Equity Premium Puzzle

Today I received a very interesting email from one of my MBA students. Here is what he wrote:

"Professor Saffi,

I noted an article of interest on one of the financial blogs I read, titled "Revisiting the Equity Premium" (http://blogs.reuters.com/felix-salmon/2010/05/20/revisiting-the-equity-premium/). The blogger advances three main points in the article;
1) most managers are not sure why they use an equity premium of 5%-8%

2) That two noted researchers indicate the premium is really 0%-2%
http://alephblog.com/2009/07/15/the-equity-premium-is-no-longer-a-puzzle/
http://falkenblog.blogspot.com/2009/07/is-equity-risk-premium-actually-zero.html

3) That we assume that equities MUST yield more than treasuries based on efficient market hypotheses, however, rather than must, we should be using the word HOPE and recognize the incentives in the system and that the past will not reflect the future.

Please let me know what you think."

Here is my reply:

At the end of the day, the magnitude of the risk premium depends on the risk aversion of investors and the future cash-flows of firms that capture productivity gains (i.e. their average returns). The idea behind using past data is exactly to try to have an estimate of its current value, which can also fluctuate over time. Is it possible that investors have been greatly exaggerating this future estimated performance? Yes, it could. In my humble opinion, this is also related to the Malthusian theory that mankind won’t be able to keep raising food productivity. People have been saying that for 210 years and we’re still going strong

To be honest, one reason why managers don't why they use 5-8% is because most have never seriously studied its determinants. This guy here probably doesn't as well:

Schrager then continues her argument with this:

“Equities are inherently riskier than Treasuries. Equity prices must ultimately reflect and compensate investors for that risk or no one would hold them in their portfolio.”

I’m not sure where that “must” comes from: maybe it’s some kind of corollary of the efficient markets hypothesis. Investors certainly hope that returns on equities will be commensurate with the risk that they’re taking. But there’s no rule saying that any given asset class will “ultimately reflect and compensate” those hopes. After all, if there were such a rule, then really there wouldn’t be any risk at all!

This has nothing to do with the efficient market hypothesis. We could still have rules for things that are inherently uncertain (just think about quantum physics or the Heisenberg uncertainty principle). There is nothing that says that the equity premium MUST be around 5% in the future, it is just our current understanding of it that allows us to forecast this. Sure, some factors are likely to reduce the premia, like taxes, transaction costs, and etc, but saying that the market premium is zero seems a bit of a stretch to me.