Mitchell’s Musings 4-3-17: Making Borderline Policy
Daniel J.B. Mitchell
Note: We resume our weekly musings with this edition. Our practice is to omit the winter quarter due to teaching obligations.
Much has happened since our last musing in late December. Some would say too much has happened. However, of particular note recently was the failure of Congressional Republicans to pass a “repeal and replace” bill for the Affordable Care Act. As numerous commentators have observed, the failure was due to an inability among House GOP members to decide on what should be the. That inability, combined with a seeming presidential indifference to the details of what the bill contained, doomed the effort. We can come back to the whys of that failure in a future musing. But, supposedly, the next big agenda item in Congress is to be tax “reform.” And, as in the case of health care, there seems to be no consensus among the Republic majority on what such reform should entail.
One version of reform, sometimes said to be under consideration and sometimes said to be off the table, is a so-called border adjustment tax. So let’s look at what such a tax might entail. “Might” is the right word, since there is no explicit proposal. The idea seems to be that a tax (tariff) would be imposed on imports of, say, 20%, and a symmetrical subsidy would be given to exports at the same 20% rate.
Note first that this tariff-subsidy arrangement bears some similarity to the (Warren) Buffett voucher plan about which I have written in the past. Under the Buffett plan (of the 1980s), exporters would be given a $1 voucher for each dollar of exports. The voucher would entitle the bearer to import $1 worth of goods and services. A voucher could be used directly by the recipient or sold in the open market. Under the Buffett plan, exports and imports would thus necessarily balance. The net effect on the combined actual exchange rate and market price of the voucher would be equivalent to the devaluation of the dollar needed to produce balanced trade.
Under the Buffett plan, the cost of the voucher to purchase imports is analogous to a tax (tariff) on imports and the value exporters obtain by selling the voucher is analogous to a subsidy of the same rate. Whether the border adjustment tax is being proposed – if it is – because of its similarity to the Buffett plan is unknown. But let’s point to an obvious feature of the Buffett plan. Although the government distributes the vouchers given to exporters, there is no revenue for the government from the plan. Essentially, there is a direct monetary transfer from importers to exporters.
If the tax version were enacted rather than the voucher scheme, and if it led to a zero trade balance, there would also be no net revenue for the government because the tax collected from importers would go entirely to the subsidy collected by exporters. If the purpose of a border adjustment tax is to improve the US trade balance and bring it to zero, the tax cannot be used to generate net revenue that would replace other taxes that the GOP wishes to eliminate or reduce. More generally, the more the border adjustment tax improves the trade balance, the less revenue it produces.
Another way to put this point is that if the purpose of the border adjustment tax is to improve the trade balance and thus create more jobs – especially in manufacturing – it can’t be a good revenue generator. If the purpose of the tax is to generate a lot of revenue to replace other taxes, then it can’t be a good job creator. In effect, the two goals, jobs and revenue, are in opposition.
There is a further wrinkle here. Some argue that if a border adjustment tax were imposed, it would actually not have any real effect because the dollar exchange rate would adjust to offset the impact. Such a conclusion might seem surprising. As noted, effectively what is being taxed NET is the trade imbalance. The tax only produces net revenue if the trade balance is negative (exports < imports). And we generally believe that if you tax something, you tend to get less of it. Getting less of the trade imbalance is a real effect, equivalent to the type of effect we associate with devaluations.
Currency devaluations “work” by changing (raising) the price of tradeable goods relative to non-tradeables. The ultimate non-tradeable is labor, so another way of looking at what is happening is that wage costs of production are lowered relative to world prices. The result is a greater incentive (due to greater profitability) for producing export goods and import-competing goods domestically. A border adjustment tax raises domestic import prices relative to labor costs due to the tariff effect. The domestic price of exportables rises relative to labor costs because exporters divert production toward world markets due to the subsidy.
So why would anyone think that a border adjustment tax would have no real effect? Note that in the case of a devaluation, there would be no real effect if domestic prices (including non-tradeables – essentially labor costs) rose to offset the devaluation effect. That is, if for every 1% the currency sank, domestic prices and wages rose by the same 1%, there would be no real effect. Those who argue that a border adjustment tax would have no real effect are assuming either that domestic prices rise to offset its devaluation-like effect or the currency rises in value in to offset the tax/subsidy impact. Were either of those offsets to occur, the trade imbalance would be unchanged and the tax would generate revenue without affecting (creating) jobs.
But why would someone think that there would be a total offset of the tax? The answer seems to be based on identities from the national income accounts. In the national income accounts, the trade balance (exports minus imports) equals by definition the gap between private and public saving and investment. So if the devaluation has no effect on saving behavior and no effect on investment behavior, then it can’t affect the trade balance. But the very same analysis tells you that if saving rises relative to investment as a result of the devaluation (or as a result of the imposition of a border adjustment tax), the trade balance must improve.
There are two main reasons why we would expect saving to rise relative to investment, whether from devaluation or a border adjustment tax. The first is the general property of money illusion and wage stickiness. Wages are unlikely to move upward on Day 1 of the imposition to offset the price effect. Put another way, real wages will decline either way.
The second reason is that we have a central bank, the Federal Reserve, which has as part of its mandatory objective function the maintenance of price stability. So if domestic prices of tradeable goods rise by, say, 20%, it is unlikely that the Fed would accommodate an offsetting 20% rise in wages. It would “tighten” monetary policy to prevent such a rise. Interest rates would rise, investment would be tamped down, and folks would consume less (save more). How smoothly such an adjustment would go would “depend.” But something like that scenario would occur, producing the real effect on the trade balance. Indeed, the less “smooth” the adjustment, the more likely you would see a quick improvement in the trade balance. There is nothing like a big recession to cause imports to drop relative to exports.
Of course, all of this analysis of a possible border adjustment tax is in the abstract. There may never be such a tax. It may never get beyond the talking stage. There is said to be an internal White House conflict between those who don’t want to monkey around with trade and those who do. So whatever tax proposals come along may never include the border adjustment tax, precisely because those who don’t want to do anything on the trade side know the tax would have a real effect.
 Economists might note that this contradiction of objectives is a case of Tinbergen’s law, which states that generally the number of policy instruments should match the number of policy goals. Here we have a case of one instrument and two goals.