Taxes and Business Costs

As noted yesterday, Canadian advocates of corporate-tax cuts have proliferated alternative measures of corporate taxes. The C. D. Howe Institute’s “Tax Competitiveness Program,” which largely consists of papers by Jack Mintz and Duanjie Chen, has focused almost exclusively on marginal effective tax rates (METRs) on capital, expressed as percentages of pre-tax rates of return.

There have been a couple of attempts to measure overall production taxes by combining the conventional METR on capital with an METR on labour. The two marginal rates are weighted by the production function’s ratio of capital to labour, making business costs (rather than the rate of return to capital) the common denominator.

In 1998, the federal government’s Technical Committee on Business Taxation, chaired by Mintz with assistance from Chen, estimated taxes as percentages of marginal input costs excluding taxes. In a 2005 Canadian Public Policy article, Mintz and Chen published figures on net taxes (taxes minus subsidies) as percentages of marginal business costs including net taxes. Last week, the C. D. Howe Institute released a paper by the same two authors (plus Andrey Tarasov) examining taxes as percentages of marginal business costs including taxes.

There are several problems with METRs on capital as a measure of tax competitiveness. However, at least it makes intuitive sense that, other things being equal, a higher ratio of taxes to pre-tax profits is less competitive in attracting investment. However, the same is not true of the ratio of taxes to business costs. Last week’s paper presents the following example:

. . . suppose it costs $2.00 to produce a widget, inclusive of all taxes that apply to income, sales, payroll and capital. If taxes were eliminated, suppose the cost of production would fall by 50 cents. This implies that taxes make up a quarter of the cost of doing business (50 cents divided by $2.00).

Suppose it costs $1.50 to produce the widget in another jurisdiction, where taxes also amount to 50 cents per widget. Clearly, this latter jurisdiction will be more competitive in attracting widget producers, even though its “effective tax rate on the cost of doing business” is 33%, as compared to 25% in the first jurisdiction.

The authors get around this problem by assuming that firms locate production so as to equalize marginal tax-inclusive costs between jurisdictions. Tax rates then determine how much a jurisdiction can produce before reaching the equilibrium marginal cost. However, this assumption is obviously shaky in a world of lumpy investments and economies of scale.

A more policy-relevant scenario would be that the first jurisdiction imposes an additional levy of 25 cents per widget to finance a public good that reduces the pre-tax cost of production by 75 cents.  The after-tax cost would fall to $1.50 per widget (i.e. $2.00 + 25 cents – 75 cents).  This policy would clearly improve competitiveness, while increasing the marginal tax rate to 50% of business costs (i.e. 75 cents divided by $1.50). In other words, measuring taxes as a proportion of business costs seems inherently biased against using taxes to finance public investments, regardless of such a policy’s effect on competitiveness.

The 2005 article attempted to estimate taxes net of subsidies, but it is unclear how the effect of public programs (other than explicit subsidies) on business costs was measured. In any case, this attempt was abandoned in the most recent paper.