In search of search theory

by John Q on July 10, 2014

This is going to be a long and wonkish post, so I’ll just give the dot-point summary here, and let those interested read on below the fold, for the explanations and qualifications.

* The dominant model of unemployment, in academic macroeconomics at least, is based on the idea that unemployment can best be modelled in terms of workers searching for jobs, and remaining unemployed until they find a good match with an employer

* The efficiency of job search and matching has been massively increased by the Internet, so, if unemployment is mainly explained by search, it should have fallen steadily over the past 20 years.

* Obviously, this hasn’t happened, but economists seem to have ignored this fact or at least not worried too much about it

* The fact that search models are more popular than ever is yet more evidence that academic macroeconomics is in a bad way

In search of search theory

Following the most recent Internet dust-up over the state of macroeconomic theory, I’ve been thinking a bit more about search models of unemployment. I first ran across these models when I was a student in the 1970s, a period of very high unemployment.

The basics of search models are simple and seemed, at least in the 1970s, reasonably realistic. Workers can look for jobs in various ways, all more or less time-consuming: looking at help wanted ads (which used to come out twice weekly in Oz), cold-calling potential employers, and asking friends and relatives to look out for openings. Since workers aren’t fully informed about the labour market, it doesn’t (at least in a good market) make sense to accept the first offer that comes along. Rather, it’s worth looking until you have a good idea what wage the market is willing to pay, then taking a job at that wage.
What was new about the search theories was the idea that this process could explain not only the inevitable frictional unemployment, but also cyclical unemployment associated with recessions. Various different ideas come into play here. First, the models imply that if there is a sudden but temporary shock leading to large-scale job losses, it will take some time to return to full employment (or ‘normal’ rates of unemployment). Second, economic shocks may create more uncertainty about market wages, leading workers to search longer.

The Internet changes all of this. It’s possible to find all the publicly advertised jobs in any given field, anywhere in the world, in a matter of seconds. With a little more effort, it’s possible to get lots of information about firms that may be hiring, even if they haven’t advertised vacancies. And, much more than in the past, its possible to get lots of information about potential employers, to assess whether the jobs they have on offer are in fact likely to be good ones.

With such a massive improvement in the efficiency of search we’d expect to see two things:

(i) shorter time spent searching, and therefore lower unemployment; and
(ii) better matches between workers and jobs, which should increase productivity and wages, and reduce subsequent quits and fires.

Both predictions were made in the early years of the Internet and, at least until 2008, the general view was that they were proving correct, though more slowly than had been expected.[^1]

But experience since 2008 has been completely the opposite of what a search model would predict. Unemployment rates have risen and employment rates have fallen. Worse still, the duration of unemployment has increased greatly. And there’s no evidence that this has been offset by an improvement in matching. Rather, lots of people have been forced to accept jobs that make little use of their education and experience, suffering wage losses as a result.

What has been the response of academic macroeconomists? As far as I can tell, almost nil. I was prompted to write this post by the debate over Kartik Arthreya’s Big Ideas in Macroeconomics. Responding to my observation that the word ‘unemployment doesn’t even appear in the index, Steven Williamson pointed out that Athreya spends several pages on the general properties of search models [^2], presenting them as the primary basis of unemployment theory. And Athreya seems to be on the money here. The Economics Nobel (yes, yes, Bank of Sweden etc, I know) is a lagging indicator, but the 2010 award to search theorists Diamond, Mortensen and Pissarides certainly does not indicate a view that such models are fundamentally flawed.

Athreya acknowledges the problem, sort of, by saying that “search is not really about searching” and observing that, if it were, the Internet should have reduced search costs. But if search isn’t about searching, what is it about? Athreya doesn’t say, and his brief discussion of housing markets doesn’t get us far – these are much more location-specific than job markets, and the Internet hasn’t changed the process all that much.

If search models aren’t the right way to think about unemployment, what is the right way? The simple answer is that unemployment is primarily a problem of macroeconomics not of labor markets. If aggregate demand is far below the productive capacity of the economy, workers will be unemployment and capital will be idle.

But there is still a puzzle here, one that search models were designed to solve. Why doesn’t competition between unemployed and employed workers work quickly to reduce wages to the point where demand equals supply and where there is no involuntary unemployment [^3]? The problem seems not be, as search models assume, that employers and potential workers don’t know about each other. Rather, it’s that employers can’t easily use the threat of new hires at lower wages to drive down the wages of existing workers (of course, this happens, but it’s clearly costly and risky in terms of worker morale). There’s quite a lot of literature looking at this, and I’ll try to post on it another time.

[^1]: Simple models of search can’t rule out cases where only one of these effects is seen. For example, the improvements in the quality of matches might be so great as to encourage people to spend even more time looking. But that doesn’t seem very plausible. It would imply a big increase in productivity and wages that hasn’t been observed. Here are a couple of references

Autor, D. H. (2001): \Wiring the Labor Market,” Journal of Economic Perspectives, 15, 25-40.

Krueger, A. B. (2000b): The Internet is lowering the cost of advertising and searching for jobs,” The New York Times, July 20.

[^2]: With his characteristic grace, Williamson also pointed out that another blogger had also observed the absence of standard macroeconomic topics in Athreya’s index and accused me of plagiarising this point for my own review. As any reviewer knows, the index is always the first place you look in a book, so its unsurprising that the oddity of Athreya’s jumped out at two of us.

[^3]: The same question applies to capital, though it seems to be asked far less often. Recessions and depressions are characterized by idle factories and farmland, empty offices and shuttered shops. Competition between the owners of these assets should drive the associated rents down to a level where supply equals demand. In fact, however, the adjustment process seems mostly to rely on scrapping or depreciating enough of the existing stock to remove the excess.



maidhc 07.10.14 at 3:36 am

A friend of mine owned a small company, and he posted an ad specifying a particular set of skills: at least an MS in optical engineering, and a certain number of years of relevant experience. He received nearly a thousand responses, NOT ONE of which had ANY of his requirements.

It’s no wonder that companies go to automated matching, but since it’s usually based on keyword matches to terms entered by people in HR who don’t know anything about the specific job, generally most of the qualified applicants get discarded. So most technical hiring is based on personal contacts.

Meanwhile at the lower end, job screening is done by people filling out multiple-choice tests (“What would you do if you found out a fellow employee is stealing a bag of potato chips from the company every day?”)


P.M.Lawrence 07.10.14 at 4:13 am

I have previously suggested that there could well be an externality in the labour market that encourages unemployment. In case anyone is interested, my most recent material on this on the internet is here.


Brett 07.10.14 at 4:16 am

@John Quiggin

i) shorter time spent searching, and therefore lower unemployment; and (ii) better matches between workers and jobs, which should increase productivity and wages, and reduce subsequent quits and fires.

I think the effects of the internet on job searching are overrated. It’s true that it has made it easier for workers to look for other jobs and higher pay, but that’s more like an expansion of what the Want Ads used to have instead of anything revolutionary. It hasn’t changed much on the “employer” end either other than some screening benefits (instead of having to keep a bunch of resumes on file, they can now have an online application system and sit on your application until they need people).


bxg 07.10.14 at 4:21 am

> The Economics Nobel (yes, yes, Bank of Sweden etc, I know)

Equivocations like this seems fairly common, but I don’t know how to interpret them. If you think it’s a Nobel price of similar prestige to the others (with a slightly different history, true), then just roll with it and omit the throat-clearing. If you think its corrupt bastardization by the Sveriges Riksbank then either say so, or if that’s off topic (probably it is here) then just don’t mention the “honor” at all.
This “a Nobel was awarded [but, I hear and give nod to you all you who think it’s not a real Nobel] ” stuff seems a bit cowardly.


Jason 07.10.14 at 4:27 am

This is not intended as some sort of complete solution, but may be of interest:

The details of the search (i.e. “microfoundations”) might not be terribly important in the sense that you can get a good model by ignoring them. It might just be dependent on the number of job seekers and the number of vacancies.


Moz in Oz 07.10.14 at 4:35 am

The internet effects on job search are only slowly grinding through Australia as far as I can see. Sure, we now have job search sites and online application processes for low-paid jobs, but it’s still a very manual process and there is a high dependence on recruiters once you’re above the median wage. IMO, obviously.

Most of the job hunting I see (and have done) is still based on the old model – make a CV, send it in response to an ad. It still goes to a recruiter who still talks to me and the potential employer and blah blah blah. Just like the 1980’s. We’ve trimmed maybe 2-3 days off the process using cellphones and email, but even that is very hit and miss.

There’s only a very slow switch to using social media like LinkedIn to actually find candidates directly. That’s where the dramatic change in search should be. Rather than knowing other people in the industry and asking if they know anyone, or fishing using an ad or recruiter, the employer can find every single eligable person within a given area who is thinking of changing jobs, and ask them directly.

I don’t see low-paid work being vulnerable to search optimisation until there’s a shortage of potential employees in those jobs. That “search” cost is determined by enemployment levels since to be counted as unemployed those people have to be actively looking for work. So the cost is really set by governments that regard 4% unemployment as a disaster to be corrected by importing labour or shrinking the economy.


Lee A. Arnold 07.10.14 at 4:52 am

John Quiggin, Do economists know how much of the anomalous response since 2008 is due to:

1. baby boomers near retirement age opting out of the workforce early, due to the crummy job prospects?

2. do-it-yourselfers saving money on repair labor by using the greater availability of technical information on the internet?

I changed the water pump on my old truck by watching a YouTube video. I have never done an auto repair in my life. You can find YouTube videos to learn how to do almost anything. I am finding that some people do all sorts of household repairs this way.


Sandwichman 07.10.14 at 5:27 am

“If search models aren’t the right way to think about unemployment, what is the right way?”

General glut.

But since such a think is not permissible just about anything else will do in a pinch. Price only equilibrates demand and supply under a host of special conditions, which of course do not obtain in the actual political economy of labor.

The anonymous 1821 pamphlet, “An Inquiry into Those Principles Respecting the Nature of Demand, and the Necessity of Consumption…” explains everything.


A H 07.10.14 at 6:41 am

”Why doesn’t competition between unemployed and employed workers work quickly to reduce wages to the point where demand equals supply and where there is no involuntary unemployment“

Isn’t it that this whole question is wrong because there is no reason to beleive that labor markets will clear, and that on a macro level the effect of the wage level on unemployment is ambigous. This is all in Keynes of course. Seems odd not to mention him.


John Quiggin 07.10.14 at 7:11 am

@4 Read the post again. I’m using the fact that the Nobel Prize in Economics (may as well not bother with pedantry) is the most prestigious prize in economics as an indicator of what the academic economics profession think is good and important work. If you accept the argument of the post, the profession is wrong in this case (and in others). As the post indicates, the general status of the prize is beside the point. I had my say on this, here


derrida derider 07.10.14 at 7:21 am

You’re right to deride the misuse of search models by the RBC school of macroeconomics, wrong to think they are therefore not useful (as in “all models are wrong, some are useful”). After all DMP are none of them even macroeconomists, and at least two of them look to be vaguely centre-to-centreleft in their political sympathies anyway (Peter Diamond is primarily an optimal tax theorist and was vetoed from a senior Obama administration job because of his open derision at the Senate confirmation hearing of the “tax cuts pay for themselves” school of Republican craziness. I’d guess from my observation that Chris Pissarides was a Blairite).

There is nothing in search theory that rules out exogenous shifts in the job offer function (such as after a financial crisis), or even that such a function is highly non-linear with an attainable corner of zero jobs at zero reservation wage (ie “no jobs available, regardless of how hard you look or how desperate you are” is a possibility).

True, the search framework may not in every case be the most useful one for exploring these possibilities, especially the second – but the point is that it doesn’t exclude them. This is an especially important point to make because today’s full matching models, at least as used by the labour economists (can’t speak for the macro guys), are much richer than the 1970s version you present here – eg sometimes marginal product cannot be estimated ex ante by one of the parties, or there are large fixed costs to both taking and leaving a job, or marginal product is a characteristic primarily of the job not the employee, or any of the monopsony-flavoured models in the “New Labor Economics” emphasising the rents generated by the job match (so, eg, workers can be exploited in the technical Marxist sense as employers appropriate that rent).

Further, I can’t see how your insight – that there should be quicker matching available with the rise of the internet – affects things much. Firstly, finding a job ad or attracting resumes was and is only ever a part of what is encompassed by “search costs” (following up those ads or resumes to assess both job and applicant quality was and is much more expensive and time consuming for both sides). Secondly, even if this was not so what we have seen – a very slow but sustained fall in unemployment followed by a sharp rise as a result of an exogenous demand shock, with only a long slow further fall as labour demand only modestly improves – is perfectly consistent with a slow but sustained fall in search costs. The Beveridge curve did slowly move in during “the Great Moderation” – sure other things were going on, but the internet could be one factor. And in many countries (most clearly in the UK and Ireland but they’re not Bob Crusoe) downward wage flexibility has, for better or worse, been greater in this recession than in past ones – which again is what search models would predict as search costs are systemically lowered (people get a handle on their prospects more quickly and adjust their reservation wage accordingly).


godoggo 07.10.14 at 7:30 am


Alan 07.10.14 at 10:01 am

My casual observations (which are in Australia and for jobs paying less than about the 60th percentile) are that very view advertised jobs quote a pay rate and that the majority of applicants do not meet the pre-requisites specified by the employer. This does not sound very efficient to me but I’m not an economist, I’m a businessman.


soru 07.10.14 at 10:46 am

Job search to me looks like it would be best modelled by queueing theory; there are a bunch of people looking for a job, and employers get to pick the entries from the queue according to their preferences. Which makes it kind of like a priority queue, assuming employer preferences between employees are reasonably correlated.

On that model, increased efficiency in the job search process corresponds to the employer being able to look more closely at a larger number of candidates at constant cost. Or in other words, a _lengthening_ of the queue.

This is the same effect as seen when cheaper RAM lead to routers being misconfigured with too-large buffers;

By basic maths (

The maximum queuing delay is proportional to buffer size. The longer the line of packets waiting to be transmitted, the longer the average waiting time is. The router queue of packets waiting to be sent also introduces a potential cause of packet loss. Since the router has a finite amount of buffer memory to hold the queue, a router which receives packets at too high a rate may experience a full queue. In this case, the router has no other option than to simply discard excess packets.

Translating the model back to being about humans should be straightforward.


Alex 07.10.14 at 10:48 am

It’s possible that the Internet has actually reduced the efficiency of search in one important way: it’s far easier to blastspam your CV to every job that matched your search, greatly increasing the numbers of unsuited CVs hitting the recruiter’s slush pile. Stories about “xxx CVs for this job!!” must be seen in the context of essentially zero-cost information replication and distribution. Conversely, it’s also easier to advertise a job into a much bigger audience. Informational friction provides a degree of natural filtering.

JQ is of course right that if your theory of unemployment isn’t macro-founded it’s worthless.


Brett Bellmore 07.10.14 at 10:48 am

Why, it’s almost as though google can’t find you a job which doesn’t exist. And, as though, with almost endlessly extended unemployment benefits, you had options other than taking a job!

Seriously, part of the problem is that both ends of the search are not looking for a perfect fit to the other, but to get a better deal than that perfect fit. So, the whole process isn’t as transparent as you’re making it out to be.

An interesting natural experiment is that they, not long ago, stopped extending the unemployment benefits. In theory, this should cause unemployment to go down. Did it?


Alex 07.10.14 at 10:49 am

Interesting application of bufferbloat there, soru!


