Thursday, March 27, 2014

Philip Pilkington — Krugman Uses ISLM to Proclaim Looming Fiscal Crisis, Denounces Those Who Don’t Use ISLM

Phil takes down Paul Krugman as a faux-liberal and an economist with a poor record at prediction.


Anonymous said...

The thing about IS-LM is that it is unfalsifiable, and can be made consistent with any possible economic circumstances whatsoever. Whatever the real interest rate happens to be, and whatever the level of national income, you can plot the point with those coordinates, and then draw a made up IS curve and a made up LM curve to intersect at precisely that point. It predicts nothing and "explains" everything.

Macroeconomics is a pseudo-science dressed up with a little elementary mathematics. Although economics is supposedly the science of the allocation of scarce resources, it's amazing how little help these guys are and how little they contribute to the national and global conversation about how to employ our resources.

circuit said...

Dan, I'm not sure about your claim that it predicts nothing and explains everything.

Take, for instance, the case of a government, via its CB, that decides to buy up the debt issued to 'finance' additional public expenditures. Both the LM and IS curves shift right to keep the rate of interest the same. In other words, teh model 'predicts' stable interest rates. And the well-informed economist/policymaker would also be able to say that the government debt held by the public would not increase.

My own view is that the model is appreciated because it is simple. It's a model with two equations! Nothing comes close to that degree of simplicity. Also, the fact that each curve can be used to represent monetary policy and fiscal policy, respectively, makes it a useful tool for policy-minded folks.

Also, I'm not sure about it being unfalsifiable. Many economists started avoiding the model when they realized the LM curve was pretty much horizontal (outside of the context in which the CB sets its rate using a reaction function, in which case an upward sloping LM is a reasonable representation of reality). ISMP was developed to account for this institutional detail.

Now, it has defects, but that's what you get from a two equation model.

Anonymous said...

circuit, why does IS-LM predict that when the government buys up public debt, both curves shift to the right by exactly the same amount?

Matt Franko said...

Whoa! pretty damning... rsp,

Roger Erickson said...

ISLMist fundamentalists.

Always attracted to what's simple ... no matter how unreal.

Perhaps the economics discipline needs a theory of fatal attractions vs adaptive selection?

circuit said...

Dan, apologies for the delay. I'm pretty unable to interact with you guys during the day.

I hope I'm understanding your questions but the answer is that when both curves shift by the same amount, the interest rate stays the same. The IS curve shifts right as a result of the amount spent by the govt on goods and services while the LM curve shift right when the CB purchases the govt securities. If these are the same amount, the rate of interest remains at the same level, all other things being equal or held constant.

Roger: A reliable, simple model enables policymakers without much technical training (eg, Mosler's political nominees to the Fed) to understand monetary and fiscal dynamics without having to go through too much technical literature.

circuit said...

...when both curves shift RIGHT by the same amount...

Unknown said...


let me start by saying that nothing you are saying makes any sense to me at all. I see ISLM as a complete waste of time and space. Its completely irrelevant and useless for describing our modern economy.

With that stipulated. I have a question for you. In this statement you make:

"Take, for instance, the case of a government, via its CB, that decides to buy up the debt issued to 'finance' additional public expenditures. Both the LM and IS curves shift right to keep the rate of interest the same."

You just described QE. The Fed buys TSY securities. Anytime the Fed contrives to add more reserves than banks need for settlement the FFR eventually goes to its natural rate (w\ a floating exchange rate) of zero. So how can the interest rates stay the same?

Do you understand that the Fed is the monopoly supplier of reserves and as such is responsible at all times for setting interest rates?

'In other words, the model 'predicts' stable interest rates."

Right, and the model is clearly wrong. Why would you use an inaccurate model?

Why wouldn't you just accept that the Fed exogenously sets the FFR as the monopolist and that the baseline rate they set is determined endogenously via private sector activity?

