Tuesday, December 13, 2016

Causality reversal is bad.

One problem with an empirical approach to complex systems is that frequently relationships can be found, but the causality is hard to determine.  If a relationship is very strong, reversing the causality can ironically lead to strong confidence about very wrong conclusions.  This is basically what I have found in housing.  There are very strong relationships between the prices of homes and the size of the mortgage market.  This relationship exists regardless of the cause of home prices, but there seems to be a strong bias to see the causality going from credit to prices.  When constrained supply is what causes prices to rise, we blame creditors.  Then, on top of supply deprivation, we impose demand deprivation.  This is what we did in housing.  Most people are supremely confident that we did the right thing, because the relationships are clear.  But since they have the causality wrong, they have no idea that public policies behind both supply and demand deprivation have been devastating to the American working class.  Hundreds of books and articles have been written about how strong those relationships are, with the causality wrong.  When I think about the mass of misidentification that is out there, taken by the public as evidence, I wonder if a response can possibly break through the fog.

J.W. Mason has a post up that seems to engage in a lot of causality reversal regarding the minimum wage. (HT: Matthew C. Klein)  Some of his causal inferences are probably conventional, so my reactions may be somewhat idiosyncratic, and maybe they are more of a comment on conventional economics than on Mason, but here they are nonetheless.
Let’s start at the beginning. Suppose there is some policy change, or some random event, that boosts desired spending in the economy. It could be more government spending, it could be lower interest rates, it could be a rise in exports.
Causality: Spending creates production?  This is, at best, a muddle.  But, the rest of the story depends on it.
Lower unemployment increases the bargaining power of workers, forcing employers to bid up nominal wages. These higher wages are passed on to prices, leading to higher inflation.
Causality: Labor utilization is inflationary?  Pro-cyclical monetary policy would lead to the same result, right?  So, now, Mason's story depends on two serial assumptions about causality - spending creates production and labor utilization creates inflation.  The simple Phillips Curve has broken down in recent decades.  But, real wage growth still correlates strongly with labor utilization - as does the Quits rate.  Might labor's "bargaining power" lead to better job matching and fewer frictions in the labor market?

See how causality does all the work here?  One interpretation says fewer frictions in the labor market has broad real benefits.  The other will say we should add frictions to the labor market.  Same evidence.  Different causality.  Diametric conclusions.

Mason does follow this with an extensive list of potential ways that an economy running near potential could generate higher real growth, many of which boil down to less friction and more efficient matching and utilization.

But, he follows with:
If we follow this a step further, we could even say that in the long run, the big problem isn’t that excessively high wages do lead to the substitution of capital for labor but that excessively low wages don’t. People like Arthur Lewis argue that it’s the low wages of poor countries that have led to low productivity there, and not vice versa; there’s a well-known argument that the reason the industrial revolution happened first in Britain and in China or India (or in Italy or France) is not that that the necessary technical innovations were present only in Britain. They were present many places; it was the uniquely high cost of British labor that made them profitable to adopt for production.
Causality: High wages lead to substitution of capital for labor?  It seems like surely the causation runs the other way - capital leads to higher wages.  It is only when higher wages are imposed legally that the causation would run the way Mason suggests, in a particular case.  Mason links to arguments for his causation, but surely all can agree that these are, at best, interesting contrarian arguments.  We should add another unlikely causation assumption to his series logical steps here.

This seems like a bit of selective observation, too.  I mean, I'm happy to argue that the idea that capital moves to places where wages are low is wrong.  But if high wages were the causal factor, I wouldn't need to argue about it.  Aren't most people complaining about capital moving to places with low wages?  It would be strange for that to be the common perception if high wages were what attracted capital.  Is capital attracted to where it can exploit weak labor or to where strong labor requires substitution?  It can't be both at the same time.

Isn't it more likely that universal legal protections for existing and new capital are the causal factor that leads to capital inflows and rising wages?

