IS-LM Keynesianism, why not and which alternatives?

I am sad that the IS-LM debate devolved into IS-LM vs. close substitutes, because I meant to raise a broader set of objections to one particular kind of technocratic curve-shifting as the foundation for macroeconomic thought.  Let me list a few alternative starting points for macroeconomics:

1. Public choice economics (still the most underrated, in today’s profession)

2. Growth theory (as distinct from the view that all business cycles are simply fluctuations in the rate of growth)

3. The New Institutional Economics of property rights and incentives

4. Financial asset pricing theory (as distinct from the view that financial markets are always efficient)

To see what this all means, let’s consider the euro crisis.  I start with #4, financial asset pricing theory, and consider whether, if “euros in Greek banks” and “euros in German banks” are fundamentally different assets, perhaps they should have different prices.  If prices cannot adjust, quantities will.

An IS-LM approach will focus on flows and be distracted — incorrectly — by claims that Ireland, Spain, etc. were not in bad fiscal positions before the crisis hit.  That’s wrong, they were writing intolerable naked puts all along.

A macro approach should then move to public choice theory and interpret the governance of the EU and ECB and the coalitions in the various European governments.  It wasn’t (for the most part) deliberately set up so that politicians could play short-run fiscal games, but that is in fact why a lot of politicians supported the eurozone.  Cheap borrowing brought a big party, but it was a ticking time bomb from the beginning, as recognized by Milton Friedman and others.  Based on Buchanan and Tullock (Calculus of Consent), the analysis can move forward with some understanding of why its governance is unworkable in a crisis, and with an understanding of why the Germans originally insisted on such a governance scheme at all.  I also wouldn’t mind a citation to Hayek and his critique of French rationalist constructivism, you won’t find that in the IS-LM model either.

We can put all that together, combined with a theory of bank runs, and then we see there will be strong and perhaps intolerable deflationary pressuresMaybe one could use IS-LM for this part of the modeling, but I’ll stick with the kind of ideas you find in Irving Fisher or Scott Sumner.  They are simpler and retain the core point that deflationary pressure can be very bad.  As Scott notes, the sort of interest rate issues raised by IS-LM are more of a distraction than anything, at this point for this problem.  And no, the eurozone for the most part has not been in a liquidity trap.  Nonetheless aggregate demand really does matter in this setting.

One often reads that Italy is the linchpin of the euro crisis.  To understand Italy we should look to growth theory, the new Institutional Economics, the theory of corruption, theories of political gridlock, and related ideas.  Toss in Edward Banfield.  Italy’s growth problems predate the immediate mess in the eurozone and they are not plausibly pinned on deflationary forces; the country hasn’t grown much in a decade.  IS-LM is absolutely silent here, but if Italy were growing at two percent a year probably the whole mess would be manageable.  Demographics matter too, and if this is a messier version of economics sign me up.

The idea that Ireland is seeing a partial recovery, piled on top of some deep structural problems in its domestic sectors, flows more naturally from institutions-based approaches than from IS-LM.

The all-important interplay between monetary and fiscal policy, critical for understanding this crisis, is forced out of the box by IS-LM.

For the final denouement, if indeed it arrives, there is Minsky and the theory of speculative attacks.  Keynes in his chapter 12 of the GT had a good understanding of how expectations can switch so suddenly, a key factor at several stages of the euro crisis.

There is more, but you get the point.  IS-LM should not be foundational for an analysis of this problem, IS-LM is not necessary, and arguably it is better not to invoke the model at all.  And if I had to give undergraduates only one point on one part of the blackboard, I would use the comparison between Greek and German banks.

IS-LM leads one to the mistaken attitude that macro is fundamentally simple, and that all would be well if only people and politicians understood the need to “get tough” with expansionary policy rather than austerity.  It’s a lot more complicated than that, yet we can understand these complications by building up from some fairly simple and intuitive models.  It’s a better approach to use public choice incentives to understand why the current situation is unworkable, rather than preaching about why the flows should be different than they are.  When I hear so much preaching, I tend to think the preacher is using a model with a missing variable or two.

Here is a good post from Philip Levy on why macro is hard, and how that relates to the recent Nobel Prizes.

Soon I’ll turn my attention to whether IS-LM Keynesianism has been making uniquely good predictions during this crisis.


Tyler, all science is hard, very hard, and as all social activity it has to be played according to rules. Social scientists, including economists, may be confused about what the rules are, but sometimes, though they know which the rules are, some are inclined to ignore them.


Very interesting, thanks for sharing.

Predictions are of course where the pudding is.

If that is so, why do people on this blog attack Krugman, Stiglitz et all, even though they were clearly the ones correct when it came to predicting the future of inflation and interest rates?

Attacking Krugman? You folks are funny.

