Tuesday, August 4, 2009

A break down of the equity risk premium

Is the equity risk premium actually zero? This post makes this claim. The author argues for a range of reasons why a risk premium is less than the typical quoted number of 6%.

I am inclined to agree that it is less than 6%, but I think it is probably incorrect to say it is zero.

I have a few issues with some of his arguments:

1.He claims that various studies have shown that the risk premium is declining, particularly since the 2000 market crash. The problem here is that I think he is confusing survey evidence of peoples expectations of the market risk premium with the implied premium that actually generates current prices. Before the market crash, in the late 90s, most investors thought stocks were great, would give high returns, and therefore would expect a high return over bonds. But of course, stocks were very expensive then and the implied premium was very low. After the crash the opposite occurs. Investors think stocks are awful, and expect low returns, but as prices are low, the implied premium is high.

2. He argues that transaction costs are ignored. But the evidence for bid-ask spread induced costs that he quotes are largely from pre-decimilization, when spreads were much higher than they are now. I agree that frequent trading will eat returns, but I don't see how this is really relevant to the equity premium. Nor do I see how poor market timing will impact the equity premium. Both of these factors will impact realized returns, but the equity premium is a fairly refined concept of the return you will earn as a buy and hold investor holding the market portfolio. In essence, someone how has an index fund in his/her 401k.

3. He states that the geometric mean will be much lower than the arithmetic mean due to volatility. No argument here, but this is a straw man as no reasonable person uses the arithmetic mean to estimate historic returns.

4. Peso problem. I completely agree that the US is a special case in that it has posted one of the highest historic premiums of any country. There is no reason to believe that the US should not mean revert to the average of the developed world.

So what is the correct premium? I sort of skirted around this issue here. But I think it is somewhere between say 2-6%. But don't quote me on that.

Finally, what would a zero percent premium mean? Well I doubt it will be along the lines of the Dow trading to 36000..