contra Krugman
Paul Krugman writes that Keynesian thought has progressed.
I disagree in comments with the usual
I do not think that Keynesians accept that there is a natural rate of unemployment. The concept is inconsistent with hysteresis. OJ Blanchard and LH Summers are definitely Keynesians. They presented a model without a natural rate in 1986. I don’t know about Blanchard but Summers and his former student JB DeLong continue to stress hysteresis (with an open mind about the sum of autoregressive coefficients of unemployment on lagged unemployment so there model might or might not be consistent with a natural rate but with certainly no evidence of any faith in policy invariant stationarity). Also there is this guy named Krugman (have you heard of him) who shows no sign of faith in a natural rate and writes about life long (OK not permanent but life long) effects of low demand on labor market experience.
Finally Wren Lewis refers to the hysteresis effects of the Thatcher deflation when he isn’t referring to New Keynesian models with no such effects as successful cutting edge models and great progress on 1960 vintage Keynesian analysis of the possible Phillips curve (by Samuelson and Solow who extensively discussed both outward shifts due to increased expected inflation and outward shifts due to cyclical unemployment becoming structural).
Note also that without using the name “Phillips” because he wrote beforee Phillips, Keynes specifically warned that there wasn’t a stable relationship between output and inflation. He noted the case of “flight from the currency” not yet named “hyperinflation” by Cagan rather anticipating the work of say Sargent.
Now it is true that Keynesians were convincedby Friedman mostly that velocity is fairly stable and predictable. But it isn’t.
I assert that Keynesian thought hasn’t made a both scientific and theoretical advance since Keynes where by scientific I mean actually useful in predicting out of sample and by theoretical I mean not just new parameter estimates made with data not available to Keynes using computers not available to Keynes.
Some semi prominent New Keynesians were amused enough to try to come up with examples. They failed.
In any case certainly not the absense of a stable expectations unaugmented Phillisp curve (as predicted by Keynes) nor the discovery of the LM curve which uh you know doesn’t exist.
Your disagreements are nit-picking. The central them of Krugaman’s piece is that there is one group that learns from experience and another that doesn’t because, like Lysenko, their minds are trapped by their search for ideologically pure ansers.
I agree with him on that. But I really don’t think that dismissing all progress in Macroeconomics in the past 76 years is nit picking.
I too was surprised at Krugman’s mention of the natural rate. My impression was that absolutely nobody believes in the natural rate anymore.
Incidentally, I am increasingly coming over to your view that far too many New Keynesians mistook advances in technique for substantive advances in their knowledge of the workings of actual economies.
The one victory PK was willing to concede to the microfoundations crowd was Friedman and Phelps. But given the fact that claims about the instability of the PC predate Friedman and Phelps ( Keynes, Phillips, S and S) and the NAIRU and the vertical PC are now widely accepted as no more than fairy tales, it appears that the microfoundations approach has had a sum total of zero successes.
Incidentally, I am increasingly coming over to your view that far too many New Keynesians mistook advances in technique for substantive advances in their knowledge of the workings of actual economies.
The one victory PK was willing to concede to the microfoundations crowd was Friedman and Phelps. But given the fact that claims about the instability of the PC predate Friedman and Phelps ( Keynes, Phillips, S and S) and the natural rate and the vertical PC are now widely accepted as no more than fairy tales, it appears that the microfoundations approach has had a sum total of zero successes.
I don’t have much to add but I think Herman’s point about advances in technique being mistaken for advances in knowledge is dead on. Incidentally, you see that in the business world all the time too. There’s something about human nature that makes putting a name on something equivalent to understanding it, particularly if the name is really cool.
I don’t have much to add but I think Herman’s point about advances in technique being mistaken for advances in knowledge is dead on. Incidentally, you see that in the business world all the time too. There’s something about human nature that makes putting a name on something equivalent to understanding it, particularly if the name is really cool.
Robert are you sayiing that the group learns from experience but makes no progress? This can be true if it’s been learning that evidence rules out many new and/or alternative assumptions and hypotheses (ones not proposed by Keynes. That can be depicted as “no progress” but is a major part of science.
“I do not think that Keynesians accept that there is a natural rate of unemployment.”
Is that natural as in God given, or as a result of some bio-chemical process, or possibly cosmic phenomenon? Or does that term reflect the recognition of a persistent economic phenomenon, one which seems to be a constant aspect of a capitalist economic structure? Persistent does not suggest natural occurrence, due to natural phenomenon.
i would suggest that unemployment is not by natural selection, but rather the result of intelligent design? In this case the designers being not godly but more likely sychphants to the natural elites.
I believe there are two extremes — low employment as at present, which is damaging, and very high employment, which I believe to be an indicator of crisis.
But perhaps what we need to look at is the employment-to-population ratio, but also the social safety net.
Suppose 100 people are of working age, and 40 are working — more factors need to be known before we know if they are in crisis. How do they access essentials? Can they move or retrain easily or not? Is the productivity of the 40 sufficient to support the 100, and if so, is it spread around or not?