I couldn't follow this at all really. It was all references without saying what any of it was. Most people on here probably know all the references. I for one don't. Ooph!

If you can't understand it, then its not for you.

Ergo, it is for nobody.

He's saying that there are other models of the economy and other fields in economics that do a better job explaining where we are, why we are here, and how we get to a better state than the tried and failed Keynesian model of IS-LM.

The Earth-centered model of the solar system did a fair job of explaining observed phenomena for a long time. It was also consistent with the religious and political dogma of its day. The heliocentric view did a better job of explaining observed phenomena, even though direct evidence wasn't available for centuries later. The model of the solar system is more akin to microeconomic models. Macroeconomics is analogous to larger aspects of astrophysics over which there remains substantial speculation and disagreememt. And human behavior is far less prediction and explicable than physical phenomena. So neat little equations like IS-LM are not likely to provide useful insights.

Keynes told politicians everything they ever wanted to hear: that government was the source of economic control and stability. We have paid a heavy price for clinging to this dogma. The geocentric theory was far less harmful to human welfare.

"Soon I’ll turn my attention to whether IS-LM Keynesianism has been making uniquely good predictions during this crisis."

It doesn't make predictions. It has enough parameters that whatever you want it to "predict", it will "predict."

Clap, clap. I found the whole IS-LM online "conversation" about as big of a time waster (though amusing) as the Amish jokes on MR yesterday. Now that everyone's egos are exercised and the inside baseball of macroeconomics (which BradyDale is not so easy or pleasant to digest), maybe we can talk about the problems in the real world?

My happy yesterday was starting "The Lost Decade"
This gives some institutional perspective to what's going on today. I have not gotten very far, but it might be a nice bridge between different macro approaches.

BradyDale, you are not alone, and that is precisely Tyler's strategy.

Brad DeLong suggested starting with simple models and adding complexity. Tyler does not address that concern: instead, he throws in 4 schools of thought without specifying a single model, without showing how they address more than any one aspect of the situation.

Tyler points out that simple IS-LM doesn't consider various important factors: yet Brad DeLong said that those could be added to IS-LM models if they are considered important. Expectations, for example. Tyler conveniently seems to forget that.

IS-LM is one tool in the macroeconomics toolbox, and an important one. Is Tyler claiming that considering a hammer an important tool means that you should never dream of using other tools? Is Tyler claiming that others think every problem is solvable with IS-LM? That's the impression he conveys.

In short, you've witnessed a huge exercise in hand-waving. He proposes "moving on" between incommensurate paradigms in unexplained ways. That's very handy for the pseudo-rationality of a propagandist, but doesn't result in a model that people can use.

Tyler's debate strategy is simply to never address the actual points of his opponents. That's why he never actually cites them. Instead, like a creationist, he simply continues to spout his worldview endlessly. In a short attention span world, that's all it takes to please your followers.

Most economists do not think IS-LM is an important tool. On the other hand, Delong and Krugman think it is foundational. Cowen disagrees. Your rudeness ensues.

Haha, it's always rude when you point out the obvious

Mike Huben is the intellectual equivalent of a child molestor.

"Expectations, for example" -- and therein lies the problem.

"IS-LM is one tool in the macroeconomics toolbox, and an important one."

Most econ graduate schools don't teach IS-LM, since they consider it outdated. So, no. It isn't very important. It mostly hangs around in undergraduate classes because it makes it easy to come up with lots of test questions and only requires relatively simple maths.

I'm puzzled by the idea that the fundamental difference lies in the value as assets of "Euros in Greek banks" and "Euros in German banks". German banks have been hardly any more responsible than Greek banks during this crisis - both took on far more Greek sovereign debt than they should have done (and, in the case of some German banks, a lot more US-originated CDOs). The difference lay surely in the value of Greek sovereign debt and German sovereign debt. Sure, the German banks are probably more solvent than the Greek ones but that is a function of the greater German resources for a strictly national bail-out. But it's not clear that the bail-outs will in the end be strictly national. It seems the Merkel government is still readier to bank on bailing out Greece in order to avoid admitting how much of the responsibility for excessive lending should be placed at the door of their own banks, and the electoral calculation behind this is not hard to understand. Being criticized for being too kind to feckless foreigners seems to them less damaging than being criticized for incompetence in having failed to oversee their own banks properly.

But I am puzzled here:

"Being criticized for being too kind to feckless foreigners seems to them less damaging than being criticized for incompetence in having failed to oversee their own banks properly."

