The Political Economy of Trade Policy Part 2: On the Presumption of Liberty

As with any model in any science, we need to ask the questions: “how well does this model reflect real-world observations? Are its assumptions likely to hold and are they key to the model?” Negative answers to these questions do not necessarily imply the model should be scrapped. All models, after all, are simplifications. Any model that could handle every possible variation would be unwieldly and thus not provide much insight. Further, as Harold Demsetz warned us, just because the real world differs from some theoretical outcome it does not mean that alternatives are necessarily better, especially when one situation is viewed through the lens of reality and the other through the lens of theory.

In the previous post, I discussed some of the political realities surrounding economic justifications for trade restrictions, primarily using the optimal tariff model as an example. Are these objections enough to recommend against using policy to try to influence patterns of trade, or am I simply making the Nirvana Fallacy?

It is certainly true that markets can fail (broadly defined as failed to achieve some optimal level or distribution) and these failures can be corrected through judicious government actions. However, these actions can cause more harm than they actually solve. Indeed, for something like an optimal tariff, even if done with the best of intentions, it can backfire and result in a much worse scenario without much effort. Further, these justifications can be misused or hijacked to give intellectual cover for essentially selfish goals.

However, the biggest issue with trade policy is mistakes can become systematic and entrenched. Aside from the reasons discussed above, most political systems, and democracies/republics in particular, are designed to be slow-moving. This means if a policy is determined to be detrimental, it may take a while to repeal or alter even if we assume no self-interest lobbying or other barriers preventing the legislation from being changed.

National trade policy also necessarily must be general. As such, it is likely to be geared toward the average person or firm. The policy may be too restrictive for some and too broad for others; it may lead to a rather substantial misallocation in resources. Consider the following example: Imagine a room with 10 people inside, five of whom are six feet tall and five of whom are five feet tall. The average height of the room would be five feet, six inches. If there is a policy to build a door in the room so people may come and go, how would the policy be structured? Maybe it is structured so that the door must be at least six feet tall, thus everyone can easily use it, but that’d mean less wall space for windows and pictures and other things, and for half the people it’d be too tall a door. Likewise, they could order the door be at least five feet, six inches (the average height), but then the tall people would find the door inconvenient to use while the short people would have no problem. The necessity of general rules is part of the reason why government action should be limited to negative rules (eg, do not steal) rather than positive rules (eg eat five servings of veggies a day).

We also need to consider the knowledge problem. The actual level of knowledge necessary to accomplish these optimal policies is both dispersed and not even necessarily consciously known to the people holding the knowledge. Acquiring both the necessarily knowledge and acquiring it in a timely fashion are impossible. What’s more, what statistical information we can gleen has some major caveats attached. Economic data is collected primarily though the use of surveys to a sample of individuals and firms and then extrapolated to the aggregate level. However, as with any survey, these surveys are subject to the same caveats, assumptions, and error terms as anything else; they may not truly represent the real economic activity they are trying to measure. Further, as additional statistical techniques are run on them, the error terms must get larger and larger, not to mention the potential for error from sampling issues, violations of various assumptions, and the like.

With all these potential for errors, which are not limited to just the political realm, we are faced with the question “how to contain damage from failures?” As mentioned above, government rules quickly become systematic and entrenched. This means that an error in government policy could quickly affect the entire society. If, for example, government bets that the Next Big Thing is going to be autonomous cars and pours subsidies into their development, and then people decide they don’t want autonomous cars for whatever reason, the American taxpayers are held holding the bag. If the various government-supported firms go under, the taxpayers will not be reimbursed, nor have anything to show for the spending of funds that could have gone toward education, health care, infrastructure, or other uses. The error would be felt (to varying degrees) nationwide. Conversely, if a private firm makes the same bet and are subsequently proven wrong, the ones who feel the loss most acutely are the firm itself: the owners, the shareholders, and those who did business with them. The losses would be generally confined to those individuals and firms. Because of these potential losses, these private firms are more likely to be careful with their spending and their projects than would government-sponsored entities.

It is from the expectation of failure and not success that I argue for the presumption of liberty in policy. With managed trade policy, while there is the potential for upside, there is a rather substantial risk of a large downside, too. Should that downside occur (a very probable event in my opinion), it’s effects would be systemic and hard to remove. Conversely, if private individuals and firms make errors, they would be the primary recipient of the downsides. Thus, I argue that free markets are more robust to error than government is. This is not an argument from Nirvana, but rather from Hell.

