The destination-based cash flow tax: A negative-revenue tax?

Our first interest in this tax, at the time it was under consideration in the US, was whether it would impose taxes on trade. An examination in one of the few peer-reviewed articles on this tax (Martin 2018) led us to conclude that it did not, but raised questions about its ability to perform the basic function of a tax, to raise revenue. Given the revenue concern, it seems important to carefully evaluate this proposal before implementation because adoption by one major country, as strenuously advocated by Martin Wolf (2019), would create perhaps irresistible pressures for other countries to adopt it as well. 

As shown by Auerbach (2017), this ingenious proposal combines features of a value-added tax (VAT) and a wage subsidy. A destination-based cash flow tax (DBCFT) mimics a standard VAT through its destination principle – exports are untaxed while domestic sales are taxed, and imports are refused a tax deduction. If we consider, for specificity, the 20% rate proposed in the US House Ways and Means Committee (2016), an importer bringing in a good valued at one dollar must sell it for $1.25 to break even. A trader able to sell a good for $1 on the export market must obtain $1.25 on the domestic market to get the same net return. As long as all domestic sales are covered, consumer prices must rise 25% above world prices, just as under a conventional invoice-based VAT. 

The DBCFT proposal follows current origin-based corporate income taxes (CIT) by allowing a deduction for wage costs – a feature that would raise wages to employees by 25% relative to wage costs to employers. 

Calculating the revenue yield of such a tax seems to require the same procedure as for a VAT. Thus, we would expect the tax base for the commodity tax part of the proposal to be private final consumption. No revenues would be raised from investment, because these expenditures would be deductible just like expenditures on intermediate inputs. Nor would any net revenues be raised on sales of goods to government. While these sales would generate tax receipts, this revenue would be completely offset by the increased costs to the government for the goods and services it purchases. 

One study for the US (Patel and McClelland 2017) and one for a wide range of countries (Hebous et al. 2019) conclude that a DBCFT would generate substantial revenues. Many DBCFT proponents argue its introduction would avoid raising consumer prices by inducing a nominal exchange rate appreciation. This is clearly possible. But an appreciation influences only nominal prices, not real outcomes such as tax revenues relative to GDP. 

Perhaps it doesn’t matter whether the government raises revenue from increases in the prices of the goods and services it buys? If private final consumption were large enough relative to wages, there would always be enough revenue to make such a tax revenue-positive. Simulations presented by Martin (2018) suggested that net revenue from a DBCFT in the US would be small, volatile, and vulnerable to turning negative. But perhaps other countries are in a more favourable situation, and a DBCFT would be a robust revenue earner there? 

To address this question, we begin by using the two key macroeconomic aggregates – private final consumption and wages – to generate a base estimate of the impact of a 20% DBCFT on government revenues. Next, we added the revenue losses associated with abolition of current origin-based corporate taxes. Finally, we examined net revenues when it is not possible to tax the financial services excluded in Hebous et al. (2019) and the services of owner-occupied dwellings excluded in Benzell et al. (2017) to find the net revenue impact of introducing a tax with these exceptions. The changes in net government revenues from implementing the tax under these three assumptions are presented in Figure 1 for ten key countries. Data for the full set of 37 countries with consistent wage share data are presented in Table A1 in the Appendix below.

Figure 1 The impact of a DBCFT on government revenues (% of GDP)

The results in the figure and the table seem to raise serious concerns about the revenue impacts of this proposal. As shown in column 1, the consumption share exceeds the wage share in only around a third of the countries considered, and the average net revenue loss from its introduction is 0.7% of GDP. Once allowance is made for the need to replace the revenues from existing corporate income taxes, net revenue is positive in less than a quarter of the countries. Assuming it is not possible to collect revenues from the financial sector and services of owner-occupied dwellings takes the average net revenue loss to over 5% of GDP.  In this case, more than 20 countries suffer revenue losses exceeding 5% of GDP. Only Lithuania and Mexico, with high consumption shares, low wage shares and low current reliance on corporate taxes, have net revenue gains.

Why is this revenue shortfall so large? How could this tax, based on returns by firms that look similar to current corporate tax returns, generate so much less revenue. The key reason is that destination-based taxes such as a VAT do not raise net revenue from sales to government, while conventional origin-based corporate taxes do. With government final demand accounting for 19.5% of GDP in our sample, counting revenues on these sales, as in the recent IMF Study (Hebous et al. 2019) would have created the impression of enough revenue to allow abolition of corporate income taxes, which raise on average 3.7% of GDP. 

These results suggest to us that there are serious problems with this proposal, especially for countries where the direct government revenue implications are strongly negative. The revenue losses are large enough that making them up through reductions in the tax deduction for wages, as in Benzell et al. (2017), would lead to enormous uncertainty about who gained or lost. Combining a large revenue earner such as a VAT with large tax-expenditure commitments such as a wage subsidy at the same rate seems likely to create net revenue losses, especially since introducing a DBCFT requires abolishing current corporate income taxes. Clearly, there are serious problems with current corporate tax systems, but it seems appropriate to consider alternative approaches (as in IMF 2019) or modifications to DBCFT proposals to avoid introducing taxes that create large, hidden net revenue losses.

Appendix

Table A1 Revenue impacts of introducing a 20% DBCFT (% of GDP)

Notes: All data in the table are for 2011 because this year provides extensive coverage and allows the data to be matched to Version 9 of the Global Trade Analysis Project (GTAP) database (Aguiar et al. 2016). The data on the share of private final consumption in GDP (NE.CON.PRVT.ZS) were downloaded from the World Bank’s World Development Indicatorson 29 March 2019. These include both private final consumption and services provided by Non-profit Institutions Serving Households (NPISHs). The assumption that services provided by NPISHs is a strong one because around half of the services provided by NPISHs are received without payment and hence unlikely to enter the net for a DBCFT (Salamon et al. 2012). Data on the adjusted share of wages were from the ALCD2 series on adjusted wage costs as a share of GDP at factor cost from the EC AMECO database. These include an estimate for the return to labour of workers in unincorporated enterprises for which the US House Ways and Means Committee (2016, p23) envisaged allowing a deduction. They also include social charges, which are allowed as a deduction against current Corporate Income Taxes. The data on corporate income tax receipts as a share of GDP were downloaded on 2 April 2019 from data.imf.org, with missing values for Canada and New Zealand obtained from the OECD database. The data on the share of GDP accounted for by the services of Owner-Occupied dwellings and non-fee based financial services provided to households were obtained from the GTAP 9 database (Aguiar et al 2016).