Tariffs and Enforcement
How can we understand the complexity of tariff regimes?
People really seem to hate complicated tax codes. The dislike for a complicated tax code seems sufficiently pervasive that politicians often feel the need to promise to reduce the complexity. Yet, for all this talk, things rarely get simplified. Why is that?
One common response to that question is that politics is messy. Some might argue that well-intentioned politicians simply find it too difficult to change things. When they get into office and try to change the tax code, special interests show up with promises in exchange for upholding their particular loophole. Others might contend that the politicians are simply engaged in cheap talk. They say what they need to say to get elected, but they have no intention of actually doing anything.
These explanations are sometimes correct. Politicians do engage in cheap talk. Politicians do succumb to special interests. However, these aren’t the only possible explanations. Those who have long read Economic Forces probably know by now that I’m willing to concede these explanations, but only after looking for an economic explanation. After all, durable policies and institutions call out for explanation. Why have they survived for so long? What purpose do they serve?
Consider a simple example. Suppose that a state has an army. The army is large and the state spares no expense in funding it. The army not only provides defense in war, but is a potential deterrent to war. If the army is a good deterrent, decades might go by without a war. The absence of war might convince some that the army is a waste of money. After all, they never fight. Of course, if those people succeed in getting rid of the army, they will discover its purpose.
Can we apply the same logic to policies and institutions? Can we apply the same logic to the tax code?
I think that the answer is often “yes.” Sometimes, even unpopular policies or those that seem arbitrary can be understood with proper economic reasoning. Today, I would like to discuss one such argument, drawing on some recent research on the U.S. experience with tariffs and how they relate to state capacity by Erik Madsen, Martin Rotemberg, Sharon Traiberman, and Shizhuo Wang.
Some Initial Background
In the early history of the United States, revenue was primarily generated by tariffs. One reason that the U.S. relied so heavily on tariffs and why many developing countries do today is the result of a lack of state capacity. This term, “state capacity,” refers to the government’s ability to collect taxes, provide public goods, and enforce the law and provide order. A country that lacks state capacity is one that does a poor job of these things.
It is not hard to understand why a lack of state capacity might lead to tariffs. Today, we think of the income tax as the important source of tax revenue. However, an income tax requires a lot of bureaucracy to collect that revenue effectively. To start, you need to know what a person’s income is in order to know how much to collect in taxes. You could ask people to self-report, but they have an incentive to lie. You must have a way of obtaining records, keeping records, collecting taxes, checking to make sure people have paid the proper amount, enforcing taxes, and punishing those who do not pay what they owe.
If you lack state capacity, you still need to raise revenue, but you might not be able to effectively collect income taxes. Countries that lack state capacity therefore tend to tax goods. And one common form of taxing goods is to tax goods when they enter a port — a tariff.
This point is well understood. But what is potentially also interesting is the relationship between state capacity and the types of tariffs that a country might use. As I have written about before, taxes on goods take one of two forms: a tax on value (ad valorem) or a tax per unit of the good (a “specific” tax). How might the use of these different types of tariffs vary with the development of state capacity?
Thinking About Tariffs from the Government’s Perspective
To think about the relationship between state capacity and the structure of tariffs, we must start with the government’s objective. We will assume that the government’s sole means of collecting tax revenue is tariffs. Furthermore, we will assume that the government wants to collect some fixed amount of revenue, net of collection costs, while also minimizing the market distortions that affect consumer behavior. These might seem like strong assumptions, but recall that we are trying to explain the historical structure of tariffs in the U.S.
If this is your objective, then it seems like the straightforward answer here would be to levy uniform ad valorem tariffs. Obviously, ad valorem tariffs affect the relative prices of imported and non-imported goods. However, a uniform ad valorem tariff would not create relative price distortions among imported goods. This is because a non-uniform ad valorem tariff would increase some prices by a larger percentage than others. The same is true of a specific tariff. Consider a tariff on a bottle of wine. More expensive (higher quality) bottles of wine would have a lower percentage tax than less expensive (lower quality) bottles, thus making expensive wine relatively cheaper after the tariff.
