By the requirements of mainstream media protection of technical economics, Peter Coy’s protection of HANK (Heterogeneous Agent New Keynesian) fashions within the New York Occasions was really fairly good.
1) Consultant brokers and distributions.
Sure, it begins with the standard misunderstanding about “consultant brokers,” that fashions assume we’re all the identical. A few of that is the usual journalist’s response to all financial fashions: we’ve got simplified the assumptions, we’d like extra basic assumptions. They do not perceive that the genius of financial principle lies exactly to find simplified however tractable assumptions that inform the primary story. Progress by no means comes from placing extra components and stirring the pot to see what comes out. (I imply you, third yr graduate college students on the lookout for a thesis subject.)
However on this case many economists are additionally confused on this subject. I have been to fairly just a few HANK seminars during which distinguished teachers waste 10 minutes or so dumping on the “assumption that everybody is similar.”
There’s a stunning outdated theorem, referred to as the “social welfare perform.” (I realized this in graduate college in fall 1979, from Hal Varian’s wonderful textbook.) Individuals can have virtually arbitrarily totally different preferences (utility features), incomes and shocks, corporations can have virtually arbitrarily totally different traits (manufacturing features), but the mixture financial system behaves as if there’s a single consultant client and consultant agency. The equilibrium path of combination consumption, output, funding, employment, and the costs and rates of interest of that equilibrium are the identical as these of an financial system the place everybody and each agency is identical, with a “consultant agent” consumption perform and “consultant agency” manufacturing perform. Furthermore, the consultant agent utility perform and consultant agency manufacturing perform needn’t look something like these of any explicit particular person particular person and agency. If I’ve energy utility and you’ve got quadratic utility, the financial system behaves as if there’s a single client with one thing in between.
Defining the job of macroeconomics to grasp the motion over time of aggregates — how do GDP, consumption, funding, employment, worth stage, rates of interest, inventory costs and so forth. transfer over time, and the way do insurance policies have an effect on these actions — macroeconomics can ignore microeconomics. (We’ll get again to that definition in a second.)
Now uniting macro and micro is vital. Macro estimation being what it’s, it could be awfully good to make use of micro proof. This system kicked off by Kydland and Prescott to “calibrate” macro fashions from micro proof can be very helpful. Kydland and Prescott could have had a little bit of grass-is-greener optimism about simply how a lot exact proof macroeconomists have on companies and other people, however it’s a good suggestion. Including up micro proof to macro is difficult, nevertheless. Right here “aggregation principle,” typically confused with the “social welfare perform” theorem comes up, extra as a nightmare from graduate college. The situations underneath which the consultant agent preferences appear like particular person individuals are rather more restricted.
Like all good theorems, this one rests on assumptions, and the assumptions are false. The essential assumption is full markets, and particularly full threat sharing: There may be an insurance coverage market in which you’ll be compensated for each threat, particularly dropping your job.
A generalized kind nonetheless works, nevertheless. There may be nonetheless a consultant agent, however it cares about distributions. The consultant agent utility perform will depend on combination consumption, combination labor provide however now additionally statistics concerning the distribution of consumption throughout individuals. In asset pricing, the Constantinides-Duffie mannequin is a superb instance: the cross-sectional variance of consumption turns into a vital state variable for the worth of the inventory market, not simply combination consumption.
All financial theorems are false after all, in that the assumptions are usually not actually true. The query is, how false? Standard macroeconomics comes all the way down to an outline of how aggregates evolve over time, based mostly on previous aggregates:
[aggregate income, consumption, employment, inflation… next year ] = perform of [aggregate income, consumption, employment, inflation… this year ] + unforecastable shocks.
That is it. That is what macroeconomics is. Principle, estimation and calibration to determine the perform. If HANK is helpful to macroeconomics, then, it should be that including distributional statistics helps to explain combination dynamics. Actuality should be
[aggregate income, consumption, employment, inflation… next year ] = perform of [combination revenue, consumption, employment, inflation, distribution of consumption, employment, and so forth. … this yr ] + unforecastable shocks.