Friday, December 2, 2016

An Economist's Cautionary Note on Free Trade

Among most economists (among whom I count myself as one), free trade is a pretty strongly favoured policy.

The reasons for this are fairly good, and fairly straightforward, in the standard case for free trade.

Under the standard theory, the main basis for the benefits of free trade is comparative advantage. If Australia is relatively more efficient in producing iron ore (that is, if it has a comparative advantage in iron ore), and China is more efficient at producing manufactured goods, then at the country level both Australia and China are better off if Australia specialises in iron ore, China specialises in manufactured goods, and the two countries trade with each other. Then both countries are able to obtain more consumption of each good than they would alone, given whatever initial resources they have. This is an economic benefit, understood since David Ricardo wrote about it in 1817.

If one thinks of the economic units in terms of countries, free trade between China and Australia is Pareto improving. Both countries are made better off, and no one (in this limited model) is made worse off. This is the Holy Grail of economic policy. The optimal level of tariffs is thus zero, as restrictions on free trade harm both countries.

But if one thinks instead at the level of individuals within a country, then free trade is no longer Pareto improving relative to tariffs. In the example above, if I'm a worker in an Australian manufacturing firm which was previously protected by tariffs, and these get eliminated, then I really do get screwed. It's not just complaining - as my firm goes broke, I lose my job, and the previous skills I have are no longer economically useful in my country. Even if I get another job, I likely will have a lower future wage for quite a while, if not permanently.

The steel workers in Ohio complaining about free trade aren't just making it up. Things really did get a lot crappier when tariffs were eliminated.

But economics has an answer here. Free trade isn't Pareto improving, but it is Kaldor Hicks improving. In other words, the total gains to the economy are sufficiently large that the beneficiaries could organise a transfer payment to those who lost their jobs which would made the Ohio steel workers also better off. 

As a matter of political economy, this transfer doesn't actually happen. You'd have to pay the losers from free trade a very large sum of money if they have to transition to years of unemployment, or a permanently lower future income. 

Of course, this isn't really a problem of economics, more just politics. Is it the economist's fault that his prescriptions don't get followed?

So much for the standard theory. It's actually pretty good, as far as it goes. Like good economic proofs, it flows from assumptions to conclusions. If it's wrong, it's because there's something in the model that's being left out, or one of the assumptions is questionable.

There are a number of possible extensions one can make, like depreciating human capital. But to me, it's the base assumptions that are the most interesting. What are they?

We have the following:

1. Consumption is a good. You're better off consuming more goods and services than fewer goods and services, all else equal.

2a. Leisure is a good, or equivalently

or 

2b. Work is a bad.

In other words, for any given level of consumption, you'd rather work less than work more. 

These are not terrible assumptions. #1 seems probably true. You may hit a point of satiation with consumption, but over most ranges of wealth that people operate on, having more stuff beats having less stuff, unless the stuff poses other costs (like screwing up your children, in which case all else isn't equal).

But what about #2?

Going from a 14 hour work day to a 10 hour work day, with the same wages and consumption, is surely an improvement in welfare.

Going from a 10 hour work day to a 6 hour work day, with the same wages and consumption, is also almost surely an improvement in welfare. 

But the big question is the following: is it still an improvement in welfare to go from a 6 hour work day, to a zero hour work day in perpetuity?

In other words, if your consumption stayed exactly the same, would you prefer to have some sort of job, or no job, ever?

You may think work sucks, but be careful what you wish for.

What if it turns out that people actually need some sense of purpose, some reason to get up in the morning?

Admittedly, having a job isn't always a fun purpose. But it's a structure, and a discipline, and a set of people you can interact with, and a routine that, if it works well, results in the satisfaction of providing for yourself.

What would life look like if you had basic consumption needs provided for you, no strings attached, without any need to work?

