« Split the Labour leadership | Main | The strange death of the business vote »

July 27, 2016


Luis Enrique

it is perhaps a minor point of intellectual history but as the intro of paper I linked to on twitter ( http://www.nber.org/papers/w8387 ) explains, Friedman's original articulation of PIH much softer than extreme version that was then shown to be flawed.

Pablo Mira

We are overlooking here a devastating critic made by Shefrin and Thaler (behavioral economics):

Ralph Musgrave

Quite agree with Chris’s above basic point. Unsurprisingly, there’s a wealth of evidence as to what people do when their income changes (up or down). Plus there’s a wealth of evidence as to what they do when they come by windfalls.

Equally unsurprisingly is that peoples’ spending goes up / down when they experience an increase / cut in income, but not to the full extent of the increase / cut. Thus people obviously think and act TO SOME EXTENT in the “permanent income” way Friedman claimed.

Deviation From The Mean

"Economists are unlikely ever to uncover universal, general-purpose models."

It depends what you believe the purpose of those those models is. If it is to uncover the inner dynamics of an epoch, say the capitalist mode of production, then a universal model, of sorts, can be devised that uncovers the inner workings of the system, and what basis that system exists upon.

If we think economics is about 'moments', such as what investment decisions we should make at any particular time or what the government should do when faced with a problem then universal models are possibly inadequate. Though someone should tell that to financial advisors who seem to present graphs showing returns over say a 20 year period and assume this can be projected forward to the next 20 years etc etc

The above article seems to always assume the latter. I.e. is viewed through the spectacles of neo-classical economics.

gastro george

"In this sense, the question: “is the PIH true?” depends upon the context in which we are asking."

As you say, many hypotheses are only correct in certain circumstances. The problem is:

1. Individuals seeking to extend them to a generalised context.

2. Whether the truthful context is knowable in advance.

Ken Houghton

The Hall/Mishkin link can be made directly to NBER, where the paper is freely available.

Ken Houghton



If models are only correct sometimes, in some scenarios, then the models should specify what those scenarios are in advance.

"All models are wrong, but some are useful" would really be better stated as "All models are wrong. Some are useful, and some are dangerous. You'd better be damn sure you know which is which."

The comments to this entry are closed.

blogs I like

Blog powered by Typepad