Microstructural Interpretation of Inelastic Market Hypothesis
(Published by JP Bouchaud | June 2022)
My paper on a microstructural interpretation of Gabaix & Koijen’s Inelastic Market Hypothesis came out last week.
Rather immodestly, I am particularly happy with it. It is the offshoot of twenty years of research at CFM on market impact and the order-driven view of financial prices. To wit, the fact that prices are mostly influenced by what people do (how much and how fast they trade), independently of whether they are “informed” or not – rather than by news about an elusive fundamental value.
This might sound like pushing open doors, but is (still) a rather heretic view in many economic departments. If there is no new information, why should the market capitalization of a company change just because of trading activity? OK, maybe buy (/sell) trades do impact prices temporarily, but after a short while, impact should dissipate and price revert to fundamental value. This is the traditional efficient market hypothesis (EMH).
Instead, long term impact of uninformed trades does not vanish entirely. In some sense, the mechanical impact of uninformed trades *becomes information* for other market participants, who adjust (in a statistical sense) their reservation prices accordingly. This, as explained in my paper, leads to a permanent impact which has exactly the magnitude found by Gabaix & Koijen’s – i.e. buying 1$ of a stock *without any “real” information* increases the market capitalization of the firm by 1$.
This is a big stone in EMH’s shoe, but it explains many long standing “anomalies” – like the excess volatility puzzle, or the trend following puzzle, bubbles, etc. Prices and value are, on the short to medium term, disconnected. When the disconnect reaches Fischer Black’s proverbial “factor 2”, slow mean reversion forces come into play and ensures that on the very long run (> 5 years) prices and value are indeed cointegrated, see our “Black was right” papers here and here.
