خیزش پایه مالیاتی و اثرات آن بر توزیع درآمد

We quantitatively analyze the way inflation alters the inequality of the income distribution in the U.S. economy. The main mechanism emphasized in this paper is the “bracket creep” effect according to which inflation pushes income into higher tax brackets. Governments adjust the nominal income tax brackets slowly and incompletely due to the rise in prices. In the U.S. postwar history, this typically happens less often than once every other tax year. In the first part of the paper, we study time series from the U.S. economy. As our central result we find that irrespective of the level of inflation more frequent income tax schedule adjustments make the relationship between inflation and income inequality more transitory in nature. In the second part of the paper, we develop a general equilibrium monetary model with income heterogeneity that is in line with our time series evidence. We find that a longer duration between two successive adjustments of the tax schedule reduces employment, savings, and output.

The “bracket creep” defines a shift of personal income into a higher tax bracket when taxable income grows over time. It occurs due to inflation. Higher inflation possibly increases tax burdens under a progressive personal income tax as taxpayers near the top-end of a tax bracket are more likely to “creep” to a higher bracket. Clearly, this effect alters inequality in after-tax income. Whether it increases or decreases inequality depends, among others, on the level and duration of inflation, the top income tax rate, and the initial distribution of income. The purpose of this paper is to assess the impact of the “bracket creep”, or rather its attenuation through inflation indexation of the tax schedule, on the distribution of income both empirically and in a dynamic stochastic general equilibrium (DSGE) model for the U.S. economy. Like in the U.S. most personal income tax systems are progressive, i.e. structured with marginal tax rates exceeding average rates and increasing with the base. Taxpayers who receive only nominal increases in wages to offset higher inflation tend to be pushed into higher brackets. This effect is considered to be particularly severe (“the cruelest tax”) in times of high inflation as was seen during the last half of the 1970s when U.S. inflation rates averaged 8.9% annually (Blinder and Esaki, 1978). To combat bracket creep in the U.S. the Reagan Administration implemented an indexation of the personal exemptions and the tax brackets based on a cost-of-living index derived from the Consumer Price Index for All Urban Consumers (CPI-U). These provisions were actually enacted in 1981 as part of Economic Recovery Tax Act (ERTA), but delayed in their implementation and did not become effective until 1985; see Altig and Carlstrom, 1991, Altig and Carlstrom, 1993 and Auerbach and Feenberg, 2000. In inflationary environments, with unchanged or loosely adjusted rate schedules and brackets, tax collections tend to rise. This raises the claim that bracket creep is strategically used by some governments to maintain tax revenues. A loose or strategically implemented “pure one-year-lag” index system can be shown to cause taxable income to be overstated by the current rate of inflation (Altig and Carlstrom, 1991 and Altig and Carlstrom, 1993). Apart from its (mis-)use as revenue instrument, the omisson of inflation adjustment of marginal tax rates is also very likely to have a considerable effect on the distribution of income. Fig. 1 shows the annual time series of inflation and the Gini coefficient of market income (before taxes) for the period from 1948 to 2004. It highlights the relationship between the two series. Both series are coined by an upward trend up to the 1980s. While this trend continues for the Gini, the inflation rate calms down and follows a slight downward trend as of the early 1980s. Overall, the two series seem to comove – sometimes more, sometimes less in phase – at business cycle frequencies. A close inspection reveals that the series get more entrained after the mid-1980s, suggesting that indexation following ERTA has led to a more contemporaneous relationship. Full-size image (25 K) Fig. 1. Inflation rate and Gini coefficient, 1948–2004. Figure options These observations are in line with the empirical part of the present study which finds that inflation has a transitorily inequality reducing impact that leads aggregate measures by about two years. The central strategy of this part, however, is to take a stand on how the effective U.S. tax system was affected during the total postwar period and then to investigate the consequences of infrequent indexation relative to the sort of system that has been in place since the mid 1980s. Methodologically, we focus on a bivariate study of the correlation structure of the inflation rate and Gini coefficient series at business cycle frequencies in the spirit of Sims (1981).1 Our methods are