Sunday, July 13, 2014

Demand in the gaps?

In a very good post, John Quiggin talks about some of the problems that have cropped up with labor search models. Personally I think labor search models are awesome, since - unlike most business-cycle models - they actually try to take into account the way economic actors behave in the real world. There's generally a "matching function" that tries to describe the technology that job-seekers and companies use to find each other - newspaper ads, human networks, websites, or whatever. Quiggin is right that changes in this technology itself are not enough to explain all the changes we see in employment and unemployment. More important might be the ways that people decide whether to take a job or to keep searching. For just a taste, see this paper by Mike Elsby and Matt Shapiro, which tries to model the fact that people get experience when they work. That is probably what Kartik Athreya meant when he wrote that "search models aren't really about search."

So search models still haven't fully explained employment patterns. What mental model of unemployment should we have in the meantime? At the end of his post, Quiggin writes:
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.
This is a simple answer, since it basically just thinks of the economy as an AD-AS model, and unemployment as being determined by Okun's Law or a short-run Phillips Curve. Economists are used to thinking in terms of supply and demand, so the AD-AS model comes naturally to mind. Inflation and output are both low, which looks like a negative demand shock on a supply-and-demand graph, so we look at the economy and say "it's a demand problem".

But on a deeper level, that's unsatisfying - to me, at least.

First of all, what causes aggregate demand curves to shift? You could have a monetary policy regime change. You could have a liquidity preference shock, possibly produced by financial market disruptions. You could have some kind of news about future productivity, or some kind of confusion about which sector is going to do well. You could have swings in sentiment - pure "animal spirits" - that become self-fulfilling prophecies. Really, you could have all kinds of stuff. This question is interesting, since different determinants of demand will entail A) different ways to predict recessions, B) different ways to prevent recessions, and C) different ways to combat recessions.

Second of all, how does aggregate demand affect unemployment? The usual explanation for this is downward-sticky nominal wages. But why are nominal wages downward-sticky? There are a number of explanations, and again, these differences will have practical consequences.

(Third of all, is an AD-AS model really a good way of modeling the macroeconomy? Supply-and-demand has its uses, and is sort of the iconic model of economics, but it has its limitations. The curves might exhibit hysteresis, for example, or randomness. The idea of abandoning the friendly old X of supply-and-demand is scary, I know, but maybe it just isn't the best descriptor of booms and recessions.)

So I'm not really satisfied by the practice of putting "demand in the gaps". If "demand" is not to be just another form of economic phlogiston, we need a consistent, predictive characterization of how it behaves - its causes, its effects, and its domain of validity.


  1. Anonymous3:42 AM

    This exercise is academic in every sense of the word. There is a simple solution to unemployment, abolish it. That's revolutionary, there should be no such thing as involuntary unemployment, it should be history. Whether this is achieved by a Job Guarantee, Basic Income or a combination of the two just do it. Then you can play with your models are bit without fiddling while Rome burns. bill40

    1. Anonymous6:42 AM

      Yeah but the problem then is that capitalists won't have power over workers, which is the entire point of capitalism (and Marx's original meaning for "Kapital").

      I mean, I'm 100% in favour of not having significant amounts of involuntary unemployment anymore via a UBI or whatever, but I'm pretty sure the people who count on the "reserve army of the unemployed" to keep profits up and labour costs down are not going to agree. Ever wonder why the government is obsessed over deficits in the middle of a downturn?


    2. The main purpose for competitive market capitalism is the level of wealth creation, innovation, and the efficiency in allocation of resources that it creates. As Kalecki argued back in 1943, eliminating things like business cycles or involuntary employment would further enhance both profits to a capitalist and prosperity to its workers. However, that'd mean less power and control over both the political process, and their ability to attract and retain human talent, so they will take power and convenience over money and more widely-diffused benefits.

    3. Unanimous11:34 PM

      You don't even need a special program, just a will to use the current tools in the right way. For about thirty years (1945 to 1975) governments kept unemployment down to historically unpresenented levels largely by spending on things that they normally do but aiming to provide a higher level of service which did involve more expenditure, and drawing taxes from very high earners who are less likely to spend.

    4. Anonymous2:13 AM

      North Korea has abolished unemployment.

    5. Arise, ye workers from your slumber,
      Arise, ye prisoners of want.
      For reason in revolt now thunders,
      and at last ends the age of cant!
      Away with all your superstitions,
      Servile masses, arise, arise!
      We'll change henceforth the old tradition,
      And spurn the dust to win the prize!
      So comrades, come rally,
      And the last fight let us face.
      The Internationale,
      Unites the human race.
      So comrades, come rally,
      And the last fight let us face.
      The Internationale,
      Unites the human race.

