You’re a policymaker in a country where people buy widgets that are produced both at home and abroad. You can set (separate) excise tax rates on domestic production and imports. (The tax on imports is, of course, what we usually call a tariff.) What tax rates should you set?
The Economics 101 answer makes two assumptions:
1. You care only about the economic welfare of your citizens (and not at all about foreigners).
2. You can’t affect foreign prices (i.e. your country is a negligible portion of the world market for widgets). The fancy way to say this is that the supply of imports is perfectly elastic.
From these assumptions, it follows that both tax rates should be zero. In fact, we can relax assumption 1) and allow you to care as much as you want about the welfare of foreigners; the conclusion doesn’t change.
But suppose we relax these assumptions in a different way:
1A. You care about both the economic welfare of your citizens and (separately) about the tax revenue earned by your government. (I continue to assume, however, that you don’t care about foreigners.)
2A. The foreign supply curve might not be perfectly elastic. Contrary to the Economics 101 assumption, this gives you some market power that you might want to exploit. (I continue to assume, though, that you take the foreign supply curve as given. In particular, this means that your policies do not affect foreign tax rates, so I am assuming away things like retaliatory tariffs.)
Now what’s your best policy? I can’t answer that because you have two competing goals (economic welfare and tax revenue) and I don’t know how much weight you put on one versus the other. But surely if I can show you that Policy A delivers on both goals better than Policy B, you’ll want to reject Policy B. The existence of Policy A leads me to call Policy B inefficient, and surely you’ll want to reject any inefficient policy.
So which pairs of tax rates are efficient?
The answer turns out to depend on a) the ratio of the elasticities of (domestic) demand and domestic supply and b) the elasticity of foreign demand. For illustration, let’s assume that domestic supply and domestic demand are equally elastic. Then the following chart shows the efficient import tariff as a function of the domestic tax rate and the elasticity of import supply:
(Click to enlarge.)
The shaded boxes indicate situations where the efficient tariff is below the excise tax rate on domesticly produced widgets. This means you should either announce two different tax rates (e.g. an excise tax of 20% on domestic production and a tariff of 13% on imports) or announce a single sales tax with an offsetting import subsidy (e.g. a sales tax of 20% and an import subsidy of 7%). Obviously there are also plenty of non-shaded boxes where you’d want the tariff rate to be higher than the domestic tax rate.
I worked this out because I was curious where the dividing line would fall between the shaded and the unshaded. I’m wavering back and forth between whether I should be at least mildly surprised by the answer. A few more details are here.
The problem with Econ 101 is that it ignores such effects as, for example, domestic suppliers being forced to relocate their production to China where they are exposed to state-sponsored intellectual property theft.
When you think about it, the mere existence of a foreign cheap labor country such as China which does not respect intellectual property rights diminishes the value of US intellectual property even where it is not stolen.
If a US company chooses to keep its production in the US in order to avoid IP theft, while its competitors (which have less IP to worry about) start producing in China, the IP-heavy company is now at a production cost disadvantage, which lowers the value of its IP.
Is it a rounding thing or should that read less than or equal to? But as the zero box is unshaded it can’t be less than or equal to.
Harold: it is indeed a rounding thing.
A tariff is equivalent to a tax on the consumption of a good plus an equal rate subsidy on the production of the same good. Unless there is an inefficiency on the production side, I can’t see how it is ever efficient to levy a tariff except to change the terms of trade.
Neil: The case where you can’t change the terms of trade is the case where foreign supply is infinitely elastic, in which case the efficient tariff rate is always below the domestic tax rate. Does that conform to your intuition?
Steve,
Not really. Adding a positive tariff is equivalent to increasing the consumption tax rate and adding a subsidy for domestic production. The latter causes inefficiency between domestic and foreign production. Diamond and Mirrlees showed, I believe, that optimal taxation requires production efficiency.
I might be missing something in how you set up the problem.
Thought more about this. IIRC, D&M assumed 100% tax on economic rents (economic profit/producer surplus). I’m guessing you did not and that might account for the positive tariff and conjoined production inefficiency.
Neil: I’m not sure, but I think you might be overlooking the fact that I’ve defined an efficient tax policy to be one that puts you on the revenue/welfare frontier (i.e. one that maximizes welfare for a given level of govt revenue) as opposed to one that maximizes welfare without any constraints.
No, I didn’t overlook that. Maximizing welfare while collecting a given level of revenue defines optimal/efficient tax policy, so that is what I assumed you did.
Isn’t the less confusing way to ask this question to assume an equal excise tax on consumption regardless of source and then ask what is the optimal import/export (the same thing under the Lerner Theorem)?
The usual way to model this is to ignore excise/tariff income since for revenue purposes you’d want a progressive consumption tax not an excise or tariff. For developing countries with limited administrative capacity to tax then revenue considerations can become important, but then the ability of influence import or export price is limited, so you’re back to wanting a uniform excise tax (except for taxing externalities like net CO emissions) and no tariff/export subsidy
Thaomas (#9): That’s certainly equivalent; I’m not sure whether it’s any less (or more) confusing.
Thaomas (#11): The problem you’re posing seems a reasonable one, but it’s clearly different than the problem that I’m posing (and that you re-stated in comment #9), yes?
I though that there is an equivalence between equal excise tax + tariff/export tax, import subsidy/export subsidy. If not I misunderstood the problem as posed.