Have you ever heard of the 1951 TSY-Fed accord? It clearly demonstrates that the Fed aka the Govt always has ultimate authority over its own interest rates? Therefore any model that doesn't acknowledge this simple fact is completely useless?

circuit said...

Auburn, look the point of my comment is to show how IS-LM can be used to explain exactly what you are saying: the Fed has the ability to set its policy rate. It also debunks the myth that government spending crowds out private sector spending by increasing the rate of interest.

Anyway, I actually do have problems with the model so I'm not the best person to defend it wholeheartedly (I wrote a post a while back). But I tend to consider its defect don't justify the amount of criticism it currently attracts from some folks.

For those who are interested, I thought the following paper from the Review of Keynesian Economics comparing different models of fiscal policy, including the IS-LM:

The paper shows that the simple IS-LM fares pretty well compared to other leading models.

Tom Hickey said...

Nothing wrong with simple models as either teaching tools or "gadgets" that summarize some point. Minsky's financial instability model can be explained in a paragraph, and the fundamentals of macro based on Y = C + I (closed economy), Y = C + I + G (closed economy with government), and ) Y = C + I + NX (open economy). It's not the simplicity of the model that is the issue but the use to which it is put. Thank heaven there are simple conceptual and algebraic model that can capture the basics of MMT or we would never get anywhere.

The problem is not so much that IS-LM is too simple. It illustrates the points it makes pretty well. It's just the wrong model to be use in explaining the macro situation in a way that is practical.

The Krugman cross is a simple model, too, and it is the one that Krugman should be using instead of IS-LM. He's already got it in his back pocket but isn't yet ready to take it out for some unknown reason.

Anonymous said...

circuit, it seems to me that IS-LM doesn't predict the effect you are describing, but that you are juts assuming the curves shift by the same amount, and that is why interest rates are unaffected.

So this juts seems like an example of what I was talking about. No matter what happens to the interest rate, the IS-LMer can retroactively draw the curves wherever they like to generate that interest rate.

Tom Hickey said...

Phil Pilkington's post on IS-LM and Krugman shows how Krugman predicted wrongly based on his reading.

Like most models, IS-LM show certain relationships that are assumed to hold cet. par. But unless theory is added to the model and the context fleshed out, there is no causality.

If independent and dependent variables are assumed, then causality may enter the picture, but only if there is some some causal explanation like a mechanism in addition to mere correlation. The so-called transmission mechanism in many economics models is (rational) expectations. That looks to me begging the question in that anything can be made to follow from (rational) expectations.

This is a problem for economics since there are no causal laws in the social sciences that resemble the laws of nature of the natural sciences. Calling something a law doesn't make it so, and appealing to subjective "forces" makes the argument weak.

circuit said...

Dan, what I described is the textbook IS-LM version of the endogeneity of the money supply when the Fed seeks to control its policy rate. I say "textbook" because it is actually found in some textbooks from the 1970s. Robert Gordon's 1978 "Macroeconomics" is an example (page 131). It is not a question of assuming that both curves shift right by the same amount. Rather, both curves shift right by the same amount because that's how you model fiscal policy with an endogenous money supply using IS-LM. As I explained before, the rightward shift of the IS curve depicts the govt spending 'financed' by bond issuance while the rightward shift in the LM curve depicts the Fed's accommodative stance to counter the contractionary effect of the bond issuance.

As for Krugman, his mistake was to forget (or ignore, or be ignorant of the fact), as many Newer-Keynesians do, that IS-LM can be used to model fiscal and monetary policy with an endogenous money supply. In his head, he applied the IS-LM with a fixed money stock, which is depicted by shifting the IS curve right. Period. The effect in this scenario is that the rate of interest increases.

IMO, this is an example of Krugman making a mistake and not properly applying a model. It's not a case against the IS-LM.

circuit said...

I also meant to thank you guys for engaging with me on this question.

Jose Guilherme said...

There seem to be two possibilities for central bank reaction, when the government raises discretionary spending.

The CB may keep the interest rate at the same level. In that case the "LM curve" is horizontal.

