Saturday, March 26, 2016

Idiopathic Tardus Augmenti

There are religious nonconformists. There are climate change deniers. And there is now a new class of political agnostic: the secular stagnation skeptic. According to a piece in Time magazine this week, Barry Eichengreen finds the issue of secular stagnation so divisive amongst academic economists that he has coined a new term to help us sort ourselves into believers and non believers: the 
Steam Boat off Harbour's Mouth: J.M. Turner
 secular stagnation Rorschach Test. I like that term. And as I look at the ink blot of incoherent theory and misinterpreted facts that is presented to us for interpretation I find myself peering at a late Turner painting. I am straining to see the ship in the blizzard.

The Time piece is supposed to explain, to the layperson, what economists mean by secular stagnation. It serves only to spread the confusion that was laid by Larry Summer’s original article in which he resuscitated the term ‘secular stagnation’, originally coined by the American economist Alvin Hansen.

The confusion that originates in Larry’s article is between economic theories of unemployment and economic theories of growth. On unemployment, economists can at least agree to disagree.

For the thirty years leading up to the Great Recession, most economists built economic models where there is never any unemployment and the quantity of labor demanded is always equal to the quantity of labor supplied. Although there are still a few rosy tinted true believers who think that this makes sense; for the most part, the breed is dying out. It has been replaced by a new religion that genuflects to the altar of the sticky price. These ‘New Keynesian’ economists agree that, in the long-run, the quantity of labor demanded will equal the quantity of labor supplied. But, in the short-run, high unemployment persists because wages and prices are ‘sticky’.

Larry is not in either of these camps. Nor am I. I have spoken at length with him about this. Larry sees high involuntary unemployment as an equilibrium situation. I have modeled that idea in my own work where I use the theory of labor market search to explain why high involuntary unemployment may persist. And Larry has written papers with Olivier Blanchard where low involuntary employment may persist as workers become discouraged and remain out of the labor force. The exact formulation of this idea is unimportant. Larry and I are on the same page. Market economies are not self-stabilizing and they do not quickly adjust to find the socially optimal employment rate in the absence of active stabilization policies.

That brings me to the theory of economic growth. On this topic, despite more than twenty-five years of intense and active research, there is little or no consensus. The dominant way that macroeconomists model economic growth was developed by Robert Solow in the 1960s. Growth, in the Solow model, is caused by an exogenous increase in an unexplained factor called technological progress.

We produce goods by combining capital, land and labor using blueprints that represent the state of knowledge. If we had combined one acre of land with twenty-five machines and one hundred people in 2000 we might have produced (by way of an example) one hundred units of output. If we had combined one acre of land with twenty-five machines and one hundred people in 2001, we might have produced one hundred and three units of output. In this example, the economy grew by three percent between 2000 and 2001. But why did this happen?

According to the Solow model, the economy did not produce three percent more output in 2001 because we used more land, more labor, or more machines (although these may also have increased). It produced three percent more output because entrepreneurs used better blueprints. And according to the economic theory that is fed into almost all macroeconomic forecasting models, the reason for the improvement in those blueprints has nothing to do with economic policy and it has nothing to do with population growth, with investment spending or with changes in the unemployment rate. It is a black box we call technological progress.

There has, of course, been a great deal of work on theories of endogenous growth. Paul Romer and Robert Lucas have both produced seminal pieces on that topic that led to reams of economic research papers that try to understand Solow’s black box. To a macroeconomist who is interested in secular stagnation, these theories are a big disappointment.

I do not know why growth is low. There are a number of promising candidate theories. My own favorite explanation is that the Fed has lost control of inflation and that firms are not creating new technologies, at a pace that is fast enough to generate high growth, because uncertainty has increased. But I do not have a good economic model that links that idea in a coherent way with economic data. When it comes to economic growth; I have very little to say about why growth is currently slow. I am not unusual in that regard. Beware of economists bearing confident assertions about the best way to increase productivity growth. This is simply an area that we know very little about.

So what do I see in Turner’s painting? If secular stagnation means that unemployment can be permanently high if we don’t do something about it: I see secular stagnation. If secular stagnation means that we will be in trouble when the next recession hits because the Fed will not be able to lower interest rates further: I see secular stagnation. But if secular stagnation means that a massive bout of government investment in roads and infrastructure will cause firms to start producing better blueprints, I say; show me the theory and the evidence that leads you to believe that that is so.

Ignorance is not a reason for embarrassment. When medical doctors do not understand the cause of a disease, they cloak their ignorance with Latin. An illness of unknown origin is ‘idiopathic’. Economists should adopt the same strategy. When growth is slow and we don't know why, the economy is not experiencing secular stagnation. We are afflicted with a bout of idiopathic tardus augmenti.


  1. Full of honesty, an ingredient of another mysterious category: the excellent.

  2. The graph on explains both why economies grow and why growth has fallen after the Great Recession.

    Each year consumers save a certain amount of money and this money is funnelled into investment. The quantum of this investment (additonal capital) is usually greater than the amount which is physically consumed due to depreciation. So the same amount of labour (or a higher amount due to population growth) is combined with a higher amount of capital than in the previous year. This is why the slope of both the disposable income and the consumption expenditure lines is positive.

    At the point of the Great Recession the line of consumption expenditure falls steeply. Aggregate income is a function of aggregate consumption expenditure (not the other way around, as both Keynesians and Friedmanians think). So the line of aggregate disposable income also falls. The graph shows that the fall in aggregate disposable income follows the fall in consumption expenditure.

    Now if consumption expenditure falls, capitalists need to invest less than before. They also need a smaller amount of labour than before since aggregate production must fall in line with aggregate consumption. So we have a smaller amount of labour combining with a smaller amount of capital than before, thus leading to lower growth.