primarily descriptive and as such imposing less assumptions than the more structural specifications used in the literature. Yet we seek to contribute to the literature by assessing whether the progressive bias of inflation is predominantly driven by the level and persistence of positive inflation or rather by an infrequent adjustment of the tax schedule. This task requires to go beyond descriptive time series analysis. To summarize our empirical results in Section 2, we find, using correlation analysis, that the relationship between the Gini coefficient and inflation rate dynamics got both more contemporaneous and statistically robust after introducing the indexation scheme. The former is confirmed by studying bivariate spectral measures. In the theoretical part of the paper, we develop a monetary DSGE model of progressive income taxation.2 In our simulations, we compare both high inflation environments (1970s) with moderate inflation environments (rest of postwar U.S. history) and infrequent schedule adjustment regimes (before ERTA) with less infrequent schedule adjustment regimes (after ERTA). In response to higher inflation or a longer duration of the bracket creep, individuals face higher income taxes, both on average and marginally. As a consequence, agents adjust their labor supply and savings decisions. Our results support the view of Altig and Carlstrom, 1991 and Altig and Carlstrom, 1993. Accordingly, the indexing scheme introduced by ERTA bounded the problem but issues of inflation and tax-system interactions are far from moot and being solved. Our results from the general equilibrium model exercise are as follows: The level of inflation has a rather small effect on income. In particular, the inflation elasticities of the Gini coefficients of wage and total after tax income amount to 0.01% and 0.15%, respectively. However, if we consider a tax policy regime that adjusts the tax schedule for inflation more frequently, we find that agents increase their labor supply by 6.3% and savings by about 4% compared to a system with less frequent adjustments. The implied Gini coefficient elasticities amount to approximately 0.2%. Hence, in this scenario inflation elasticity increased. This insight can be reconciled with our findings from the empirical part: A more significant and contemporaneous relationship between the Gini coefficient and inflation rate dynamics for a more frequent indexation (annual adjustment since 1985) implies a more immediate and concerted reaction of private households’ behavior to inflation. A more infrequent adjustment dilutes this direct reaction due to the possibility of spreading it. The former results in a higher, the latter in a lower elasticity as found in our model’s simulations. We conclude that the inflation rate exceeding some threshold (e.g. 5% or 10%) is less problematic for the effects of the bracket creep compared to the duration of creeping up brackets. The remainder of the paper is structured as follows. Section 2 presents empirical evidence for U.S. time series. Section 3 introduces the OLG model with two assets: money and equity. The model is calibrated with regard to the characteristics of the U.S. economy in Section 4. Our numerical results are presented in Section 5. Finally, Section 6 concludes.

“Bracket creep” has often been cited in the literature as one of the major distortionary effects of inflation. However, whether moderate levels of inflation also affect income-inequality through bracket creep has virtually not been analyzed since the 1970s. In a basic correlation analysis in the time domain we find for time series on income inequality and inflation in the U.S. that the relationship between the Gini coefficient dynamics and the inflation rate got more contemporaneous since the introduction of inflation indexation in 1985. Estimating bivariate spectral density measures additionally reveals that the average duration of association between the inflation rate and the Gini coefficient shrank from more than three years in the pre-1985 period to less than two years in the post-1984 period. Both our empirical and theoretical analysis suggest a progressive effect of inflation. In terms of size, it amounts to about one fifth to one third of the absolute value of the mean change in the Gini coefficient for one decade. The quantitative effects depend not only on the level of inflation but, in particular, also on the indexation system that is in place. In sum, we find that the duration of the bracket creep, i.e. the time period between two successive income tax schedule adjustments, is more important for equilibrium values of aggregate savings and average labor supply than the annual change in the tax rates due to bracket creep. A shorter duration of bracket creep results in higher equilibrium labor supply and output. In this sense, our results suggest the change in U.S. tax policy after 1985 and the inflation–indexation under ERTA to represent a successful change.