I tend to be most cynical about politics but try not to lose sight of many things going on at the same time. You've reduced Germans to one cynical and self interested thing with a primitive motive and a Machiavellian choice. And this in the land of the master of multi causality, and most eminent spokesman of austere bourgeois values, Max Weber. You've also obliterated history and institutional lags. German banking regulation dates before Merkel. Not to mention that Germans are now criticized for dragging their feet full stop. I don't buy this game theory of the German predicament. I like to think there may be some ideologies lurking behind Germany's low tolerance for public deficits and their irritation towards the chronic rent seeking banking and banking regulation of their southern neighbors. For example, did you read recent lecture by the influential chief economist at Deutsche Bank titled 'I Am an Austrian in Economics'?;jsessionid=F15203FD27E9D8C039EBC2589EE44B52.srv12-dbr-de

And then this today. Wow, they don't want bail out. 

"Details of the initiative emerged after Josef Ackermann, chief executive of Deutsche Bank, the largest German bank, issued a blunt warning on the European Union’s plans to recapitalise lenders, saying he would rather sell off valuable pieces of his institution rather than accept state funds."

The banks say they will accept their losses, take it on the chin, find private sector solutions. We don't need no keynesianation, they seem to sing...

Some points:

- The reason why Italy might change my 'idea' of capitalism: productivity hasn't increased for almost twenty (not ten) years. How did they do this! It has some world class companies, leads the EU instance in parts of aviation industry, it's car manufacturers are outstanding, it's food industry is first rate. My 'idea' always was that, whatever, "capitalism (however defined) + technological progress = increase of productivity". No, I do not say "prosperity" or "low unemployment", I say "increase of productivity". And the historical record has quite some evidence to back this up just look at the transition countries post 2000. But for some reason, Italy defies this idea (productivity increase in Greece was clearly above the EU average and therewith very much above the Italian increase, post 2000!). Weird. It surely requires an institutional explanation. By the way - Italy is in a much worse state than many people think as the difference between U3 and U5 unemployment (about 2-4% in most European countries) is about 8% in Italy. U5 unemployment in Italy is very, very high.

- It's not just the naked put thing. Look at this graph of the current accounts in Europe (click on the enlarged version 'here'): Germany really seems to have some serious internal reflation to do (in terms of the naked put - that would indeed make the German Euro less valuable...). The main difference between Germany (which after about 1996 had, on average, larger government deficits than either Ireland or Spain or the Baltics or Iceland! ) is the current account, which after about 2000 moved into surplus. Larger government deficits in combination with a surplus on the current account of course mean: low consumption and/or investment growth, which is exactly what happened, in Germany. It was beggaring its neighbours al along, while at the same time exporting capital to finance their imports of German goods. These surpluses are the root cause why Grmany is not dependend on foreign inflows of capital, while the 'deficit on the current account countries' are, in the 'gold standard' environment of the Euro area (including the Euro-peg countries). Well, I guess you read Kash Mansori on this. At this moment, France is going the way of Spain and Italy, too, by the way, despite the fact that average productivity in France is higher than in Germany. And, by the way, consumption is already down, in the 'peripheral countries' - but this demand void has not yet been filled (as this fall in demand mainly concerns consumer durables like cars this implies a leftward shift of the IS - see Solow on this). Look at these graphs:

- So, you're right, capital flows did play a role - as they financed deficits on the current accounts. Private investors by now try to get their money back. But in terms of the IS/LM model that means that M, which is supposed to be constant in the 101-model, diminishes. I.e, the LM curve shifts to the left. But as these money flows financed investments and as the Minsky moment you rightly invoke led to a demise of the investments, the IS also shifts to the left... leaving high unemployment, low production and a low interest rate.

The point: I agree with your basic views. It's rather stupid to think that interest rates (especially low interest rates) guide entrepreneurial behavior in any kind of predictable way. And institutions do matter (U5 in Italy!). But IS/LM can be used to explain the pain in Spain, in a very simple way. LM shifted to the left - but IS shifted even more (in the end I'm a National Accounts guy - I understand these lines not as behavioral functions but as estimates of the total amount of monetary expenditure/use of money).

But the shifting led to some other changes. The IS became (even) more vertical when it comes to the influence of the interest rate on investments but AT THE SAME TIME horizontal when it comes to the influence of investments on the interest rate. In fact you have to draw two IS lines, one solid and and horizontal and one dotted and vertical, in the graph. Increasing M again won't make much of a difference as LM has become more or less horizontal (ooops, liquidity trap) unless you change the vertical IS line - which can not be done by changing the interest rate. M and I have to be increased at the same time, but you can't use the interest rate to do this. German consumers will have to start spending more.

Seems like a false choice discussion. IS-LM model is relatively simple and illustrates certain important points about interplay of Savings, Investments and Liquidity Preferences given level of output. I don't see how understanding this setup in any way stops you from using all the things that Tyler mentioned. You should use all those things AND understand IS-LM.