The Political Economy of Trade Policy (Part 1)

Writing at EconLog, Scott Sumner makes the following point in his excellent blog post entitled “Keynesian Fiscal Policy is Dead“:

Many non-economists do not understand fiscal policy; they view it as something that can be applied on a sort of ad hoc basis. But things don’t work that way, as Keynesian fiscal policy requires a countercyclical (full employment) budget deficit.  It’s a full-fledged policy regime that must be maintained over time, not a gesture to be employed at a point in time.  You can’t say “let’s do fiscal policy this year”.

Scott’s point can be expanded to include trade policy as well.  Trade policy is, likewise, not something that can be applied on an ad hoc basis.  Consistent and predictable rules are necessary to maintain trade.  What’s more, theoretical economic trade management (eg, an optimal tariff) requires an extreme level of precision and consistency.

But can we assume such consistency in policy?  I think not, especially in a democracy/republic.  If one of the parties does not buy into said policy regime, then maintaining it is virtually impossible, unless that party is deliberately kept out of office.  Or, if the two parties have differing points of view on the goal of trade policy.  If one group wants an optimal tariff and the other wants tariffs to match international taxation regimes (both, in theory, legitimate tariff goals), those two policies will be at odds with one another.  How the regime is established will depend on who is in power at any given time.

How likely is it, even with the same party in power, that the policy remains consistent?  That is not so clear.  Politicians, especially in a system where they face voter pressure, are subject to various whims; as Adam Smith puts it: “[statesmen] whose councils are directed by the momentary fluctuations of affairs (WN 468).”  Even if we assume consistency in policial parties, inconsistency in various affairs should give us pause when assuming consistency over time in policy goals.

But, even if we assume consistency in policy goal over time, we run into the issue of policy adjustment.  For example, an optimal tariff depends on how sensitive domestic consumers are to changes in price; if they are highly sensitive, then an optimal tariff would be extremely low.  If they are not very sensitive, then an optimal tariff could be relatively high.  But, this sensitivity adjusts over time (Second Law of Demand).  This would mean that the optimal tariff level would need to be adjusted periodically (ideally, constantly) to compensate for this changing sensitivity.

However, firms and individuals adjust to policies.  With any policy, we run into Gordon Tullock’s transitional gains trap.  Firms and individuals will fully capitalize the monopoly gains they get from policy protections, leading them to do no better than before the policy was initiated.  What this also means is that if the policy were to be changed or removed, these firms/individuals would face the potential for major losses.  It would be strongly in their interests to keep the policy from changing.  Furthermore, since these losses to the entrenched firms/individuals would be highly visible and concentrated, but the benefits dispersed among the entire society, it would be a relatively easy pitch for these entrenched interests to keep the policy in place.

Another point worth quickly mentioning in relation to the previous paragraph: most trade policy models assume no resources spent on lobbying.  If we relax that assumption and assume that firms/individuals do indeed lobby, either for or against various trade policies, then any potential gains from these policies are quickly eaten up in the costs of lobbying.  Further, because of the aforementioned dispersed costs and concentrated benefits, the entrenched interests would be willing to spend more on lobbying than the people harmed by the policy,* which could potentially cannibalize all potential gains from the policy and then some.

*This point may require some explanation.  Assume a society of 10 individuals.  A policy is proposed that could improve the wellbeing of one member by $200, but only through the cost of the rest of society by $300 (or $33.33 per person).  Therefore, the one member would be willing to spend up to $200 to lobby for this policy, but each individual may only be willing to spend up $33.33 to lobby against.

The Ambiguity of Purpose Leads to Abuse

The case for free trade is not absolute.  There are many reasons, theoretical and practical, where an exception to the general rule of free trade may be desirable.  Reasons of national defense, of terms of trade, of national welfare, can all be justifiably given for tariffs.  As I and others have discussed elsewhere, these exceptions have high hurdles to clear and rely on some rather strong assumptions, but they remain, at least in theory, justifiable.

But it is important to note with these different justifications is they are all mutually exclusive.  The national defense justification is a tariff high enough to prevent competition into the domestic industry.  The terms of trade argument, on the other hand, is a sufficiently low tariff designed to generate welfare gains and force other countries to lower their prices.  In other words, one tariff is designed to decrease imports and the other to increase imports.  These are, obviously, at cross purposes.