Ebenezer Scrooge 07.10.14 at 10:52 am

The Internet has the potential for being an efficient search methodology, but it often isn’t used that way. Nobody wants an efficient market except economists. Efficient markets occur, to the extent they exist, despite the best efforts of sellers, not because of them. (And employers, although formally buyers, are sellers as a practical matter, in all but the tightest of labor markets.) As Alan @ 13 pointed out, the Internet is usually devoid of price information. Just ask the ghost of Steve Jobs about labor monopsony, or any HR professional on truth serum about withholding information from employees.


Ben J 07.10.14 at 12:10 pm

Prof Quiggen, I wanted to make two points.

Firstly on cyclical unemployment – I thought Diamond motivated search theory as a way to ‘microfound’ structural unemployment, rather than to explain cyclical fluctuations. If you let the slow decline in unemployment from late 2009 onwards be down to poor policy responses to large nominal shocks, then you can explain a good chunk of the slow recovery. In the US, insufficient activism from the Fed and ongoing fiscal drag, and we all know the story of the ECB. I don’t know why we can’t use sticky wages, nominal shocks and poor policy response to explain why the US and the ECB have had slow recoveries, and Australia and say, Israel, have performed relatively better.

Structurally, we still know that the overwhelming majority of job openings are filled without job advertisements. I have seen estimates of as high as 80% of job openings are filled through personal contacts (a friend of a friend, etc). I don’t know what the most recent estimates are, but my prior would be that it is still a similar number. This is a process which websites like Seek cannot improve on. So perhaps only a small subset of job seekers would benefit from information technology efficiency improvements.

From a broader perspective, I don’t know why you think Solow’s productivity paradox shouldn’t apply here. We know information technology has not improved productivity by what it was expected to (unless we assume productivity would have essentially flat-lined without the IT revolution, which is hard to believe). So if IT doesn’t improve overall productivity, why would you expect it to have such a dramatic effect on labour matching? Sure, some people were hopeful, but I would argue that they were just wrong, not that search models don’t give us at least some insight, even if this is a low bar.

So perhaps we should attribute the failure of IT to improve search matching as just another facet of the Solow productivity puzzle, rather than a failure of search matching as a concept.


Nichol 07.10.14 at 12:48 pm

Maybe internet has helped companies set up schemes to find people that fit in any given box they like to define. But it didn’t help the unemployed to have a chance to explain that although they don’t quite fit that list, they might be a great person for that job. Especially in the longer run .. as job specifications won’t stay the same for ever.


Ed 07.10.14 at 1:22 pm

” is the most prestigious prize in economics as an indicator of what the academic economics profession think is good and important work. ”

So-called Nobel Prize in Economics or the Swedish economics prize. I read somewhere that Nobel arranged for his prizes to be given out by Norway, not his native Sweden, in part because he suspected the Swedes would pull this sort of chicanery.


Ed 07.10.14 at 1:28 pm

Its been decades since the private sector has been able to create enough jobs to absorb new labor market entrants.

Explosive growth of world wide population in the late twentieth century, meaning alot new labor markets entrants each year worldwide for decades than the economy can really absorb, plus the world hitting peak liquid fuels production in 2005 and its resulting effect on growth, plus growing automation, and people wonder why there are declining real wages and increased unemployment. I remember reading predictions in the 1970s that this would happen. Countries with low birth rates and low rates of immigration, such as Japan, have been able to limit the effects, but of course can’t escape the increased automation and the resource constraints on growth.


Kieran 07.10.14 at 1:44 pm

Probably orthogonal to your main point, but there’s an old and I think under-exploited connection here to Granovetter’s (1973) work on the actual social network structure of job searches. Maybe more on the micro side … anyway, Scott Boorman wrote a little-known paper that tried to model what an optimizing actor searching for a job would do in a given social network structure. He published it in the Bell Journal of Economics—here’s the paper.


Justin 07.10.14 at 1:49 pm

Having done a job search in 2012, I am not so sure about the premise of increased efficiency. In my field (programming), there’s a huge proliferation of job sites for posting both resumes and job offerings, which creates a huge amount of noise. Companies are completely inconsistent in what they ask for. Some require you to upload a resume and then fill out a custom application on some terrible web app that repeats what you just sent them. Then there’s the problem of companies whose requirements are just lies: they ask for 10 years of iPhone development experience. And as a searcher, your incentive is to throw your resume at them, because there’s no way for them to signal seriousness–over half my interviews were for positions where I had not fully matched their requirements.

There’s some efficiency to being able to submit job applications from my living room, but as soru points out, this can just increase the load on the employer to screen applications.

In the end, I got a job because a friend of the family sent my resume to a friend of hers. Anecdotally, these kind of network effects still seem to have a huge influence, which doesn’t fit with the premise that searching has gotten more efficient.


jake the antisoshul soshulist 07.10.14 at 2:02 pm

@Ed #19.

That makes sense. But why is there a panic among some about falling birth rates?
Other than the burden of aging boomers on Social Security and Medicare, shouldn’t this be a good thing. Basic logic would indicate that falling birthrates would lower future unemployment as long as demand kept pace. And, at least as long as the boomers are alive, they will provide some sorts of demand.


Barry 07.10.14 at 2:04 pm

Brett Bellmore 07.10.14 at 10:48 am

” Why, it’s almost as though google can’t find you a job which doesn’t exist. And, as though, with almost endlessly extended unemployment benefits, you had options other than taking a job!”

Coherence never has been your strong suite, has it now?


Layman 07.10.14 at 2:15 pm

“An interesting natural experiment is that they, not long ago, stopped extending the unemployment benefits. In theory, this should cause unemployment to go down. Did it?”

No, not in any meaningful way. At least not here in the Land of Opportunity.


Barry 07.10.14 at 2:28 pm

Thanks, Layman. I don’t expect Brett to learn from that, but in his favor, that merely puts him at a Chicago-macro level of honesty :)


J Thomas 07.10.14 at 2:36 pm

” Why, it’s almost as though google can’t find you a job which doesn’t exist. And, as though, with almost endlessly extended unemployment benefits, you had options other than taking a job!”

Coherence never has been your strong suite, has it now?

You imply that Brett should have a problem with putting these two together. Like, if there aren’t any jobs, you *should* have other options than taking a job!

But if both of his points are true in themselves, why is it his responsibility to reconcile them?

Still, here’s one way to put those together. The way it is now, there aren’t nearly enough jobs to meet desire for jobs. To see that, don’t look at the unemployment rate but at the employment rate.

The economy is a game of musical chairs and there simply aren’t enough seats for everybody. Brett is right about that.

And people on unemployment aren’t as desperate as people whose unemployment has run out. He’s right about that too. It’s a Thoreauvian quiet desperation.

One possible choice would be to do away with unemployment insurance. Then desperate people would commit crimes and be incarcerated. They would provide more jobs for prison guards than they do for unemployment agency workers, so that would be a net plus. Also while they are in prison they don’t count as unemployed. So the statistics would look better.

It would create jobs — security jobs, and legal-system jobs, and prison jobs. Keeping the unemployed out of prison does not create nearly as many jobs.


mpowell 07.10.14 at 3:28 pm

As with other commenters, I don’t think you have a smoking gun here. In fact, I think you’re just plain wrong. There’ s been plenty written about the problems with internet job search. There is very little evidence that it is helping. Until you establish otherwise, you haven’t proven anything.

Not to say that the search based theories are correct. Just that job search experiences in an internet search age says nothing about it one way or the other.


MPAVictoria 07.10.14 at 3:40 pm

” Keeping the unemployed out of prison does not create nearly as many jobs.”

How does paying people unemployment, which they then spend on good and services not create jobs?


Brett 07.10.14 at 3:51 pm

Where are you getting the idea that we’re not creating enough jobs? During boom conditions (like the late 1990s), we were and are – unemployment was below 4%, labor force participation was ticking upward, and that was all happening at the same time as a massive increase in illegal migration from Mexico for work.


Peter Dorman 07.10.14 at 4:04 pm

A lot of the replies have already spoken for me, esp. Derrida Derider. I just want to add two additional bits. First, don’t overlook the literature on how greater wage flexibility can exacerbate macro fluctuations. Second, search “theory” is not really a theory but a framing apparatus that can be outfitted with many different specific theories. In that sense, it is a big improvement over, say, supply and demand. (Actually it’s a kind of superset of supply and demand.) On a mundane level, it’s elementary to show that S=D has no particular claim to equilibrium status in a model that incorporates flows of workers and jobs and some sort of matching process. (The 45 degree line in a Beveridge Curve is not presumed to be an equilibrium locus.) That’s important.

One more bit. This is not just about macro. Earlier this year I submitted some testimony to the US Occupational Safety and Health Administration about the methods they were using in cost-benefit analysis, specifically the Viscusi-inspired hedonic value of life stuff. One of the points I made is that the theory behind the claim that workers are paid compensating wage differentials for dangerous work, among its many other faults, has been developed only in supply and demand frameworks. But it’s been shown that, in general, it doesn’t hold in search. The fact that search has become the norm in the profession ought to mean something in this context.

I guess the main thing here is that the search approach is not just S&D with friction, where the internet could be seen as friction-reducing. It’s a much more general, flexible framework.


Brett Bellmore 07.10.14 at 4:19 pm

My points are hardly inconsistent: Some people can’t find jobs because the jobs they need aren’t there to find. Other people ‘can’t’ find jobs, because the difference between unemployment and having a job isn’t enough to motivate them to get a job. Both contribute to the high rate of unemployment, they don’t have to represent the same unemployed people.

My personal opinion is that the government shouldn’t pay people to not work. But maybe the government could reasonably be the employer of last resort, as with the CCC.


Lee A. Arnold 07.10.14 at 4:21 pm

JQ, I also think that internet search may not apply to one of the biggest employment markets affected by the downturn, and which is barely returning even now: residential housing construction. Basically, any able-bodied teenager or young adult can walk onto a jobsite and be useful. How are they found? By word of mouth, or they walk right in off the street. Internet usage, even literacy is not required.


Lee A. Arnold 07.10.14 at 4:27 pm

Brett #28: “Where are you getting the idea that we’re not creating enough jobs?”

Because the employment-to-population ratio hasn’t budged. In the U.S. the unemployment rate has dropped predominantly because people stop applying for benefits, so they aren’t counted as unemployed. The inverse of this graph would be the unemployment-to-population ratio. Here is the U.S. evidence:


Lee A. Arnold 07.10.14 at 4:29 pm

In other words the “unemployment rate” is useless in this discussion, and almost meaningless in any discussion.


P O'Neill 07.10.14 at 4:34 pm

In the same way that “search is not really about searching,” the entire edifice of New Keynesian economics (a la Mankiw) surely has the same problem. The line was always “menu costs are not really about menu costs” (i.e. the direct cost of printing new price lists) yet that’s how the models were always motivated. In the same way that the Internet should have lowered search costs, it should have lowered the direct costs of price adjustments, the supposed 2nd order benefits of changing prices (compared to the cost) then disappears and the whole story about sticky prices is gone.


Tom Slee 07.10.14 at 4:54 pm

Ignorant question: What has happened to asymmetric information-based theories of labour markets, as in efficiency wages?

My daily impressions fit with other commenters about the Internet and search. It may make some open positions easier to find, but getting any useful information about the job (or about the candidates, given CV-spamming, as Alex points out) is no easier, so I don’t think it’s helped much with the matching part of the problem.

Also impressionistically, the Internet may have led to the exploding credentialism that seems to have affected the labour market as the floods of CVs make the use of simple filters (eg must have Masters degree) unavoidable.


Brett 07.10.14 at 6:02 pm

@Lee Arnold

Because the employment-to-population ratio hasn’t budged. In the U.S. the unemployment rate has dropped predominantly because people stop applying for benefits, so they aren’t counted as unemployed.

The actual growth rate in the economy isn’t fantastic, though – real GDP growth has been below 3% for the past two years, and it’s way down from the 4+% growth we had from 1996-2000. You can see that with this FRED graph as well.

That was my point. There’s no broader “lack of jobs” crisis beyond the fact that we’re in a weakly growing economy. The trend line on labor force participation rate mostly matches that.


Thornton Hall 07.10.14 at 6:03 pm

@7 Your number two is what this lawyer is trying to turn into a full-time job.

Re: OP. You describe the problem thusly:
But there is still a puzzle here, one that search models were designed to solve. Why doesn’t competition between unemployed and employed workers work quickly to reduce wages to the point where demand equals supply and where there is no involuntary unemployment?

Simple answer: equilibrium (demand equals supply) is probably not a good way to understand a dynamic, chaotic, system characterized by violent reversals.


The Raven 07.10.14 at 6:44 pm

This is all very erudite, and I feel my limits here. But…John, have you ever worked in private industry? Because search isn’t the problem and the internet only helps a little, any HR manager will tell you that.

The risk of failed hires and the difficulty of letting people go are big factors at work here. It’s hard to hire new people, and it’s expensive when a hire doesn’t work out. Conversely, it’s hard to let people go.

Look, let’s say a business has enough extra business that they need to create a new position. First, the company looks around internally for someone they can promote. The advantage there is that they’ve got someone who is already a known quantity, who they know they can work with. If necessary, that person’s position can be filled by a more junior hire—less expensive and less risk. If they can’t fill the position that way, they start the process for an outside hire. The job is posted and the hiring manager and HR look around for someone outside the company they already know to hire.

You can kind-of see that search is the least problem here. In hard times, job seekers will search very hard. Demand is the problem.


Kurt H 07.10.14 at 6:52 pm

Several of the prior comments have zeroed in on the key rebuttal to your post. The crucial part of search and matching is not the search costs but the matching function. The real reason that matching is an imperfect model in most macro scenarios is that the labor market is modeled as a mass of interchangeable workers and the difficulty of matching workers to jobs they can actually do is abstracted into the efficiency of the matching function. This is obviously going to be of limited application to the real world.

Clearly, we would get better results if we modeled human capital with multiple categories (i.e. you can’t get a job as an economist with a degree in biology). Firms and institutions don’t just need generic laborers — they need people with specific skill sets. Thus we would expect disruptions in employment to take time to correct.

However, there is a fair criticism that search models would explain the persistence of unemployment but not the magnitude. If any macro scale variables impact the matching function, then you could have spikes in unemployment.

I think current macro does have some serious issues, but search theory is actually one of the better areas (and as others have pointed out, it’s really a technique rather than a theory). The big problem in macro, as I see it, is that we don’t have a good handle on what parts of the economy have emergent properties arising from heterogeneity, and which parts can be modeled with a representative agent. In fact, it’s entirely possible that representative agent modeling is never a good idea.


JW Mason 07.10.14 at 7:13 pm

academic macroeconomics is in a bad way

Does this mean you’re feeling more positive about heterodox economics?


Glen Tomkins 07.10.14 at 7:47 pm

What exactly are you trying to do by discrediting a downstream consequence of the mistaken idea that unemployment is caused by mismatch? Sure, easier search should have made unemployment drop if mismatch had ever been an important contributor to it. But demonstrating the failure of that to happen seems unlikely to convince people who ever gave that idea any credence. They can easily escape the force of your refutation by claiming that upstream factors other than easier search have changed. I imagine that they would (already do, really) claim some combination of the job market being more fragmented into ever more specialized skill sets, plus workers are just more lazy these days — both of which would offset any increased ease of search.

I don’t see any effective alternative to attacking faulty ideas as far upstream as possible. To imagine mismatch as the cause of unemployment, you had to simply take demand for granted. Perhaps that was a more understandable error in the 70s when the dread wage-price spiral loomed large. But it was an obvious mistake even back then. People who still believe mismatch is the problem and search the answer have as their real problem the basic assumption behind that belief. If you fail to get at the assumption, they will simply generate fresh maggot theories to explain high unemployment based on that faulty assumption. They’ll just say it’s the same damn lazy, entitled, all-powerful workers that were the problem back in the 70s. QED.


