Potential Output: Why Should We Care?

Brian Romanchuk has a characteristically thoughtful post making “the case against growth and stimulus.” He’s responding to pieces by Larry Summers and John Cochrane arguing that macroeconomic policy should focus more on output growth.

Brian has two objections to this. First, environmental resource constraints are real. Not in an absolute sense — in principle a given throughput of physical inputs can be associated with an arbitrarily high GDP. But in our economies as currently organized there is a tight connection between rising GDP and increased use of fossil fuels. Even leaving aside climate change concerns, that means that faster growth may well be cut off by a spike in oil prices. [1] The second objection is that the link between higher growth and better labor-market outcomes may not be as tight as Summers suggests. In Brian’s view, things like public investment may not do much for incomes at the bottom because the

U.S. labour market is obviously segmented. The “high skill” segments are doing relatively well… Non-targeted “demand management” (such as infrastructure spending) is probably going to require creating jobs for college-educated workers. (You need an engineering degree to sign off on plans, for example.) It is a safe bet that the job market for college graduates would become extremely tight before the U-6 unemployment rate even begins to close on its historical lows. This would cause inflationary pressures…

This suggests that the focus should be on direct job-creation programs for people left out of the private market, rather than policies to raise aggregate demand.

Since I am (very slowly) making an argument that there is space for more expansionary policy, evidently I disagree.

Before saying why, I should add one other argument on Brian’s side. One reason to be against “growth” as a political project is that higher GDP does not increase people’s wellbeing. In my view this is clearly true for countries with per-capita GDP above $15,000-20,000 or so. This is a moderately respectable view these days, though obviously a minority one. For most economists the case for growth is still so obvious it doesn’t even need stating — having more stuff makes people happier.

I don’t believe that. But I still think it’s worth arguing that there is more space for expansionary policy to raise GDP. For three reasons:

First, I think Summers and Cochrane are right (!) about the importance of tight labor markets to raise wages, flatten the income distribution and increase the social power of working people more broadly. I don’t think you would have had the mass social movements of the 1960s and 70s (even on such apparently non-economic ones as feminism and gay rights) if there hadn’t been a long period of very tight labor markets. [2] The threat of unemployment maintains the power of the boss in the workplace, and that reinforces all kinds of other hierarchies as well.

Corollary to this, I’m not convinced that the labor market is as segmented as Brian suggests. I think that in many cases, people with more credentials get to the front of the queue for the same jobs, as opposed to competing for a distinct pool of jobs. It seems to me the historical evidence is unambiguous that when overall unemployment falls there are disproportionate gains for those at the bottom.

Second, I think the idea of a hard ceiling to potential output is an important part of the logic of scarcity that hems in our political imagination in all kinds of harmful ways. Yes, infrastructure spending, and sometimes also increased social spending, even a basic income, can be presented as measures to boost demand and output. But you can also look at it the other way — these are good things on the merits, and the claim that they will boost output is just a way of defusing arguments that we “can’t afford” them. To me, the policy importance of saying we are far from any real supply constraint is not that higher output is desirable in itself (apart from its labor-market effects); it’s that it strengthens the argument for public spending that’s desirable for its own sake.

Third, on a more academic level, I think the idea of a fixed exogenous potential output is one of the most important patches (along with the “natural rate of interest”) covering up the disconnect between the “real exchange” world of economic theory and the actual monetary production economy we live in. Assuming that the long-run path of output is fixed by real supply-side factors is a way of quarantining monetary and demand factors to the short run. So the more space we open up for demand-side effects, the more space we have to analyze the economy as a system of money claims and payments and coordination problems rather than the efficient allocation of scare resources


[1] As it happens, this was the the topic of the first real post on this blog.

[2] The best discussion of this link I know of is in Armstrong, Glyn and Harrison’s Capitalism Since 1945. Jefferson Cowie’s more recent book on the ’70s makes a similar case for the US specifically.


(I wrote this post a month ago and for some reason never posted it.)


