I’ve done a big update of my notes on computation for heterogeneous-agent macro. They start with the endogenous grid method and end with by solving a HANK model using sequence space Jacobians. Codes are in Julia.
Forward guidance is a big part of modern monetary policy, but how effective is it? Johannes and I explore this Q in a model of durable consumption subject to fixed costs. FG Is much less powerful than contemp. rates. Short thread 1/n
In our model, the optimal monetary policy is essentially to target non-housing inflation and ignore housing. We show that a policy of targeting CPI inflation generates a large recession with negligible benefits from cooling down demand in the housing market. 5/9
We think a better description is that the consumption of housing services is supply determined with rationing occurring through search. Even with sticky rents, the effective price of finding housing services can rise and ration demand. 3/9
@JonSteinsson
It's hard to simultaneously signal an intention to respond aggressively to off-equilibrium inflation without responding strongly in equilibrium. I think this complicates the policy response when the model is telling you to look through inflation.
The premise of our paper is that the standard New Keynesian model is not appropriate for thinking about housing services. The NK model assumes production is demand determined, but if consumers want more housing, the economy cannot immediately create more of it. 2/9
Much of the discussion for down weighting the role of shelter inflation has been based on measurement issues. Our model here is not about measurement but rather about the conceptual role of housing rents in the inflation rate targeted by monetary policy. 7/9
Relatedly,
@neilmehrotra
and I recently wrote a research note arguing that CPI shelter inflation is likely to prove quite persistent owing to long estimated lags in the reset of contract rents. 9/9
Housing inflation largely reflects a change in the price of a good that has already been produced so the scope for misallocation is much smaller than its CPI weight would imply. 6/9
We consider the optimal monetary policy response to a shock for housing in both a static search model and in a dynamic two-sector New Keynesian model with search frictions in housing. Arguably, the pandemic represented a large increase in demand for housing. 4/9
A common Q: People finance durables at long-term interest rates so doesn’t FG still work through those rates? Actually no. What matters is the cost of financing today vs tomorrow. If short rates are low, we expect long-rates to rise so financing today is better. n/n
@BrzozaBrzezina
@ecmaEditors
@ben_moll
@jfwieland1
Yes, there is a theory of the bust as a response to a durables overhang from the boom (others have made this point). The economy will then have a period of low demand and monetary stimulus cushions the fall but elongates the period of low rates.
Why? Buying a durable good is partly a short-term decision. Do you do it today or tomorrow? The opportunity cost of buying today is the interest between today and tomorrow. This intuition is reflected in the smooth pasting condition for adjustment thresholds. 4/n
The main figure. x-axis is years until real rates are cut by 1% for 1 quarter. y-axis is effect on current output. The 3-eq model says real rates at all horizons are equally effective. The fixed cost model says FG is much less powerful. 2/n
Most of the action is from the extensive margin. This figure decomposes the output response into intensive and extensive margin contributions from durables. The extensive margin is much more sensitive to contemporaneous rates than FG. 3/n
@GeorgeSelgin
The exact driver of the relative price movement isn't crucial. We also do a "catchup" experiment in sec 5 that you might prefer. But to your Q, remote work = home office, etc. See Wieland and Mondragon's paper