Thursday, May 10, 2012

How do economists know that the effects of demand fluctuations on output are temporary ?

Warning*

Update: a link a veritable link

That is
http://d.hatena.ne.jp/himaginary/20120513/How_do_economists_know_that_the_effects_of_demand_fluctuations_on_output_are_temporary

By the way have I ever mentioned that I hate captcha. I found the comment box as clearly labeled in Japanese (without using goigle translate) but, on my first try, I mistook a v for a V in the captcha and had to coment again (without reading the captcha erroe message in Japanese so I wasn't sure I wasn't double commenting)

My use of "know" is ironic. I don't believe in knowledge and I especially believe that macroeconomists such as myself don't know jack shit (sorry abstract away from jack shit). What I mean is that standard advice to monetary authorities is based on the assumption that, in the long run, monetary policy only affects inflation, so the best rule is an inflation target. Also it is agreed that DSGE models are important and useful and macro DSGE models eal with fluctuations around an exogenous balanced growth path. I we don't have good reason to assume the effects of demand shocks are temporary, we have wasted a lot of time.

I will attempt to give an entirely fair summary of the case that demand has temporary effects on GDP. Any suspicion that I am setting up straw men or parodying arguments should be stated in comments preferably with a cite which provides a trace of evidence. It will be news to me (unless the cited authors are co-authors of mine).

I honestly believe that the argument is
1) in the long run GDP per capita is determined by technology.
2) technological progress is exogenous and hasnothing to do with economic conditions
3) QED.

In this post, I will ignore DeLong and Summers and accept claim 1, because claim 2 is necessary for theargument and obviously nonsense.

It is not based on the belief thattechnology really comes from outside of theeconomy. Also it is not based on any economic model of technological progress. Standard models of R&D do not imply this property. Models of learning by doing sure don't imply this property.

I think theargument isthat theory tells us something very important (must be wrong theory predicts and must be tested - it doesn't answer) when theory implies no such thing.

OK why do macroeconomists make an assumption which directly implies our most important conclusions (I honestly think our only conclusion but I am trying to be more than fair) and which doesn't fit the most nearly relevant theory ?

1. The only country which counts is the USA. US GDP keeps touching thesameexponential growth path (except for the great depression and WW II). So that's theway things are. The claim that this is acoincidence implies that thepattern should not be found for any other country ever. That happens to be true, but really who knows if these alleged "other countries" exist.

2. We are the tools of our tools. We can program computers to simulate linear models and we can linearize around a steady state so there must be a unique balanced growth path which we can normalize to a steady state. (thereaderis now thinking "what a tool" and is asked not to deny itas I can rationally forecast the reader's reaction).

3. Solow assumed exogenous technological progressand he won a Ricksbank Nobel memorial prize.

I think this is it. If some economist assumes something for. Convenience and gets cited a million times, then it is a standardassumption which is better than being merely true.

We assume things, because weassumethings. Only one deviation from the standardmodel is allowed per article. Nominal rigidities *and* endogenous technology arenotallowed in the same model. Therefore monetary authoritiesought to simply target inflation.


*This post will not be up to my usual standards of intellectual humility and diplomacy, because iAm typing on an iPad. IHate typing. On an iPad but my non iComputer broke and my new one hasn't arrived.

No comments: