The Mortensen and Pissarides employment model

I have been seeing and hearing misunderstandings of it.  Here is one link to the piece and here are a few points:

1. Simulating the model, it is possible to generate something approximating the time series of U.S. employment changes, at least up through the 1990s.

2. This can be done by combining shocks to labor productivity, as measured (not calibrated to get the required result), combined with changes in search and matching profitability.  No bizarre assumptions are required about the intertemporal substitutability of leisure, as the matching function does the work in this regard.

3. The results do not require sticky wages or AD shocks.  Do not.

4. Mortensen and Pissarides do not explain "structural unemployment" in the traditional sense.  Instead the model shows that one overall negative shock — to all sectors — can make good matches harder.  In other words, unemployment can look like it is structural when it is not and in this regard the model can be viewed as a threat to traditional structural explanations.  Furthermore unemployment can look like it is AD shock-induced even when it is not.  Another way to read the model (my words, not theirs) is that it deconstructs the traditional distinction between cyclical and structural unemployment.

5. The model is built upon earlier theoretical contributions by Peter Diamond, but it is a mistake to speak of a unified Diamond-Mortensen-Pissarides model.  I read Diamond's view as closer to Keynesianism and the Mortensen and Pissarides story as much more Schumpeterian.  It is worth reading Paul Krugman's MIT-centric account of the prize, though I would stress the diversity of views among the winners.

6. Mortensen and Pissarides do not analyze monetary and fiscal policy in their basic model, but it is not obvious that such reflationary policies will solve the employment problem, which is defined in real terms, not nominal terms.

7. In the pure model, the Beveridge curve can have either slope and so we should not infer much from an observed Beveridge curve.

The Mortensen-Pissarides model is daring and subversive and it is impressive that the committee would award a prize for what is essentially a single paper.  Yet I don't see too many bloggers trying to come to terms with it.


A good summary, although the Shimer Puzzle is worth mentioning as a point against the model:

"The textbook search and matching model [ie, Mortensen-Pissarides] cannot generate the observed business-cycle-frequency fluctuations in unemployment and job vacancies in response to shocks of a plausible magnitude." (Shimer 2005 AER:

There are fixes to this problem, of course, but they tend to rely on strong assumptions about the value of leisure.

I am a little surprised at your characterization of Schumpeter and Keynes, putting Schumpeter in the Austrian school and placing them in diametric contrasts with Keynes. For the less informed readers of economic history on this site, this is a bit, well...the following says some of it

This is from an essay by Peter Drucker:

"The two greatest economists of this century, Jospeh A. Shumpeter and John Maynard

Keynes, were born, only a few months apart, a hundred years ago: Schumpeter on

February 8, 1883, in a provincial Austrian town; Keynes on June 5, 1883, in Cambridge,

England. (And they died only four years apart - Schumpeter in Connecticut on January 8,

1950, Keynes in southern England on April 21, 1946.) The centenary of Keynes's birth is

being celebrated with a host of books, articles, conferences, adn speeches. If the

centenary of Schumpeter's birth were noticed at all, it would be in a small doctoral

seminar. And yet it is becoming increasingly clear that it is Schumpeter who will shape

the thinking and inform the questions on economic policy for the rest of this century, if

not for the next thirty or fifty years.

The two men were not antagonists. Both challenged longstanding assumptions. The

opponents of Keynes were the very "Austrians" Schumpeter himself had broken away

from as a student, the neoclassical economists of the Austrian School. And although

Schumpeter considered all of Keynes's answers wrong, or at least misleading, he was a

sympathetic critic. Indeed, it was Schumpeter who established Keynes in America. When

Keynes's masterpiece, The General Theory of Employment, Interest and Money, came

out in 1936, Schumpeter, by then the senior member of Harvard economics faculty, told

his students to read the book and told them also that Keynes's work had totally

superseded his own earlier writings on money.

Keynes, in turn, considered Schumpeter one of the few contemporary economists worthy

of his respect. In his lectures he again and again referred to the works Schumpeter had

published during World War I, and especially to Schumpeter's essay on Rechenpfennige

(that is, money of account) s the initial stimulus for his on thoughts on money. Keynes's

most successful policy initiative, the proposal that Britain and the United States finance

World War II by taxes rather than by borrowing, came directly out of Schumpeter's 1918

warning of the disastrous consequences of the debt financing of World War I.

Schumpeter and Keynes are often contrasted politically, with Schumpeter being portraye

as the "conservative" and Keynes the "radical." The opposite is more nearly right.

Politically Keynes's views were quite similar to what we now call "neoconservatice." His

theory had its origins in his passionate attachment to the free market and in his desire to

keep politicans and governments out of it. Schumpeter, by contrast, had serious doubts

about the free market. He thought that an "intelligent monopoly" - the American Bell

Telephone system, for instance - had a great deal to recommend itself. It could afford to

take the long view instead of being driven from transaction to transaction by short-term

expediency. His closest friend for many years was the most radical and most doctrinaire

of Europe's left-wing socialists, the Austrian Otto Bauer, who, though staunchly

anticommunist, was even more anticapitalist. And Schumpeter, although never even

close to being a socialist himself, served during 1919 as minister of finance in Austria's

only socialist government between the wars. Schumpeter always maintained that Marx

had been dead wrong in every one of his answers. But he still considered himself a son of

Marx and held him in greater esteem than any other economist. At least, so he argued,

Marx asked the right questions, and to Schumpeter questions were always more

important than answers."

Here is the link:

I have always enjoyed Schumpeters work. It embodies competitive market economic dynamics that is always the threat to monopolies and a friend to a competitive economy.

"3. The results do not require sticky wages or AD shocks. Do not."


We need to tie UI benefits to registration in a worker re-training web site. If called, you have to show.

Then let companies write off 100% of their costs to run training programs as long as they hire 25% of folks who make it through.

This recognizes that we need specialized low cost workers. This is an outcome of productivity gains.

So glad someone finally pissed on sticky wages.... the only place they exist is on college campuses and public employees. Even the UAW guts the workers when they own the plant.

I can see how M-P is Schumpeterian, in the sense that the human capital of an individual worker is creatively destroyed, by, say, a crash in the housing market, which no longer supports the same number of carpenters. But, why is it anti-Keynesian?

If M-P are saying that the loss of attachment to the labor market leads to structural unemployment, then wouldn't Schumpeter suppport the investment in human capital (that is, retraining, and even support for internship) to reattch the person to the job market and convert the human capital to something more valuable?

Isn't that Keynesian? I mean you are supporting a person during retraining; or you may simply be waiting, in some cases for the market to adjust to a sustainable equilibrium.

Rather than trying to put living economists into bottles and labeling them "Schumpeterian", "Keynesian", "Austrian" based on however the writer wants to classify them to conform to their own view of the world,

WHY NOT ASK the living economists M-P or Diamond what they think would be the appropriate job policy for the current state of the economy.

I mean, M-P and Davidson are not dead yet. You can ask them.

That may also be why we should be grateful that they do not posthumously award a Nobel Prize.

Tyler, thanks for the points. Here is my understanding of the paper.

1) To reproduce the negative correlation between the destruction rate and the creation rate during a recession an aggregate negative shock in the value of labor is required. This COULD be the result of an AD shock, even though it does not have to be. If it is not, however, then what could it be? RBC people who are willing to argue that the decline is the result of a decrease in overall productivity (negative technology shock) will have to identify the source of this shock and highlight its significance in reality.

2) The increase of unemployment in the current recession, and pretty much every other, cannot be the result of "creative destruction" (a new distribution of values for the various jobs) as idiosynchratic shocks cause an increase in both job creation and job destruction, the exact opposite of what is being observed. This is what Krugman was trying to emphasize. But is this not what some Austrians have been arguing?