Well, as it turns out, we have many decades of data on that question. They're on display in a housing estate or ghetto near you. And the results ain't pretty. Ask Theodore Dalrymple, who wrote about this extensively

Every few months, doctors from countries like the Philippines and India arrive fresh from the airport to work for a year's stint at my hospital. It is fascinating to observe their evolving response to British squalor.
At the start, they are uniformly enthusiastic about the care that we unsparingly and unhesitatingly give to everyone, regardless of economic status. For a couple of weeks, they think this all represents the acme of civilization, especially when they recall the horrors at home. Poverty—as they know it— has been abolished.
Before very long, though, they start to feel a vague unease. A Filipina doctor, for example, asked me why so few people seemed grateful for what was done for them. What prompted her question was an addict who, having collapsed from an accidental overdose of heroin, was brought to our hospital. He required intensive care to revive him, with doctors and nurses tending him all night. His first words to the doctor when he suddenly regained consciousness were, "Get me a fucking roll-up" (a hand-rolled cigarette). His imperious rudeness didn't arise from mere confusion: he continued to treat the staff as if they had kidnapped him and held him in the hospital against his will to perform experiments upon him. "Get me the fuck out of here!" 
My doctors from Bombay, Madras, or Manila observe this kind of conduct open- mouthed. At first they assume that the cases they see are a statistical quirk, a kind of sampling error, and that given time they will encounter a better, more representative cross section of the population. Gradually, however, it dawns upon them that what they have seen is representative. When every benefit received is a right, there is no place for good manners, let alone for gratitude.
By the end of three months my doctors have, without exception, reversed their original opinion that the welfare state, as exemplified by England, represents the acme of civilization. On the contrary, they see it now as creating a miasma of subsidized apathy that blights the lives of its supposed beneficiaries. They come to realize that a system of welfare that makes no moral judgments in allocating economic rewards promotes antisocial egotism. The spiritual impoverishment of the population seems to them worse than anything they have ever known in their own countries. And what they see is all the worse, of course, because it should be so much better. The wealth that enables everyone effortlessly to have enough food should be liberating, not imprisoning. Instead, it has created a large caste of people for whom life is, in effect, a limbo in which they have nothing to hope for and nothing to fear, nothing to gain and nothing to lose. It is a life emptied of meaning.
"On the whole," said one Filipino doctor to me, "life is preferable in the slums of Manila." He said it without any illusions as to the quality of life in Manila.

I skipped the most striking descriptions of the problem, because if I started, I'd end up quoting the whole thing. Read it all, if you haven't before.

A question, dear reader.

Do you think the problems of the people described above stem from a lack of consumption? They don't have to do any work, so in a standard model, the only problem left is that they must be consuming too little.

Suppose that Dalrymple is describing his subjects in the above article honestly, and you have two policy choices to consider for the above recipients.

Option A - Increase their welfare payments by 50%

Option B - Find them a not unpleasant job for 6 hours per day, and require them to do honest work in order to receive the same welfare payments as before.

Which of these two policies would result in a larger improvement in human welfare for such people?

In the standard model, the answer is obvious. Given our assumptions, Option A is far preferable. Do you believe that?

Would it change your mind to find out that lower class whites in America (especially in rust belt parts of the US that have been worst hit by job losses from free trade) in recent decades have been so despondent that their life expectancy has actually been dropping as they kill themselves with alcohol, opiates, and suicide?

The standard answer to this is that we have an opiate problem. And a drinking problem. These are "substance abuse" issues. But why now? Alcohol was always there. Why is it only now that people decide there is no other purpose or hope in their lives, and start drinking themselves to death?

To turn these concerns back into the language of economics, the Holmes conjecture is that if leisure is not always a good, and work is not always a bad, then it is no longer obvious that the optimal level of tariffs is zero.

Sometimes, you might prefer to have some restrictions on trade in order to keep jobs in America.

But you have to be honest about why you're doing this.

Targeted tariffs won't raise consumption. They won't spur economic growth. They will lead to more expensive goods, and less consumption. David Ricardo was right on all that. Comparative advantage still exists, and be very wary of anyone who talks about free trade without acknowledging this.

But they might also lead to more employment. And this may well be worth it in terms of the quantity that the economist's social planner is meant to care about, namely total welfare.