Even if these justifications are intellectually impeccable, they remain tactically useless for a simple reason: when scarcityism admits multiple justifications, any number of rent-seeking firms can exploit these various justifications for personal gains.  In other words, multiple conflicting justifications can be given by various firms, and all hope of improving national welfare goes out the window; the more exceptions carved out, the more rent-seeking individuals and firms will seek to exploit them.  As rent-seeking increases, more and more resources are devoted away from satisfying consumer wants and toward rent-seeking.

The presumption of liberty, that is a tendency to allow free trade vs controlled trade, remains strong even if it is, in theory, the option that has lower welfare because of the potential of rent-seekers to exploit ambiguity in a national policy to promote welfare.

Today’s Quote of the Day…

…is from pages 101-102 of Paul Krugman’s 1997 book Pop Internationalism:

There is no question that in many cases comparative advantgae arises from self-reinforcing external economies rather than as a result of underlying national resources.  In such cases international competition may exclude a country from an industry in which it could have established a comparative advantage, or drive a country from an industry in which comparative advantage could have been maintained.  In these cases, a [sic] intellectually respectible argument can be made for government policies to create or preserve advantage.

The fact that an argument is intellectualy respectible does not mean it is right.  Concerns over competitiveness that are valid in principle can be and have been misused or abused in practice.  Competitiveness is both a subtler and a more problematic issue than is generally understood.

Absolutely.  Some folks like to justify Trump’s economic policies based on obscure or particular economic arguments: optimal tariffs, or increasing terms of trade, or forcing other nations to lower tariffs and subsidies, or national defense.  All of these arguments are intellectually respectable (if not consistent with one another).  The internal logic of them holds.

But just because something is possible does not mean it is probable.  It is possible that a tariff could improve the well-being of a nation (subject to some key caveats).  But how probable is it that government could effectively create and enforce such a tariff and not face public choice concerns?  How probable is it that there would be no additional costs to the process, or that it won’t get hijacked for self-interested uses?

It’s trivially easy to come up with some theoretical reason why something can happen.  Translating that into reality, however, is a totally different beast.

The Transition Costs of Tariffs

The conversation around tariffs tends to focus primarily on gains and losses.  Domestic protected producers are said to gain because of higher prices and returns.  Domestic consumers are said to lose because of said higher prices; the consumers now need to expend more effort to acquire the same or a fewer number of goods.

But the standard “welfare analysis” story of tariffs is, while correct, incomplete.  We live in a world of transaction costs and people do not instantly and costlessly move from one state of being to the other; that is, people face “transition costs.”  These costs include: searching for new suppliers and all the transaction costs therefrom (if they’re buying the imported product), tax compliance (if they are the importers), enforcement costs (for customs agents), lobbying costs (if they are lobbying for/against protection), etc etc.  These costs are not directly captured in the textbook model of tariffs but are costs nonetheless.

The implication of including transition costs of tariffs into the analysis is it makes the argument for tariffs even weaker; the true costs of the tariffs are likely considerably higher than typically anticipated.  As resources are devoted to compliance (or avoidance) of the tariff, to finding new suppliers or paying higher bills, that means fewer resources devoted to “building a better mousetrap.”  It means fewer resources dedicated to expanding and hiring and innovating.  It means fewer resources being devoted to satisfying desires.  It means we should be extremely skeptical of anyone who claims tariffs will “make America great again.”

Along the same lines as this post, check out David Henderson’s recent EconLog post on the costs of the TSA.

Trade-Offs and Public Policy

This semester, I have been studying Law & Economics with Robin Hanson at GMU.  In class, we have been discussing the legal system, how it is structured, and other ways to structure it.  Questions we’ve pondered include: why can one appeal on matters of law and not matters of evidence?  Why are rules of evidence what they are?  Should all contracts be enforced or what limits should be placed on them?  Why are property taxes structured they way they are?  Why common law in the US as opposed to civil law?  Etc.

