Now at a panel on microfoundations with @jasonfurman and @Claudia_Sahm!!
Furman kicks it off.
Asks: Were microfoundations just a blizzard of useless algebra, or do they really help us think about policy?
Recently, more and more research shows you can't think about macro without thinking about what's going on with individuals and firms.
Asks: Were microfoundations just a blizzard of useless algebra, or do they really help us think about policy?
Recently, more and more research shows you can't think about macro without thinking about what's going on with individuals and firms.
Furman: Inequality matters for macro and we can't think about inequality if we ignore microfoundations.
Amir Sufi says: Macro data *are* super useful. Gotta think about aggregates.
BUT, applied micro methods yield very convincing evidence about phenomena that have macro importance.
BUT, applied micro methods yield very convincing evidence about phenomena that have macro importance.
For example, Sufi says, higher debt leads to a much larger contraction in household spending in response to unemployment. This is now well established.
This has big implications for macro.
It also means representative agent macro models shouldn't be used for business cycles.
This has big implications for macro.
It also means representative agent macro models shouldn't be used for business cycles.
Another finding from micro: Monetary policy affects households a lot through the mortgage refinancing channel.
So macro models need to include some representation of the housing market in order to predict the effects of monetary policy.
So macro models need to include some representation of the housing market in order to predict the effects of monetary policy.
Next is @Claudia_Sahm.
Sahm: Are U.S. households ready for another recession? NO. But that's typical. Many many U.S. households have low liquidity and little savings.
Sahm: Are U.S. households ready for another recession? NO. But that's typical. Many many U.S. households have low liquidity and little savings.
Sahm talks about a survey asking people what they would do in response to an unexpected $400 expense. 40% of households wouldn't be able to pay without borrowing or selling something (or couldn't pay at all). When the survey started, it was 50%.
When households think of credit cards as their financial buffer, it means they could be really hurt by recessions!
Sahm: We see low liquidity all up and down then income distribution. Recessions happen when middle-class people pull back on their spending.
Sahm: Models that assume consumption smoothing have big problems.
Furman then quotes a tweet of mine, embarrassing me thoroughly. 😂
Next up: Janice Eberly.
Micro data has much more variation, she notes. Even without heterogeneity in models, you need variation in the data to estimate lots of aggregate relationships.
Micro data has much more variation, she notes. Even without heterogeneity in models, you need variation in the data to estimate lots of aggregate relationships.
Eberly: Investment didn't bounce back after the Great Recession. Why?? Residual effects of the crisis?? Actually, weakness in investment pre-dated the crisis! Investment was weak starting around 2000!!
Eberly: Maybe investment was weak because companies were shifting to intangible capital instead of physical capital.
(In other words, this is about digitization/ephemeralization of the American economy.)
(In other words, this is about digitization/ephemeralization of the American economy.)
Eberly: Intangible capital just isn't included in macro data. We need micro data to measure it.
Stiglitz was scheduled to appear next, but he couldn't make it and has been replaced with @ThomasPHI2.
This is all to the good. 😊
This is all to the good. 😊
Philippon: Micro data might be able to tell us why productivity is slow, which is super important for macro.
My thought: It's really interesting how "microfoundations" used to mean "behavioral and technological assumptions that were probably bullshit but which let us solve models easily and/or reach our desired conclusions", and now it means "micro data".
This is good, this is good.
This is good, this is good.
Philippon is listing reasons why micro data is hard to use for macro. Good that there's at least one skeptic on this panel! Otherwise there would be no real disagreement. There are no Minnesota macro holdouts on the panel. 😊
Rabbits say hello 
Philippon: Micro data is limited because of lack of price data.
(Seems fixable, no?)
(Seems fixable, no?)
Furman: "I agree with everyone about everything"
The discussion quickly turned from recessions to the issue of market concentration and market power.
A small but eye-opening illustration of how priorities in the econ world are rapidly changing.
A small but eye-opening illustration of how priorities in the econ world are rapidly changing.
The focus of this panel is entirely on micro DATA. Not on micro ASSUMPTIONS. This totally changes the meaning of the term "microfoundations"!!
It means Milton Friedman's terrible "pool player" argument is on the way out.
noahpinionblog.blogspot.com/2016/06/the-po…
It means Milton Friedman's terrible "pool player" argument is on the way out.
noahpinionblog.blogspot.com/2016/06/the-po…
Friedman argued, basically, that micro assumptions in macro models shouldn't be checked for realism, because macro models are "as if" models, and thus whether the microfoundations correspond to reality doesn't matter.
This, of course, is a recipe for crappy, broken models.
This, of course, is a recipe for crappy, broken models.
But all these economists are on stage talking about...checking the assumptions in macro models against micro data!!
The intellectual hegemony of Friedman's bad argument seems to be slowly disappearing, and that's a very very good thing.
The intellectual hegemony of Friedman's bad argument seems to be slowly disappearing, and that's a very very good thing.
Think this is a good place to end the thread! Thanks for tuning in...
(end)
(end)
