efalken wrote:You appear to agree geometric vs. arithmetic average as an appropriate adjustment, so that's 2%.
Yes, but when you say: “Per the ERP being 5%: I think it's more like 2%, which is near what Fama/French, or Ivo Welch, think (they are around 3%)”, do you refer to mean or geomean?.
efalken wrote: You state "The ERP is a forward looking concept. The historical ERP does not reflect the expected ERP priced by the market over this period." That seems rather anti-empirical. 100 years and 17 countries is a lot of data. I imagine another lifetime of data would also be insufficient, so it seems non-falsifiable via this reasoning.
If we assume normally distributed returns with an annualised SD of 20% and an annualised expected return of 3% the chances are about 62% that the actual outcome will fall outside the range 2-4% after 100 years. We basically have one 100 year independent data point (data from the 17 countries are not really independent) from an unstable distribution. We cannot infer with any confidence from the one 4% historic geomean ERP for the world index whether the market priced a 1% or 7% ERP. So I don’t think empirical research into the ERP tell us much about the forward looking ERP.
More importantly, if there actually was proof that the priced ERP = the realised ERP, it would imply that the market has psychic powers. The stock market’s purpose is not to predict the future. Its purpose is to move capital from those who have it to those who can use it productively, and it certainly does not need to predict the future to accomplish this.
efalken wrote: On taxes and transaction costs, you argue smart investors pay much less. I agree. But even smart investors could not avoid significant transaction costs until the 1970's, so all that data prior to that periods needs such an adjustment. Average returns to average investors are the data that supports or rejects asset pricing theory. Taxes too are generally ignored, but Blum and Gannon went through a pretty transparent and reasonable simulation and found the 3% adjustment.
You state "the market correctly prices only for taxes and costs incurred by the smart money to eliminate this arbitrage opportunity." Perhaps this is why the historical return on stocks since 1990 has been so meager, because prior to this, costs were significantly higher to even the smartest investors (low cost mutual funds were pretty insignificant prior to this time).
On taxes. You think it reasonable to assume that historically the average $ invested in the US stock market lost 3% per year to taxes. The Shiller data gives us a nominal geomean return for US stocks of 8%. If we assume your numbers are correct we have to deduct 6% from this to account for costs and adverse market timing. So investors would have reported a 2% return to the IRS which you believe was taxed at a rate of 150% (3%/2%).
The research you cite seems to ignore the fact that a huge proportion of invested funds is tax advantaged (pensions, insurers, foreign investors etc).
I would like to know if your argument that the stock market prices on a risk neutral basis applies to the cost of capital as well. Returns = cost of capital, so do you contend that companies expect that the cost of debt and equity capital are equal?
I don’t see why only returns to investors should count towards the ERP. Returns which end up in the pockets of other market participants (the middle men and helpers) is still a reward for the risk incurred by the investor – these participants have to either provide a useful service to investors, or find a way to extract gains from foolish investors.
Your argument about a correlation between lower costs and meagre recent returns would be persuasive if valuations went up as costs came down and stayed there. The fact that valuations have come down from their absurd levels counters this argument. The S&P500 real return 1990-2010 was 5.7%- not exactly meagre.
I believe a successful market has to return the initial dividend yield + the growth in dividends to long-term buy and hold investors less a bit of dilution. It also has to pay arbitrageurs and liquidity providers a fair wage for performing useful functions like tightening spreads and eliminating inefficiencies. Other investors are engaged in a zero sum game.
Those long term investors always incurred low costs, they only pay trading costs on their initial investments, when they reinvest divs and rebalance. This has always been an order of magnitude less than 3% per year. ETF’s and index funds has only benefited the small buy and hold investor which has always been a small proportion of the market.