Contrary to Intuition, a Tariff Would Indeed Make Mexico Pay

For a whole 24 hours, it looked like Donald Trump might be considering paying for the border wall by imposing a 20% tariff on Mexican imports. In the face of near-universal outrage, he seems to have backed off of that idea, although he claims to still be considering it, and I believe him. It’s nice that he’s at least reconsidering the policy, because it would be monumentally stupid. But, while the policy would be monumentally stupid, it would indeed succeed in making Mexico pay for the wall, at least in theory.

If that statement took you by surprise, you’re not alone. After the policy was revealed by Trump’s Press Secretary, social media was abuzz with furor over the fact that Trump’s attempt to make Mexico pay for the wall would raise the prices of the goods that Americans buy. How could that be? How could taxing American product purchases result in Mexico paying for the wall? Clearly, Trump is just a liar (in fairness, he is) and a cheat (in fairness, he is), and he’s secretly trying to stick Americans with the bill while claiming that Mexico will pay.

Except that Mexico would pay. The exact reasons why are more than a little counter-intuitive, but they are very much sound. The answer, as it often does, comes from economics. I’m going to make an attempt to explain why, but I should start with a few caveats:
1) This explanation is still more than a little simplistic. It isn’t really designed to be a full in-depth review of tariffs and international trade, but rather a high-level overview of how who “pays” a tax on paper and who really pays a tax aren’t necessarily the same party. But I’m ignoring a bunch of stuff, like the quirks of the specific markets in which Mexican businesses operate, and the possibility of repercussions (like a trade war), and playing a little fast and loose with my math. That’s not really fair, but helps make the fundamental point clearer without changing the result.
2) No part of this analysis should be construed as support for a 20% tariff on Mexico. Just because it would succeed at making Mexico pay for the wall doesn’t make it any less stupid as a policy, for reasons that I will broach below, but that are worthy of a much longer essay than this.

To begin, let’s look at the way that money changes hands in the status quo today. In the general case, whenever Americans buy a product from a market in which Mexican businesses compete, they could buy it from a Mexican business, or a comparable US business. In equilibrium, they will do both, spending a total of X dollars on Mexican products and Y dollars on US products. Then, the US Government collects some tax from US businesses, but not from Mexican ones(1). Let’s call this case “the baseline,” and it looks like this:

What happens today

Now, let’s introduce the effect of the tax. If Mexican businesses did nothing in response to the tax, all Mexican exports to America would immediately get 20% more expensive. Consumers, responding to the increased price, would buy fewer Mexican goods. So, we would reasonably expect Mexican businesses to reduce their base prices by some amount to compensate. It’s hard to say by how much they would reduce prices exactly, but we can safely say it will be somewhere between 20% (which absorbs the full cost of the tariff) and 0% (not at all). Let’s call these Case 1 and Case 2 respectively, and they would look like this:

In case 1, the prices that consumers see would be identical to the baseline, so we would expect consumers to buy the same relative amounts of each good — in other words, X and Y stay the same. But now, rather than a full X dollars flowing to Mexican businesses, only .8*X dollars would, and .2*X dollars would be intercepted by the US Government to help pay for the wall. By laying the flows in Case 1 on top of the flows from the baseline, we can get see the “net flow,” which is one way of thinking about the impact of moving from one case to the other. For Case 1, net flow looks like this:

Expressed like that, it’s very easy to see how Mexican businesses wind up paying for the wall.

But it’s pretty unlikely that Mexican businesses would absorb the full 20% tariff right out of their profit (in many cases, they probably didn’t even have 20% profit to start, anyway). So, let’s look at the other extreme, Case 2. Here, the prices of Mexican goods rise relative to the baseline, so consumers’ buying habits are going to shift. They will buy strictly more US products relative to the baseline, because they are relatively cheaper, and strictly fewer Mexican ones(2). Another point worth mentioning is that in Case 2, the average price of goods across the market rises, so we would expect consumers to buy fewer products total. Case 2 is a little more complicated than Case 1, but it is still very possible to make a net flow graph, which looks like this:

Here, we see that the US government is still extracting money from Mexican businesses, but there’s a side effect: US consumers leak money in both directions — to US businesses and the US Government. Consumers are undoubtedly worse off in this case, but if you look at the American side of the house put together, collectively, there’s no loss, because those dollars are kept in America.

And that, in a nutshell, is why a 20% tariff would indeed make Mexico pay for the wall. We may not know what the final outcome would be, but it must fall somewhere in between cases 1 and 2, and any way you slice it, the US government would extract the necessary money from Mexican businesses over some length of time. The only variable is the degree to which the policy would cause money to leak from US consumers to US businesses and the Government, but Trump likely does not consider that a loss — just a campaign promise kept.

Why the policy is still stupid:
Tariffs in general are a bad idea. They are a created inefficiency. In effect, the government is making it 20% more expensive (technically, more than 20%) to produce Mexican goods, and the taken in the aggregate, the market is never better overall when the costs of production rise. We can see that effect manifest in the relative values of A + B and X + Y. When the average price of goods rise, people buy less overall, resulting in fewer transactions, and in any elastic market, this reduces the total value of transactions and lowers the GDP.

(1) The government does actually collect taxes on Mexican products, like sales taxes. Because those taxes are uniform on both sides of the house, it doesn’t change the fundamental dynamics at play here, so I’ve neglected it. Feel free to object, though.
(2) I’m doing a Bad Thing here. The graphs were supposed to represent flows of money, but we can’t actually predict if the flow of money will decrease just because prices rise. There will be fewer transactions, sure, but they will each be more valuable than before. I’m choosing to operate under the assumption that the market is elastic in these cases because a) I think that’s most likely true given Mexico’s major exports, and b) even if it were not true, it would just make the math more complicated without changing the final result. Once again, feel free to object.