Layman 07.10.14 at 7:48 pm

“There’s no broader “lack of jobs” crisis beyond the fact that we’re in a weakly growing economy.”

I think it’s a bit more serious than that. I doubt we’ll see the job participation rate in the US improve to pre-2008 levels for many years. For that to happen:

– The number of jobs created has to exceed the long term trend in population growth; that requires some 3 million new jobs per year or 250,000 per month;
– On top of that, we need the jobs to fill the gap between the current labor participation rate (about 63%) and the target (about 67%). With 150 million people in the potential labor force, that’s another 6 million jobs just to recover our losses. To close that gap in 5 years requires 100,000 new jobs per month. However, CBO forecasts that the labor force participation rate will not improve by 2021.
– The last year in which we achieved anything like this growth (350k per month) was 1994.

With US businesses focused on extracting greater productivity (and profits) from relatively fewer employees, I think this is the new normal. And M. Picketty has done a good job of convincing me that real per-capita growth rates of 3 or 4 % are an aberration.


Witt 07.10.14 at 8:00 pm

I am genuinely confused by some of the statements (or rather, the unspoken assumptions) in the OP. Speaking as someone who has spent parts of the last 15 years in workforce development and small business services, there are several things that don’t ring true to my (US) observations.

First, is there any data (there must be) on the percentage of job openings that are posted online? At least in my experience, the MAJORITY of small businesses and nonprofits that I have dealt with either never publicly list their jobs at all, or list them only on their own websites (rather than CareerBuilder, Monster, Idealist, etc).

This is not an isolated phenomenon. Sure, the big universities and hospitals and so forth are all online — and the chain restaurants and hotels. But the manufacturers? The independent retailers? The arts and culture orgs? The environmental groups? The human services nonprofits? Much less likely.

In addition, my state has tried all sorts of strategies to get employers to list openings with their statewide job-search system (which Unemployment Insurance claimants are required to use in their job search). Their overtures have been overwhelmingly rejected by all but a handful of (largely politically connected) employers. My own organization has never once, in my eight years there, listed a job on the statewide website. We know from our peers’ experience that it will just generate a high volume of utterly useless resumes.

That’s not even to touch on the problem of phantom job listings. These were a problem in the old Help Wanted ads too — as a naive 19-year-old, I fell victim to a temp agency that posted a teaser ad for a nonexistent permanent job, just to get candidates in the door and give them a sales pitch.

But these days, phantom job listings are everything from political appeasement (local officials want to believe that Big Name Employer X is hiring) to legally-mandated ads from lawyers who are going through the hoops to pretend they’ve advertised for an American worker so they can hire a H1B candidate, to phishing scams, and more. And because so many of them don’t have a phone number or physical address, it can actually be harder to suss out the scams than it was in the old days.

So right off the bat, there’s an enormous segment of the job market that isn’t participating online, and of the remaining listings there is quite a lot of “noise” (I didn’t even get into companies with with misleading zip codes, etc. which can be quite challenging for jobseekers who want to look for jobs in particular towns or cities).


J Thomas 07.10.14 at 8:08 pm

” Keeping the unemployed out of prison does not create nearly as many jobs.”

How does paying people unemployment, which they then spend on good and services not create jobs?

When I rethink it, that may be wrong. It’s a complicated thing that’s hard to measure, and I was trying to imagine how it could go. When somebody gets arrested they provide jobs for the legal system, but maybe not all that much. Then after maybe a year or so, if they wind up in maximum security it costs something like $50,000/year to keep them there. That’s an old number, it might be different now. I don’t think it’s very many people who get $50,000/year in unemployment compensation, but I could be wrong. I guess I could look that up.

OK, as of 2010 the goal was to provide half the employment income. That would be $50,000 for a $100,000/year job. But in practice unemployment checks were about 40% of the goal. That’s enough to say it’s mostly less. However, lots and lots of people don’t wind up in maximum security.
This is an estimate for California in 2008-9. $47,000 per inmate. California cost is likely to be high, but then costs have likely risen considerably since 2008. The breakdown appears not to include anything for prison construction, only maintenance. There are various ways this number could be wrong in either direction but I’ll go with it for now.

Mean annual wage was $46,440 in May 2013. So we spend as much on behalf of each inmate as the average worker gets as gross income, considerably more than unemployment benefits. And the ones that wind up in prison were likely making less than average pay before they lost their jobs.

I’m not sure whether it matters that unemployment insurance is yet another insurance scam. Employers and employees pay into the fund, and then when the inevitable unemployment comes some money comes back. If employers instead just gave the money to their employees while they worked, the employees could save it or spend it, and probably spending would be higher than it is.

But then, there’s no particular reason to think businesses would give that money to their employees rather than keep it for themselves, maybe put it into T-bonds or give it to owners to put into T-bonds. When we look at what would inevitably happen if we didn’t do it this way, I lack confidence I know what the alternative would be.


Thornton Hall 07.10.14 at 8:15 pm

This is the classic blogosphere debate about economics. The OP tries to connect theory to empirical reality and say WTF? All the “informed” commenters then explain: that’s not really the theory, and that’s not really the data.

Long story short, professional economists get away with defending a “just desserts” theory of the 1% while millions look for jobs that don’t exist.

A long series of posts over at inter fluidity about welfare econ ends by saying, “so positive econ should stop making normative claims”. That, is of course, true. But when is he going to point to a single piece of positive econ that could even inform those claims?


Tom 07.10.14 at 8:36 pm

Some thoughts:

– while the searching technology improved, the reservation wage may have gone up. E.g. recent college grads who can live at home with their parents with a fair amount of amenities as opposed to decades ago when this option was less palatable. Unemployment benefits were also extended in length in US.

– I am not sure that search was used to explain cyclical fluctuations. What is the reference here? I have always thought of search as explaining mainly the structural part of unemployment, not the cyclical one. After all, as you say, to explain cyclical unemployment you still need an exogenous shock.

– The above point also explains the problem with search in macro. If search is the main tool to explain unemployment and search can mainly explain the structural part, cyclical unemployment falls off the picture. It is stunning to me how many mainstream and otherwise talented macro-economists just can’t bring themselves to recognize the involuntary component of unemployment, which seems pretty standard in a keynesian approach. Instead the focus becomes the utility maximization of the unemployed so that the unemployed are always choosing to remain such.

– Higher uncertainty increases unemployment if the mean of the wage distribution remains the same but not necessarily otherwise. A higher variance won’t make you wait longer for a better offer if the expected value of that offer has become lower, as it happens in recessions. You may have just to settle for a so and so job.

– It is a nice post, I wish there were more wonkish posts here.


Sandwichman 07.10.14 at 8:38 pm

Glen Tompkins @45: “I don’t see any effective alternative to attacking faulty ideas as far upstream as possible.”

Anon., 1821:

Where division of labour is well developed, the skill of the labourer is available only in that particular branch in which it has been acquired; he himself is a sort of machine. It does not therefore help matters one jot, to repeat in parrot fashion, that things have a tendency to find their level. On looking around us we cannot but see, that they are unable to find their level for a long time; and that when they do find it, the level is always lower than at the commencement of the process.


Layman 07.10.14 at 8:48 pm

“If employers instead just gave the money to their employees while they worked, the employees could save it or spend it, and probably spending would be higher than it is.”

Of course, employees who didn’t save it, or didn’t save enough, and then lost their jobs, would starve. Every alternative scheme to avoid social safety nets eventually runs into this question, what to do when personal responsibility falls short. The answer is either “eff those lazy people” or “have a social safety net”. Maybe you choose the former, I don’t know. But if you don’t, then the proper realm of debate is over the form of the safety net, not whether it exists at all.


Lee A. Arnold 07.10.14 at 10:17 pm

Brett #40: “There’s no broader “lack of jobs” crisis beyond the fact that we’re in a weakly growing economy.”

We don’t know that. The last several recoveries have been “jobless recoveries”.


Jason 07.10.14 at 10:41 pm

All of the recoveries in the US postwar period have a striking regularity:

I can’t imagine a story that involves the internet** (having only existed for the last couple of recessions in the US) has much to do with how quickly the unemployment rate falls unless every effect of the internet exactly cancels (increased queue size per comments above, reduced search costs and times per John’s original post). Occam’s razor would suggest that you could just leave the internet out of any explanation.

** In place of internet, you could put unemployment benefits, government spending, inequality … or any of the other factors that have changed since the 1940s to today.


Collin Street 07.10.14 at 10:42 pm

Of course, employees who didn’t save it, or didn’t save enough, and then lost their jobs, would starve.

More than that, it should be obvious that any scheme based on individual savings will be more capital-intensive than a risk-sharing/insurance scheme, since:
+ in a savings-based scheme each person needs to individually accrue enough savings to cover them for the desired period of coverage, while a risk-sharing scheme can discount the money needed per-person
+ in any case an insurance-based scheme can be largely funded out of current money, not capital, because the flows in and out are fairly predictable.

So individual savings accounts will be vastly more capital-intensive, and because the money will be needed at fairly short notice it would need to be kept in fairly liquid form, unuseful for investment. Individual savings accounts for unemployment[1] are bad economics, risks need to be pooled for best outcomes. And that means insurance, which means obligatory insurance, which means power-of-the-state either directly or delegated.

[1] Same argument applies to healthcare, btw. Retirement’s more predictable so the need for liquidity doesn’t apply: there’s a non-batshit-insane case to be made for individualised savings as a large part of retirement planning.


J Thomas 07.10.14 at 10:44 pm

Of course, employees who didn’t save it, or didn’t save enough, and then lost their jobs, would starve. Every alternative scheme to avoid social safety nets eventually runs into this question, what to do when personal responsibility falls short.

Brett Belmont is concerned that when society takes too good care of people, they won’t take good care of themselves. So if they enjoy living on unemployment they won’t start looking for work until the unemployment money is about to run out.

This doesn’t make sense to me when I think of it from the point of view of an unemployed person. I can hope to get 40-50% of what I was paid at my last job, but often it’s less. This is not a lot of money. It’s pleasant not to have to work, but there’s the worry what happens next. The longer I don’t have a job, the harder it will be to get one. It’s bad for people who fall into depression, who get up and play video games or something, eat a pizza and go back to bed, don’t wash don’t shave feel worse and worse with that sense of dread growing…. I don’t want to be him.

But thinking as an employer it makes perfect sense. If my employees think they’ll starve if I fire them, then they will be *good* employees. If they’re desperate for work from the first day they lose some other job, then I can hire them cheap and they’ll be grateful. Anything that makes my employees less scared is a moral hazard for them. What they need to do is whatever I want them to, and anything which gives them the illusion of freedom is evil. If you pay them to not work, then they won’t work. It’s as simple as that. Maybe a very few of them will starve, but those will serve as an example to the rest and all the people who do the right thing as a result more than make up for them.

Speaking for myself, I think people are answering the wrong question. I don’t want people to be that desperate unless we have a famine or something. And when people get stuff that isn’t wages, that isn’t them mooching off society.

I think we should give the basics to everybody, just for existing. Particularly with our high-fixed-cost low-variable-cost automated stuff, what it costs society to double production is far less than what it cost to set up the production lines and make the first 50%. It just isn’t so expensive to do that.

And we can offer people reasonable rewards to work. If we can produce all the stuff people want with 40% or 50% or 60% of the workforce working, then that’s fine. They value their leisure more than they value what we’d give them to work more than they do. No problem.

If we come up with ways to produce more stuff with less work, that’s fine too. Fewer people get special rewards for working, more people get more stuff, it’s fine.

The economy is here to serve us and not so much the other way round.

What happens if there’s a crisis and we can’t be sure the economy will produce enough of something essential, like food? We can come up with plans to get more, and ask people to help out, and maybe offer rewards starting with social recognition but others too. If too many choose not to, if they’d rather have their leisure along with a chance to starve, then OK, them’s the breaks.

We do not have a responsibility to survive by working cheaper than machines. The economy is here to serve us, and we should cherish it according to how well it serves us.


Layman 07.10.14 at 10:44 pm

You should rework this chart using the labor participation rate.


Collin Street 07.10.14 at 10:46 pm

Mind, if you have a lot of capital you probably want other people to want capital. Which explains a lot of things.


Bruce Wilder 07.10.14 at 11:03 pm

Why doesn’t competition between unemployed and employed workers work quickly to reduce wages to the point where demand equals supply and where there is no involuntary unemployment?

A simple hypothesis: The full-employment equilibrium requires a higher, not a lower wage.

Labor markets, faced with a surplus of unemployed workers can grind out a lower wage, but if the full-employment equilibrium is only achievable at a higher wage, then the labor market is faced with a problem it cannot solve — at least not in any reasonably finite period of time, by itself. To arrive at a higher real wage by a process of lowering all prices and wages would require the Walrasian auctioneer to work very, very hard indeed, as such a deflationary process assaults the processes of money-financed investment and past financial commitments, fixed in nominal terms, to debt-financed ownership, control and investment.

There’s a basic asymmetry involved: the full-employment equilibrium is a ceiling, and all so-called “sh0cks” that knock the economy off of its equilibrium path in the direction of lower activity, place many (though conceivably not all) markets into a state, in which there’s downward pressure on prices, wages and rates.

If a shock were to throw the economy off its equilibrium path in an upward direction, toward greater activity, putting upward pressure on prices, that pressure could be satisfied quickly by inflationary adjustments.

One essential difference is that upward shocks do not throw resources out of the circular flow, out of employment and out of the generation of income, while downward shocks very well may. Deflationary adjustments, because they require the prolonged unemployment of resources and raise the money-financed costs of investment, can be very costly.


Thornton Hall 07.10.14 at 11:14 pm

@59 This seems to be a situation where the interests of CEO’s and companies diverge. CEO’s, the one’s who only care about themselves, want, above all else, lower marginal income tax rates. This puts them in political alliance with the crazies. So, to keep the crazies happy, they basically have to agree with the rest of the crazy platform, which includes a plank strictly opposed to the minimum wage. Anyone with a brain can see that Wallmart would profit from a higher minimum wage in a way that is questionable if they unilaterally raise wages. But then the crazies come out.

This became starkly obvious when I considered the fact that Sears, which should try anything at this point, is not loudly calling for sales tax on Internet transactions? Why not? Because it is in the interest of Sears, but not in the interest of CEO’s generally.


Thornton Hall 07.10.14 at 11:15 pm

Why the apostrophes in CEO’s? I wish I knew.


J Thomas 07.10.14 at 11:16 pm

Why doesn’t competition between unemployed and employed workers work quickly to reduce wages to the point where demand equals supply and where there is no involuntary unemployment?

Because when wages go down, so does demand. That requires wages to go down more.

Given perfect competition, all the workers would find themselves making less than a living wage. The first of them to starve would make things easier for the remainder.

Given more people than the economy needs, the Iron Law of Supply and Demand requires that the supply of people must drop to meet the demand for labor.

No sane person advocates this sort of thing. One reason is that if you tell a bunch of people that the economy doesn’t need them so they should starve — they should let the economy work them to death to get the last little bit of value out of them — they will be irate. They may likely get violent.

This is not a viable economic theory because it incites too much violence against its proponents.


Sandwichman 07.10.14 at 11:27 pm

“Why the apostrophes in CEO’s? I wish I knew.”

Because CEOs are innately possessive?


