I then compare these reforms to a reform that shifts the tax base from income to consumption. In this case, it is difficult to maintain the level of progressivity that exists under the current income tax although ways exist by which the regressivity of the reform could be blunted. Whether the long term growth gains from consumption tax reform would offset the initial increase in regressivity remains to be determined. In addition to a wholesale replacement of the income tax with a consumption tax, I also consider a partial shift.
Should we raise environmental taxes? This question has been asked increasingly in the face of widespread environmental problems including global warming, air and water pollution, and a host of other environmental problems that we face today. A “textbook” answer to this question would be yes, we certainly should raise taxes to the point where the tax equals the marginal social damage from pollution. Unfortunately the real world is more complicated than a textbook world. Measurement problems abound: how do we measure marginal social damages? Concerns about economic efficiency intrude given the widespread prevalence of other taxes. Finally, distributional concerns come into play. Environmental taxes tend to be regressive: poor people pay a disproportionate share of their income in these taxes relative to rich people.
This paper attributes an association between insurers’ premium-to-loss ratios and state taxes to the allocation of premiums and losses by multistate insurers. Insurers likely find this form of income shifting efficient because it is largely unimpeded by stringent state guidelines or enforcement. To our knowledge, this study is the first to document income shifting using firm-level data, collected from the publicly-available accounting reports used to compute tax bases and filed with multiple governments. Although documentation of income shifting is not unique, the quality of the data and the ability to specify how income shifting is accomplished distinguish this paper from prior income shifting studies.
The findings in this study are robust to a battery of additional specification checks. One, some states permit insurers to credit guaranty fund assessments against premium taxes. In those states, the statutory tax rate likely overstates the premium tax burden, and insurers have an incentive to overstate premiums. To assess the effect of guaranty fund assessments on the findings, the regression model was reestimated excluding insurers from six states (Arizona, Florida, Georgia, Hawaii, Louisiana, and Texas) that account for 82 percent of all assessments in the U.S.19 Inferences are qualitatively unaltered.
If the tax management hypothesized in this paper exists, i.e., tax-motivated cross-state allocations of premiums and losses in the statutory reports filed with state officials, then the coefficients on the tax variables should not be negative if the tests are conducted using a set of insurers that file with only one state. Forced to report all premiums and losses from all underwriting in a single state, these insurers could not avail themselves of a tax management strategy that requires multistate filings. Column F reports the results from estimating the regression equation using 180 insurers who only underwrite in their state of incorporation and are unaffiliated with any insurers in other states. Electronic Payday Loans Online
Columns D and E report the results from two tests of the robustness of the results to different specifications. Conclusions are qualitatively unaltered by the results in these two regressions. Column D uses a separate intercept for each insurer instead of insurer-specific controls (DIRWRITE, SIZE, CAPITAL, PORTRISK, and lines of business). The tax measures remain negative and highly significant. The coefficient on ETR is -4.07 with a t-statistic of -3.07. The coefficient on INCDUM is -0.063 with a t-statistic of -3.73.
With one exception, the signs of the other regression coefficient estimates are as predicted. The coefficients on RATEREG and RESTRICT are negative, consistent with both prices being suppressed and profits being underreported in states with more restrictive product sales regulation. The coefficient on NOFAULT is negative, consistent with prior studies indicating lower prices in states with no-fault automotive insurance. INSPOP’s coefficient is negative, consistent with prices falling in the presence of increased competition. The coefficient on DIRWRITE also is negative, confirming prior findings that insurers who directly sell their policies offer lower prices. The coefficient on SIZE also is negative, suggesting larger insurers offer lower prices. CAPITAL’s coefficient is positive, consistent with less capitalized firms having to accept lower prices in response to buyers’ concerns about their insolvency.
Column B shows the coefficients when EASYETR is included as a regressor. Recall EASYETR is intended to identify insurers that underwrite primarily in lines with the most opportunity for avoiding taxes through cross-state allocation of multistate policies. Sixteen percent of the sample involves insurer-states with a positive value for EASYETR, i.e., write greater than 75 percent of their business in lines other than automobile, farm, fire, home and workers’ compensation.
These estimates assume the formation of implicit taxes exerts no upward pressure on prices when taxes are increased. To estimate implicit tax effects, let A be the after-tax return all insurer-states face, let P be the pre-tax return for insurer-state i, and j be the tax rate in insurer-state i. In all insurer-states, A = Pi (1-Ji) and *P / = P / (1-j). Evaluating P at the mean value for PRICE (1.61) and j at the mean value for ETR (0.015), the derivative is 1.63. Thus, a one standard deviation increase in ETR (0.005) would be expected to increase PRICE by 0.008 (1.63*0.005).
The regression results in Table 3 present evidence that is consistent with property-casualty insurers managing their accounting reports to mitigate tax costs.13 In the tests conducted using multistate insurers (columns A-E), the regression coefficient estimates on ETR, PREMRATE and INCRATE are always negative, as predicted, and significant with t-statistics ranging from -1.90 to -4.41.