It might lead to fewer rust belt whites killing themselves with opiates, because their communities are totally hollowed out with everybody sitting around on welfare without any purpose in their lives.

If steel products cost slightly more as a result, personally that doesn't strike me as the end of the world.

Of course, this is a cautionary note, not a case for tariffs-a-go-go. To say that the optimal level is not zero does not imply that the optimal levels is high, or across-the-board. And it's also not clear that tariffs versus free trade is the only solution to this, or even the best one.

I personally think that automation is a much bigger worry in this regard than free trade. I have similar questions about automation, which also doesn't strike me as everywhere and always welfare improving.

These aren't straightforward questions. If you ban the automobile, we get stuck with horses and carts forever.

And yet... and yet...

The Deaton and Case finding seems to me to be one of the most important findings in social science in recent years, and portends an enormous and growing problem. There are lots of workers who simply do not seem to be economically useful anymore, and in communities where lots of these people have ended up on welfare as a result of the endless grind towards replacement by robots, life is purposeless and miserable.

There are many other purposes that can be fostered - community, charity, art, religion, family.

But until we have a handle on how to solve the torrent of lives being sucked into the abyss of misery, as large as the AIDS epidemic, I remain open to a range of different policies in response.

13 comments:

  1. Convincing. It worries me, however. What I am worried about is that solving the problem with protectionism and technology bans leads to stagnation, i.e. a nation that found a different way to foster a purpose in its subjects will overcome the first one. But, there is one big but... instituting bans and tariffs is easy, whereas rebuilding culture, community, reviving religion, etc. is hard.

    ReplyDelete
    Replies
    1. I actually share your unease, especially if you start thinking about the problem in a dynamic setting. There is a real worry that anyone that doesn't choose the economically optimal path is liable to get outspent, and perhaps ultimately out-muscled, by those that do. It has a bit of the flavor of the argument about Moloch that Scott Alexander put forward - something can be awful, and still selected for.

      The Dalrymple point though suggests something a little harder - it's not just that there has to be something to do, but there also has to be some risk of failure or negative results which tie actions to meaningful consequences. Work has a struggle aspect to that, which not every other form of "something to do" has in quite the same way. Community, for instance, is important, but I don't know that it fills quite the same role.

      That said, if all those other factors hadn't declined so much, I don't think unemployment alone would be so crushing. Your last point is spot on - to an economist, fixing the other things is preferable, but probably quite a bit harder.

      Delete
    2. I don't think rise in productivity would be problematic if Industrial Revolution happened in say 12th century social setting. To be more precise, I think the main problem is that people lack hierarchy and guidance on the one hand, and have too much individual freedom on the other.

      Sports, arts, hobbies, and certainly life challenges itself (diseases, childrearing, etc.) seem to provide risks, but less and less people seem to be doing all that stuff these days. It is in this area that I think ban on technology would be even more useful than in case of automation. Entertainment industry, in addition to being great contributor to alienation, acts analogously to addictive substances...

      Born and raised under socialism I lived in what is very much a different, in every way, setting than America. As a consequence of socialism we have even less remains of community and religion. Many problems in my country are very different than problems in the USA, some are the same, or similar, like everywhere these days. Now, even thought average income in my country is eleven times smaller than that of the USA, I think I've managed to recognize some pathologies of modern man that are harder to notice in the West, namely insatiable appetites and lack of contentment no matter how good you have it. People here often complain for not having German standard of living, and I always answer them that Germany is not the default mode of human existence, and that therefore they should not compare their standard of living to that of Germans, but to that of hunter-gatherers. I think that this problem exists even in Western countries. People in the US may not compare themselves to Germans, but I could bet that many people are miserable simply because they can't afford something that isn't a life necessity. Recognizing the fact that we are living in incredible luxury, even in what is today considered poor countries, compared to the default mode of human existence (there's no need to even go that far, we can just compare to our grandparents), I think goes a long way toward learning gratitude, and thereby gaining some inner peace. Unfortunately culture today doesn't teach gratitude and humility, but entitlement and victimhood.

