Kevin Murphy models the stimulus–and the results aren't pretty

Geoffrey Manne —  20 January 2009

A great video from the University of Chicago here with comments from John Huizinga, Kevin Murphy and Robert Lucas. John Huizinga also wonders if we’re calculating the costs. Robert Lucas is skeptical.

But Kevin Murphy’s discussion is (not surprisingly) worth the price of admission (I only wish the video showed the slides). He puts some mathematical meat on the bones-answering my earlier question with a lot more variables than Krugman was able to muster.  His model itself raises a range of interesting and essential issues, seemingly glossed over by the stimulus fundamentalists (it’s their turn thusly to be tarred)–including (among others):

the relative inefficiency of government in allocating resources;

the conflict between stimulus and investment (between deploying underemployed resources and improving productivity);

the prospect of drawing otherwise productive resources into less-productive stimulus (Murphy: “You have to remember, even with 7% unemployment, 93% of the resources out there are being employed somewhere else. So when I try to draw things in, chances are I’m going to draw a lot of resources out of other activities.” (one might call this an error cost, but the Krugmans and DeLongs of the world don’t believe that governments can commit errors, unless led by Republicans, of course.));

the problem of diminished private savings (Ricardian equivalence);

the fact that even unemployed resources have value greater than zero (even unemployed people value their own time and reallocating otherwise-inefficiently employed resources to other sectors is a value);

deadweight cost of distortions from taxation.

At bottom, Murphy has this equation (I don’t know how to write Greek letters in WordPress, so bear with me):

f(1-L) > a+d

f = how much comes out of idle (as opposed to currently employed) resources

L (Lambda) = the value of idle resources (so 1-L is the gain from moving idle resources into production)

a = government inefficiency cost

d = deadweight loss

Iff the left-hand side is greater, the stimulus would be worthwhile

Murphy’s bottom line, given his parameters (f=.5, L=.5, a=positive, and d=.8):

Unless government is 55% more effective, more productive, than private output, it’s not looking good.

As Murphy notes, others like Christina Romer and Paul Krugman may believe that government is in fact much more efficient. I think the evidence is decidedly against them, but it is helpful to see what it comes down to. So what do you think?

Seems to me that Murphy has produced an enormously careful and valuable analysis of the sort that Krugman will dismiss as ideological, as he does with everyone who disagrees with him.  It is an embarrassment that Krugman (and not Murphy, et al.) has sway over the new US government and the intelligentsia (read: New York Times readers) that will enable the coming stimulus debacle.

UPDATE: O frabjous day, Callooh! Callay!–Murphy’s slides are available here (pdf). (HT: David Henderson).

UPDATE II: Brad DeLong agrees with me!  So does Matt Yglesias (link at Brad’s URL).  I mean, they don’t agree that Kevin is right, just that he presents the best framework thus far.  I would just like to go on record as having claimed this well before those jokers got there.  (HT: Tom Smith). While you’re visiting Brad, do take a look at the comments for a sense of what happens when people drink the DeLong/Krugman Kool-Aid.  People who dismiss Kevin Murphy as a hack are petulant, ignorant fools not worthy of the slightest consideration (beyond this). At least DeLong himself has better sense. (Although he esimtates alpha at zero! At zero!!!! Hahahahahahah!!!!! Oh, that’s rich.)

Geoffrey Manne


2 responses to Kevin Murphy models the stimulus–and the results aren't pretty


    I posted this on econlog so I might as well post it here as well. Obviously I think the core assumptions of textbook macro are pretty awful and near worthless in analyzing the current situation. And if you start with bad assumptions, your superstructure’s not gonna be much better.

    Here goes nothing….

    Well, I have about as much faith in your models as I do in medieval medicine. Remember, to be a science you must attempt to be descriptive. It is not enough to drag around historically revered models and look for confirming evidence.

    Savings – investment = Net exports? Oh really? Only if you conflate separate processes that are best viewed separately. Savings does not equal primary financial flows does not equal business investment.

    Divide the economy into real and financial layers. In the real economy you have payments for goods and services, wages, taxes, etc, between four nodes: [H] [G] [F] and [W] (households, government, firms, and outside world. If you like you can label payment flows from [H], [G], [F], and [W] to [F] as C, G, I, and X, illustrating PY = C + I + G + NX. On this we all agree.

    Now one unavoidable reality of such a diagram is that while any one node can run a surplus or deficit, aggregate surplus is zero. Moreover, domestic surplus equals world deficit (or vice versa, in fact).

    So domestic savings equals net exports. Without investment. Investment is a payment flow from firms back to firms which at least on the aggregate level is self-financing. Which is why it is no surprise that in the real world investment is such a volatile a component of GDP and in fact has very little to do with savings.

    Surpluses at individual nodes flow through the financial market as primary flows. Or not – when markets freeze up there are no flows. Moreover, financial flows can be used to support business investment (confirming evidence!!!) but can also be used to support deficit spending as well as speculation.

    Maybe, just maybe, imbalances in the real economy led to unsustainably high primary flow volume which led to both overinvestment in housing and over-leveraging on Wall Street. The growing trade deficit was part of the problem – Bernanke’s global savings glut. But don’t forget household income polarization — if the top 1% of households receive 24% of income, of course there’s gonna be a helluva lot of money looking for a creditworthy borrower.

    But if you assume savings = investment, you have just assumed away the whole problem.

    Don’t even get me started on your equilibrium assumption or your production function assumption. Capital is a constraint on output? On which planet? Obviously there’s too much capital chasing too few opportunities. At the margin, capital is available instantly at very reasonable rates. And Demand is not a bottleneck? True only for plant managers in the former Soviet Union.

    Shame on you for ignoring how the world works! Don’t obfuscate your mediocre assumptions with a veneer of mathematical sophistication. And certainly now is not the time for smug victory laps!

    Tear up those crappy models and attempt descriptive realism.

    It’s easy if you try!


    That’s an excellent video. And post.

    The list of economists skeptical of the stimulus is starting to look more and more impressive though I remember reading there weren’t any out there) but I wouldn’t hold my breath for any of the stimulus supporting economists to address it on the merits. Indeed, the ease with which Professor Krugman is able to glibly slap around Bates Clark and Nobel winners without addressing their ideas on the merits is disturbing and, I think, sad for the economics profession.