3) No bizzare assumption are made about the substitutability of leisure and no wage stickiness is assumed, but the model makes plenty of other bizarre assumptions, as I see it, to make up for the fact and generate plausible dynamics. For one, it is assumed that all jobs draw values from a distribution at the same frequency. Hence every job that receives a low value after a draw can bounce back when a new draw is made (and this possibility leads to labor hoarding). Now, is there any worker in textiles whose job has been outsourced or replaced by technology who thinks that in a future draw the value of that job is going to go back up, at least in the reasonable future? Secondly, what causes a job to be eliminated is the unwillingness of a worker to accept a wage below the opportunity cost of going back to the job parket, which depends on overall labor market conditions. This is a very strange assumption that would hold only if people can fill jobs that require any skill (completely mobile among jobs). If, as seems more realistic to me, the skill of the individual is tied to the skill of the job then the opportunity cost of employment for the worker should be related not to the overall labor market conditions but to the conditions of the market for that particular skill. In this case, if a skill becomes less valuable so all jobs that require that skill are being eliminated then the person's oportunity cost of employment should be equal to his utility of leisure because he can do no better by going into the job market. Therefore, the model would almost degenerate to a New Classical model, in that every worker whose job is destroyed by a new draw would become unemployed simply because they would rather sit on their behind doing nothing than accept a lower pay. Hmm, do we really believe that?

Of course, it is likely that I am misreading the model, so I would welcome any corrections.

Andrew writes: "Quickly?" Who says things should happen quickly? Seriously. By analogy, thermodynamics (which uses terms like 'equilibrium') doesn't usually say anything about the rate at which things happen (that's kinetics and statistical mechanics).

You have pointed out the greatest weakness in microeconomics.

In physics, within two weeks of beginning a classic mechanics class, you are forced to deal with the effect of time. In physics, nearly everything has a factor t. Newton is such an important figure in physics because he realized calculus was critical to dealing with the effects of time. In your second week, you are forced to see time as a factor that must be known precisely, but that can never be measured precisely.

In economics, we learn time is a hand wave. In the "short run" and in the "long run" - in the "short run" a shortage of workers will drive up the wages of new hires and wages will be inverted with low productive workers earning more than the more productive experienced workers.

In the "short run", an economic contraction will result in increased unemployment and no change in wages, but in the "long run" high unemployment will drive down wages, but fail to "clear the market" and eliminate unemployment, but in the "long run" we are all dead and no longer unemployed or employed earning too much or too little.

The first reports I heard explaining their work used the word "friction" a lot, but no one explained what "friction" is in the labor market. In physics we study friction, and while it isn't a sexy topic, it is studied with different theoretic and pragmatic models - static, dynamic, rolling, aerodynamic, etc.

In the labor market, the friction preventing a person from taking a low wage job at a different hospital or burger joint is the failure of the market to provide a Star Trek transporter with a $1 fare to replace the bus from inner city home to the closed hospital or store, with transport from inner city home to the hospital or store in the suburb.

Clearly, there needs to be a factor t in all economic models that, for example, incorporates the time required for an inner city to go from the bottom of decline to the growth of gentrification so the workers with homes in the inner city find the friction of getting to work decline. Or to invent and bring to market the transporter that economists assume the market will supply to make their theories work. Obviously it is liberals and big government that has denied the world the Star Trek transporter which solves all unemployment problems.

very helpful summary.

it seems obvious to me that cyclical downturns can exacerbate matching problems, but what drives this in their model?

Ok. The model. Now is there evidence that it applies to something?

"it deconstructs the traditional distinction between cyclical and structural unemployment."

I don't think so. Structural unemployment is not really about search. Search assumes you have a round peg searching for a round hole in a sea of square holes - it's just a question of time and information friction. Structural unemployment is trying to chisel a round peg so that it fits into a square hole - it needs not just time, but also retraining, etc. The two diagnoses have vastly different policy implications.

I thought only real business cycle theories focused on shocks to productivity.

Keynes had a "passionate attachment to the free market and [...] desire to
keep politicans and governments out of it"? I think not. He called for the socialization of investment and "euthanasia of the rentier" and the "oppressive power of the capitalist". Before he wrote his book he said "The decadent international but individualistic capitalism, in the hands of which we found ourselves after the war, is not a success. It is not intelligent, it is not beautiful, it is not just, it is not virtuous – and it doesn’t deliver the goods. In short we dislike it, and we are beginning to despise it. But when we wonder what to put in its place, we are extremely perplexed." Post Keynesians, who adhere most closely to the economics of Keynes himself, typically do not have those passionate attachments you speak of.

Comments for this post are closed