UPDATE: There’s another argument I meant to mention. When I look around I see a world full of energetic, talented, creative people forced to spend their days doing tedious shitwork and performing servility. I find it morally offensive to claim that a job at McDonald’s or in a nail salon or Amazon warehouse is the fullest use of anyone’s potential. When Keynes says that we will build “our New Jerusalem out of the labour which in our former vain folly we were keeping unused and unhappy in enforced idleness,” he doesn’t have to mean literal idleness. In a society in which aggregate expenditure was constantly pushing against supply constraints, millions of people today who spend the working hours of the day having the humanity slowly ground out of them would instead be developing their capacities as engineers, artists, electricians, doctors, scientists. To say that most of the jobs we expect people to do today make full use of their potential is a vile slander, even if we are only measuring potential by the narrow standards of GDP.

3 thoughts on “Potential Output: Why Should We Care?”

  1. in principle a given throughput of physical inputs can be associated with an arbitrarily high GDP.

    !!. Sounds like the violation of some physical law.

    1. But it’s not! Because GDP isn’t a physical quantity.

      As a practical matter there is a fairly strong link, especially in the short run. As the post says.

  2. In my opinion, what we call “growth” should be divided between 3 different main components:

    1) average productivity (due to technology);
    2) total size of the population.
    3) percentage of the population that is working;

    aggregate production is the product of these three components, so any change in one of these results in growth (or shrinking in case of negative changes).

    1) Change in productivity is ambiguous, because it might mean increased oil extraction per hour worked, but it could also mean increased efficiency in the use of oil. I think that this problem should be addressed by nudging technological progress to the “more efficient use of resources”, for example by taxes and subsidies. However this component in my opinion is unrelated to the business cycle, so it isn’t really the point.

    2) Total population growth obviously is a strain on the planet but I’m not going to tell people to not have sons, so this is also beside the point.

    3) Then we get to changes in the percentage of the population that is working, that is more related to the business cycle, and also to the “bargaining power” of the workers.
    Obviously if the economy “expands” so that, for example, 20% more of the people are working, it will need 20% more factories, it will use 20% more resources etc., so this kind of expansion certainly strains the planet.
    Also in my opinion the additional demand for building more factories when the economy is expanding is what drives booms.

    If we see booms in these terms, it is obvious that “potential output” is reached only when 100% of the people are working, so that it actually never reached but only approximated at the top of the business cycle (this is different from the usual definition of potential output).

    Keeping the economy closer to potential is desirable for two reasons: because more stuff is produced, and because there are less unemployed workers (who are somewhat marginalized by society) and workers will have more bargaining power VS capital, hence more income equality.

    However when we speak of the % of people who are working, there are some additional factors that should be counted:

    3.1) non-accounted-for work, for example in many situations an housewife is “working”, it’s just that her work is not counted in official calculations, so that in a situation of cultural change from a society where women stay home to one where they are employed confronting rates of employment to population could be misleading;

    3.2) the amount of hours normally worked in a week. In a situation where workers have small bargaining power, some workers will do a lot of overtime because they need more money or to comete vs. other workers, while other will be forced in part time jobs. This is something that in my opinion is happening a lot today.

    3.3) Also there might be an increase in the quantity of years one is supposed to work, for example in Italy, while we have high unemployment, we are also raising the retirement age.

    Thus in an ideal world, the problem of equality could be solved by placing a limit on the lenght of the workweek, and/or anticipating retirement, with a fixed quantity of aggregate hours worked/real output.

    However in the real world we are going in the opposite direction, and in my opinion the main reason is this:

    1) increased bargaining power for workers means higer wages relative to productivity (higer wage share), that means less profits. Therefore, either because of international competition, some minimum level of profits imposed by the financial system, or sheer ideological drive, we go in the opposite direction of lowering the wage share whenever is possible. Thus higher employment can be reached only as long as it doesn’t actually cause an increase in wages.
    For example, Romanchuk: “The latest reading of the U-6 rate is 9.7%, which is about 3% above the minimum reached during the relatively tight labour markets during the tech bubble. Under the current economic structure, the labour market was overheating.”
    On what basis does he say that the labor market was overheating? I don’t think that the wage share at the time was particularly high.
    I’m convinced that in the world of today, unemployment is used as a tool to keep wages low, even though often this is hidden through various layers of economic wordings so that the proponents don’t really understand what they are doing. Classic example: Greece. Paradoxically this is also the reason I’m skeptic when the various problems of the EU are blamed on the Eurocrats: if the crisis is a way to lower wages, even a country that is not inside the eurozone like the UK will magically enter in a crisis to lower UK wages.

Comments are closed.