Site Meter

Monday, July 15, 2019

Milton Friedman and the Keynesians

Brad DeLong noted that new Keynesians are actually Friedmanites. On all points where 1960s Keynesians disagreed with Friedman, they concede. The fusion of Keynesians and Monetarists would seem odd to anyone who can't conceive of real business cycle theory (really anyone who can't conceive of a human being taking real busines cycle theory seriously, that is, all normal human beings and most economists). Paul Krugman notes a return of Paleo Keynesians. Then discuses it more.

By the way, Krugman's link to DeLong doesn't work any more The Washington Center for Equitable Growth returns a 404 error. I went to Krugman to get the link to Brad, noted Krugman said brilliant things (no surprise) *and* that his link is rotten (that's news).

I'm going to get to the point soon but I have to discuss Krugman now.

Oh boy this introduction ran away from me. I am going to put the whole rant after the jump. the bottom line is I say Friedman was always a Keynesian except for his insistence that the effect of the nominal interest rate on money demand is more or less pretty much negligible. This means that I argue that he differed from Keynes because he was a monetarist. In the 80s the difference between monetarists and Keynesians (which always was a matter of a paremeter estimate and not any fundamental disagreement) was dwarfed by the difference between them and the fresh water new classical ratexians. But the point, if any of the rant is that Friedman is determined not to be trapped among the Keynesians and that he bases his efforts fundamentally on the importance of i.

More followup on the state of Keynesian economics. In the Brad DeLong post I cited, he mentions as one of the key planks of New Keynesian (as opposed to old Keynesian) macroeconomics the rejection of the old-fashioned notion of a stable relationship between unemployment and inflation.

Now, you can argue that the notion of a long-term usable unemployment-inflation tradeoff was never really part of Keynesian economics, that it’s a caricature of what 60s Keynesians actually believed. Nonetheless, the stagflation of the 70s was a decisive moment in economic ideology. Stagflation seemed to confirm the Friedman-Phelps notion — based loosely on “microfoundations”, i.e., notions of rational behavior — that sustained inflation would get built into wage and price setting, so that historical correlations between unemployment and inflation would disappear. And this in turn gave a huge push to the anti-Keynesian revolution.

Put it this way: when I was in grad school, I remember lunchtime conversations that went something like this; “I just don’t buy the Lucas stuff — it’s not remotely realistic.” “But these people have been right so far, how can you be sure they aren’t right now?”

Yes indeed, I can argue that it is a caricature of what 60s Keynesians actually believed. An extensive search of the literature revealed exactly one 60s era Keynesian article which contained that assertion (not written by a promineant paleo Keynesian). Also I trust Krugman's report on intellectual history of MIT lunchtime conversations (which are more important than published articles which merelay record the new consensus and not how it developed). The point is that the caricature of paleo Keynesians was successful and was a decisive moment in economic ideology. The macro profession changed methodology entirely as a result of a caricature, based on fake intellectual history.

If anyone doubts these claims I propose he read the work of James Forder . Those who are like me, to stingy to pay 28.60 for a kindle version can see if Forder's Oxford working papers are still on line (possible dead link warning).

The history of macroeconomic methodology turned on successful intellectual history fraud and a successful charicature. I insist that this is not an exaggeration.

Actually let's go through the points where Friedman disagreed with the Paleo Keynesians.

1. Price stickiness causes business cycle fluctuations: You clearly need price stickiness to make sense of the data. However, there is now widespread acceptance of the point that making prices more flexible can actually worsen a slump, a favorite point of Tobin’s.

This is also the theme of Brad DeLong's first publication (with Larry Summers) if you count NBER working papers as publications. It was written in around 1983 at the peak of Fresh Water dominance when Cambridge MA new Keynesians were especially obsessed with price stickiness. Of course it was also argue by Keynes in the General Theory (about wage stickiness Keynes basically assumed prices were flexible with 7 words of concession that they aren't always "except for the case of administered prices" Keynes 1936. Here importantly Keynes Tobin, Summers and DeLong are discussing whether flexibility increased a bit would be better and not about the effect of perfect flexibility. Keynes did argue that flexibility didn't matter at all in a liquidity trap but he didn't understand the importance of the liquidity trap for this argument. Summers, DeLong and Tobin assumed that the possibility of a liquidity trap was negligible just as everyone did from 1960 through 1998 and almost everyone until 2008.

