Some stream-of-consciousness note-taking from the research for my upcoming Equity Risk Premium video. ⠀ Thanks @icallbullshit for recommending this video. Really fascinating watch: ⠀ https://www.youtube.com/watch?v=YDDAKWtMeD0 ⠀ One thing that sent me down a rabbit hole was learning what Mehra and Prescott actually meant by “risk” in their original 1985 paper on the Equity Premium Puzzle. ⠀ Their model was specifically focused on consumption risk. Very roughly, the logic was something like: ⠀ Risk premium ≈ how badly returns covary with bad consumption states, scaled by how risk averse people are ⠀ In other words, an investment should be particularly painful if it performs badly at the same time that your consumption is already being squeezed and an extra dollar is especially valuable to you. ⠀ Then they actually looked at U.S. consumption data from 1889 to 1978, specifically real per-capita consumption of nondurable goods and services. ⠀ And this wasn't exactly a peaceful 90-year stretch. It included the Panic of 1907, World War I, the Great Depression, World War II and the 1970s oil shock. ⠀ Yet aggregate consumption was surprisingly smooth. Annual consumption growth had a standard deviation of only around 3.6%. ⠀ That was fascinating to me. Their model could explain some equity premium, which makes intuitive sense. Risky stocks should command a premium over safe assets. The problem was that measured consumption risk wasn't remotely large enough to explain the size of the historical premium without assuming people were implausibly risk averse. ⠀ And that's the actual puzzle. Not why an equity premium exists, but why it has historically been so large. ⠀ Of course, this was 1985. We have much more nuanced models now, and decades of research have proposed things like rare disasters, survivorship bias, myopic loss aversion and other explanations. There still doesn't seem to be one generally accepted answer. ⠀ Still learning, but I think it's useful to understand what the initial efforts were actually trying to explain before getting into all the models that came after them.
You are welcome Kar, us fellow nerds have to stick together.
Michael Conroy@conroy119 · 9h
One thing that bothers me about the puzzle is how much confidence gets placed in “real” historical returns and consumption data. Neither is directly observable, both depend heavily on estimates of inflation and purchasing power over very long periods. Then layer in monetary debasement. Cash and Treasuries are nominal claims, so persistent inflation and currency expansion can quietly destroy their real value while productive assets can reprice with the currency. This 1889–1978 sample crossed multiple major monetary regimes: the classical gold standard, creation of the Fed, suspension and devaluation of the dollar against gold in 1933–34, Bretton Woods, and finally the collapse of gold convertibility in the 1970s. So I wonder if part of the “equity premium puzzle” is less “why did stocks return so much?” and more “are we overstating how safe and economically neutral the supposedly risk-free asset really was?” Especially when the monetary system itself changed so dramatically underneath it.
Canadian Investor@canadianinvestor · 12h
I don't need more rabbit holes haha but if you want to build a new calculator I have an idea ....lol
See the full comment section 👀Sign up for the full Blossom experience!
