John Campbell of Harvard has a nice paper published in a Canadian Econ Journal that estimates the equity risk premium. His conclusion: the world (and US) equity premium is around 4% currently.
The article isn't free, but if you are have access to a university library you can probably download it for free.
Given a 30 year bond rate of about 4.3%, this implies a long term return to stocks in the US of 8.3%. Why does this matter??? Well if you are assuming a 40 year investment horizon (someone who is, say 25 now) and you contribute $1,000 a month, 8.3% return will give you about $3.8M in your portfolio at age 65. But if you were using 11% (the long run historical return on equities) you would be expecting $8.6M. Given that most people are not saving enough, a lower return on stocks is not going to help.
RE the equity premium -- back around 2000-2001 when all these articles were coming out about the equity risk premium (specifically that it would be lower going forward), I read an interesting piece comparing the various findings. It pointed out the big differences in methodology -- chiefly, whether the premium was calc'd arithmetically or geometrically and what risk-free rate was used.
ReplyDeleteWhich brings me to my question for you -- what's the "correct" Rf to use for an equity premium? I see you used 30-year T-Bonds here, but other places (including the firm where I work) use the 1Mo T-Bill. Strictly speaking, is the long bond really free of risk? What about interest rate risk? And by using the long bond as your base building block for nominal equity returns, aren't you assuming that same interest rate sensitivity into your equity model?
Personally, I think the 1moUST makes more sense, but I'm interested in your response.
PS: found your blog thanks to FinancialRounds.