Not to mention that (to a non-economist) most of the items on Tyler's list appear to be fairly removed from the most fundamental questions of macro. If we're talking about where to START macro analyses, why would any of these 4 be on the list? 1,2, and 4 sound more useful as micro theories than as macro theories that could tell you something about GDP.

I would only suggest that, much more importante than public choice is political science. Why stick with public choice to udnerstadn the interplay bewteeen economy and politics when you can add political science. We political scientistas are much more competent to understand the importance of institutions than economists...

Don't get me started. What you say is just absurd. There was no science in political science until economics showed the way. There was no rent seeking theory until Buchanan and Krueger. There was no institutional theory until Doug North. What went before was muddle headed and indefinable in the terms relevant here.

The exception is Max Weber, it goes without saying, property rights, closure, monopolization in politics as in economy, he formalized i.e. theorized the study rent seeking and the transition from informal to formal institutions. The economists are indebted to him (how can they admit it, they can't of course, yet everyone knows).

But in political science, before the economic imperialists showed the way in regulation theory and analysis of government-business relations and intra-state bureaucratic politics, there was only Mosca, Schumpeter, elite theory, rosy-tinted American pluralists, Lindblom, German corporatists, Marxists of the relatively autonomous state, and, of course, the misunderstood and misused Max Weber, who was a sociologist and professor of economics to boot!

Really? It doesnt matter if economics influenced political science. My whole point is that political science now has a lot to contribute, much more than economics (about institutions).

Bold claim. Evidence?

ISLM IS NOT AN ECONOMIC MODEL! Why can't you damn people get this through your heads. It is a statistical model with no economic content whatsoever, and it can fit anything.

Of course, the crap Cowen sells is just a different color of turd.

Is that what the other side is saying?

Yeah, do you?

Andrew, you know Tyler does not do book recommending!

Sure. Don't take any risks at all. Whenever making a deposit, whether in Germany or Greece, always wear a condom so you won't catch any eurological diseases. (... did I hear a groan?)

IS-LM leads one to the mistaken attitude that macro is fundamentally simple...

Nailed it, discussion over.

Why does using simple models have to imply that something is simple? I am not sure how that makes sense. You use simple models where appropriate, regardless of the actual thing in question being simple or complicated.

Nailed it, discussion over.

Because that simple model leads to the mistaken belief that the solutions to economic problems can be identified and implemented by a bunch of lawyers, bankers, and government employees in the nation's capital.

Nailed it. Discussion over.


nailed it. discussion over.

Islam does mean "submission".

This is a good demonstration of why Austrian-type thoughts will be relegated to the precincts of the economics blogosphere, orthogonal to serious academic thinking.


As long as we are tossing out non sequiturs, I have a question on the liquidity trap idea. It seems like it is based on the assumption that a reduction in interest rates will always incentivize lending/borrowing and that when that doesn't happen, you don't question the assumption, you create a new name for a special case.

Tyler, can you please provide some sort of a syllabus for items 1-4 that might be accessible for those of us without access to academic journals (books or free articles) Sometimes even us lawyers need to keep our feeble macro skills in tune.

I must be missing something. Cowen and Krugman both seem to basically follow Fisher. I think it's right to say Krugman is using ISLM to better explicate that view. Oh, and that's what I believe.

Steve Keen's models are the ones that most closely resemble the actual economy.

Steve Keen is a classic crank.

Tyler what books would you recommend to familiarize oneself with tfinancial asset pricing theory?

Will you recommend some books on the basics of New Institutional Economics?

Both CAPM and EMH would say that the risk adjusted rate paid by banks in Germany and Greece should make investors indifferent on the margin about where to make their deposits. You have a better theory?

I think I can speak pretty well to New Institutional Economics and Asset Pricing Theory.

By what stretch of the imagination are they macroeconomics rather than simple extensions of core micro?

Situation: the patient is very sick, and far too many doctors are prescribing medicine that will make her worse. Dr. Krugman has a simple argument to refute those dangerous quacks. Dr. Cowen says: "That argument is oversimplified. You can get a better prescription by an ad-hoc policy of switching between these other four arguments according to criteria that I won't explain right now."

It seems to me that Dr. Cowen could be right on every single fact (something I'm certainly not qualified to judge), and still be reaching the wrong conclusion. I certainly don't see that his Frankenstein of a theory has the power and simplicity to overcome the very bad ideas out there. If, as another commenter says, IS-LM "has enough parameters that whatever you want it to “predict”, it will “predict.”", then Cowen's four theories, put together, have enough that they will support any course of action you want them to.

In other words: Krugman may not be 100% right, but if we're judging the men by their enemies, then he certainly has the advantage over Cowen in this argument.

This is the economics equivalent of 'if you don't know machine language, you have no business using a computer'.

The problem is not that IS-LM is too simple to be very useful, it's that not enough people understand even IS-LM.

Comments for this post are closed