Thornton Hall 07.10.14 at 11:41 pm

The more I think about it, the more it bothers me that Sears and the whole Chamber of Commerce are not beating down the doors of Congress and demanding that Amazon pay sales tax. It really is the poison injected by Ronald Reagan. What else could explain it?


Eggplant 07.10.14 at 11:48 pm

Rather, it’s that employers can’t easily use the threat of new hires at lower wages to drive down the wages of existing workers (of course, this happens, but it’s clearly costly and risky in terms of worker morale).
They should still be able to drive down the wages of new hires, right? Shouldn’t this, in the economist’s world, be enough to clear the market?
Isn’t the obvious conclusion that the supply of jobs is inelastic over the applicable wage range? Just because wages get driven down doesn’t mean employers need more workers, whom they are already, on average, profiting from quite handsomely, as the wages for the existing workers haven’t tracked productivity in decades. As Wilder notes, driving down wages means employers may need fewer workers.


Bruce Wilder 07.11.14 at 12:37 am

I pay sales tax on Amazon purchases; I live in California. As I understand the current state of law, a State can collect if the seller has a physical-presence in the state, and California enforces that rule. A giant business like Amazon has a physical presence in many states, and so, they really couldn’t avoid sales tax in very many places. I would think Amazon would actually benefit competitively vis a vis much smaller rivals, although, of course, Amazon store-fronts a great many smaller rivals, so maybe that’s a wash. Afaik, Amazon does not have to collect sales tax on the pass-thru sales, if Amazon is not fulfilling the order. The proposed reforms I know of seek to exempt small mail-order houses in any case.


Bruce Wilder 07.11.14 at 12:55 am

JThomas @ 62 Because when wages go down, so does demand. That requires wages to go down more. Given perfect competition, all the workers would find themselves making less than a living wage.

I cannot tell if you simply have no idea what “perfect competition” means, or if you really cannot see what kind of factors might govern wages.

I don’t know that the line between fundamental insight and fatal oversimplification is so thin that you should miss it completely.


J Thomas 07.11.14 at 2:41 am


Bruce Wilder, it sounds like you have some sort of complaint.

Why doesn’t competition between unemployed and employed workers work quickly to reduce wages to the point where demand equals supply and where there is no involuntary unemployment?

I’ll try to say it simply: Two women, one job. Any woman without a job starves.

Woman A gets the job for $50/hour.
Woman B offers to do the work for $35/hour, and she’s capable of doing it. The boss agrees.
Woman A offers to do the work for $20/hour and gets the job.
Woman B offers to do the work for $15/hour.
A $10
B $8
A $5
B $3
A $2
B three meals a day
A two meals a day
B one meal a day

If they can’t organize, if they can’t make an agreement with each other that one works and both eat, so they give the employer one bid with no competition, then they lose.

This is how it works when there are two bidders who both must win — or else. They bid until the one with the fewest resources can’t bid any more, and loses.Then the winner takes the prize along with whatever she has left.

What keeps that happening today is mostly social structure. To some limited extent it takes time to learn how to do a job, so there’s a limit to how fast you can hire and fire. Also whenever you hire there’s a chance the new employee won’t be able to do the work, and then you lose until you can get somebody new. It doesn’t matter how cheap a bargain you get if you don’t get the work done. You have to spend time around some of your employees and it’s no fun if they despise you. It’s no fun if all your wife’s friends tell her she should despise you. If various of your employees look for sneaky underhanded ways to sabotage you that won’t be pinned on them, they can cost you a lot more than their pay.

So instead of paying below-starvation wages, it makes sense to pay people enough to get by and a little more. Keep them unless they do something wrong, they know you could replace them with cheaper people but you don’t, you pay more than you have to, and they feel pathetically grateful and do their very best for you.

So without the “friction” (you usually don’t fire somebody until you’ve had time to see whether they can do the job etc) and ignoring the social factors, when there are more people competing for jobs than there are jobs and the people have no alternatives, wages will quickly fall. If workers find some better choice than accept very low wages, then wages stop falling. Failing that, wages will keep falling until enough of the competitors die off. When supply of workers has fallen to meet demand, then there is no more involuntary unemployment.

Of course it’s oversimplified, but it’s economics so what would you expect?


John Quiggin 07.11.14 at 3:17 am

“Does this mean you’re feeling more positive about heterodox economics?”

It depends what you mean. As I’ve said before, I don’t see any value in self-conscious heterodoxy, and I don’t think the main heterodox schools of thought have contributed much to the debate since the GFC. Austrians have claimed it proved them right, then gone back to talking about methodology and libertarianism. Some Marxists have had interesting things to say, but at the expense of Marxist orthodoxy as laid down by, say, Andrew Kliman, who claims that the crisis reflects the declining rate of profit in the face of a successful workers movement. The post-Keynesians can at least claim vindication for Minsky, but again they don’t seem to have done much with it.

OTOH, there’s an obvious sense in which criticisms like the one I’m putting forward above, and those of Krugman, Stiglitz and others constitutes heterodoxy particularly with respect to the idea that mainstream economics is a successful scientific enterprise, and that divisions within it entirely reflect intellectual disagreement rather than class interests. See, for example,


K F 07.11.14 at 3:24 am

I think of other search areas – has the Internet reduced the number of unmarried people? Even with eHarmony and and Tinder, people get married at later ages. Well, maybe Tinder has something to do with it… Effect of Internet on search (for work, for love, for …) is complicated.


dahl 07.11.14 at 3:46 am

“The efficiency of job search and matching has been massively increased by the Internet”

The internet perhaps means a day or two less in the context of a 6-8 week process (and often longer) – ‘massive efficiency increase’ is overselling it to say the least.

I’m not saying that the searching model is perfect. I am saying that the rumours of increasing efficiency for job search and placement through an, at best, moderately faster advertising process, have been greatly exaggerated.


Unlearningecon 07.11.14 at 9:54 am

“The problem seems not be, as search models assume, that employers and potential workers don’t know about each other. Rather, it’s that employers can’t easily use the threat of new hires at lower wages to drive down the wages of existing workers (of course, this happens, but it’s clearly costly and risky in terms of worker morale). ”

Is it even the case that workers themselves could negotiate a position they would not previously have been given for lower wages? Though a standard assumption in economic theory, this seems absurd in reality. If you’re right for the job, you get it at the wage being offered; if you’re not, you get nothing. And if demand increases, businesses hire more people at the going wage. As you imply – and as Keynes pointed out some time ago – there is no ‘labour market’ to speak of.


JW Mason 07.11.14 at 12:07 pm

Some Marxists have had interesting things to say, but at the expense of Marxist orthodoxy as laid down by, say, Andrew Kliman

Andrew Kliman is a crank who represents no one except himself. Indeed, I’ve been struck by how many people feel the need to bring up Kliman as an example of Marxist as religious dogmatic, because he is literally the only one left. What your quoted sentence means is “some Marxists have had interesting things to say.” Why not leave it at that?


bob mcmanus 07.11.14 at 1:01 pm

73: Michael D Roberts and G Carchedi are also in the “declining rate of profit” camp, although I don’t pretend to understand the nuances.

And I follow a link over at Thoma’s (one link per comment, Cacchetti and Schoenholz, “Is the Fed Behind the Curve”)

“Turning to the second possibility, it is now commonplace to argue that the neutral or equilibrium real interest rate has declined. Economists expect the real rate to reflect the marginal return on capital in equilibrium. To the extent that the trend rate of growth has slowed – as the FOMC and many observers (including ourselves) believe – the equilibrium real rate should be lower than it was prior to the crisis.”

And I wonder if a long-term declining real rate of return is completely ridiculous, although I am sure I am misunderstanding everything.


Ronan(rf) 07.11.14 at 2:06 pm

What (non technical*) Marxist or other heterodox economists (or political economists) would people recommend ?(who arent Andrew Kliman – although I’ve never read him, just it seems to be the consensus here) Is Wynne Godley and Marc Lavoie’s book on Monetary economics good ?

* ‘non technical’ as in articles and books that are written in an accesible way, rather than a description of methodology.


In the sky 07.11.14 at 2:36 pm

Why doesn’t competition between unemployed and employed workers work quickly to reduce wages to the point where demand equals supply and where there is no involuntary unemployment?

Because when wages go down, so does demand. That requires wages to go down more.

You state this as a certainty, but it is not.

For example if wages are $12, a firm might hire 1 person. If wages are $10, a firm might hire 2 people.


bob mcmanus 07.11.14 at 2:52 pm

75: Oh Christ. There are just too many, as someone who avoids orthodoxy, I still find a library of political economy published every year.

At the right, Jacobin, New Left, and New Inquiry are good blogs by my radical standards, and I give this place much credit for linking to them. They can, especially if you search, lead you to good resources.

David Harvey is accessible in his recent work, and prolific.

Samir Amin is something of a hero to me, and the “world systems” school (Braudel and Annales, Wallerstein, Andre Gunder-Frank, Giovanni Arrighi, Brenner?) is winding down, at least the first generation, but is fun and useful.

James C Scott and studies of peasant political economies, “industrious revolution” and “Smithian Capitalism” is an area of interest for me right now.


Charles Peterson 07.11.14 at 2:53 pm

1) Internet searching has limitations and flaws not obvious to some. It’s easy to get a lot of stuff similar to what you want. It becomes increasing hard to find exactly what you want, at some point an old fashioned library might be better…even simply traveling through the shelves as I often have done, pre-judging books by their covers.

2) Searching seems a poor model of extended above trend unemployment. The Keynesian (and before) lack of demand seems like a much better explanation. Has that fallen completely out of macroeconomics fashion now? Pure superstition likely better than that enterprise then.

3) One of the worst ideas ever is that of economic equilibrium to which things return. Better would be a sense of history and a realistic view of the world we live in. Exponential growth cannot continue forever. But capitalist economies have relied on that from the beginning. Krugman gives the great bicycle analogy

From where we are now, I don’t see the bicycle righting itself ever again. Economists wave their hands and say things like “substitution” but I don’t see it. Those things which aren’t tied to the physical are optimally free.

Time to start thinking about how to build a decent communist society since it doesn’t require endless exponential growth. And how to get there in the least painful way.


Layman 07.11.14 at 3:19 pm

“For example if wages are $12, a firm might hire 1 person. If wages are $10, a firm might hire 2 people.”

I hear this argument against minimum wage hikes all the time, but in ~30 years experience in business I’ve never seen it work that way. If you have budget to hire one additional person at $1o, you can hire 1 @ $10, and in a pinch you can probably hire 1 @ $12, as the adverse budget impact at that level could be manageable. But what you can’t do is hire 2, even if your budget was $12, because that’s a huge budget impact.

There could be an impact at firms who hire lots of people, because scale, e.g. someone with budget to hire 5 @$12 could hire 6 @$10. Even that’s doubtful, though – if you could depress the hiring wage to $10, you’ll be allowed to hire the planned 5, and the excess will be ‘saved’, i.e. taken to the bottom line.


Trader Joe 07.11.14 at 3:45 pm

Right on layman.

Indeed more practically, before a company even thinks about hiring the frist guy a@10, theyve already started by paying some incumbent person some amount of overtime, or employed a part-time person or even a temp. The conversion to a bona fide new job @10 or any other price comes after the part time person was converted to full time and overtime is no longer efficient.

Most companies have built sufficient slack into existing work forces that there is some amount of flexibility before they need to even approach the labor market for the lengthy and messy task of hiring and usually training a new employee.


Thornton Hall 07.11.14 at 3:57 pm

@75 As a philosophy major who has been reading 2/3rds of Thoma’s links for about a year now, and as someone who has agreed with everything Krugman has said (that I understood) in his column, I came to the conclusion that something was deeply flawed in the academic discipline. I followed by objections first to the Freshwater, Saltwater debate but found both sides equally wrong on a philosophy level. Paul Krugman’s response in the “accounting identity” debate was a glaring red flag. No one trained in philosophy could read Krugman’s response and not see that he was committing the exact same error he was criticizing. I moved to the Orthodox/Heterodox debate and found that while the heterodox had the humility befitting a story-telling non-science, they obviously had no clue on how to climb out of the coal pit.

One think I kept coming back to was the word “model”. I knew the science of weather forecasting had improved by leaps and bounds thanks to models run on powerful computers. It took me a while to figure out that “models” in Econ weren’t what scientists call models. The more I read about equilibrium analysis the more I couldn’t believe that anyone would begin to imagine it might say something interesting about human behavior. What was worse, I found out that in practice, nobody could point to an example of a equilibrium model that matched the data. Oh sure, there were models that were backwards compatible with the data, but you can scour the Internet for the rest of your life, there is not one example of an equilibrium hypothesis validated by data.

In science, that’s call falsification. In Econ it’s called “the best we can do.” But what about real models, the kinds used by scientists?

It turns out, not surprisingly, scientists had reached all the above conclusions already. The book Forecast, by Mark Buchanan is an easy read and a wonderful antidote for philosophers like me who looked into the blogosphere and thought: “Either that’s totally fucked up or I am batshit crazy.” (I’m not.).

See, here:


J Thomas 07.11.14 at 4:28 pm

“Because when wages go down, so does demand. That requires wages to go down more.”

You state this as a certainty, but it is not.

Pardon me for the turn of phrase. Of course it is not a certainty. I didn’t even say it well enough to show what I meant, and what I meant depends on a whole collection of things that cannot be certain in the real world. And yet it can make sense to talk about tendencies and “drives” and such, and discuss whether our ideas about such things are all wrong.

For example if wages are $12, a firm might hire 1 person. If wages are $10, a firm might hire 2 people.

It looks like you’re saying that when wages go down, total employment might go up and so the total paid goes up.

Let me describe the model I’m working from, which of course is oversimplified and not in any way inevitable.

Why do wages go down? Because competition among workers has gone up, while competition among employers looking for workers has not gone up. Supply and demand. If employment goes up enough to threaten the safety cushion of unemployed, then wages will go up and not down.

Is production going up while employment is not going up? Probably not. Why would demand increase when employment does not increase, and wages do not increase? (Well, but it could happen. More exports. More government spending. More rich people buying lots and lots of stuff.) The old rule of thumb was that labor was about a third of costs. But of the other costs, for example you pay for raw materials but labor is a fraction of the other guy’s costs too. He pays for various things and labor is a fraction of the cost of the people he buys from. Summing the series, it would make sense to figure a rule of thumb of labor as maybe half the total cost, the rest being for rights and privileges and such. But then, total cost of human capital is likely to be more like 70% of operating expenses, so….

If labor costs across the whole economy aren’t going up, how is labor going to pay for more stuff? Why make more stuff unless somebody will buy it? When total labor costs go down, demand goes down some. But other demand might rise.

Wages go down. So people buy less. So less needs to be produced. Producing less, employers can lay off employees. Now there are more people looking for fewer jobs so it’s plausible that wages will go down more.

Of course you can argue it other ways. If there are lots of employers, they can go ahead and produce lots of product, and compete and drive down prices below cost. Lots of cheap stuff for people to buy. But the old rule of thumb was that you need at least eight competitors or they will do unconscious price-fixing collusion without even meaning to and certainly not in ways they can get caught and prosecuted for. When some of them get real big, they get economies of scale in demanding preferential treatment from suppliers and distributors, so the small competitors (who must offer price discounts due to being off-brands) don’t wind up very profitable and don’t get much opportunity to expand. Yet another rule-of-thumb, when there are four big competitors in an industry, the biggest one or two will probably do well. So there is usually far more competition among employee candidates for jobs than among employers for customers. So no, prices don’t go down. Instead production goes down, and prices stay up.

