Matt Rognile does not like the LM curve (nor does David Romer)

Read the whole post, here is one excerpt:

…many countries now operate under an inflation targeting framework, in which responding to inflation is the key feature of the policy rule. In this environment, depicting policy as a relationship between “Y” and “i” misses what’s really going on—better to abandon the upward-sloping LM curve altogether and use a simple horizontal line to depict the current policy rate. I’m not alone in this sentiment.

David Romer wrote an entire piece for the JEP in 2000 called Keynesian Macroeconomics without the LM Curve. (As the title suggests, he shares my feelings on the matter.) Tyler Cowen puts this at #6 on his list of grievances. It’s a pretty obvious point—yet, for reasons I don’t entirely understand, we still print thousands of undergraduate textbooks a year with LM front and center.

Matt is an economics Ph.d student at MIT and an expert in macroeconomics.

Here is a good quotation from the above-cited David Romer piece (Romer, by the way, is not a member of my tribe, in fact he is a tenured professor at Berkeley):

In short, recent developments work to the disadvantage of IS-LM. This observation suggests that it is time to revisit the question of whether IS-LM is the best choice as the basic model of short-run fluctuations we teach our undergraduates and use as a starting point for policy analysis. The thesis of this paper is that it is not.


Matt is an expert, huh? High praise for a second year PhD student.

I was thinking the same thing. But it's true in some fields at the best schools.

There are lots of people with expert knowledge and no PhD.

There are lots of PhDs with no expert knowledge.

Tyler, don't you have a textbook? Does it include IS-LM?

Having kind of followed this debate as an outside, marginally loyal reader of various blogs, I conclude this: econ professors and grad students are delivering nonsense to paying undergraduates on a platter of authoritative knowledge. And some such professors and grad students knowingly peddle this wastefulness.

And Bryan Caplan wants them to run this country because the rest of us are inadequate.

Theories, models, and even facts that are just plain wrong aren't always unwelcome in education. We tell many lies when we teach principles because the correct story IA too complicated to understand.

For example, in math we teach students that a continuous function is one that can be drawn without lifting your pencil off the paper. Simple and useful enough. But then you get to Advanced Calc and realize you have to suspend belief in all the garbage you previously learned and embrace esoteric definitions and proofs that would seem impractical and unnecessary if there wasn't further insight from the inquiry.

How about, 'Events with a probability of zero can't occur.' Yet probability zero events occur from every draw in a continuous distribution.

How many assumptions of standard economic models are fictions? We simplify models to be mathematically tractable, and simplify empirical analysis to work with available data.

We lie all the time. The important part is to remember when we're lying. Some people have lied for so long, they believe the lies.

When the practical insights of a lie are displaced by a simple model closer to the truth, we take a step forward. This debate is suggesting that the LM lie has outlived its useful purpose.

I suppose it would always be right and ethical to tell students not to believe too much in the insights of flawed models. But then we'll get students who believe in nothing we teach.

Are academics all basically intellectual frauds?

Judging from the failures of well known and respected academically to talk us out of economic morass, it may be true. I know that pilots can fly planes and land safely most of the time. I know that plumbers can unclog my drain. But I'm still not sure any economists can do anything to fix the economy.

i am reminded of a saying,
"disputes among professors are bitter precisely because the stakes are so small"

Given that the Fed can have a pretty big impact I am not so sure that these disputes are all that small. Economic policy matters dude.

policy is set by people who don't give a shit what ISLM is called.
the austerity people just want to support the bankers

"the austerity people just want to support the bankers"

No, those where the pro-bailout people. Although there is some overlap, particularly in europe, don't conflate the two groups.

Okay, so ISLM is dead in the water. The best one can say is that it explains some epiphenomena. So please can we get back to analyzing underlying first order phenomena?

“Evidence and models coincide in their conclusion that the process of creative destruction is an integral part of economic growth and fluctuations. Obstacles to this process can have severe short- and long-run macroeconomic consequences.”

That was the head of the economics department at MIT writing before the crisis in praise of Schumpeter. From memory I think the MIT professor was referring mainly to the obstacle of rigid labour markets.

Ideologically dominant Keynesian activist monetary and fiscal policy, which panders to the political cycle of welfare states, and which treat symptoms rather than causes of crisis, is plainly an equivalent but even larger obstacle to creative destruction.