      Delete
  2. You are also omitting a few key points of the standard economic analysis:

    1. Your explanation of Pareto's argument is unconvincing to anyone not already familiar with it.
    I recommend David Friedman's explanation of foreign trade as a form of technology:

    There are two ways we can produce automobiles. We can build them in Detroit or we can grow them in Iowa. Everyone knows how we build automobiles. To grow automobiles, we first grow the raw material from which they are made – wheat. We put the wheat on ships and send the ships out into the Pacific. They come back with Hondas on them.

    From our standpoint, growing Hondas is just as much a form of production – using American farmworkers instead of American autoworkers – as building them. What happens on the other side of the Pacific is irrelevant; the effect would be just the same for us if there really were a gigantic machine sitting somewhere between Hawaii and Japan turning wheat into automobiles. Tariffs are indeed a way of protecting American workers – from other American workers.


    2. More importantly, standard economic theory also explains why displaced workers are unlikely to benefit from tariffs (even if we assume they can benefit from them).
    Political favors will go to concentrated interest groups because they can be counted on returning a favor. Displaced workers are a dispersed interest group.
    Is there a political model where these incentives are reversed?

    I agree that some important variables are not considered in the standard economic analysis. Are you offering an alternative framework?

    ReplyDelete
    Replies
    1. Only the sketch of a framework, alas. But if you want to know what it would look like more formally, write down a model with multiple inputs, one of which is labor. All inputs go into good production. Utility is positively increasing with consumption of goods, but also with utilisation of labor (the "people have being idle" hypothesis). I'm pretty sure the result outlined would be obtained from a formal proof - you may choose to tilt away from consumption only and towards more labor utilisation, depending on relative utility from more goods versus more labor at certain points of the curve.

      As for the other stuff, I totally agree - I got lazy and presupposed a general understanding of free trade arguments, just to avoid having a bazillion word post.

      Delete
  3. like_wut from redditDecember 22, 2016 at 3:10 AM

    But this raises the another problem. Economists generally believe in a broadly liberal-libertarian assumption that the satisfaction of your preferences is good i.e. giving you what you want is good, and generally forced to do something you don't want is bad. But this is obviously not that case. Most people surely don't want to work, because if they wanted, then all the welfare clients would simply donate time to charity, there are always these unpaid volunteering options, and yet the data is they are not really doing it much. So it seems being forced to work, obviously not by a whip but by economic necessity, may be good for us. This means, preference does not equal what is good for us.

    I think economics currently cannot deal with this. On a common sense level we do understand, yeah if you prefer hard drugs, it is still not good for you, but there are no models for this, no framework. Economics is strongly preference based.

    This is a deep problem. Obviously we cannot go the other extreme and say if philosopher-kings would completley ignore individual preferences and force everybody to live a healthy and virtuous life they would be happy. But it seems, happiness sometimes correlates with preference satisfaction and sometimes it does not.

    We can have opinions about it. But how to have anything like a science about that?

    ReplyDelete
    Replies
    1. It's a good question. Perhaps there's some hope for formalisation in Behavioral Econ, if we can understand the specific form of bias that's generating this - hyberbolic discounting makes people prefer slacking to work in the short term, even though they'd like to work eventually. I think there's something to the fact that stated preference does seem to be for work - people in distressed areas frequently complain about the lack of jobs. If you don't even have stated preference on your side, you have to be a lot bolder to start implementing mandated policy requirements.

      But I agree, it's a big challenge. I think this is why I tend to rely on big picture warning signs like suicide and drug addiction - it's pretty hard to claim that people all just suddenly got a preference for suicide, and more likely that something went deeply wrong somewhere along the way.