As people have been pointing out, this model is unrealistic about the job market. Even when there are a thousand resumes submitted for each job, that does not mean a thousand competitors because most of them are ignored. The actual competition for jobs is much less.

There’s a distinction between “special” jobs that are highly specific and require special arcane skills, versus more common jobs. So if you are expert in some weird computer language and a company has software written in that language that they need to translate into something that a regular code monkey can maintain, there may be only a dozen people in the world who have qualifications and documented experience, and only one who lives nearby. But if you want a job using a forklift on a loading dock, there’s usually somebody standing around they can choose if you want more than the going rate. One way, your only competitor has some high fixed costs tacked on. The other way, there are thousands of local job markets and you might get a job just by standing around at the right time and place. But if you offer to work at 3/4 the regular rate to get the job, they’re likely to think there’s something wrong with you. Is saving a few dollars worth working with somebody who has problems?

Since “the job market” doesn’t actually work like a free market, you can’t depend on it to get free-market results.

But if it did, then when there were more people who desperately needed work than there were jobs, the unskilled labor market and also any particular specialties that had more qualified specialists than needed, would find wages reduced to starvation levels. Supply and demand. Reduce the supply of labor by death until it is no larger than demand for labor. That’s how supply and demand works. That’s how it is supposed to work.


bob mcmanus 07.11.14 at 4:31 pm

81: Current reading, among others, Penelope Francks Rural Economic Development in Japan, 1875-1945. A long quote, with apologies, from my notes, from around the middle:

Although the concern here has been to present nineteenth-century
economic change on its own terms, rather than as a stage towards industri-
alization, the differences between Japanese and European proto-industri-
alization experience do suggest that a development path which has its
origins in the kind of rural industrialization now widely advocated for
contemporary developing countries may travel a different route from that
followed by much of the Western developed world.

Thus, the emergence of specialized agricultural and manufacturing
regions and of labour-force specialization prefiguring the factory organi-
zation of ‘real’ industrialization were not characteristics of Japan’s nine-
teenth-century rural growth. Although there was regional specialization
into different commercial crops or manufactured goods, households every-
where continued to practise agriculture alongside whatever else they
diversified into and the growth of rural manufacturing simply helped to
redress the balance between agriculturally well-endowed regions and the
more marginal areas within and beyond them. Moreover, rural manufac-
turing in Japan was carried out on the basis of a (usually gender-based)
division of labour within the rural household, rather than by means of a
proto-factory workforce increasingly differentiated from agricultural
workers. The complementary technical and organizational changes in agri-
culture and rural industry which made these developments possible consti-
tuted the mechanisms of the ‘industrious revolution’ through which rural
household members were enabled to work harder and longer, but also to
produce and consume more.

The Japanese historical case also reveals something of the conditions
that might lead to this kind of development path. High population density
(at least in the habitable areas) and relatively good means of communica-
tion greatly enhance the scope for rural growth linkages of the kind that
underlay the nineteenth-century virtuous circle. The nature of the Toku-
gawa system, with its fragmented markets and decentralized political
control, created scope for local rural producers which a national mass-
market might not have done. The Tokugawa seclusion policy, though not
as tight as once thought, certainly acted to (and may have been adopted in
part in order to) promote import substitution and restrict competition
from overseas, in ways that ultimately favoured the rural economy. By the
time that Japan was once again opened up to international trade, rural
producers were able to benefit from it, whilst at the same time retaining
their position as manufacturers of the differentiated Japanese-style con-
sumer goods for which a considerable domestic market had already been
built up. It might also be the case, as some have suggested (e.g. Bray 1983,
1986; Palat 1995), that irrigated rice cultivation, with its ability to support a
large population on a small land area, its marked seasonal peaks and
troughs in labour requirements, and its proclivity towards small-scale,
intensive and often family-based production organization, provides
particularly favourable conditions for the development of the kind of
pluriactive rural household that drove Japan’s industrious revolution.

Now what? This isn’t economics? Not enough math? Un-theoretical? Not useful or furthering the theory? Not helpful to poor and desperate people? One thing it is, is data-based and empirical.

Now, back to “search theory.” Sorry, I think I am too stupid, ignorant, and anti-social to follow the conversation.


Bruce Wilder 07.11.14 at 4:56 pm

Thornton Hall @ 81: As a philosophy major who has been reading 2/3rds of Thoma’s links for about a year now, and as someone who has agreed with everything Krugman has said (that I understood) in his column, I came to the conclusion that something was deeply flawed in the academic discipline . . .

I read Thoma’s blog faithfully, following links (probably never more than half, though), for at least five years, before I ran, screaming soundlessly into the night. But, I was trained as an economist and worked as one for some years, so I had more to overcome.

I liked your blogpost very much. It deserves to become an internet classic. It is incisive and on point. I liked that you choose that particular quotation from Krugman:

So how do you do useful economics? In general, what we really do is combine maximization-and-equilibrium as a first cut with a variety of ad hoc modifications reflecting what seem to be empirical regularities about how both individual behavior and markets depart from this idealized case. And people using this kind of rough-and-ready approach have done really well since 2008, on everything from inflation to interest rates to the effects of austerity.

That was such a shameless display of the full horror, which is mainstream economic “reasoning”, yet Krugman fans (of which I admit I am one, too) seem too often not able to see it.

J Thomas @ 68: Of course it’s oversimplified, but it’s economics so what would you expect?

You’ve got me there.


Thornton Hall 07.11.14 at 5:02 pm

@83 I can’t quite read your tone, so I am a little confused. I think that what you posted is a good example of non-science Econ, and a very interesting one. It uses the language of the social sciences to explain how humans interact in a way that is applicable to a more general understanding of ourselves and our social behavior.

It strikes me as exactly the kind of thing economics should have been doing over the course of the 21st Century. The economists known to the public shouldn’t be Laffer, Friedman, and Krugman, but rather those who could transform the above passage into popular page-turner prose ala David McCoullough.

Instead, we got math. But when you look at the accuracy of the instruments of Econ, the significant digits should have been zero.

The coal pit I reference isn’t to say heterodox is lost, it’s to say that it can’t stop the Orthodox from using Friedman’s “prediction” criterion as the definition of “useful economics” and then proceeding to harangue the public for not heeding predictions which are patently false. To beat them down we need to show there’s a better way to make predictions. And I think now, with massive computing power, it might for the first time, be appropriate to use mathmatical modeling as a tool of economics.

To summarize: all good Econ to date is like what you quote, and in the future that will still be true with the addition of insights from Agent Based Models.


Thornton Hall 07.11.14 at 5:03 pm

Edit: 20th Century for “21st”


Thornton Hall 07.11.14 at 5:13 pm

@Bruce. Wow. Thanks.


The Raven 07.11.14 at 5:47 pm

Witt@47: “At least in my experience, the MAJORITY of small businesses and nonprofits that I have dealt with either never publicly list their jobs at all, or list them only on their own websites (rather than CareerBuilder, Monster, Idealist, etc).”

Just so. The main result of listing on one of the big web sites is your HR person gets spammed by hundreds of eager desperate job-seekers. mainly gets job-seekers spammed and scammed.

It depends what type of jobs you are talking about. IT firms do post online, but not at the big sites; they’re more likely to list at boards where people they want hang out. Engineering disciplines naturally post at engineering society sites. In the trades there is still no substitute for good old-fashioned pavement pounding. And so on.


Thornton Hall 07.11.14 at 6:10 pm

@84 I’m not the only person to get jarred by the quotation about “useful economics”. A little while after my post came this from the head of the Philosophy Department at Duke:


Eric Titus 07.11.14 at 6:32 pm

Manning’s neo-monopsonistic framework is a significant theoretical improvement on search models (although the empirical results have been mixed). The basic idea is that a higher wage –> more applicants and better matching. That is, according to a number of optimization functions.

As a sociologist, I’m a little skeptical of the way economists model search costs. Companies find employees in all sorts of ways: networks and word of mouth; recruiting at universities and technical schools, help wanted posters. I’ve seen little indication (anecdotally or in the literature) that labor markets are the best way to understand jobs, unemployment or wages. Yet one can be assured that we’ll continue to get variations on these models from economists regardless of what the data says!


Tom Slee 07.11.14 at 6:35 pm

There is one benefit of the mathematical approach that I don’t think has been mentioned, which is that it generates arguments with a high standard of internal consistency: it creates a very precise path from a set of assumptions to a set of conclusions. Personally I find this a real plus (and one, while we are generalizing about fields of study, some philosophers could benefit from) and it has nothing to do with validity of predictions or significant digits.


Thornton Hall 07.11.14 at 6:48 pm

@Tom Slee Can you give an example?


Bruce Wilder 07.11.14 at 7:07 pm

@ 89 (7.11.14 @ 6:10 pm)

I saw the Alex Rosenberg piece earlier — the tireless Lars P Syll, a Swedish economist, pointed to it on his blog. The first commenter on the Syll post pointed to a review of Rosenberg’s book, by Deirdre McCloskey, who is a recovering Chicago economist, and a very sharp critic — she was . . . unkind to Rosenberg. But, that led me to her essay, The Two Secret Sins of Economics, which is available in various forms and drafts across the internet, including a published pamphlet (but don’t buy it, plenty of pdf downloads scattered about; never confuse price with value — economics lesson number 1 ;-). Such is the serendipity of the world wide web.


John Quiggin 07.11.14 at 7:13 pm

@JWM I’m glad to hear that’s a general view of Kliman. I had the impression that he was widely regarded as the standard-bearer for Marxist economics as a school of thought (the kind of thing I don’t like) rather than economics influenced by Marxian perspectives (the kind of thing I often do like).


Thornton Hall 07.11.14 at 7:36 pm

@93 I’m at page 6 and already it is clear that the committees in whatever Nordic country, in honoring whatever inventor of mass destruction you please, have quite certainly made a critical omission.


J Thomas 07.11.14 at 7:40 pm


The more I read about equilibrium analysis the more I couldn’t believe that anyone would begin to imagine it might say something interesting about human behavior. What was worse, I found out that in practice, nobody could point to an example of a equilibrium model that matched the data.

There was a time when equilibrium theory was useful in chemistry. It still has uses, but chemistry has moved on.

Economics started as interesting and useful fantasies. Nobody could watch the various shopkeepers, bankers, artisans etc doing their work — they wouldn’t stand for it. Lots of them had secrets. To get any sense of how it all fit together you had to imagine it.

People imagined feedback loops that tended to stabilize things. Watts’s steam engine governor came in 1788, compared to Adam Smith’s Wealth of Nations in 1776. The idea that things could be set up to naturally adjust themselves without requiring God’s active intervention was in the air.

It’s such a powerful idea that some people tend to get hypnotized by it. It’s so beautiful they think it has to be true. It explains everything. It must be happening all the time, optimizing everything.

Meanwhile, the discipline of economics has no grand projects. If engineers build dams that fail or steamships that blow up, people get the sense they might not quite know what they’re doing. After each war, surviving nations that feel like losers revise their military doctrine while the winners tend to sit tight. There is not really any failure mode for economics. People say that economists did not predict 2008, though lots of economists say they did predict it. But what were they supposed to do about it? It was nobody’s job to predict 2008….

If there was some big task that economists were responsible for, that teams of economists could do and then everybody could judge the result, that might give us some progress. While their most visible task is to make up stories to explain the economy that some people like, we can’t expect them to do much better than psychiatrists who make up stories to explain individual people that the people like.


John 07.11.14 at 7:55 pm

I’ve enjoyed reading this and many of the comments are excellent. I’ve collected a nice reading list for later — thank you.

My question: is it a crude over-simplification (or cop out) to say that unemployment is first and foremost an outcome of deficient demand (“aggregate demand”, or firms’ demand for labour) and that everything else (search, matching, information assymetry, Granovetter’s networks) are secondary* in that they speak to the ordering of job seekers or the process of job allocation?

*Students of undergrad labour economics are often bemused that time spent on these interesting topics does not bring them any closer to “solving” the problem of unemployment!

(Written on a phone so perhaps not not fully elucidated.)


Bruce Wilder 07.11.14 at 8:11 pm

Tom Slee @ 91

Unlearning Economics had a post yesterday, which addresses that point.


Sandwichman 07.11.14 at 8:11 pm

“People imagined feedback loops that tended to stabilize things.”

And they were half right!


Bruce Wilder 07.11.14 at 8:26 pm

J Thomas @ 96

Back in the day, to derive an equilibrium analysis, you had to find a constraint, such as a conservation law. That’s the way Euler did it. When economists find a constraint, they usually get it wrong, as when John Cochrane proclaims, “The money has to come from somewhere . . .”


Thornton Hall 07.11.14 at 8:37 pm

@Bruce Is it just unquestionable conventional wisdom around economics departments that Smith and Hume were libertarians? I am not quite finished w/ McCloskey’s pamphlet, but I can’t help but wonder how much better it would have been but for the slander against the Scottish Enlightenment. No wonder she didn’t like Rosenberg. He read Hume with a mind unclouded by the UofC.


William Timberman 07.11.14 at 8:50 pm

I kinda forgive her, inasmuch as she appears to be the kind of libertarian who shows up for a ditch-digging job with a shovel rather than a bullhorn and a copy of Atlas Shrugged.


J Thomas 07.11.14 at 9:05 pm


“People imagined feedback loops that tended to stabilize things.”

And they were half right!

Yes! And now we know how to design feedback systems in a variety of circumstances. Given permission, we could design feedback systems that actually would tend to stabilize our financial systems.

The existing systems are designed to fluctuate because that provides opportunities to profit at somebody else’s expense.

For example stock markets are designed to destabilize prices, maybe partly because they were designed by and for brokers. And brokers make money from commissions. Commissions come when customers churn their accounts. When there’s lots of buying and selling. And stability does not promote that.

We know how to design stock markets that would tend to be more stable, that would be good places to park retirement money etc. We know how to buffer those markets against arbitrage coming from destabilized markets. We don’t yet have the political will to do it.

We know ways to increase or decrease liquidity without requiring large amounts of fractional-reserve bankloans. We have no consensus that would be a good thing.


Tom Slee 07.11.14 at 9:12 pm

Thornton Hall at #92 asks, if I read correctly, for an example of the value of equilibrium analysis.

You might well already know this example, but George Akerlof’s “Market for Lemons” (there’s a PDF here) is a classic example of good equilibrium thinking. He shows that when one party to an exchange knows more about the quality of a good (used cars in his example) than the other party, the market for that good may have problems. It turns out (because Equilibrium!) that it the owner of a good used car cannot get the price that the car would justify, even if there are potential customers who would be prepared to pay it. One might think that the party with more information is in the better position, but it turns out in this case that’s not true – asymmetric information is a problem for both.

The cases where the model applies are open for debate — including whether the used car market itself is an example — but what the equilibrium analysis shows is a mechanism that would be difficult to tease out in other ways.

If you are really bored, I wrote a long piece about another of Akerlof’s ideas (an equilibrium analysis of identity) a couple of years ago here.


Thornton Hall 07.11.14 at 9:14 pm

Wait. The claim is that scams are better understood thanks to equilibrium analysis?


Sandwichman 07.11.14 at 9:15 pm

“…now we know how to design feedback systems in a variety of circumstances.”

I’ll let Jonathan Swift reply:

“So nat’ralists observe, a flea
Hath smaller fleas that on him prey;
And these have smaller fleas to bite ’em.
And so proceeds Ad infinitum.”

The feedback systems we design also generate unintended byproduct feedbacks whose consequences are not as well understood. Case in point: carbon emissions.