Is the Keynesian failure to listen to Schumpeter’s central message about economic cycles and competition dynamics to be etched in history as the reason for cumulative economic devastation in the early 21st century? Is MIT today merely producing chief economists at the international monetary authority who placidly go along with plumbing the entire world economy with drip feeds rather than boldly crafting the structural fundamentals?

You know one of the smartest things Steve Jobs ever said?

“Death is very likely the single best invention of life. It is life's change agent. It clears out the old to make way for the new.”

Gloomy, yep, but there speaks a true Schumpeterian entrepreneur, full of life and light. He may have read ‘Business Cycles’ in his spare time. His words are eerily appropriate for the economic times.

I imagine the next Nobel Prize Winner in Economics will win it for his/her solid defense of ISLM.

This is really getting tiring. So much debate about so little. I honestly believe the problem is not IS-LM per se, but the fact that some people believe they know the model even if they don't.

It is a simplification of reality, obviously. The LM curve simply represents all the combinations of Y and r (or i) such that the money/asset market is in equilibrium. In this version, the CB affects the economy by changing the amount of liquidity (here money supply for simplicity), and hence impacts the interest rate. What gives a bad rash of frustration to some is that it sounds like the CB is targeting Ms. But Ms can simply be interpreted as an instrument, something along the lines of open market operations. And through these operations, it impacts the interest rate. So now some will say "yes but most CBs target inflation!!!". Well, I don't have problem with that, but that's a moot point: there is no inflation in the basic IS-LM model! It assumes that prices are sticky in the short run. You better not try to go too far in your discussions of inflation dynamics and inflation targeting with this model, move on to some simple dynamic AD-AS model à la Mankiw intermediate textbook instead.

Now, does it mean I hate IS-LM because it is not a perfect description of reality? No! Instead, I am careful to point out the shortcomings and shortcuts to students. And this has the advantage of make them reflect on macro models in general, and leads them to figure out what would be the next logical step (dynamics). I have seen many implementations of som IS-MP model, and all of them look forced, very very forced. I fail to see how this should be considered the Holy Grail of undergrad econ.

So please, can we move on to another debate?

Hmmm. The European Central Bank targets inflation (1,9%; how about fine tuning...). But it does this by explicitely targeting a 4,5% M3 money growth rule... in the long run. As M3 growth was higher n the 2004-2007 period, it implicitely a much lower target at the moment, something like 1,5% (as a 2% increase triggered the rate hikes). Some people might need some catching up on what Central Banks actually do.

At MIT we have more than a few drinking songs with the chorus "I go there too" so don't be blinded by the light. Buyers of the IS LM model who think we are down below the x-axis have made pretty specific predictions about unemployment, interest rates, effects of austerity, etc. If you think the model is bunk then say how they got to their predictions when a lot of people lost a lot of money disagreeing and lay your model for the last few years out. Did IS LM put you in some kind of wilderness during the Asian financial crisis, aregentina, the last couple of years?

Isn't LM "how the interest rate would behave when M was constant and Y increases or decreases?". From the reactions and articles above I get the impression that people see it as a function which describes how the Central Bank tries to influence the interest rate.

I notice that Tyler is still attempting to distract from the fact that he hasn't said what he himself thinks. He doesn't even admit that he's suspending judgement (the pompous way of saying he hasn't thought it through or doesn't know.)

Krugman made some of his published predictions based on IS LM models. Why isn't Tyler pointing us to predictions made with models he prefers? This is the stage of the argument where Tyler could look more and more wrong if he said anything concrete, and so "Look, somebody else criticized IS LM!" That's all his claque really needs.

It's fun watching him squirm. Doubtless he's hoping that Krugman and DeLong will grow bored and leave him alone until his next piece of propaganda.

Here's a question. There's been a lot of talk about how the ISLM isn't the best simple model to 'map' the world. But missing from all these conversations is a short list of other good, simple models (with links to explanations) that fulfill a certain explanatory function better than ISLM. Surely setting ISLM side-by-side with other models is the only way to judge it.

An ARMA model models GDP very well.
And I am pretty sure I could code up a neural net with the same inputs as IS-LM that does a better job.
Or is that not what you meant?

Comments for this post are closed