How to think about taxes
By Scott Sumner
Because tax reform is currently in the news, I thought it would be useful to describe what economists know, and don’t know, about taxes. I’ll start with what we know:
1. Legal tax incidence doesn’t matter. If a tax is imposed on a market, it makes no difference whether the buyers or sellers are legally obligated to pay the tax. Thus the payroll tax is legally split 50-50 between the employers and the employees, but the take home pay of workers would be exactly the same if the tax were paid 100% by employers, or 100% by employees. Ditto for a gas tax. If the original price of gasoline is $1, then a 25-cent gas tax might cause the price to rise to $1.24. If the gas station paid the tax, they’d just boost the price to $1.24. If the consumer had to pay the tax, the gas station would charge $0.99, and then add the 25-cent tax on at the pump (as sales taxes are collected.) This is one issue on which all economists agree, at least on the long run. (Due to sticky wages and prices, the short run outcome may be different.) If you find an economist that does not agree, I’ll reply with the “no true Scotsman” argument.
2. The true economic burden of a tax depends on the elasticity of supply and demand. In most industries, supply curves are very elastic, especially in the long run. Thus most of the burden of sales taxes probably falls on the consumer. If we take the 25-cent gas tax above, it would lead to slightly less driving. This would slightly depress world oil prices. This would slightly depress wholesale gasoline prices. In my example above, I assume the wholesale price of gasoline fell by 1 cent. If so, the 96% of the gas tax would be borne by consumers, and 4% by suppliers (probably oil producers, not gas retailers). This is also an issue on which economists agree.
Here’s an excellent graph from John Taylor’s textbook:
3. A more difficult question arises when we look at various types of income taxes. In that case, the long run elasticities are difficult to estimate. Economists do not have a good sense of how much of corporate taxes (or even top bracket personal income taxes) are borne by the rich, and how much are passed on to the general public via higher prices and/or lower wages. We also don’t know much about the Laffer Curve. No one really knows the maximum amount of revenue, per capita, that a developed country can raise via taxes.
4. We do know that subsidies are negative taxes, with effects that are the exact mirror image of taxes. Taxes raise revenue but reduce quantity of what’s taxed, whereas subsidies cost revenue but increase the quantity of what is being subsidized. Again, it makes no difference whether a subsidy is paid to the provider or the consumer.
5. We also know that an equal tax and subsidy exactly offset one another. Thus if gas stations must pay a 25-cent gas tax, but also receive a 25-cent subsidy for each gallon sold, it’s as if nothing happened. No effect at all. If we combine this result with point one above, we know for certain that a 25 cent gas tax on producers combined with a 25-cent subsidy to gas consumers has absolutely no effect on anyone. This is not controversial. However, the price at the gas pump would rise to $1.25 in the second case. After consumers got their 25-cent rebate from the government, they would still be paying $1.00 per gallon. So it would look different.
6. When we move to the realm of international trade, economists see exports as the way of paying for imports (money is just a veil, trade is actually all about barter.) Thus economists believe that a 25% tax on all imports, combined with an equal subsidy to all exports, would have zero effect, for reasons identical to the gas example above. But there is one complication. The exchange rate would rise by 25%, so that the net price paid by importers, and received by exporters, would not change at all.
7. In contrast, if there were only a tax on imports, or only a subsidy to exports, then trade would be distorted. The exchange rate would rise by less than 25%. Importantly, both sides of the trade equation are impacted by tariffs and subsidies, as exports are the way we pay for imports. If we import less, then we export less, at least in the long run. Thus a 25% tariff would appreciate the dollar by less that 25%, and both imports and exports would decline. Protectionism would hurt West Virginia coal, Iowa farmers and Seattle jet makers. And an export subsidy (like Ex/IM Bank) boosts both exports and imports, hurting firms like US Steel, which compete with imports.
Bob Murphy asked me to address three questions:
I think you might also clarify–are you saying the following? (Because it’s very counterintuitive.)
1) An import tax by itself will reduce the trade deficit.
2) An export subsidy by itself will reduce the trade deficit.
3) An import tax coupled with an export subsidy will not affect the trade deficit.
So far I’ve been ignoring trade deficits. When we run a trade deficit, it means we import stuff now in exchange for exporting stuff later. If it seems like we’ll never have to pay for the imports (as trade deficits seem to run on year after year) it’s because the accounting is flawed. To some extent we are paying for imports by earning big profits on overseas investments. That’s the whole “dark matter” debate, which is worth a blog post on its own. Or we might give them some of our property. But the key point is that we must give other countries something of real value (not just paper) for all the cars they send us, unless the other countries are essentially giving us Lexuses and BMWs as gifts.
Tariffs and subsidies don’t have major first order effects on the deficit. Tariffs reduce both imports and exports, and subsidies raise both imports and exports. To the extent they matter at all, it is due to the impact on national saving and investment. Thus tariffs might boost national saving, which would reduce the trade deficit, while subsidies might reduce national saving, which would increase the trade deficit. I say, “might” because there are many factors to take into account, including Ricardian equivalence. Most economists believe that Reagan’s expansionary fiscal policy boosted the US trade deficit. If so, then you’d expect Trump’s likely fiscal policies to do the same. But of course it also depends on what’s happening in the rest of the world, not just the US. To answer Bob’s three questions: yes (a little bit), no, yes.
8. Let me end up on a point where I’m not well informed. Although in theory the proposed border adjustment tax/subsidy should be completely neutral to trade, there are some real world complexities that I don’t fully understand. Suppose part of our goods imports are paid for by UK tourists at Disney World. That service export probably won’t be subsidized. Also suppose part of our goods imports are paid via high overseas profits earned by US multinationals. Is that going to be subsidized? What about sales of LA homes to Chinese buyers? In other words, if the actual plan differs from the textbook assumptions of 100% tax and subsidy, do we still get a 25% appreciation in the dollar? And if we do, is it possible that it occurs not through changes in the nominal exchange rate, but rather the real exchange rate. Thus because of countries like Hong Kong that stubbornly peg their currency to the dollar (nominally), might the dollar only appreciate by 12.5% in nominal terms? In that case, the real adjustment would have to occur via a combination of higher than normal inflation in the US, and deflation in places like Hong Kong, or indeed much of the world. Central banks play a key role here.
Even with all that uncertainty, it’s important to know that economists do understand an awful a lot about taxes. I see many commenters who seem unaware of even points 1, 2, 4, 5, 7, which are all extremely well understood.
PS. If you think that a subsidy of 20% on sales of LA homes to Chinese buyers would be more controversial than the subsidy on goods exports, you are probably right. Which shows just how irrational we are when it comes to trade. We claim to love it when American blue collar workers build jets and bulldozers and sell them to the Chinese, but freak out when America blue collar workers build high rise condos in LA and sell them to Chinese buyers.