    The extent of the fall in investment may be seen from the second graph on

    Although investment is just a fraction of GDP compared with consumption, the absolute fall in investment was greater than the absolute fall in consumption.

    1. Philip
      You are giving a perfectly good explanation for why low aggregate demand may lead to high unemployment. That does not help us to understand why productively growth is low.

  3. Slow growth is no mystery to anyone with a tad of historical perspective. When we had demand side policies that got workers higher pay, we had higher growth. When the government got a bee in its bonnet and decided to fight a war or build a canal or highway system, we had higher growth. Demand drives growth.

    When we had supply side policies that cut workers pay, we had slower growth. You can trace this back through the Middle Ages and into the Greco-Roman era. It's hard to go further back since the records get rather sketchy. Even today, you can see the negative impact of pro-business policies by comparing red states and blue states.

    Now and then one hears the phrase "money burning a hole in one's pocket". That pocket is always a consumer's. Someone wants to buy something. It is never someone running a business eager to invest in some new production. Businesses respond to demand. If you've ever run a business or studied anything about accounting, you'd realize that running a business requires revenue, that is, someone spending money. There is no way to do this with just business firms investing in response to higher demand. The multiplier on that is too small if anyone is to show a profit.

    We actually do know exactly why growth has been slower over the last three or so decades, but there are powerful political forces intentionally obfuscating the reason.

    1. Ah, the political.

      Kaleberg, you said "When we had supply side policies that cut workers pay, we had slower growth. You can trace this back through the Middle Ages and into the Greco-Roman era."

      I'm fascinated by that kind of history. Can you pass along some links on slow growth in the middle-ages and Greco-Roman times, arising from the equivalent of supply side policies? Thanks!

    2. Once again: this is the same confounding of low growth with high unemployment. If high demand really did lead to high productivity growth, Zimbabwe would be overtaking China in terms of GDP per person.

  4. Often lost in the discussion is that "secular stagnation" is mere the name for the effect. Often people use it as if it is itself a tangible cause.

    Regarding Roger's statement: "But if secular stagnation means that a massive bout of government investment in roads and infrastructure will cause firms to start producing better blueprints, I say; show me the theory and the evidence that leads you to believe that that is so."

    This is fine to say, but not to do, as it is at the nexus of a classic asymmetric payoff game. We need the investment to be made, as the products of that investment are good for us regardless. If the addition to demand lifts us off, that is better, and free, therefore we must do it.


    1. This comment has been removed by the author.


    2. Roger FarmerMarch 28, 2016 at 7:46 AM
      We could certainly use some new roads and bridges and, more public infrastructure might even help us to absorb idle resources. If you believe that this will also lead to higher productivity growth then 1) spell out the chain of reasoning that makes this happen and 2) show me the empirical evidence that it has happened in the past. The evidence for a link from demand to productivity growth is weak to non existent.

    3. Well, as they say, economists differ...

      Rather, say the link from demand to productivity has not much data. Conclusions one way or the other are difficult to draw from the lack of data.

      However, the link has a really good story, and stories are important.

      But I would rather let Gerald Friedman do my talking for me, as this issue is the crux of the Sanders' plan kerfuffle:

      The investment decision, however, is clear as a bell, as previously illustrated. Any good investor would do it in a heartbeat.


    4. > spell out the chain of reasoning that makes this happen
      (this is pure speculation)
      - high unemployment / underemployment and low wages lead to a reduced pressure on companies to become more productive. - The creation of a large low-wage sector in germany produced a lot of jobs with very low productivity, since business models that rely on low wages suddenly became profitable (every 4ths german employee). Once you have had such a job for a few years it is very hard to get a better one (de-qualification, signalling).
      - it is easier to start a business that is based on low-wage / low-productivity job than a business that has high productivity per employee (kind of crowding out)
      - historically, technological advances were not exploited due to the availability of very cheap (slave) labour (e.g. roman empire)
      - high wages of british workers created pressure on companies to become more productive during industrialization (???)

  5. I've been peeking in your economic window. Something about your post keeps me coming back. The topic, probably. Good title, too.

    Near the end you wrote:
    "But if [X] means that [Y], I say; show me the theory and the evidence that leads you to believe that that is so."

    You also wrote of "incoherent theory and misinterpreted facts".

    And even your own "favorite explanation" is still only an explanation because it lacks "a good economic model that links that idea in a coherent way with economic data".

    I'm wondering if you would expand on that a little. To satisfy the standard you set, what measure of theory and model and evidence and coherence would be required? (Yeah, I'm not asking a lot.) Thanks.

    1. Economic theory is based on the assumption of diminishing returns. More labor and more capital generates more output. But at a decreasing rate. It follows that, according to neoclassical theory, building more capital will not lead to faster productivity growth unless labor is also increasing.

      There are exceptions. If an economy starts from a very low base, China is an example, there will be a period of very rapid growth, achieved by building capital, while the economy catches up to the technological frontier. Once an economy, like the United States, is at the frontier, growth becomes much harder. The only way for US to grow faster is by inventing new technologies at a faster rate so that we may use labor and capital more efficiently.

      The work by Paul Romer and Robert Lucas sought to explain the process of growth endogenously. They argued that neoclassical economics is wrong to assume that a one percent increase in all inputs leads to a one percent increase in all outputs. Technically, that argued for what economists call "increasing returns to scale". The evidence for that is mixed but, after thirty years of research, I think it is fair to say that we still have a very poor grasp of what causes steady increases in labor productivity. My beef with Larry, is that he did not distinguish clearly between high unemployment and slow growth.


Note: Only a member of this blog may post a comment.