In any case, Friedman and the paleo Keynesians agreed prices were sticky. The "only that which is rational is real" Hegelian flexible price insanity came later.

2. Monetary policy > fiscal policy: Not when you face the zero lower bound — and that’s no longer an abstract or remote consideration, it’s the world we’ve been living in for five years. And Tobin, who defended the relevance of fiscal policy, is vindicated.

Bingo. Friedman definitely argued that a liquidity trap was impossible. This in spite of the fact that all major economies had clearly been in a liquidity trap a few decades before he began insisting that. On the other hand, Krugman disagrees with Keynes who often assumes a liquidity trap without explicitly saying so. Krugmanian macro is different from Keynes' macro, because Krugman stresses that he thinks everything changes in a liquidity trap. (here I agree with Krugman).

I mention in passing that Archetypal Paleo Keynesians Samuelson and Solow asserted thagt a liquidity trap was impossible in their extremely influential 1960 article which no one seems to have actually read (I didn't until I was over 50 years old so roughly after the 50th anniversary of it's publication). They argued that because they dismissed Ricardian equivalence. But I should avoid going further on this tangent because it's an elephant.

A revolution based on demand for money at i = 0 is rather odder than a revolution based on a thrupence a pound tax on tea. In any case, the key point is that Keynesians assert that i is fundamental and Friedman asserts it isn't.

3. Business cycles are fluctuations around a trend, not declines below some level of potential output: This view comes out of the natural rate hypothesis, and the notion of a vertical long-run Phillips curve. At this point, however, there is wide acceptance of the idea that for a variety of reasons, but especially downward nominal wage rigidity, the Phillips curve is not vertical at low inflation. Again, a very Tobinesque notion, as Daly and Hobijn explain.

There is therefore wide acceptance of the plucking model which asserts that GDP is like a guitar string which moves up from the neck of the guitar then back and hits a ceiling. Now who came up with the plucking model ? Friedman. If it is a key point of disagreement of Friedman and the paleo Keynesians, then Friedman is a paleo Keynesian. Someone who presented both the plucking model and the natural rate hypothesis has a great capacity to accept cognitive dissonance in the anti Keynesian cause. Friedman knew perfectly well that output is not well approximated by a symmetric stationary distribution around a trend.

Also Paleo Keynesians detrended when doing empirical work. Okun not Friedman introduced the concept of a NAIRU and paleo Keynesians (with one Canadian exception whose name I don't remember) asserted that there was not a stable inflation output tradeoff (this is very very clearly stated in Samuelson and Solow 1960 (I can't think of any way to make it more explicit so I link to a post with a link

4. Policy rules: Not so easy when once in a while you face Great Depression-sized shocks.

I don't recall Paleo Keynesians criticizing the idea of policy rules. I think they criticized Friedman's proposed rule (adopted for a few years by Volcker then abandoned because it became useless as soon as it was adopted (I think by coincidence)).

Notably applying his rule requires constant monitoring and intervention with open market operations. He described it in a way that it sounds like leaving the private sector alone, but he blamed the Fed for not taking decisive action in 1929. He knew that he was proposing constant intervention so Say's law was true in practice. He disguised this fact with his usual brilliant rhetoric so many economists (including me) accepted his proposal for macro stabilization as opposition to macro stabilization. Brad's post here is way better than my post which you are reading.

5. “Low multipliers associated with fiscal policy”: Ahem. Not when you’re in a liquidity trap.

Notice that this only works if one defines fiscal policy, as Friedman did, as temporary tax cuts. If one defines it as Keynes definitely did as temporary increases in government spending. Friedman and Ricardo without one Ricardo if common sense (a Ricardo is a unit of common sense equal to the lowest level ever observed far from a Great Lake) have no argument that this would have a small multiplier. One trick was to redefine fiscal stimulus as temporary tax cuts (at the instant that one found an argument that the effect of temporary tax cuts is zero). Another is to use "small" to mean 1 when one is arguing that multipliers are small and zero when one is arguing against fiscal policy.