However, when one looks at the history of tariffs in the United States, one sees extensive use of specific tariffs. Furthermore, even among the goods subject to ad valorem tariffs, they are not uniform. One might simply blame that on the ignorance of policymakers, but before falling back on that explanation, one might ask whether there is some justification that relies on economic reasoning.
Verification Costs
Recall that the government’s objective is to maximize revenue, net of collection costs. Herein lies an important insight. If the government were to institute a uniform ad valorem tariff, importers would have an incentive to misrepresent the value of their cargo. Such misrepresentations are likely to vary by product. This makes the tariff costly to enforce. Customs officers have to inspect cargo and punish those caught in such deception. Specific tariffs are not subject to this undervaluation problem because these tariffs are a tax per unit imported. Since the government’s objective is to maximize revenue net of collection costs while also minimizing market distortions, it now faces tradeoffs on a number of margins. A specific tariff creates a relative price distortion relative to an ad valorem tariff. However, if the enforcement cost of the ad valorem tariff is sufficiently large, then the distortion of the specific tariff might be preferable to the revenue lost via the collection costs of the ad valorem tariff.
Of course, for states with low state capacity, there is a problem: you cannot possibly manage all these different margins. Thus, what Madsen, et al. predict is that as the U.S. developed state capacity, we should expect the tariff regime to become increasingly complex. In other words, separating goods by categories and adjusting tariff policy by category requires a significant amount of administrative competence. The complexity of the tariff regime is therefore a function of state capacity. They show that this is indeed the case. As the U.S. developed state capacity, the complexity of its tariff regime increased.
As a corollary to this result, we should also expect a more complex regime to correspond with greater use of specific tariffs for the reasons discussed above. Certain goods are likely to have high verification costs. A state with more advanced administrative capacity is therefore capable of not only levying different tariff rates on different categories of goods but also different types of tariffs. In other words, we should expect that specific tariffs are used more often as the tariff regime becomes more complex.
Furthermore, since specific tariffs create distortions, we would expect to see specific tariffs levied on relatively homogeneous goods (goods that do not vary much in terms of quality), but that are subject to value detection costs. The reason is that these types of specific tariffs minimize the market distortion and collection costs.
By the same logic, all else being equal, it should also be the case that specific tariffs are likely to be more common among cheap goods than expensive goods. The reason is straightforward. High-valued goods are likely to trade in much lower quantities than cheap goods. Lower quantities of trade make detection of deception and verification less costly. Thus, one is more likely to rely on ad valorem tariffs.
Finally, recall that the choice between types of tariffs isn’t the only relevant margin. The government could vary the ad valorem tariff rates as well. Their theory predicts that for the set of goods subject to ad valorem tariffs, cheaper goods are likely to face higher rates than more expensive goods. Like with the logic above, this is due to the fact that less expensive goods will tend to trade in higher quantities. Since the cost of verification is increasing in quantity, one should expect higher ad valorem tariffs on cheaper goods.
How do these particular predictions hold up when compared to U.S. history?
What they are able to show is that as the complexity of the tariff system increased (i.e., there were more categories of goods with different tariffs), specific tariffs became more common for the most narrowly defined categories of goods. This suggests that not only did the tariff regime get more complicated as state capacity grew, but also the use of specific tariffs grew because of the ability to exploit this more complicated regime.
In addition, they exploit variation produced by tariff legislation in the 19th century to examine the prediction that ad valorem tariffs are higher on low-valued goods. This is indeed what they find. A one standard deviation increase in the value of the import reduces the tariff rate by one percentage point.
They also use modern data to exploit the variation in administrative capacity across countries to test their predictions. Again, they confirm that countries with lower administrative capacity have fewer tariff categories and a lower share of specific tariffs. They also show that more homogeneous categories of goods are more likely to have specific tariffs.
A Quick Summary
As I alluded to at the beginning, when we observe longstanding policies that seem complex or obviously wrongheaded, it is easy to dismiss them as ignorant and in need of replacement. Before we are so quick to do so, we should consider whether there is an economic rationale for the policy. The authors make a compelling argument that the complex nature of the U.S.’s historical tariff regime — and the similar patterns that persist in the world today — are largely driven by revenue optimization given the constraint of state capacity. There is much to learn through the lens of price theory.