    6. cesium627:21 PM

      Refusing to measure unemployment is not the same as abolishing unemployment. Abolishing unemployment does not require that all decisions about how people will be employed be centralized.

  2. Daniel5:59 AM

    different ways to predict recessions

    Last time I checked, you were a fan of the EMH. One of the implications of the EMH is the impossibility of economic forecasting.

    So which way is it ?

    1. One of the implications of the EMH is the impossibility of economic forecasting.

      Actually, that's not true at all. EMH just says that asset prices incorporate the best available economic forecast.

  3. foosion6:52 AM

    One of the implications of the EMH is the impossibility of economic forecasting.

    No. At most, EMH says you can't consistently make above market profits using economic forecasting.

  4. Is it so hard to explain the demand gap of 2008-2013? Households borrowed like crazy:

    from 2000 to 2008, and then were forced to deleverage when it turned out that no matter what the ratings agencies claimed, it actually was possible for all real estate values to move down simultaneously, devastating the supposedly AAA-rated tranches of CDO's. This sudden imperative for increased saving across a huge number of consumers has to be offset by increased spending somewhere in order to sustain total overall spending. Where was that supposed to come from? Some increased spending came from the federal government in the form of automatic safety-net spending and the 2009 stimulus, but this was not enough to offset the collapse in spending from consumer deleveraging.

    I feel the above story is pretty compelling and supported by the data, like, the fact that dramatic increases in government spending and the monetary base led to zero inflation. I don't see how labeling "demand" as just another "economic phlogiston" helps push the conversation forward. We have a demand shortfall. The conversation we need to have is:

    1. will something like NGDPLT help us now and in the future;
    2. to what extent should we offset an AD shortfall with increased federal spending;
    3. to what extent can we exploit the demand crisis to make supply-side improvements like reducing unhelpful/expensive government regulation;
    4. how do we get the right-wing hard-money idiots to cut it out already. You have no god-given right to charge high rents on your piles of cash, interest rates are not "artificially low", money creation is not "debasing the currency", etc.

    Kenneth Duda
    Menlo Park, CA

  5. Anonymous11:33 AM

    Demand needs money. If you backtest the recessions they all start with negative yield curve i.e. a lack of money. In every case the Fed is tightening liquidity too much. Their data is lagging, in Q1 2008 they were still worried about a 3-4% inflation for the year. They are overshooting on the downside, and thus likewise also on the upside.

  6. The answer is not aggregate demand but effective demand. You will see later.

  7. To a first approximation isn't demand basically spending?

    Third of all, is an AD-AS model really a good way of modeling the macroeconomy? Supply-and-demand has its uses, and is sort of the iconic model of economics, but it has its limitations.

    Show us a plausible alternative. :^)

  8. " If aggregate demand is far below the productive capacity of the economy, workers will be unemployment and capital will be idle."

    The real problem with this statement is that it's a completely uninformative tautology. Output equals spending, output below capacity must be some mix of unemployment and underemployment of labor and capital.

    To say anything mire meaningful than output is below capacity one must break up demand and supply into their components.

  9. One thing that I noticed is that there is a remarkable regularity to recoveries from recessions:

    It would be hard to imagine a theory that included dependence on e.g. the internet, unemployment insurance, even monetary policy in such a way that all of the details canceled out to leave you with the same behavior across almost 70 years of data (in the US).

  10. In his 2001 text on Macroeconomics published by MIT press, Barro spends a couple of pages explaining why AS-AD model does not make sense. For the AS curve, his argument goes like this- an upward sloping aggregate supply curve implies the firms are able to sell their desired quantities at the going price. But then this does not make sense with basic Keynesian idea that producers and workers are constrained by aggregate demand in their ability to sell goods and services (p.788). The aggregate demand on the other hand only responds to changes in price level and not to changes in relative prices (p.279).

    The way I have tried to understand unemployment is through multi market perspective- In the macroeconomy all the markets are interconnected. If these interconnections work well- that is markets are able to accommodate changes in demand and supply- then there might be some temporary adjustments required but eventually things will balance out. For example if housing market suffers a setback and construction workers are unemployed then eventually they will find jobs elsewhere. Prices and wages will adjust to make this happen. In this case, even if prices are flexible, there will be some unemployment till adjustments are complete. In case of sticky prices, the adjustments will be delayed, demand for commodities might be affected causing more unemployment and hence more people might stay unemployed and for a longer period. There still remains the problem of figuring out the degree of price stickiness or level of adjustment costs that could bring about such chain of events. This is not totally from my head- I have been using Russell Cooper and Andrew John’s text on macro for my principles classes and they build their apparatus to understand the macroeconomy in this way. Either way, thinking about adjustment costs brings us the labor search models.