      Delete
  4. The problem is that economics is also wrong about comparative advantage and blind to absolute advantage. As Ian Flrtcher notes , pretty much none of the preconditions for comparative advantage are true:
    No externalities
    Nations trade only goods and services, not debt and assets
    Factors of production are domestically mobile
    Factors of production are not internationally mobile
    Long-term growth is caused by short-term efficiency
    There are no economies of scale
    There is no cross-border investment
    "In the 1950s, when these assumptions were much closer to reality, free trade may have been a winning move for America, but those days are long gone. It is free traders, not protectionists, who are living in the past and sticking their heads in the sand."

    There is a wider problem of economists pretending that incredibly crude models are reality, and the pretense always seems to be most psychotic when the interests of globalists and financial parasites are threatened by people noticing that the end effect of this supposed economic growth for nearly everybody is being left without any reason to live: no work, no family, no community, no nation, no country -- atomized consumer krill for the Wal-Marts of the world. Economics is fundamentally about selling that, pretending that it's not just inevitable, but good, shilling for the system that is doing it, like priests of a cynical imperial religion pacifying subject peoples. The conclusions have always come first in economics, with the rationalizations following after, the preferred conclusions have always been those that suit the entities that are most likely to pay economists with money, influence, and prestige.

    ReplyDelete
    Replies
    1. Interesting. I agree that economists sometimes tend to take their models too literally. That said, having somewhat drunk the economics Kool-Aid, I think there is a value to the formal modeling. Formal models help make sure that the "then" really does follow from the "if", as a professor of mine once put it. Sometimes, intuition gets complicated, and maths helps you keep track of whether all the results will necessarily hold. In other words, the ideal response is not to just point out that the assumptions don't hold, but to show how the results change under the alternative assumptions. Some parts, like the environmental externalities, result in a pretty simple fix - just impose a pollution tax on all goods (foreign and domestic) in proportion to the pollution created producing them. Other ones, like the drivers of long-term growth, are more complicated.

      In other words, the response to bad models should be better models, not no models at all. Though in the end, all models are simplifications of reality, not reality itself.

      Delete
  5. Free trade and autarky are not the only two options. You know this, but the fact gets confused during public debate among nonspecialists.

    Though not an economist, I grasp how Heckscher, Ohlin and Samuelson analytically find untariffed trade to be optimal. I also grasp how tariffs imperfectly substitute (according to Heckscher and friends) for the income tax. I can quantitatively distinguish a second-order effect from a first-order effect, so am not wholly unequipped to form a knowledgeable layman's opinion in such matters. I even appreciate Gomory and Baumol (the first half of whose book I have read) while finding their anti-Heckscher-Ohlin-Samuelson conclusions ultimately unconvincing.

    And yet I still want Donald Trump's 35 and 45 percent tariffs.

    The science of economics seems to have missed something important when it comes to international trade. I don't know what it is, but something.

    ReplyDelete
    Replies
    1. I think when you get to a conclusion that seems odd or unsatisfying, what's valuable is at least trying to figure out which assumptions you find questionable. I think a great example of this is the Mankiw and Weinzierl paper on the optimal taxation of height. They're not saying they advocate it, but that our standard assumptions mean it's optimal.

      Delete
  6. Amusing. Since you have taken the trouble to link the paper, I have read it half an hour (that is, I have not tried to trace out all the paper's mathematics, nor to penetrate every point such as the paper's elasticity of labor supply; but otherwise I have read the whole paper).

    Having read, I wish to take one of Mankiew's and Weinzierl's sentences deliberately out of context, from p. 11: "In practice, however, such Pareto improvements are so small as to be uninteresting."

    That's sort of how I feel about the Heckscher-Ohlin-Samuelson optimum: its Pareto improvements are so small as to be uninteresting. Indeed, my private suspicion is that Heckscher's Pareto improvements are so small as not practically to exist.

    The smallness of the improvements Heckscher, Ohlin and Samuelson theoretically predict has not been enough remarked upon during policymaking, in my view. You've got to jack up tariff rates pretty high, like 60 or 70 percent as I recall, before Heckscher really starts delivering.

    Taxation of height! I'll remember that.

    ReplyDelete
    Replies
    1. (I seem to have broken the thread. This was not deliberate.)

      Delete