BHI Policy Study Executive Summary

An Economic Model of Massachusetts Tax Policy

(September 1994)

Executive Summary

A proposed graduated income tax (GIT) for Massachusetts would cause a loss, in 1995, of 78,000 jobs, of $1.231 billion in labor income, and of $87.448 million in state tax revenue. This would happen as a result of the way in which the GIT reduces the economic reward for taking jobs and for creating job opportunities in Massachusetts.

The GIT would reduce the take-home pay that a Massachusetts worker gets to keep, on average, by taking an extra or temporary job, by working extra hours, by becoming a family's second-income earner, or by working and living in Massachusetts rather than some other state. By thus reducing work incentives, the GIT would increase the cost of labor to Massachusetts employers. The result is the above-estimated reduction in Massachusetts jobs, labor income and tax revenue.

The November 1994 election ballot will contain two measures that relate to the proposed GIT. The first of these, Question 6, would amend the state constitution to require the adoption of a graduated income tax. The second, Question 7, would implement Question 6 with a law that scraps the existing "flat tax" of 5.95 percent on earned income and of 12 percent on certain unearned income. The new law applies a single rate to all income, rising from 5.5 percent to 9.8 percent, depending on tax bracket and filing status.

An often-cited report, published in October 1993 by the Massachusetts Department of Revenue, contains a thorough and persuasive discussion of the importance in tax analysis of accounting for the behavioral effects of tax-law changes. Economic models that account for such effects (so-called "dynamic" models) make it possible (1) to determine how a tax-law change would affect economic indicators like jobs and earnings and (2) to determine how the tax-law change would affect tax revenues in light of its effects on these indicators.

Economists are sometimes forced by data limitations and other problems to ignore behavioral effects in analyzing tax-law changes. When doing so, they might offer a judgment of what change, if any, will take place in tax revenue under a new tax law, given the admittedly unrealistic assumption that the new law has no effect on taxpayer behavior.

Economic models based on an assumption of this kind (so-called "static" models) are necessarily limited in value. While they can be reliable for very limited changes in tax law, where the behavioral effects are likely to be small, they are not reliable for sweeping tax-law changes like the GIT, where the behavioral effects are likely to be large. Because they assume away behavioral effects, models of this kind are unable to determine what effects the new law would have on economic indicators like jobs and labor income.

The Department of Revenue Report was required, as it happens, to rely on a static model in performing its analysis. Using this model, it concluded that the GIT would cause tax revenue to rise by only $44 million, less than .5 percent of state tax revenue. It further concluded that "the economic impact due to the aggregate increase in tax liability of this magnitude is expected to be quite modest." Proponents of the GIT have interpreted these conclusions to mean that the GIT is revenue "neutral" and economically harmless.

Before, however, one can safely accept such an interpretation, it is worth taking another look at the GIT and the possibility of estimating its behavioral effects. In economics, where one attempt to estimate such effects might fail, another might succeed. And, anyway, there are dozens of highly-regarded studies showing that state tax increases have negative effects on employment. Why not look a little more closely at this issue before automatically accepting the reassurances of grad tax proponents?

That is what we set out to do in this study. In taking up this challenge, we assembled historical data on key Massachusetts economic indicators such as employment, capital, labor and capital income, and federal and state tax rates. We then developed a dynamic econometric tax model for the purpose of determining the effects of the GIT on Massachusetts employment, labor income and tax revenue.

Two key findings emerge from our analysis: (1) Massachusetts employment is significantly and negatively related to the state "average marginal tax rate" (AMTR) on labor income (the average state tax that Massachusetts workers pay on an additional dollar of labor income) and (2) the GIT raises the AMTR. Our predictions concerning job and tax-revenue losses are based on a statistical determination of how changes in the AMTR affect jobs and how the GIT would change the AMTR. It turns out, as we would expect, that, when examined in the context of a dynamic model, the GIT has significant negative effects on economic activity.

A copy of this study can be purchased by e-mailing the institute at BHI or by calling (617) 573-8750.