Haftime 07.11.14 at 9:59 pm

On the chemistry equilibrium vs. economics equilibrium – chemistry has a few fortunate advantages – things happen on a faster timescale, and there are a lot things equilibrating (glass of water contains ~3,000,000,000,000,000,000,000,000 molecules), so you have you lot of averaging to benefit from.

Even then, there are a lot of problems (glasses, proteins leap to mind) that are somewhat tricky to deal with using straightforward unthinking equilibration, let alone continuum modelling.


Ronan(rf) 07.11.14 at 10:06 pm

Thanks for the recommendations Thornton and bob.

Bob – the little(only because it’s dense and complicated) I’ve read of the ‘world systems school’ I liked a lot (primarily Arrighi) I’ve been meaning to read something by Samir Amin, anything specifically you would recommend ?

Thornton – I ordered that book you noted (it looks pretty good) and I liked your blogpost a lot. One thing though (bear in mind I’m not an economist/social scientist etc, so this is more a caveated observation rather than a definitive statement) but this:

“Notice what never happens: no one ever looks at the world without having already decided how it (the world) works.”

Afaict, most people who use quant methods in their research or try to model real life would respond (something along the lines of) “YOU cannot look at the world without having already decided how it (the world) works” .. ie even when you are using your own powers of observations backed up by deep historical/sociological expertise, you are still viewing the world through a specific prism (complete with your own biases, gaps in knowledge etc), or in other words your own mental model.
Perhaps a (specific) model might be flawed, but are you objecting to the fact that economists models don’t have *perfect* explantory/predictive power and so the entire enterprise is worthless, or just that their usefullness has been overstated ?


Ronan(rf) 07.11.14 at 10:13 pm

Just to add, fwiw, I’d agree with Tom Slee’s comment here ..
” There is one benefit of the mathematical approach that I don’t think has been mentioned, which is that it generates arguments with a high standard of internal consistency..”
Not because I’m any good at maths (I’m basically innumerate)but I do think it does give you a *useful* (and different; not superior,but useful) way of thinking/looking at problems etc. (I won’t say any more than that – because I couldn’t articulate it properly – but I do think it does.)


Thornton Hall 07.11.14 at 10:18 pm

@107. Thanks Ronan! Bruce above recommended the blog of Lars P Syll. His post today is along the lines of “The Secret Sins Of Economics” that Bruce also mentioned, with the advantage of being in direct response to Simon Wren-Lewis, a prominent member of the Econoblogosphere (from whose blog, Mainly Macro, I find myself banned). Here’s a snippet:
And applying a “Lucas critique” on Lucas own model, it is obvious that it too fails. Changing “policy rules” cannot just be presumed not to influence investment and consumption behavior and a fortiori technology, thereby contradicting the invariance assumption. Technology and tastes cannot live up to the status of an economy’s deep and structurally stable Holy Grail. They too are part and parcel of an ever-changing and open economy. Lucas hope of being able to model the economy as “a FORTRAN program” and “gain some confidence that the component parts of the program are in some sense reliable prior to running it” [Lucas 1981:288] therefore seems – from an ontological point of view – totally misdirected. The failure in the attempt to anchor the analysis in the alleged stable deep parameters “tastes” and “technology” shows that if you neglect ontological considerations pertaining to the target system, ultimately reality kicks back when at last questions of bridging and exportation of model exercises are laid on the table. No matter how precise and rigorous the analysis is, and no matter how hard one tries to cast the argument in “modern mathematical form” [Lucas 1981:7] they do not push science forwards one millimeter if they do not stand the acid test of relevance to the target. No matter how clear, precise, rigorous or certain the inferences delivered inside these models are, they do not per se say anything about external validity.

Long story short. Yeah, I get the postmodern constructivist bs, like the man said (kicking a stone), “I refute it thus.” The question isn’t whether our world view shapes our perception (it does). The question is whether you allow for the possibility that your perception could change your worldview. By hammering at equilibrium no matter how often it fails to map our experience with the world, academic orthodox Econ simply refuses to notice a problem.


Thornton Hall 07.11.14 at 10:19 pm


J Thomas 07.11.14 at 10:44 pm

“…now we know how to design feedback systems in a variety of circumstances.”

The feedback systems we design also generate unintended byproduct feedbacks whose consequences are not as well understood. Case in point: carbon emissions.

Sure, it happens. But we’re better off with open-source systems designed for stability that get a lot of testing, than we are with commercial systems that are designed on purpose to be unstable.

Everybody knows that steam engines blow up, but nowadays they very seldom blow up inside their design specs. Partly by long experience, we mostly know what works.

Electronic circuits are inherently unstable in some ranges, but your computer chips work because they keep them within workable ranges and also run them slow enough to stabilize. If computer glitches happened often enough to worry about, we’d all be a lot more worried.


Thornton Hall 07.11.14 at 10:54 pm

@108 The best quick take-away nugget (to my mind) from Dierdre McCloskey’s “Two Secret Sins Of Economics” (which I read this afternoon at Bruce’s suggestion) is her analogy to newspaper chess problems on the value of using math to test internal consistency:

I have expressed admiration for pure mathematics and for Mozart’s concertos. Fine. But economics is supposed to be an inquiry into the world, not pure thinking. (If it is to be justified as pure thinking, just “fun,” it is not very entertaining. No one would buy tickets to listen to a “theory” seminar in economics. Believe me on this one: as mathematical entertainment the stuff is really crummy.) The A-prime/C-prime, existence- theorem, qualitative-only “work” that economists do is like chess problems. Chess problems usually do not have anything to do even with playing real chess (since the situations are often ones that could not arise in a real game). And chess itself has nothing to do with living, except for its no doubt wonderful purity as thought, á la Mozart.


J Thomas 07.11.14 at 11:00 pm

Afaict, most people who use quant methods in their research or try to model real life would respond (something along the lines of) “YOU cannot look at the world without having already decided how it (the world) works” .. ie even when you are using your own powers of observations backed up by deep historical/sociological expertise, you are still viewing the world through a specific prism (complete with your own biases, gaps in knowledge etc), or in other words your own mental model.

It’s a big long complicated process when you try to verify your model against the real world.

But the model itself is likely to be more consistent than it would be without the rigorous detail.

Like, a little while ago I wrote about the job market as if it was some sort of consistent market, and I made completely ad hoc implicit assumptions about how employers would deal with their own problems. I assumed without a second thought that when demand dropped that employers would reduce supply of their products. That’s like assuming that when the job market gets tight, a lot of job hunters just give up and don’t try to get jobs — which I didn’t assume.

If you have to explicitly say what your assumptions are you won’t necessarily make better assumptions, but at least you have more chance to think about them. That has to be a plus.

Plus the model will tell you about things that your assumptions predict that you may not have been thinking about. So for example if you assume that free markets optimize something, when you look at an actual model you might find that two different free markets are competing to optimise different things, each of them influencing the parameters for the other. One might be fast enough to determine the conditions the other must adapt to, and the other then slowly changes the environment of the first. Or they might do chaotic interference with each other. It’s easy to assume everything will work out, and then find that things you assumed away actually do interact given your assumptions, in ways you did not notice they would.

If the models give you a richer understanding of the consequences of your assumptions that’s a great big plus, independent of the relationship of your assumptions to reality. When you don’t really understand the consequences of your assumptions then it’s harder to tell how well they fit reality.


Tom Slee 07.12.14 at 12:02 am

Thorton Hall #104: The claim is that scams are better understood thanks to equilibrium analysis?



Thornton Hall 07.12.14 at 12:08 am

@114 I apologize for the snark. A fairer statement would be this: it seems like the problem solved by “seeing how asymmetric information causes problems in markets” would not exist, but for some crazy nonsense that you have to agree to as the price of admission to serious Econ journals. In other words, equilibrium provides partial guidance on how to get out of the (rabbit) hole dug by equilibrium.

I mean, ask any idiot on the street about the used car buying process and you’ll get a very good real world description of market failure.


Bruce Wilder 07.12.14 at 12:46 am

Analysis can be very powerful. It is how, as in the Akerlof example, we can identify the necessary and sufficient elements of a system and work out how a mechanism functions. But, Thornton Hall, philosopher, is correct: analysis is a priori.

The problem in economics starts very early, in the classroom, when the professor lectures on the model of market price under perfect competition, and turns to wave out the window at “markets” or industries, which supposedly approximate perfect competition, because of the large numbers of buyers and sellers or the homogeneity of a commodity product, and so on, and, then, not subtly, suggests that perfect competition constitutes an ideal, and the corrupt world outside the classroom window might be improved by making actual markets more like the ideal of perfect competition.

There’s nothing particularly wrong with the analysis of perfect competition. Its simplifying definitions help students to understand very basic concepts. But, there’s something very wrong about supposing that exercise in conceptual imagination describes any actual market or industry, even approximately, never mind that it describes some sort of desirable ideal!

It’s that ill-considered hand-waving, which is the problem in economics. It’s that ill-considered hand-waving, which insulates the core analysis as an ideal, from critical thinking and prevents the serious development of operational models and measurements of the actual, institutional economy. Instead of field work and case studies informing the profession with facts, the typical economist learns an analytic toolkit, like the canonical “market failures” litany, which can be used to generate a pontifical explanation, with almost no basis in factual knowledge.

That’s why I was horrified by Krugman’s cavalier “a variety of ad hoc modifications reflecting what seem to be empirical regularities about how both individual behavior and markets depart from this idealized case. . . . using this kind of rough-and-ready approach” — there’s no methodology there, and no recognition that a methodologically informed construction of an operational model might be necessary. And, the consequence is that there’s no real feedback going on. The “idealized case” — “maximization-and-equilibrium” — remains untouched by anyone’s encounter with the world.

Now, there’s no doubt that Krugman is a reasonable guy, and I’m sure all his “ad-hoc modifications” seem reasonable to him. New Keynesian “sticky prices” seem reasonable enough, for example. Model them as motivated by menu costs, knowing that menu costs are not about menu costs, search is not about search, and 75 years after the General Theory, Krugman is using a toy model kludge, IS / LM analysis, to argue ineffectually.

Nothing is ever wrong, and no one ever changes his mind, because facts do not really matter.

As an operational model, maximization-and-equilibrium isn’t ideal — it doesn’t make any sense at all. In a world of radical uncertainty, where people do not know what they do not know, “maximization” can not defined. “equilibrium” is even more vacuous — in practice, it comes down to what pedants call, the ergodic hypothesis, a crazy idea that history doesn’t matter — particularly when you consider that capitalism is one strategic investment after another in making history matter.

On one level, I get it. Economists feel they have a powerful analytic apparatus, which gives them a vocabulary of concepts, which they can use to produce insight about a variety of problems. The problem is that most of them know very little, outside of their own specialized research, and the profession as a whole is not particularly receptive to the results of specialized research. Economists are highly trained in bad taste, when it comes to interpreting what is going on around them in the economy, what is important and what it implies. (Many conservative economists take as a badge of honor that they know that they do not know anything — it helps them sincerely recommend laissez faire.) The discipline is structured around resistance to learning. And, it is harmful, this ignorance. Civilization may very well be facing collapse, because economists cannot be bothered to understand the foundational dynamics of the industrial revolution.

P.S. On “maximization-and-equilibrium”, an operational model might use rent-seeking as a behavioral rule, and cycles as a dynamic ordering scheme — it is not as if no economists have ever thought about these things semi-realistically. I don’t know how the profession could get there, but I do think it could begin by euthanizing everyone, who insists on pontificating about the real interest rate, as if there can usefully be thought to be one, or as if money is just a disposable assumption in the toy model of the week.


Unlearningecon 07.12.14 at 8:42 am

@Tom Slee re: ‘Market for Lemons’

One thing that’s always struck me about that paper is the fact that used car markets don’t actually fall apart. In fact, I’m not sure of any markets that actually follow the mechanics outlined by Akerlof. For me, that’s a primary example of how a paper in economics can become ‘seminal’ without actually having evidence behind it.


J Thomas 07.12.14 at 9:30 am

I don’t know how the profession could get there, but I do think it could begin by euthanizing everyone, who insists on pontificating about the real interest rate, as if there can usefully be thought to be one, or as if money is just a disposable assumption in the toy model of the week.

Try out this related question — why is it that philosophy has not achieved a consensus about reality? There are all these philosophers, and they disagree, lots of them talk as if there’s no way to really find out the truth, and they never seem to get anywhere. How come after all this time they don’t all agree about things?

Meanwhile we have a bunch of amateur philosophers who think they have a handle on things. I’ve met amateurs who like Nietzsche about the way a lot of amateur economists like Ayn Rand.

When I took beginning philosophy courses they taught about the classics. We read little snippets of Plato and Aristotle, Kant and Hume, and they tried to teach the fundamentals of each set of ideas. I’m reasonably sure that more modern philosophers have improved on each of these, but what we got were the originals with all their flaws. It makes a kind of sense to do it that way, to some extent they did it like that in physics too. We learned some Galileo and some Newton before Einstein and a botched description of quantum mechanics. Ontogeny recapitulates etc.

So beginning economists learn the classics too, and they particularly learn economic logic. Scientific method has particular sorts of patterns to it, the sort of logic that lawyers do has different patterns, economists use a different logic still that’s sort of related to evolutionary thinking. What do economists have in common? They’ve learned to think like economists. They might go on to do scientific method or sophisticated mathematical modeling, but those techniques won’t resonate with the profession because they aren’t connected to the common understandings that economists share.

Far too much, the goal of economic endeavors is to come up with a satisfying explanation of something. In one sense there’s nothing wrong with customer satisfaction, but look at the directions that goal has sent astrology….

I thought some about the more goal-oriented approaches I’d like economics to take, where measuring the achievements would give some sense how useful the methods are. But as I developed the idea it struck me that what I wanted economics to do is just exactly what operations research does.

So now my goal is to increasingly ignore economists and direct attention to OR guys instead.


Tom Slee 07.12.14 at 2:45 pm

UnLearningEcon #118 and Thorton Hall #116. I posted the Lemons paper as an example of a case where a bit of mathematics within a maximize+equilibrium framework gives a mechanism that has the benefits of being internally consistent (#91). If you then want to object that it has other problems (eg any idiot can see this, or it doesn’t correspond to the actual used car market), well OK, but that’s shifting the goalposts.

The benefit, to me, is that if the mechanism doesn’t fit a particular reality (eg used car market, and the market has at least been thin at various times) then at least we know where the problem lies — the assumptions. In fact, it cuts down the overall level of goalpost-shifting, which seems to me like a good thing.


Thornton Hall 07.12.14 at 6:52 pm

@12o If you then want to object that it has other problems (eg any idiot can see this, or it doesn’t correspond to the actual used car market), well OK, but that’s shifting the goalposts.

It may not have been clear before, but those are now, and forever will be, the goal posts.

If academic econ wants to be something other than a parlor game it needs to provide insights that we wouldn’t otherwise have into the world we actually live in.

Thanks for helping me focus on that. I’m pretty proud of that sentence.


John Quiggin 07.12.14 at 8:11 pm

As regards the lemons paper, it does not predict that used car markets cannot exist. The empirical evidence consistent with the model is that near-new used cars sell at a larger discount to the new price than can be explained by depreciation alone.


Sasha Clarkson 07.12.14 at 9:03 pm

I don’t like the metaphor of “shifting goalposts”: I prefer to think of “first approximation” etc. Of course, if subsequent approximations involve complicated and implausible layers of epicycles, then perhaps it’s time to consider the (once heretical) paradigm shift of ditching circles in favour of ellipses.