So here Friedman set up a staw man -- fiscal policy is temporary tax cuts and he can argue that they have no effect with the absurdly strong assumptions that only that which is rational is real (which he uses when it is convenient). His highly influential argument is based on setting up a straw man plus an absolutely absurd assumption about what people do and don't know and understand (which is demonstrated false by polling data which shows that the vast majority of US adults (all but 8 % or 12% depending on the poll) didn't know about the ARRA tax cut (which the vast majority personally received) let alone that it was temporary and not to mention the absurd idea that they had the Federal Government intertemporal budget constraint in mind and knew that the temporary tax cut had to be balanced by tax increases or spending cuts in the future (oh and also assumed it would be tax increases not spending cuts). Utterly totally insanely absurd. Too crazy even for Ricardo. And again highly influential when combined with a lie about what one's opponents proposed.

But aside from that it was a useful contribution and recognized as such by contemporary Keynesians such as myself.

So what did Friedman contribute aside from caricatures of the thought of the economists he wanted to criticize ? Well we have a good bit of Friedman V Friedman who assumes a natural rate when not looking at data and a plucking model when looking at data, and who assumes rational or irrationality depending on what serves his argument. We have an assertion of Ricardian equivalence. Ricardian equivalence is found in New Keynesian models (at least until 2008) but basically no New Keynesians were convinced any more than Ricardo was. It was conceded for the sake of argument with fresh water economists.

At this point, I think Friedman is as Friedman presented himself -- a quantity theorist. His key point was the claim that velocity is roughly constant. The data collected in the early 1980s (during the Fed's very brief period of focusing on M not i) appeared devastating until the solid proof that Friedman was wrong was dwarfed by the overwhelming utterly devastating proof that he was wrong available since 2008.

All of this is, in fact, focused on his insistence that a liquidity trap is impossible. He seems to have understood (maybe better than Keynes) how key the liquidity trap was to The General Theory (which is therefore not general but analysis of one of two possible regimes). Top paleo Keynesians Samuelson and Solow agreed with Friedman. The data dare to disagree with three winners of the Sveriges Riksbank Prize.

Friedman mostly disagrees with them (and not his caricatures) about the slope of the LM curve not the importance of the IS LM expectations augmented Phillips curve model.

Here I abandon actually citations needed intellectual history and try to read his my and psychanalyse him in this armchair (I really am in an armchair right now). I think that Friedman basically agreed with Keynes and Samuelson and refused to admit it. All three have great respect for the market and think that it should generally be left alone except for macroeconomic stabilization. All think that macroeconomic stabilization requires constant monitoring and intervention.

Friedman differentiated himself from Keynes et al with brilliant (and unscrupulour rhetoric) and insistence that the nominal interest rate i isn't very important. I think he felt a tribal and ideological need to be anti-Keynesian and relied on the smallest disagreements to differentiate himself. The point of all this was to introduce a post which I haven't written.

In contrast, the fresh water new classical economists are, as economists, completely utterly different from Friedman. They identify him as a founding figure by ignoring almost everything he said, just as the crusaders claimed to be Christians. I see no way to doubt that shared policy proposals, general idealogy and tribe are more important to them than the substance of models, methodology, evidence, empirical work or anything else. They belong to his tribe which is not united by a research program.

1 comment:

Anonymous said...

On your Section 4 about policy rules,I think the gap was much wider than you indicate.

Friedman's proposed rule may have involved intervention via open market operations to keep M on its predtermined fixed path, but the crucial objection of the Keynesians was that it did not allow for any feedback from the current state of the economy, let alone the forecast state of the economy. Most also felt that given the number of possibly relevant variables, any fixed rule would be inferior to sensible discretion. (Obviously, they advocated the use of fiscal as well as monetary policy.)

I am now retired, so do not have free access to libraries and journals. Therefore I cannot give an exact reference, but as far as I remember a very good example of the Keynesian objections was Modigliani's AEA Presidential address. For the same reason, I cannot document my comments generally, but they are based on my memories from graduate study at MIT in the second half of the 1960s folowed by years as an acadamic economist.

ALMAR