are just random fluctuation around an average return. Stock returns would not be predictable.
Instead suppose that while the long run average is 8%, the expected return in any given year will vary over states of the economy. When times are great and the economy is booming perhaps the expected return to stocks over the next year are 5%. That is, stocks don't seem that risky, so stock prices are already pretty high and expected returns are a little bit low. When times are terrible like the end of 2008 and prices go really low perhaps expected returns rise much higher, say 12%. Stocks are risky. No one wants to hold risk, as times look pretty bleak, so in order to hold on to stocks investors have to be compensated with a higher expected return.
Intuitively, I think predictability makes a lot of sense, but it's still an open question in finance as the empirical evidence is mixed. Stocks are quite volatile, so the amount of data required to answer the question is immense. Not to mention, dividends move as well. Perhaps, it isn't the expected return that's changing it's the expected future dividend stream. That is, the price of the market didn't go down, because the expected return went up, it went down because future profits/growth are likely to be low.
Suppose that the long run average for stock returns is 8%. The question, then, is the expected return on stocks over the next year always 8%. This would be true if stocks followed a random walk, which for a long time was a basic assumption in Finance research. If stocks follow a random walk, then all the ups and downs that prices take are just random fluctuation around an average return. Stock returns would not be predictable.
Instead suppose that while the long run average is 8%, the expected return in any given year will vary over states of the economy. When times are great and the economy is booming perhaps the expected return to stocks over the next year are 5%. That is, stocks don't seem that risky, so stock prices are already pretty high and expected returns are a little bit low. When times are terrible like the end of 2008 and prices go really low perhaps expected returns rise much higher, say 12%. Stocks are risky. No one wants to hold risk, as times look pretty bleak, so in order to hold on to stocks investors have to be compensated with a higher expected return.
Intuitively, I think predictability makes a lot of sense, but it's still an open question in finance as the empirical evidence is mixed. Stocks are quite volatile, so the amount of data required to answer the question is immense. Not to mention, dividends move as well. Perhaps, it isn't the expected return that's changing it's the expected future dividend stream. That is, the price of the market didn't go down, because the expected return went up, it went down because future profits/growth are likely to be low.
No comments:
Post a Comment