If a plethora of high wage options for labor is the cause of capital intensification, then I guess we don't have to worry about laborers dislocated because of capital induced declines in manufacturing or agricultural employment.  Capital was attracted to agriculture because ag labor had so many opportunities for high wages?  And, all that labor moved to the cities because of those high wages, and it was only after they moved there that capital was induced into the cities for the new manufacturing production?  So, there was some other cause of high wages in both the newly industrialized cities and in the depopulating rural areas, and capital infusion was a lagging result of the higher wages that had come from this outside cause?  And, now production is moving from the rust belt to the developing world because rust belt wages are too low and high wages in China are attracting capital as a substitute?  Are there people making these arguments?  Is there a literature on this that lurks in the shadows outside of the public conversation?  Are there Marxists out there who complain that CEOs are moving capital into developing economies because wages there are high?

Going back to the beginning of Mason's post, he seems to think that when McDonalds' CEO claims that the minimum wage leads them to replace workers with capital, leading to unemployment, that this is a problem for minimum wage critics, because higher productivity should lead to general expansion, mitigating the cost of unemployment.  Here, Mason has a great post showing how pro-MW arguments about efficiency wages and anti-MW arguments about rising productivity are both basically the same argument, selectively observed to ignore the mitigating side.

But, I think his general assumptions about causality lead him to a sort of false equivalence about this.  There is a difference between capital utilized in the face of a price floor and capital utilized because of other causes.  There is no reason to believe that the positive effects of a minimum wage will be strong enough to create employment for the workers who are unemployed in that efficicency wage story.  On the other hand, on the issue of capital-induced productivity, that capital had to be taken from somewhere.  Capital doesn't just pop up out of nowhere.  There are tradeoffs.  Capital wasn't deployed somewhere else so that it could be deployed to replace sub-MW workers.  There is no reason to believe that the net change to productivity will be positive.  And, especially given the central place of propensity to consume in the pro-MW arguments, where the income is supposedly transferred from employers who would be more likely to invest to workers who would be more likely to spend, there is no reason to expect this capital to just pop up out of nowhere.  Spending causes production, Mason told us.  How exactly did that capital show up if this whole ball started rolling because we induced more people to buy hamburgers instead of saving their incomes to build kiosks?

6 comments:

  1. Hi Kevin.

    Thanks for the response. I suspect our starting points are too far apart to reach agreement on much, but for what it's worth, I think you've got my argument right. The key point is, as you say, that spending causes production. We don't live in a world of scarcity, and in particular, we don't live in a world of scarce capital. The investment in labor-saving machinery at McDonald's, or in industry in a developing country, does not deplete a fixed pool of savings that would otherwise be invested elsewhere. I don't expect you to agree with either of those claims, I'm just confirming that they are my assumptions here.

    In general, I think the best approach is to be prepared to tell a variety of causal stories. In some contexts it makes sense to think of spending driving income, in other contexts it makes sense to think of income driving spending. Old-school development economists used to talk about "three gap" models of development, where investment might be held back by a lack of demand, or a lack of saving, or a lack of foreign exchange. Depending which of those constraints is binding in a given setting and over a given timeframe, it may or may not be right to say that capital deployed in a certain way is necessarily withdrawn from somewhere else.

    ReplyDelete
    Replies
    1. You're right. We're probably not going to agree, but your framework does present a nice exercise for me regarding priors and how these ideas come together.

      Thanks for the comment!

      Delete
  2. Well, I think I fall into the spending creates supply camp.


    The experience in Japan in the Great Depression with their brilliant central banker Takahashi Korekiyo certainly indicates that simply printing money to finance government outlays can work to beat a depression.

    WWII in the US comes to mind also.

    Sure, too much money can be printed.

    ReplyDelete
  3. Well, I think I fall into the spending creates supply camp.


    The experience in Japan in the Great Depression with their brilliant central banker Takahashi Korekiyo certainly indicates that simply printing money to finance government outlays can work to beat a depression.

    WWII in the US comes to mind also.

    Sure, too much money can be printed.

    ReplyDelete
    Replies
    1. Well, I don't necessarily disagree, but it seems like there is a difference between liquidity and real consumption that is important here.

      Delete
  4. On a macro level, our capital surplus indicates more foreign investment in the US than vice versa. Yet the public is focused on cases like Carrier when the reality is that there is more foreign investment in the US than vice versa. I wonder to what extent this misperception is magnified by our federal gov't debt (ie, some portion of the foreign investment in the US is just in debt instruments, not factories (like BMW in SC, etc))?

    ReplyDelete