As for “nearly new” cars: do models of depreciation take into account non-rational factors like the vanity of the purchase who needs the latest registration number or “special edition” marketing paintjob? I seem to remember reading in a car magazine that the latter lost their value more quickly than regular models. But perhaps its a different kind of rationality, depending upon how one measures one’s self-worth?

The “latest-of-the-latest” model fetishist can revel in a display of (at least relative) wealth, and I can be smug about the bargain I got by buying a three-month old ex-demonstrator model: which, incidentally, by now has had most of its post-manufacture glitches fixed.


Walt 07.12.14 at 9:12 pm

It’s crazy to me that people are arguing about the merits of the lemons paper. The only way you can think that the conclusion of the paper is objectively obvious is if you are from a parallel universe with no Republicans in it. “Markets can solve all coordination problems in society” is the dominant idea of our age. Akerlof showed that this doesn’t work as soon as you have different quality goods trading in the same market, and then pointed out this case neatly explains several real-world examples. The only purpose of the math is to answer a certain kind of skepticism that the details can’t possibly work out.


A H 07.12.14 at 10:12 pm

@122 As regards the lemons paper, it does not predict that used car markets cannot exist. The empirical evidence consistent with the model is that near-new used cars sell at a larger discount to the new price than can be explained by depreciation alone.

This is just false, the paper explicitly states that the model predicts the market will break down,

“However, with price p, average quality is p/2 and therefore at no price will any trade take place at all: in spite of the fact that at any given price between 0 and 3 there are traders of type one who are willing to sell their automobiles at a price which traders of type two are willing to pay.”


Tom Slee 07.12.14 at 10:30 pm

A H: In the Lemons model, there is no trade for good quality used case. But that is not the same as a prediction that, in the real world, there will be complete market breakdown. Akerlof and everyone else knows that the real world is more complex than the very simple Lemons model: that’s why it’s a mode.

What Akerlof does say is “..there tends to be a reduction in the average quality of goods and also in the size of the market” (Para 2), and also “The automobile market is used as a finger exercise to illustrate and develop these thoughts. It should be emphasized that this market is chosen for its concreteness and ease in understanding rather than for its importance or realism”.

Sasha Clarkson #123: Akerlof does not take into account the factors you describe. In fact, he says “The usual lunch table justification for this phenomenon (excessive depreciation) is the pure joy of owning a ‘new’ car. We offer a different explanation.” So he’s not saying that what you call the irrational aspects don’t play a role, just that there are alternatives to think about.

One of the reasons I chose this as an example is because of the modesty of Akerlof’s goals, which contrasts with much of the economic imperialist approach to the social sciences.

And I agree with Walt at #124.

And finally, just in case anyone cares, I am not now nor have I ever been an economist, so I don’t speak for them and really don’t have a car—of any quality—in this race.


Tom Slee 07.12.14 at 10:31 pm

On line one it should say “for good quality used cars” of course. Aaargh.


Thornton Hall 07.12.14 at 10:55 pm

@Walt 124: you say a world with “Republicans” in it. But really you mean economists. That is the problem. The field confuses the obvious and then good guys like Akerlof and Krugman come in and get us close (but not all the way) to where we started.

One thing that never gets modeled: the opportunity cost of doing economics. What if we never went down the rabbit whole in the first place?


john c. halasz 07.13.14 at 1:54 am

JW Mason @ 73:

I haven’t read Kliman lately, but you’re bordering on bad faith there, the fact that JQ feels relieved notwithstanding.


A H 07.13.14 at 4:35 am

Tom Slee @ 126

So what is the point of the model in the paper?

The equilibrium model adds nothing that a more ad hoc approach couldn’t add. Say a heuristic based risk premium model.


JAFD 07.13.14 at 5:56 am

Trying to find link to McCloskey’s review of Rosenberg’s book – Googl not helpful – is it online ?

Thanks for help with this.


Bruce Wilder 07.13.14 at 6:10 am


Chris Bertram 07.13.14 at 6:23 am

Rather hard to take seriously a review that refers to “Marxists like Martin Hollis” as McCloskey’s does.


Tom Slee 07.13.14 at 12:08 pm

A H. To uncover a mechanism that is, along with other factors, at work in markets where quality is uncertain. To show that markets with asymmetric information will share a number of characteristics. To explain one economic role of guarantees, brand-names, chains, and licensing practices. To impose some order on our map of how the world works.


Tom Slee 07.13.14 at 12:10 pm

A H. I have zero idea what an ad-hoc heuristic based risk premium model is, and so I can’t comment on whether it would provide the richness that thousands upon thousands of people (albeit some of them economists) have found illuminating about the world around us.


Jonathan 07.13.14 at 5:13 pm

“That’s why I was horrified by Krugman’s cavalier “a variety of ad hoc modifications reflecting what seem to be empirical regularities about how both individual behavior and markets depart from this idealized case. . . . using this kind of rough-and-ready approach” — there’s no methodology there, and no recognition that a methodologically informed construction of an operational model might be necessary. And, the consequence is that there’s no real feedback going on. The “idealized case” — “maximization-and-equilibrium” — remains untouched by anyone’s encounter with the world.”

I have seen version of this criticism elsewhere (e.g. Dani Rodrik has made similar comments, which I’m too lazy to google up) and it’s a good one. But in a certain sense it’s futile. As far as I can tell, no one knows how. Krugman’s methodology is to apply his best judgement. The problems with this are obvious. But telling people they should do something they can’t do is not going to help.

To be useful it is necessary to produce actual results. For instance, if you think fiscal stimulus is called for, it is necessary not only to give a qualitative justification but a numerical estimate of how much reduction in unemployment will be produced by a given amount of stimulus. It that context, this (from Mr. Hall, above):

“equilibrium (demand equals supply) is probably not a good way to understand a dynamic, chaotic, system characterized by violent reversals.”

is pointless. Aside from its simplistic characterization of equilibrium, what economists actually do is to linearize non-linear models, fully aware of the possible pitfalls, and try to apply judgement to the results. No one here has proposed an alternative.

This procedure is not falsifiable (passing over the well known (at least philosophy circles) problems with falsifiability as a criterion) because you can generally tweak the parameters of the model to get more or less anything. And if tweaking the parameters won’t do it then changing the functional form probably will.

I’ve mostly enjoyed reading this thread because if the points made from actual experience. But I must say that there’s a self-congraduatory tone to some of it that makes me uneasy.

I wish I could say that it’s a good idea to remember that economists are not, on the whole, stupid, and they know what you know. On the whole this is true. OTOH, credentialed economists often do say thing that I believe are just crazy. For instance, there is a Youtube of Larry Summers saying that rules against sexual discrimination are unnecessary because market forces will eliminate it. He conspicuously failed to mention how long it would take. This strikes me as a claim that could only be made by someone who has run limits to infinity a few times too many. Examples could be multiplied easily and endlessly; I picked a Democrat to cite but Republicans are more productive and further off the wall.

So, what should actually be done? Sorry; if I knew that I would be smart. I spent several years studying Philosophy of Science. (For that reason I know what’s wrong with the falsifiability criterion in gory technical detail.) At the end I sometimes thought that I’d principally learned ways to avoid saying silly things. Avoiding being silly sounds easy; it’s not.


Jason 07.13.14 at 6:09 pm

@Bruce Wilder (#100)

“Back in the day, to derive an equilibrium analysis, you had to find a constraint, such as a conservation law. That’s the way Euler did it. ”

You may find this interesting. There is an information theoretic approach that doesn’t require constraints (it’s not a “free lunch”, the dynamics are limited) … but in the course of applying it, I may have found a proper constraint (it’s related, but not identical, to the quantity theory of money):


Thornton Hall 07.13.14 at 9:23 pm

@Jonathan 136
Krugman has produced a version of this response.

Here is Krugman:
Specifically: we have a body of economic theory built around the assumptions of perfectly rational behavior and perfectly functioning markets. Any economist with a grain of sense — which is to say, maybe half the profession? — knows that this is very much an abstraction, to be modified whenever the evidence suggests that it’s going wrong. But nobody has come up with general rules for making such modifications.

Now maybe, someday, someone will find a way to do something truly new — integrate neuroscience into economics for real, not as a marginal research topic, or turn agent-based models into a useful tool. I’m for it! But merely noting the foolishness of some economists and calling for a new paradigm in the abstract won’t get us there.

The problem with the response is that it imagines a distinction between his “textbook” economics and those producing “foolishness”. But notice, he has already told us that there is no distinction: “nobody has come up with general rules for making such modifications.” On his own description of the field, all the work is done by the “ad hoc modifications”.

Policy makers say “How do we get from here to broadly shared prosperity?”

Mainstream academic economics says “Here’s this self driving car.”

After much criticism, folks like Krugman come in and say, “Ok, ok, it doesn’t drive itself. But put me behind the wheel and I’ll drive us where we want to go.”

But he’s answering the wrong question. Before the economists came along with a car, we were content to walk. We just needed directions. 

“No, no, no,” says Krugman, “this is the only car we have!”

But Krugman himself just explained that the car isn’t doing the work. 

When you read Krugman, he talks a lot about IS/LM, but the directions, the actual policy advice, comes not from the model, but from the ad hoc modifications. And where do they come from?

The ad hoc modifications are not Econ, they are history: at the end of the day, the navigation is simply a matter of following the lessons of the Great Depression and other crisies. 

But it’s a really fancy car and very expensive. Entire Universities are built for the purpose of maintaining it. We should just walk. But the UofC crowd will tell you: skilled workers will fight like crazy to protect their industry from change.


Thornton Hall 07.13.14 at 9:25 pm

The Graf after the ellipsis is still Krugman.


bob mcmanus 07.13.14 at 11:50 pm

129: Well, yeah. Kliman is far from an exception in the school of orthodox Marxist economists, as anybody who followed my link to Michael Roberts can see.

The problem I have with the orthodox Marxists is their retention of models emphasizing industrial capital and wage labour; differences between constant and variable capital; differences between real and formal subsumption; between production and reproduction; between economics and politics.

IOW, all that classical economics categorization that connect them directly and sympathetically to mainstream economics as represented by DeLong, Krugman, Thoma, and Quiggin.


Ogden Wernstrom 07.14.14 at 4:39 pm

To be useful it is necessary to produce actual results.

I see Econ as a social science – but it deals with largely-measurable quantities, which gives many people – even practitioners – the idea that there should be a higher level of quantitative, predictive accuracy.

I do not ask Sociology or Psychology to tell me how many school shootings there will be in the US this year, then compare their results to see who is most correct. Nor do I expect that they correlate, say, classroom size with rates-of-shootings.

I don’t know that Political Science has good predictive power in specific instances, either.

Most of the concepts I learned in first-year Econ classes were first-order effects of change in a single variable. The phrase I heard the most from the professor was, “all else being equal”. Later, in Econometrics, it became clear that the noise in most real-world measurements added up to a lot of models with poor predictive power in the short term.

Forty years ago, I studied “Newtonian Economics”, while today’s economics studies may have moved a little beyond that. Maybe we will reach Quantum Economics someday? Will Economics go more-Schrödinger, -Heisenberg, or -Schön? (Or are we already in Schön territory?)


Marshall 07.15.14 at 6:42 pm

I wrote a column responding to this excellent post.


Thornton Hall 07.15.14 at 7:33 pm

@142. Marshall, I’m not done with my blog post response, but I’m curious how you would respond to the following:

Marshall is playing semantic games here. Economic actors face a indetermine future yet economic modeling is based on fixed, pre-specified functions. This is a serious problem. So Marshall says: wages aren’t set by assuming the pre-set mechanism of Nash equilibrium, the way wages are set is “indeterminate” in the search model. But he doesn’t mean indetetminate the way the world is. He actually means “a variable that the modeler can change”. That of course is simply another way of saying, unlike the world, the model assumes a pre-set, mechanical rule for how wages are set. But it doesn’t have to
Be the Nash rule. Pick a different rule and generate different data. (Sigh)


Bruce Wilder 07.15.14 at 8:11 pm

I like Thornton Hall’s snarky metaphor of the driver-less car, but I’m not at all sure that we can simply walk. If we walk, we still have to, more or less, walk together. Public policy requires a large measure of shared understanding.

The issue posed by the OP was whether search models, used in macroeconomic theory to model unemployment, should have any connection to the actual institutionalized technologies of employment search. This is classic Popper. Here’s an a priori analysis. Here’s an operational model measuring parameters and variables in a concrete situation. Should there be feedback from what we learn, operationally about the performance of institutions, to the idealized theory? What does that feedback say about what we keep or discard, modify or extend, in the theory?

From there, we meandered to Krugman’s loose marriage of “maximization-and-equilibrium” to seat-of-the-pants IS/LM analysis, as a way to argue macro policy. Alex Rosenberg, in the essay Thornton Hall linked to, upthread, argued that Krugman, by swearing ritual obeisance to “maximization-and-equilibrium” is conceding the legitimacy of his opponent’s arguments, thus ensuring that he can never win the argument.

Rosenberg’s position is that Krugman should exchange “maximization-and-equilibrium” for “uncertainty-and-reflexivity”. That cannot be a straight-up trade. There’s no axiomatic deductive system exemplifying “uncertainty-and-reflexivity” and I suspect there never will be — it is simply not possible. But, it certainly is possible to adopt “uncertainty-and-reflexivity” as one’s “vision” of how the economy works as a system, and to assemble analytic insights into the kind of mechanisms and behaviors the economy manifests. Ultimately, though, that would be an economics in which operational modeling, institutional and historical observation and interpretation would have much more weight. And, there would have to be an adherence to methodology in the operational modeling for that to work; the ad-hoc IS/LM analysis Krugman learned in undergraduate intermediate macro at Yale in 1972 could not be presented as ad-hoc “state-of-the-art”, nor could its efficacy be held to have no implication, constitute no Popperian test, for the “maximization-and-equilibrium” theory held by the orthodox faithful as revealed truth.

When you take the axiomatic, deductive idealization as your “first-cut”, the ad-hoc modifications become hand-waving about “frictions” when they should be serious investigations of mechanism. A classic example is the New Keynesian “sticky price” assumption, which is then attacked by conservatives as apostasy, a denial that prices adjust and resources are re-allocated. It is a fruitless, endless argument, in which no useful confrontation of concepts and/or fact takes place. The actual economy is organized around administrative hierarchies, and many “markets” (allowing a broad application of the metaphor) do not have a market-clearing equilibrium in price, so the convention that the economy can be modeled as behaving as if it is seeking a Walrasian general equilibrium appears highly questionable. And, developing a vision of how the economy works from laboring with such a model does not do much for anyone’s interpretive intuition, because there’s no correspondence between the abstractions and the observable, the kind of correspondence that would be worked out in operational models, operational models that would inevitably make the case for discarding “maximization-and-equilibrium”, a case the devotees do not want made.

It seems to me that the search theory, which is the subject of the OP, falls victim to the quicksand of this intellectual morass. Search theory seems to be an analysis of a mechanism, a way of elaborating a diffuse and formless friction into something more specific. But, if it is something specific, something operational, then there might be feedback from observable fact. “Feedback to what?”, you might ask. Well, to involuntary unemployment, the policy justification for Keynesian economics in all its flavors.

It’s probably impossible to get a coherent concept of “involuntary unemployment” into an axiomatic “maximization-and-equilibrium” model. That’s why the conservatives, who want to legitimate laissez-faire in the interests of plutocrats, insist on a methodology that features a privileged place for a model of that kind, and deprecation of operational modeling, open to “uncertainty-and-reflexivity”, where involuntary unemployment might emerge from the inevitable inefficiencies of 2nd best institutions, institutions that require public management. Why Krugman continues to play the mug in a mug’s game must be left as an exercise for the reader.

