Site Meter

Sunday, December 06, 2015

Is Okun's Law OK ?

I think this is the 4th thing I've written about this Krugman post "Anchors Away"* while alarmed by my obsession with that application of 1960s macro and 1960 (but not 61) macro to current issues, I do see some hope. I have made it to the last two paragraphs.

How much output growth would this involve? An Okun’s Law type relationship also works pretty well for the euro area as a whole, and gives a coefficient close to the US level:

[graph] So a 4 percentage point decline in unemployment would, if the historical relationship holds, mean 8 percent more in real GDP — that is, this naive calculation puts the euro area output gap at 8 percent, which is huge.

Should we take this seriously? If not, why not?

Glad you asked Paul. First I don't have a problem with the calculation that Euro block unemployment is 4% greater than it could and should be. The astonishing implication is that Euro block economies could endure unemployment as low as 6% without suffering ever rising inflation. I see no reason to doubt this. Even the famously dysfunctional Italian economy managed it (the stability of inflation is even more striking if one looks at wages not consumer prices -- the ups and downs of the consumer price inflation rate in the graph in the linked post correspond to sharp increases and decreases in the price of oil -- no new graph as this is not this post's topic).

But I am not so convinced by the Okun's law estimate and calculation. Krugman's estimate ( and Okun's) is a simple regression with no lags or leads. The regression is valid if unemployment is a function of current output so the immediate effect of a downturn is equal to the long run effect of a prolonged depression. However, in practice employment adjusts slowly in recessions with the peak of unemployment typically occurring after the trough of GDP. I think the coefficient of GDP growth on the change of unemployment reflects this slow adjustment of employment. I'd guess a long run Okun's relationship would give a number closer to labor's share so two thirds, maybe 1 but not 2. I won't estimate this coefficient using changes from 2007 to 2014 for different Euro group countries (although I should). It implies an output gap of roughly 3 to 4% which I find highly plausible.

I think this implies an average allowed deficit of 4.5 to 5% of GDP (IIRC the Eurocrats calculate that a 1% output gap causes the deficit to increase by 0.5% of GDP based on increased transfers as well as reduced tax revenues). *The pun in the title doesn't fit the content of the post and set a very bad example, which may have influenced me and led me aweigh from the balanced prose path appropriate for serious topics)


Blissex said...

«alarmed by my obsession with that application of 1960s macro and 1960 (but not 61) macro to current issues»

I am more alarmed with your dad not explaining to you when you grew up the facts of life with economists and politicians; while the photo of you with a child may mean that they explained you the less obvious ones about the flowers and bees. :-)

The facts of life on economists and politicians start with the realization that politicians don't go to economists asking for the best policies based on the most valid theories and evidence.

They go to economists (or central bankers) in one of three ways:

* "My voters or donors wants X, I can give them X, find me some kind of law or theorem Y that makes X sound scientific so I can make the opposition sound stupid when they object to X".
* "My voters or donors want X, but doing X costs Y, find me a way to shift most of Y on the voters of the opposition or on the next administration".
* "My voters or donors want X, that's not going to happen, so give me some good scientific sounding excuse to try and pacify them".

Politicians and their sponsors are usually cunning sharks, and it does not happen often that they genuinely need to ask what to do, what they need is cover.

Economists and central bankers that provide that can become quite wealthy, whether they provide that cover because that's actually what they believe or whether because they are responding to market demand.

Peter said...

"while alarmed by my obsession with that application of 1960s macro and 1960 (but not 61) macro to current issues,"

Should it be "macro and ... micro"?

Robert said...

@Peter I meant macro and macro, that is Okun's law which is 1960s macro and an expectations un-augmented Phillips curve which is macro from 1960 (but not macro from 1961 or later).

JW Mason said...

I'd guess a long run Okun's relationship would give a number closer to labor's share so two thirds,

I'm fascinated to learn that the level of capital stock is fixed in the long run, regardless of the growth rate of output. How did you make this remarkable discovery?

If you are working with a production function framework, which you are, then if a sustained deceleration of Y is accompanied by a comparable deceleration of K, then the long-run Okun coefficient will be one regardless of labor's share.