Percent -10 J-08 J-10 S-09 N-09. Percent J-08 J-10 S-09 N-09

Transcription

1 Tax Reforms and Intertemporal Shifting of Wage Income: Evidence from Danish ly Payroll Records Claus Thustrup Kreiner Søren Leth-Petersen Peer Ebbesen Skov Online Appendix

2 1 A. [P8,P9) 1 B. [P9,P9) S-8 N-8 S-9 N-9 S-8 N-8 S-9 N-9 3 C. [P9,P99) 3 D. [P99,+] S-8 N-8 S-9 N-9 FIGURE A1: SHARE OF INCOME SHIFTED BY INCOME PERCENTILES Notes : The figure shows the difference between wages in a given month compared to the same month in 28 (as a percentage of the average monthly wage in 28) for the T group and measured relative to the C group. In Panel A, the T-group includes employees with average monthly earnings within the 8th and 9th percentile of the wage distribution, cf. Figure 2. The T-group selection moves upwards in the wage distribution as we move from Panel A to Panel B and forward. The construction of the T group and the C group is described in the note to Figure S-8 N-8 S-9 N-9 Source: The monthly payroll (eincome) register from the Danish tax authority (SKAT).

3 6 A. Shifting indicator by standard treatment status C-group T-group 6 B. Shifting indicator by treatment status and information C-group T-group informed F-9 A-9 A-9 S-9 O-9 N-9 D-9 F-1 M-1 A-1 M-1 A-1 S-1 O-1 N-1 D-1 J-11 FIGURE A2: SHARE OF INCOME SHIFTERS AMONG SURVEY RESPONDENTS Notes: The left panel replicates the graph in Panel A of Figure 4 but includes only the 88 employees in the treatment group and the 29 individuals in the control group who participated in the survey described in Table 4. The right panel is constructed in the same way as the left panel but the T-group is confined to those respondents who are informed about the opportunity to shift income, defined as the group with Q1 = 'After New Year' & Q2 = 'Legal' (see Table 4). F-9 A-9 A-9 S-9 O-9 N-9 D-9 F-1 M-1 A-1 M-1 A-1 S-1 O-1 N-1 D-1 J-11 Source: The monthly payroll (eincome) register from the Danish tax authority (SKAT) and telephone survey information obtained by Capacent Epinion.

4 TABLE A1 INCLUDING COVARIATES IN ESTIMATION OF THE ELASTICITY OF TAXABLE INCOME All months All months Only D9 & J1 Excl. D9 & J1 Excl. N9, D9 & J1 Income group All individuals Non-shifters All individuals All individuals All individuals (1) (2) (3) (4) () Full sample.8 (.7,.1).3 (.2,.).8 (.76,.83).2 (.,.3). (-.1,.2) income P8.1 (-.1,.3). (-.2,.2).18 (.12,.23) -.1 (-.3,.2) -.1 (-.4,.2) P8 income < P9. (.3,.7).2 (.,.4).48 (.43,.2).1 (-.1,.3). (-.2,.2) P9 income < P9.11 (.9,.13). (.3,.8).91 (.8,.97).4 (.1,.6).2 (-.1,.4) P9 income < P99.16 (.13,.18).6 (.4,.8) 1. (1.42,1.7).3 (.1,.6).1 (-.2,.4) P99 income.26 (.2,.32).1 (.4,.1) 3.24 (2.93,3.) -.1 (-.7,.) -.7 (-.13,-.1) Notes: This table of ETI estimates is identical to Table with the only exception that we have included a large number of control variables in the regressions thereby allowing for differences in income growth based on observables. The table reports the ETI estimate of each regression and the 9% confidence intervals on these estimates in the brackets. All reported estimates include the following covariates: age dummies, gender dummy, marriage dummy, regional dummies, 1 sector dummies (see Table 2) and 4 company size dummies (see Figure 7). Source : The monthly payroll (eincome) register from the Danish tax authority (SKAT) and administrative registers at Statistics Denmark.

5 TABLE A2 ESTIMATES OF THE ELASTICITY OF TAXABLE INCOME: SENSITIVITY ANALYSES All months All months Only D9 & J1 Excl. D9 & J1 Excl. N9, D9 & J1 Income group All individuals Non-shifters All individuals All individuals All individuals (1) (2) (3) (4) () A. Baseline.1 (.8,.11). (.3,.6).86 (.82,.89).3 (.1,.4).1 (-.,.3) B. Wide C-group.13 (.12,.14).8 (.6,.9).92 (.89,.9).6 (.,.7).4 (.3,.) C. Narrow C-group.8 (.7,.9).3 (.2,.4).81 (.77,.86).1 (-.,.3) -. (-.2,.1) D. Baseline year 29.9 (.7,.1).4 (.3,.) 1.8 (1.4,1.11) -.1 (-.2,.1) -.3 (-.4,.2) E. Doughnut sample.12 (.11,.14).6 (.,.8) 1.1 (.97,1.).4 (.3,.6).3 (.1,.4) F. Post reform year (.6,.1). (.3,.7).46 (.42,.49). (.3,.6).3 (.1,.) Notes: The table reports ETI estimates using specification () and 9% confidence intervals in brackets. The first row repeats the baseline estimates from Table. The baseline construction of the T-group and C-group is described in the note to Figure 3. The rows labeled 'narrow' and 'wide' refer to C-group selections with an average monthly 28 earnings level in the range 32,-3, and 24,-3,, respectively. The row labeled '29 income' refers to estimates where 29 is used as the baseline year to separate individuals into treatment group and control group, and where income differences are measured relative to 29 income. The 'doughnut' sample excludes individuals with an average monthly 28 earnings level in the range 33,-37,. Finally, the row 'Post reform year 211' refers to estimates based on post reform income observations from 211 instead of from 21 used in the baseline. Source: The monthly payroll (eincome) register from the Danish tax authority (SKAT).