Simultaneously, I am evaluating a book for my course this summer: Trade-Offs by Harold Winter.  Trade-Offs is a public policy-focused look at economic reasoning.  In the book, he points out one of the dangers of public policy analysis (Page 5, original emphasis):

Even if there is agreement on the broad objective of maximizing social welfare, policy objectives may differ due to differences in the definition of social welfare.  A good example of this can be found in the economic analysis of crime.  To deter crime, we must use resources for the apprehension, conviction, and punishment od criminals.  But should the benefits that accrue to individuals who commit crime (also members of society) be added to social welfare?  If yes, this may suggest that fewer resources can be used to deter crime, because crime itself has offsetting benefits.  If not, crime is more costly to society, and more resources may be needed for deterrence.  Notice, however, that it is a fact that a criminal reaps a benefit from commiting a crime (or why commit the crime?), yet it is an opinion as to whether that benefit should be counted as social welfare.  Policy objectives and definitions of social welfare are subjectively determined.

What is also subjectively determined, as explained by Carl Dahlman in his 1979 Journal of Law & Economics article The Problem of Externality, is the effectiveness of the policy change proposed.  When a policy proposal is made, the proposer implicitly assumes that whatever institution he is invoking (government, market, etc) can necessarily solve the problem he’s subjectively identified better than the status quo (otherwise, why would he make such a proposal?).

All this subjectivity means that discussing “optimal” policy gets really tricky.  Optimal tariffs, Pigouvian taxes, optimal forms of law, legislation, etc are going to depend greatly on how we measure social welfare.  When discussing tariffs, should the welfare of foreign producers and consumers be counted?  If so, why?  If not, why not?  When discussing Pigouvian taxes, should the welfare of clean-up companies be taken into account (eg, the laundromat who loses business because fewer people are washing soot-caked clothes) and is government necessarily the best solution?  What makes sense given a certain accounting of social welfare doesn’t with a different accounting.

Answers to these questions can go a long way in helping us consider supposed market failures: whether something optimal or suboptimal will depend a lot on how these trade-offs and welfare are measured (to Winter’s point above, if the welfare of criminals is taken into account, there may be too much police activity.  If the welfare of criminals is not, there may be too little).  In this sense, optimality is in the eye of the beholder.

I’d argue that the subjective nature of social welfare policy suggests a strong presumption of liberty for people to choose their own way.  Indeed, there is no initial reason to believe any given action taken by an individual is somehow sub-optimal given the subjective nature of social welfare.  Even something like pollution is subject to these conditions.  This realization also should force economists (and their consumers) to ask the question “what are we assuming?” and “how are my biases affecting this analysis?”

Economists rarely argue about data.  It’s somewhat rare that someone made a math mistake or jumbled data (ideally, that gets caught long before publication).  Outcomes are not in question, but the subjectivity of trade-offs are.

Is Government Part of the Economy?

The question posed in the title of this post is key to understanding the relationship of governmental policies to individuals and the economy.

Traditionally, governments are considered as agents outside the system (in technical terms, exogenous).  For example, the following comes from Jack Hirshleifer’s 1970 textbook Investment, Interest, and Capital (page 11):

Following the standard tradition of economic analysis…government will be treated as an agency outside the social-exchange relationship, purportedly acting in the interests of society as a whole rather than the interests of the particular social grouping who happen to constitute the government.

A lot of welfare economics, from Pigou to modern, tend to treat government the way Hirshleifer describes. It’s an impartial spectator that can, with proper knowledge, set things right. Government can just come in and, with some perfectly crafted policies, fix any “market failure.”

But one of the things Public Choice teaches us (and this goes back to Bastiat) is that government is not separate from the system, but indeed part of it (in technical terms, government is endogenous). It is populated by the same people as those who make economic decisions. Governments are populated by people, and those people have hopes, dreams, desires, senses of right and wrong, just like the rest of us. They’re self-interested (which is not the same as selfish or self-centered), just like the rest of us. They respond to incentives, just like the rest of us.  Once we incorporate this simple fact, then the expectation of government “serving the public interest” becomes problematic.  Even if we assume away the knowledge problem, to expect government to be “molded from finer clay”, so to speak, and to be able to adjust the system from afar without any impact on the government itself is the height of foolishness.

So yes, in theory, with extremely strong assumptions, one can construct a tariff or tariff system “free” from cronyism. But if we weaken those assumptions, if we take government as endogenous (internal) rather than exogenous (external), then simple welfare economics like Pigou goes right out the window.