There are a number of problems with Professor Charles Swenson’s paper, which was the subject of a July 15 press conference, that claims changes to California’s corporate income tax will result in job growth. The most glaring is the fact that Swenson compares a change adopted by California in February 2009, which allows multi-state corporations to choose between two formulas when allocating profits to California for tax purposes, to policies in other states that mandate the use of a formula based solely on a corporation’s share of sales. We’ve previously written about the impact of California’s adoption of “elective single sales factor” apportionment here. The new California law, which would take effect in 2011, lets businesses choose between a “sales only” formula and California’s current formula which allocates income for tax purposes using a formula based half on sales and one-quarter each on the corporation’s payroll and property in California. The Swenson study, he calls it a “natural experiment,” looks at five states – Georgia, Louisiana, New York, Oregon, and Wisconsin – that adopted a different policy: one that mandated a “sales only” formula. The difference is substantial. Proponents of the “sales only” approach traditionally argue that works by using both a carrot and a stick: the carrot is lower taxes for businesses with large shares of payroll and property in the state that export products outside the state, and a stick – higher taxes – for those that locate property and payroll outside the state and sell into California. California, however, adopted a “carrots only” approach that allows businesses to choose the approach that results in the lowest tax bill. This apples to hamburgers comparison of “elective” to “mandatory” single sales factor apportionment, alone makes the paper irrelevant to any discussion of the impact of the 2009 change in law or its potential repeal in Proposition 24.
But there’s more. Swenson’s state-by-state analysis finds that in each of the five states he examined, overall employment declined. Moreover, employment in each of the states examined declined more after the enactment of single sales factor apportionment than it did in the period immediately prior to the change in state policy.
It is also worth noting that Swenson uses a dataset – the National Establishment Time Series (NETS) – that he previously faulted PPIC researchers Jed Kolko and David Neumark for using in their study of the effectiveness of California’s enterprise zone program that came to a different conclusion that his own. On the merits, we’d argue that the PPIC paper is the stronger of the two, but regardless, Swenson can’t have it both ways: the data is either reliable or it isn’t. Swenson’s critique, which includes the statement that he has “very serious concerns with the data source,” is available here at 1 hour 22 minutes into the testimony.
Finally, yet importantly, as we noted earlier this week, if something seems too good to be true, it probably is. The Swenson paper doesn’t take into account the fact that lower corporate tax payments will require offsetting reductions in state spending and that means fewer jobs for teachers, homecare workers, and college professors. And, as we’ve noted before, that just because a study appears on the surface to have the elements of economic research does not mean it actually follows sound methodology and results in reasonable findings. We suspect that we haven’t seen the last of this one. May the buyer beware!
— Jean Ross