Conversely, suppose those 100 are all employed. I think of England during the blitz — crisis indeed. But even in less drastic circumstances, if all working age people are employed, where is the capacity in case a further emergency arises?
In order to decide what might be a natural rate of unemployment, we need to know all these factors. But look at the 50s and 60s. About half the working age population were not working (lets call them “wives.”) http://rwer.files.wordpress.com/2011/02/us-employment-to-population-ratio-1948-2011-january.png
And yet the nation was prospering and growing better than any time in living memory.
It’s not unemployment, but unsustainment, we need to worry about.
Noni
If only. Macroeconomists don’t abandone hypotheses which are inconsistent with the evidence. we hae come up with many ideas and hypotheses which are not in the General Theory … . They make it possible to make worse predictions. Theory papers based on such ideas are published in top journals. Very little has been tossed out after it fought the data and the data won. so very little learning from experience.
By natural rate of unemployment, economists mean that, even if one eliminates all cyclical or structural headwinds, a predictable number of people will be either choosing to quit and search for another job, or performing in such a poor way that they will be fired. It has nothing to do with deities or biology, it is just a prediction based on observation of large numbers of humans over time.
I’d say its a prediction based on not observing humans. Economists also believe that the long term unemployed lose job skills and work habits making them less likely to be hired, more likely to be fired and more likely to quit. Also different firms have different approaches to human resources — there are lifetime employment firms which invest in training and hope to get loyal workers by being loyal employers. Other firms rely on being able to lay workers off when demand is slack. A sharp downturn helps the second set of firms. This can have extremely long lasting effects (trust does not develop easily or quickly).
My problem with the idea of “the natural rate of unemployment” is with the word “the” the claim that it is a constant. Of course we know it is very different in different countries. But the argument made by Friedman and Phelps must be that, within a country, it can’t ever change — in any case that it can’t be affected by demand management.
This is a very strong claim. Without it, there is no natural rate hypothesis properly speaking, because the claim that there is always some natural rate which might change and we have no idea how or why is not a falsifiable hypothesis. The hypothesis, to be useful and to be a scientific hypothesis, must include the assertion that the natural rate is approximately exogenous. The claim becomes that disinflation does not cause high unemployment for a long time. OK still not there yet until we define “a long time” so that it isn’t infinite. I’d say that, by the terms of the debate started by Friedman and Phelps, 20 years is definitely a long time and the long period of much increased unemployment in Europe following the determined disinflation places the natural rate hypothesis in the ash heep of the history of thought along with say the Ptolomeic model of the planets (the comparison is an expression of profound respect).
So clearly it might take a long time to create a peer-reviewed list of post-Keynes hypotheses and theories that have proven inconsistent with the data. It might help to stop wasting everybody’s time–or at least tell folks when they’re out on a limb instead of building on knowledge–and be a source to counter politicians’ claims. Somebody should start compiling (maybe Volume I: The Chicago School) and figuring out how peer review could possibly succeed.
“It has nothing to do with deities or biology, it is just a prediction based on observation of large numbers of humans over time.” mark
Aha!! Just as I thought. “Observation of a large number…” humans or workers or work place phenomenon? Who knows? What we do know is that that large number of observations, which is what might add a bit of scientific objectivity to the issue, will have many inter-related factors to take into account. As I said, though apparently a bit too sarcastically for mark to realize, there is nothing natural about unemployment. it is merely a persistent characteristic of an economic system. If birth control were more ubiquitous there would probably be a reduction in unemployment at some later point. At the end of a global war there is likely to be a reduction of unemployment. On the other hand these social events or changes can be offset by technological change. I repeat, there is nothing natural about the level of unemployment. What seems to be natural is an economist’s reluctance to observe and differentiate the multitude of factors impinging upon economic phenomenon.
Difficult to sustain the claim that *the* natural rate has anything to do with observation. It was a purely faith-based construct based on bad theory.
Friedman’s definition:
“The “natural rate of unemployment”, in other words, is the level that would be ground out by the Walrasian system of general equilibrium equations, provided there is embedded in them the actual structural characteristics of the labor and commodity markets, including market imperfections, stochastic variability in demands and supplies, the cost of gathering information about job vacancies and labor availabilities, the costs of mobility and so on” [Friedman, 1968a:8].
This is just bonkers.
Even in the idealized Arrow-Debreu world, convergence is problematic. Then, the model is very fragile. eg. Minor deviations from mkt completeness and only a remarkable fluke gets to an optimal equilibrium[1]. And even minor information asymmetries can send the economy far away from the optimal equilibrium. The time-complexity of the optimal decision problem …
Now look at all the deviations from the ideal that Firedman mentions and then claims convergence to a unique, stable, presumably Pareto-optimal equilibrium. Only a complete ignoramus, a madman, a devout faithful, an ideologue or a propagandist could have made the claim Friedman did.
[1] Robert Waldmann once heroically tried to explain this result to the non specialist: Geanakoplos and Polemarchakis