Or the CB may raise the interest rate. In that case, the "LM curve" slopes upwards.

In both cases the IS/LM model or - even better, as Krugman has tacitly acknowledged in recent postings - the IS/MP model does a pretty good job at describing economic reality.

An economic reality with an endogenous money supply, of course.

So the attacks on Krugman because he uses that model are just a waste of time and energy, IMHO.

Krugman should be denounced every time he strays off course, as a result of his close links to the Democratic Party establishment. That does happen once in a while, and shows him at his worst.

But he should be supported most of the time, because he has stood for a consistent opposition to the Malthusian, austerian mindset that has plagued the western leadership in the past decades.

IS/LM may have methological problems, questionable epistemological foundations, yes. It may well be "impure".

But why care? It is useful for summarizing, explaining and - yes - predicting economic outcomes. One could hardly ask for more from such a simplified gadget.

Tom Hickey said...

The reason that IS-LM is a problem as used by Krugman et al is that he (they) focus on monetary policy as "the answer" rather than fiscal.

Krugman only admits preference for fiscal in a liquidity trap and has said that once the US is out of the trap, then the test of the MMT will come and based on his reading of IS-LM, he thinks MMT will fail and I-LM will prevail.

So in Krugman's view, IS-LM is the right model and the MMT-Godley model is wrong, when taken generally.

This is where the debate now stands as far as Krugman is concerned, waiting for the test he has proposed after the US has exited the "liquidity trap" he has described.

Tom Hickey said...

The more specific thing wrong with using IS-LM as it is being done is that it is being marshaled to promote negative rate setting as a solution the lagging demand that is slowing the recovery — which from the MMT POV is bonkers.

Jose Guilherme said...

IS/LM (or IS/MP) can be drawn always with a flat LM curve in order to correspond to Mosler's "permanent zero interest rate" rule.

In that case only fiscal policy remains and IS\LM will be fully compatible with the MMT/Godley POV.

As for the empirical test suggested by Krugman, one risks having to wait many years for it, considering the apparent intention of western Central Banks to keep rates low for a long, long time.

Tom Hickey said...

Right, IS-MP replaces IS-LM in an endogenous money context in which the cb sets the rate. IS-LM is suited to an exogenous money context with the cb adjusting quantity.

David Romer (spouse of Christina), Keynesian Macroeconomics without the LM Curve

William Nordhaus, A Brief Exposition of the IS-MP Curves: A Replacement for the Traditional IS-LM Curves

Nordhaus now edits the Samuelson text on Economics.

Romer and Nordhaus are cutting edge New Keynesians, so Krugman should be able to be persuaded to accept the IS-MP instead of IS-LM as better reflective of the context.

Jose Guilherme said...


Krugman has already introduced MP through the back door, since many of his recent postings exhibit a MP line instead of the usual LM curve.

He has even mentioned (approvingly) endogenous money in the framework of a "Tobin approach" to macro.

So Krugman is in practice much closer to MMT than he cares to admit.

Anyway, it's a great thing that he is providing free marketing on his blog for MMT and the likes of Syll, the Lord Keynes blog and other expressions of so-called heterodox (in fact, plain realistic) economic thought.

circuit said...

Tom, you wrote:

"The reason that IS-LM is a problem as used by Krugman et al is that he (they) focus on monetary policy as "the answer" rather than fiscal."


Later, you wrote:

"The more specific thing wrong with using IS-LM as it is being done is that it is being marshaled to promote negative rate setting as a solution the lagging demand that is slowing the recovery"

This is also correct, but the IS-LM is not the problem. As I explained above, the IS-LM allows for another policy prescription: expansionary fiscal policy with an accommodating CB (monetary-fiscal stimulus).

Krugman seems to prefer a "monetary policy only" prescription. This is not surprising given that his '98 paper on Japan is exactly that: a monetary policy-centered solution to the liquidity trap. IMO, this is THE problem with Krugman's overall approach to policy.

Finally, I agree with the spirit of Jose's comments.