The Inelastic Markets Hypothesis
Why do stock markets exhibit so much volatility? This is the question that Gabaix and Koijen seek to answer in this paper. They develop a new model and provide new evidence suggesting that this is because of flows and demand shocks in surprisingly inelastic markets.
In the simplest “efficient markets” model, the price is the present value of future dividends, so the valuation of the aggregate market should not change. However, we find both theoretically and empirically, using an instrumental variables strategy, that the market’s aggregate value goes up by about $5 (our estimates are between $3 and $8, and we will use $5 for simplicity in the theory and discussion). Hence, the stock market in this simple model is a very reactive economic machine, which turns an additional $1 of investment into an increase of $5 in aggregate market valuations.
They use a model with a representative consumer who can invest in a purely bond fund or in a mixed fund (which invests in both bonds and equities according to a mandate that specifies, for example, a ratio of 20%, 80% bonds and stocks respectively). They then trace out what happens when a consumer sells $1 of the pure bond fund and invests this $1 in the mixed fund. The mixed fund must invest this inflow into stocks and bonds: but that pushes up the prices of stocks, which again makes the mixed fund want to invest more in stocks, which pushes prices further up, and so on. In equilibrium, the total value of the equity market increases by $5. The authors call this high sensitivity of prices to flows the “inelastic markets hypothesis” whereby flows in the market and demand shocks affect prices and expected returns in a quantitatively important way.
In the empirical part of the paper, they provide a quantification of the market’s aggregate elasticity by using a new instrumental variables approach the “Granular Instrumental Variables.”
In Search of the Origins of Financial Fluctuations: The Inelastic Markets Hypothesis
Authors: Xavier Gabaix, Ralph S. J. Koijen
From: Harvard University, University of Chicago
What are the Granular Instrumental Variables
Granular Instrumental Variables (GIV) are a new set of instruments that allow researchers to establish causal relations in a wide variety of economic contexts. In constructing them, Gabaix and Koijen assume that many decisions are taken by a few large actors, such as firms, industries, or countries, whose idiosyncratic shocks (e.g., productivity shocks) affect the aggregate ones. These idiosyncratic shocks at the firm, industry, or country level are valid instruments for aggregate endogenous variables such as prices. They lay out econometric procedures to optimally extract idiosyncratic shocks from the data in order to create GIVs, which are size-weighted sums of idiosyncratic shocks.
[W]e show how GIVs allow for a novel estimation procedure: they yield an instrument that allows us to estimate the elasticities of both supply and demand. Indeed, idiosyncratic demand shocks to large firms or countries give a valid instrument for demand change – and thus allow one to estimate the elasticity of supply. They also allow us to estimate the elasticity of demand: the idiosyncratic demand shock of a large firm impacts the price, which changes the demand of other firms.
Granular Instrumental Variables
Authors: Xavier Gabaix, Ralph S. J. Koijen
From: Harvard University, University of Chicago
