QuoteOriginally posted by: gjlipmanCan I raise an idea that I had while reading the article, and maybe someone can tell me that it has already been considered to death?I am an investor. I can either invest in risk free bonds or equity. Equity has higher uncertainty than risk free bonds. Now there are two sources of this uncertainty as far as I am concerned - there is uncertainty because stock prices have always moved, and then there is the risk that all the market parameters move. The analysis that is usually done (and I admit I haven't read the Mehra and Prescott review) assumes that the future moves similarly to the past. If someone was willing to guarantee this to me, I would feel much happier putting 100% of my money into equity (and lowering the equity premium I require). I maintain some fear however (which cannot be proven invalid) that there is a chance, maybe 5%, that at some stage in the next 10 years, something will happen to cause equities to lose 95% of their value. And while I have this fear, I am always going to be cautious putting too much into equities.Any comments?You are thinking like a Bayesian. There is a version of the Capital Asset Pricing Model which does not assume all investors agree on the probability distribution of future returns, but that each investor has the degree of uncertainty appropriate for someone without strong prior belief studying the historical evidence.If I buy stock for a year, my main risk is the volatility of the market, say standard deviation of 25% for a diversified portfolio.If I decide at age 25 to put all my retirement savings in stock, I still have a significant volatility risk, it's about 3% per year. At this time scale the fact that I don't know the long-term equity risk premium is a risk roughly the same order of magnitude of my volatility risk.One solution is to mash these together, like Bayesians do. This makes the long-term argument for stocks less compelling, but does not affect the short term argument much.Another approach, along TrendFollower's lines, is to forget about the premium. What are your alternatives? History shows that it's at least plausible that stocks will give you a 4% per year real return, and if so are likely to be the best easily available no-work investment. Say that's not true. Even at 1% real return, stocks are probably better than alternatives.At 0% or negative real returns, you might prefer to be in something else. But what? If stocks return 0% it means public corporations are losing money (since stocks pay off on the winners but are capped at $0 for the losers, a net no return means more companies are losing money than making money). That means corporate bonds will show default rates far too high to give investors positive returns. Zero corporate profits mean zero corporate income tax (although the tax code, like shareholders, collects from the winners without paying the losers), it also means layoffs and low wages that hurts tax collections and demands government spending. That makes government bonds unattractive either because the government becomes financially shaky, or prints so much money that the real return on government bonds is negative. Commodities and other real assets will suffer due to reduced business demand, and also reduced consumer demand because people are making less money.Therefore, I can convince myself that if stocks turn out to be a bad investment, there were no easily available, no work, good investments except maybe loading up my basement with canned food and guns.My conclusion in this is not that stocks have to be a great, or the best available, long-term investment; just that there's no great security without work. I sock away a lot of my extra money in the stock market, but I put most of my faith in investments that are not easily available and require work. Then you can throw out the general, long-term argument, your return will depend on your skill and effort. I don't mind that, it puzzles me that people should expect anything different.

Aaron said: "You are thinking like a Bayesian"Sort of, except I have two key differences: I am always very unconfident with my prior distributions, and I am always sceptical of the power of historical data to give information about the present parameters. Which leaves me with not much knowledge at all!I accept what you are saying about the problems of other instruments, yet can I raise the following scenario:I am investing with a finite (let's say 20 year) time frame. Am I better to invest in equities for 20 years or floating rate debt. I expect equities to do better over 20 years, but what is the probability with which it will do worse? This probability is certaintly real. If there is poor corporate performance, equities will go down, and while other stuff will do badly too, if you're in floating rate debt at least you won't lose your capital. So I certainly wouldn't say equities are the best investment, entirely because of my uncertainty in the parameters.

This topic really puzzles me so I went to read the papers ... They are great - full of historical numbers. NONE of the papers I looked analyzed what is the actual market. There are SP futures contracts going out 2 years and there are OTC equity derivatives markets where I have heard of 30-years deals. I am not equity derivatives expert but the futures contract appears to be priced on the basis of Libor swap minus the anticipated divident yield. For example the Sep 04 contract (906.3 while spot is 898.5) corresponds to increase of 0.5% while Bloomberg shows dividend yield 1.7% giving us two year the swap rate of 2.2%. I suspect the same pricing will apply longer dated OTC deals. The market is pricing the forward contract as if there was no equity premium. I guess this is in line with the risk neutral aspect Black & Scholes since the forward contract can be created from a call - put. Those who believe there is 4-5% equity premium should just buy the "cheap" forward contracts - so why don't they?

I am not an expert on the academic work in this area, but one question that I have had is what about survivor bias? Are these studies indicating outperformance including money invested in bankrupt companies or are they investing in an index with rolling membership?And would that be an argument for "indexing" -- i.e., indices do not go bankrupt?AP

Most recent JoF has an article about survivor bias and the equity premium puzzle. Basically it says that survivorship bias likely only accounts for 1% of the excess returns in the US market. It's not bad.The question I find FAR more interesting than the equity premium:why are stocks so volatile?My intuition tells me that when we are able to explain this by starting from the market microstructure we'll be able to account for the equity premium as well. I think niether problem is solvable without greater microstructure sophistication first.

"why are stocks so volatile?"My simple (not incorporating too much reading of overly theoretical journals) answer to this is that more and more people make their stock picking decisions based on "what do I think the price will be tomorrow" rather than "what is the company worth". Even those analysts that do attempt to use fundamental analysis are really just trying to come up with their excuse on "I think it's going to go up" or "I think its going to go down". This is the sort of mindset that leads to Amazon, Tulip mania, Enron, etc (although huge riches in the short term), and volatility that on all accounts exceeds the volatility of DCF analysis.

boy that was a little bit of a shot, wasn't it?most people on this board would save themselves quite a bit of trouble if they spent more time reading and less time writing/positing halfhearted behavioral explanations.

Apologies B2 if you took my comment personally. To be honest - I would never have picked JoF if I had actually noticed that you had referred to it in your post (and so have subsequently edited the reference). I'd actually love to read more of that sort of article if I was around the university more. Yes, knowledge of what other people have written (and often proved) is very worthwhile, but I don't feel that not having read it should prevent one from expressing thoughts - in many cases thinking outside the square and using experience (not that I'm claiming to have much!) can lead to insight. And some people (although thankfully few on Wilmott.com) tend to rely on quoting academic papers rather than actually understanding what the authors are saying. So what I meant by my initial phrase was that I wasn't an expert in what rigerous studies had shown, but I did feel that my gut feeling might be of interest to some. (I would have sent this private message if you had it activated - once you have read it please send me a private message and I will delete this clarification).

In the website of the AIMR (www.aimr.org) you can find a forum about the Equity Premium Puzzle with an excellent selection of articles.