Paying For The Wall/Fence: Capital Controls Edition

Can we fund the partly see-through wall by taxing worker remittances?

Given that tax reform cannot really force Mexico specifically to finance the wall/fence [0], it pays to see what other options are available. One option mentioned by President Trump on the campaign was to impose restrictions on worker remittances back to Mexico.

Worker remittances to Mexico were $24.8 billion in 2015 (CRS (2016)) and $27 billion in 2016 [1]). Suppose we applied a 10% tax on such remittances; what would be the impact?

Figure 3 from CRS (2016), modified by author.

There are several reasons why remittance taxes are bad from a general efficiency and development context. However, in terms of the question at hand, here are two particularly salient concerns (from Mohopatra (2010)):

A remittance tax would also drive these money flows underground. A shift of flows to informal channels can hurt efforts to leverage remittances for increasing access of recipients to formal financial services (financial inclusion) and to raise financing for infrastructure and other development projects …

Such a tax is difficult to administer as remitters can resort to using informal channels. Also such a tax is highly regressive. And they produce huge deadweight losses as remittances are highly cost-elastic.

How much can a tax on remittances raise? In estimating the revenue gain, a key question surrounds the elasticity of formal sector remittances with respect to the tax. A low estimate of the elasticity of substitution between formal/informal remittances ranges from a low at 4, and a high at 8 (Light and Lewandowski (2015), pp.13-15). Note that this is not an overall remittances elasticity, just that for formal remittances that can be taxed. From Light and Lewandowski (2015):

The elasticity of substitution described here should not be confused with the price-elasticity of demand for remittances overall. The price-elasticity of demand for remittances is the change in total remittance demand compared to a unit change in price. Such an elasticity has been found to be relatively low, less than one. The elasticity of interest here, is the demand elasticity for (taxed) formal remittances only. This elasticity is much higher, because it includes both, the own-price elasticity of demand, plus the elasticity of substitution between alternative methods.