The coordination of society in an economy of specialization and trade requires a shared vision of how it works, and how it should work, and what it means when it fails to work, what our expectations should be of those assigned to continually reform and manage it, and what technical knowledge those assigned to reform and manage the economy, should have. Mainstream economics — at least the portions devoted to macroeconomics and financial theory — are failing us.

Neoclassical economics, with the axiomatic, deductive system that Paul Samuelson codified for it, is a powerful way to think, and, consequently, the basis for powerful rhetorical engines, capable of spewing out political argumentation and legitimating propaganda with ease. (witness the pitiful career of Matthew Yglesias or the vast chorus of neoliberalism) As a foundation for a scientific enterprise, its strong immunity to criticism and factual observation, makes it potentially paralyzing for a democratic society, which may be part of its continuing appeal for the powers-that-be.


Thornton Hall 07.15.14 at 8:30 pm

@144 Thanks for the work that went into that. Want to go through it a few more times, but one comment right away.

I agree there’s no deductive system that incorporates uncertainty and reflexivity. In my car analogy, the “deductive system” is in the role of the car. But I think there are other vehicles besides deductive analysis. Weather prediction has gotten much more accurate by increasing the resolution of the data and the processing power of the computer. My understanding is that Agent Based Models follow the same approach. Imagine the progress if brains like Krugman and DeLong would join the new style?


Bruce Wilder 07.15.14 at 11:01 pm

Yes, I’m labeling a class of alternative vehicles as “operational” modeling, without getting into the merits of particular techniques or methods, to emphasize that whatever its content, it wouldn’t be pure theoretical analysis. I think economists would still do theoretical analysis; it just couldn’t assume the same doctrinal weight or immunity from criticism. Pursuing as a holy grail, a grand unified analytic model that resembles the economy, would be ruled out of court as an impossibility, and therefore, also as inappropriate as the basis of a legitimizing “vision” of a just ordering.

I don’t know think the point I want to make would be enhanced by exchanging an humble agnosticism for an enthusiasm for some investigatory technique, which I could never back up. I have Mark Buchanan’s book on my Kindle, and I like what little I’ve read so far; I’m not refusing to consider. I do think reflexivity poses a stronger challenge for economics than it does, say, for weather forecasting or biological sciences. No matter how complex a phenomenon a tornado may be, a tornado is not smarter than the forecaster, nor strategically aware of the forecaster (or other storms). That’s not by any means an insurmountable problem, and the behavior of a greedy genius may be a simpler and more predictable (institutionally channelled) form of malevolence than a tornado, but it does suggest a different explanatory style and ambition will be required.

Although it might seem, on many practical levels, a technical problem, I don’t think that’s why, despite the rising torrent of criticism against mainstream economics, criticism continues to seem “futile”, as Jonathan put it. Even acknowledging the immense difficulty of, say, DSGE modelling or some abstruse econometrics, economics, at its core, economics is easy, too easy.

That is, what Dani Rodrik has dubbed “1st best economics” — and, the kind of bad moral philosophy based on it that Thornton Hall’s blogpost cited — is easy to spin out. Once you’ve mastered Samuelson’s Foundations or Principles, or Friedman’s Chicago gloss on it, you’re probably a living example of the admonition that a little knowledge is a dangerous thing. Write, “If” and follow on with idle and ill-conceived, counterfactual speculation, rinse and repeat, with great moral confidence in the Freedom to Choose, wages equal to marginal product, or similar nonsense, and you’re good to go.

It’s the power of that easy, no-need-to-know-anything analysis combined with moral confidence — and believe me, the most corrupt and hypocritical conservative can be convinced that they are champions of a moral power to choose “freely” arrived at as the insight of the innocent in a college classroom he once was — that constitutes the foundational appeal of neoclassical economics. You can see it in the passionate rantings of John Cochrane. (Krugman’s occasional admonition that “economics is not a morality play” could be read as unwillingness to engage in a contest to shape economics, qua legitimating dogma. [Note to Paul: it most certainly is a morality play, if you know how to work it, loser])

1st best economics is known to be wrong, even by the standards of 1st best economics. The Theory of the Second Best is a canonical result of longstanding in mainstream economics. Wrongness doesn’t stop many, or even slow them down. Doing second-best economics requires an acquaintance with particular and uncertain institutional facts and an acceptance of moral ambiguities or shortcomings. It doesn’t make the practitioner feel good in quite the same easy way.

A better economics will be a self-consciously second-best economics, and it will need both professional standards and a moral narrative suitable as legitimating doctrine, as well as more powerful technique.

It will need a way to make sense of facts and distribute them to professionals, and not just in the stylized form stripped of context that is common now. Evolutionary biology shows that it can be done. Biologists often earn their degrees doing field work, becoming experts in mollusks or ants, tidal pools or isolated desert oases — very particular species or ecologies or whatever. But, the Darwinian hypothesis combined with classification and a deep tradition of investigating mechanisms now deepened by molecular genetics gives them a way to make claims on generality, and biologists do learn a remarkable amount about creatures and ecologies outside their personal ambit. I don’t think it’s possible to just transplant a misunderstanding of evolution to economics, and expect it to grow organically. Biology’s precedent is not a solution, just a way of outlining the problem.


Tom Slee 07.16.14 at 12:24 am

Bruce Wilder: I could not agree with you more about a little knowledge, and particularly knowledge of idealized free market economics, being a dangerous thing. Incidents like the reorganization of the post-Soviet economy, which Stiglitz described in one of his books, are clear examples of the kind of ideological overconfidence you talk about.

But then you make two steps that I find harder to agree with. You slip from this naive econ 101 to “mainstream economics” and to “neoclassical economics” as if they are the same thing (at least, that’s how your paras 3 to 5 read to me), and they just aren’t. There are many mainstream economists who are far from the Cochrane’s of the world — maybe Rodrick is a mainstream economist by many measures. And surely Krugman, Stiglitz, DeLong et al are “neoclassical economists” who don’t share the outlook you implicitly attribute to them.

At some point, the dislike of the Chicago school seems to have overlapped with a dislike of mathematics and formality (the movement previously known as post-autistic economics for example). I just think they need to be kept separate.


J Thomas 07.16.14 at 1:34 pm

You slip from this naive econ 101 to “mainstream economics” and to “neoclassical economics” as if they are the same thing (at least, that’s how your paras 3 to 5 read to me), and they just aren’t.

Let me try out some definitions, and see who they work for.

“Mainstream economics” consists of everything about economics that 90% of economists would both say they understand, and would not label as bullshit.

So Econ 101 is mainstream, 90% of economists will not call it bullshit even though they (for reasons which may vary among economists) consider it oversimplified and in many ways wrong.

Things labeled “Austrian School” are mostly not mainstream because less than 90% of economists agree.

Probably 90% of the details of “Austrian School” economics are mainstream when taken in isolation, because they really aren’t that weird, it’s only a few basic ideas and a few highly important derivations that are all that different from Econ 101.

A “Mainstream Economist” is somebody who gets called an economist and who mostly agrees with the mass of economists about what’s mainstream, and who does not publicly disagree with mainstream economics.

If you say “Mainstream economics is bullshit” then you are not a mainstream economist even if you fail to disagree with 99% of it in detail.

If you make a point of talking about the ways that Adam Smith was wrong, or Ricardo was wrong, or Samuelson was wrong, you are not a mainstream economist.

If you are an economist who fails to publicly disagree with mainstream economics, and whose own work is specialized in a way that less than 90% of economists would claim to understand, then you are a mainstream economist if you say you are, and you are not a mainstream economist if you say you aren’t. In this context being a mainstream economist is a lot like being a Christian. You are if you say you are, provided you don’t spout such controversial heresy that a lot of other Christians argue that you aren’t.

Say you teach Econ 101. You can’t just not-teach the old, outmoded stuff in the textbook. Future economists need to know about it, it’s part of the tapestry of fundamental ideas that economists must be aware of. You could say “This is all approximations that have been superceded by better things” like physicists do when they teach Newton or Special Relativity etc. A lot of people will believe it anyway.

Over time, the size of mainstream economics can only increase. There’s no way to throw any of it away, except to gradually forget it. Incompatible ideas can both be mainstream unless mainstream economists start publicly disagreeing about them. And 90% of economists won’t do that.

I say, just put it aside.

This link is to a private company that is attempting to make a profit doing economic modeling for customers. Each customer has specific needs, and the company attempts to meet those needs, providing also ways to measure how well they meet the need. They use operations research methods.

Eventually this sort of thing will deserve a new name. Maybe it will be called econology or possibly econometrics. It will be like astrology and astronomy, or alchemy and chemistry.

Did it do any good for astrologers to argue whether their discipline got real-world results? How scientific it was? The difference between mainstream and heterodox astrology? Not much.

Look for work that can contribute to a science of econology, an art, a craft, an engineering discipline of econology. Ignore economists except when they get usable results.

There was a time (extending at least into the 1980’s in the USA) when high government officials were influenced by astrologers. Probably they will be influenced by economists for centuries longer. Live with it.


Dave Timoney 07.16.14 at 5:22 pm

Reading the comments, it is obvious that opinion divides into two camps: those who theorise what should be and those who have direct experience of what is. The gulf between them is large. The fundamental problem is twofold: the standard academic theory is incomplete in its model of the labour market and misguided in its appreciation of the impact of technology on search and matching.

First, the labour market is not a simple dynamic between employers and candidates. There are usually third-parties involved, whether recruitment agencies, internal HR functions (quite distinct from the “employer”) or mutual friends, who all have their different motivations and incentives. Rather than asking why online search has not reduced unemployment, you might profitably ask why it has not disintermediated recruitment agencies.

For example, internal recruiters are usually inefficient due to excessive caution. This is because of their fear of making a bad hire, as they’ll have to face an unhappy hiring manager on an ongoing basis (though the latter will have made the decision to hire, they’ll blame the recruiters for a poor shortlist). Similarly, the high success rate of personal recommendations depends on innate caution (if you recommend everyone, however inept, you quickly lose credibility). Agencies do not have this problem because of their transactional dynamic, so they tend to place candidates more quickly (more carelessly, critics would say).

Though paid by the hirer, agencies make their money by securing exclusive representation rights for valuable candidates (contracts allow for fees to be charged if the employer hires, independent of the agency, a candidate whose CV was already submitted by the agency). For this reason, most online advertising is actually placed by agencies, not by employers directly. This is not just limited to high-skill roles in IT, banking or engineering, but applies at lower levels as well.

Most online adverts are not for real jobs. It is illegal to simply make them up, so most are written to be sufficiently generic to avoid challenge (for obvious reasons, you don’t include the employer’s name, or even the precise location if this is a giveaway, unless you have an exclusive representation deal with them, as other agents will use this lead to cold-call the client). Agencies get CVs, speculatively email them to employers (based on known hiring interests, which flushes out imminent hiring decisions), and then punt interviews (i.e. they sell the candidate to the employer, not the job to the candidate).

The challenge agencies faced with the arrival of the Internet and the growth of job sites, like Monster and Jobserve, was that advertising was no longer expensive. Previously, large agencies could dominate the page advertising in newspapers and specialist mags, keeping insurgent competitors out through high prices (incidentally, the loss of recruitment advertising explains much of the troubled economics of newsprint today). With low costs for online ads, they could only preserve their dominant position (in terms of hoovering up candidates coming onto the market) by massively increasing the number of adverts they placed.

The global coverage of the Internet also resulted in the ballooning number of “inappropriate” CVs (e.g. candidates from India etc without a work permit). This proved to be a good thing for agencies as they justify their fees in part by “screening” these out. The more noise, the better. In other words, the increased efficiency of search has been offset by the massive degradation of the (virtual) database.

The final point to make is that “matching” has proved to be a process that cannot be easily automated, less because of limitations in the technology and more because of: a) the unreliability of CVs (which is the result of no standard semantic format rather than simple misrepresentation), which necessitates inter-personal qualification (i.e. at least a phone interview by an agent); b) uncertainty on the part of the employer as to what they actually want (far more common than you might think – the recruitment process if often an unstructured needs analysis); and c) the reality that all hires are about negotiating compromises – fitting squareish pegs into roundish holes is where agents actually earn their corn.


Thornton Hall 07.16.14 at 5:43 pm

@147 The absolute heart of the problem is the way “good guys” like Krugman and DeLong slip out of the net and combine sensible policy advice (informed by history, not models) with finger pointing at Chicago. There’s a fantastic “post-autistic” paper about how the flexible definition of “neoclassical” allows Econ an easy out. “You say Rational expectations is wrong so I point at this model that doesn’t use them. You say that model has a different problem and then I point to Simon Wren-Lewis, and so on.”

And then something truly bizarre happens like Krugman disavowing “equilibrium economics”!!!!
And it remains true that Keynesians have been hugely right on the effects of monetary and fiscal policy, while equilibrium macro types have been wrong about everything.


Thornton Hall 07.16.14 at 6:23 pm

Just to be clear about what Krugman’s comment about “equilibrium macro” is…

From blog: Uneasy Money
In discussing modern macroeconomics, I’ve have often mentioned my discomfort with a narrow view of microfoundations, but I haven’t commented very much on another disturbing feature of modern macro: the requirement that theoretical models be spelled out fully in axiomatic form. The rhetoric of axiomatization has had sweeping success in economics, making axiomatization a pre-requisite for almost any theoretical paper to be taken seriously, and even considered for publication in a reputable economics journal.

So does this imply Krugman is dismissing all of modern macro as wrong?


Bruce Wilder 07.16.14 at 11:02 pm

Tom Slee @ 147

The problem is not that Krugman and Cochrane have different outlooks or views; its that their differing views are not open to systematic reconciliation, because no one involved appears to have a methodology worthy of the term.

Thornton Hall @ 150, 151

It is not like Krugman’s macroeconomics is particularly sophisticated or informative; at best, it is a series of analytic sketches, without much grounding. Krugman’s macro rarely leaves James Tobin’s Intermediate Macro class, Yale circa 1972. That’s a pretty confined space.

In taking a swipe at “equilibrium macro” for failing to predict, Krugman is once again failing to appreciate either the methodology or politics involved. The macroeconomics of RBC and the New Classical theorists was always about what Deirdre McCloskey termed, proving qualitative theorems. They never wanted to predict recessions; they merely wanted to show that recessions were Pareto-efficient adaptations to changing circumstances. They wanted to undermine the case for policy intervention or management.

Krugman’s macro can predict only in a very limited sense, and since it is ungrounded seat-of-the-pants prediction, it is not like there’s an expansive view of the economy gaining credibility. Chris House is right that not even having a pre-2007 view of the financial sector integrated with the New Keynesian ad-hoc-ery is something of a fatal weakness.

In the back-and-forth on inflation and deflation, I don’t see any notice taken of the Winter-Spring 2008 global commodity price spike, which made a mess of Bernanke’s attempt to foam the runway with liquidity ahead of the economy’s recessionary landing. It was the brief bit of deflation that inevitably followed the bursting of the commodity bubble that triggered the crisis. (Using YoY inflation can make that bit of deflation disappear from some series and create other distortions; and it’s a prime example of why “core inflation” is not always the index of choice.) That global price spike indicated a policy constraint lurking in the shadows that might be a little more relevant than a Philips Curve formulated when the AFL-CIO represented one-quarter of the workforce.

Whatever it is that Paul Krugman and Brad DeLong are doing, in additional to lacking a discernible methodology, it is not that rich in interconnected systemic or historic detail. It’s not anywhere near the standards of Minsky’s John Maynard Keynes or James Galbraith’s The Predator State

Comments on this entry are closed.