The Effect of the Change in the income Tax Deduction System on the Labor Supply in Korea

Transcription

1 The Effect of the Change in the income Tax Deduction System on the Labor Supply in Korea Heonjae Song, Woori Shin Department of Economics, University of Seoul November 10, 2017

2 Intro: 2013 Tax Law Amendment Change to the tax credit from the income tax deduction This is an important tax revision that is expected to have a significant impact on the tax burdens of low-income workers and high-income workers. There has been no study of the impact of these changes on the labor supply of workers in Korea.

3 Intro: Effect of system change on the tax burden of workers Before Income tax deduction Special deduction Medical expenses Education expenses Donations Insurance and pensions After Tax credit 15% of the expenses : Medical, Education, Donations 12% of the expenses : Insurance, Pensions Assuming workers with spending 1 million Korean Won (KW) only for medical expenses Worker with gross wage income of 10 million KW (marginal tax rate of 6%) Income tax is reduced by 6% of the medical expenses Worker with gross wage income of 60 million K W (marginal tax rate of 24%) Income tax is reduced by 24% of the medical expenses The tax credit amounts are 15% of the medical expenses irrespective of marginal tax rate. If there is no change in working hours and wage rates, tax burden on low-income workers will decrease. tax burden on high-income workers will increase.

4 Previous research In the U.S and Europe, general consensus has been reached about the effect of taxes on the labor supply of male and female workers (Hausman(1981), MaCurdy, Green and Paarsch (1990), Blundell, Duncan and Meghir(1998), Aaberge et al(1999), Devereux(2004), French(2005), Meghir and Phillips (2008)) For men, the impact of taxation on labor supply is relatively small For women, the possibility of participating in the labor market and the working hour increase if they receive tax benefits The effect of single mother is bigger

5 Previous research (Cont d) There is no consensus on the results of studies conducted with data from Korea. Method Data Result Na, Nam and Moon (2002) Estimating labor supply before and after the decrease in the marginal tax rate of the income tax in 1994 Economic activity census and the Household income and expenditure survey Reduced labor supply Nam (2007) IV estimation method Economic activity census and the Household income and expenditure survey Increase in labor supply Jeon and Hong (2009) Double-difference method (using reduction of the income tax rate in 2004 and 2005) KLIPS (Korea labor and income panel study) Low-educated(low-income)workers: no change High-educated(high-income)workers: Increase in labor supply Kim and Sung (2007) 2SLS KLIPS (Korea labor and income panel study) Married women: Increase in labor supply

6 Differences from previous research Korea s previous studies have analyzed the effect of the changes in which the tax rates of all workers are lowered Researcher s arbitrary judgement can be involved in constructing a treatment group and a control group in the process of applying the difference in differences estimation method

7 Differences from previous research (Cont d) In this study, we construct a treatment group and a control group using the characteristics of the system. tax burden is reduced if worker s marginal tax rate is less than 6%. tax burden is increased if worker s marginal tax rate is more than 24%. tax burden is not changed if worker s marginal tax rate is 15%. We can examine both the effects of the increased and the decreased tax burden on labor supply at the same time.

8 Data: NaSTaB (National Survey of Tax and Benefit) NaSTaB has been completed for nine years until 2016 after the initial survey started in NaSTaB trys to increase the accuracy of the data by collecting the individual income tax filing documents.

9 Targets of analysis NaSTaB s 7 th ~9 th wave Workers who provided one s income tax filing documents Workers whose marginal tax rate has not changed

10 Targets of analysis The proportion of low-income workers is considerably high for female workers. The number of female workers whose marginal tax rate (mtr) is more than 24% is much fewer compared to male workers. <The number of workers by income tax bracket> Low-income worker (mtr=6%) Control group (mtr=15%) High-income worker (mtr>=24%) Male Female

11 Effective Tax Rate (=Income Tax/Yearly Wage) Income tax burden of low-income workers is reduced. Income tax burden of high-income workers is increased more compared to the control group. Low-income worker (mtr=6%) Control group (mtr=15%) High-income worker (mtr>=24%) Male Female Male Female Male Female Note: The table reports means with standard deviations in parentheses (0.29) (0.22) (0.35) (0.44) (1.62) (2.68) (1.39) (2.03) (3.33) (3.39) (0.92) (1.51) (unit : %)

12 Working Hours For male, working hours of control group decreased while working hours of both low and high-income workers increased. For female, working hours of low-income workers and control group increased Low-income worker Male 2,052 2,090 (684) (570) Female 1,933 1,967 (488) (417) Control group Male 2,124 2,074 (408) (351) Female 1,989 2,002 (237) (216) High-income worker Male 2,061 2,064 (474) (310) Female 1,920 1,920 (0) (0) Note: The table reports means with standard deviations in parentheses. (unit : hour)

13 Summary Statistics of Male Workers(2013) Low-income worker Control group High-income worker Age (12.2) (7.7) (6.8) Education (2.4) (2.2) (2.4) Married (%) Household income net of own labor income (million won) (13.9) (17.7) (20.9) Net assets of household (million won) (317) (291) (415) Total salary (million won) (8.87) (15.5) (38.78) Number of Observations Note: The table reports means with standard deviations in parentheses.

14 Summary Statistics of Female Workers(2013) Low-income worker Control group High-income worker Age (10.4) (8.7) (9.4) Education (2.8) (2.6) (2.2) Married (%) Household income net of own labor income (million won) (26.1) (37.4) (20.1) Net assets of household (million won) (264) (351) (127) Total salary (million won) (6.67) (11.93) (14.35) Number of Observations Note: The table reports means with standard deviations in parentheses.

15 Estimation: Fixed effect model Male workers : worker s annual working hours : control variables (age, education, marital status, etc.) : indicator variable 1, if workers affected by the system change 0, if workers who are not affected : time-fixed effect : individual fixed effect that is not observed including one s preference for work

16 Estimation (Cont d) Female workers Check the selection bias for female to participate in the labor market. We generate an inverse Mill s ratio through the probit estimation for labor market participation decision for females who are more than 20 years old. : worker s annual working hours : indicator variable 1, if workers affected by the system change 0, if workers who are not affected : Inverse Mill s ratio : control variables(age, education, marital status, etc. : time-fixed effect : individual fixed effect that is not observed including one s preference for work

17 Results: The effect on the labor supply (annual working hours) Male Female Low-income worker 98.47** (42.88) (38.30) High-income worker 100.8** (50.77) (29.88) Education 22.09** Household income net of own labor income (million won) (11.08) (30.79) (1.224) (1.064) Net assets of household (million won) ** (.9.20) (9.31) Inverse Mill's ratio 802.3** (322.6) Number of Households Number of Observations 1,863 1,002 Note: the table reports coefficients with robust standard errors in parentheses. Each regression also controls for age, marital status, year dummies and job type dummies. Asterisks denote significance at the 1%(***), 5%(**), and 10%(*) levels.

18 Discussion For lower income male workers, substitution effect seems to dominate income effect. For high income male workers, the reverse appears to be true.

19 Discussion (Cont d) Considering that the effective tax rate of low-income male workers decreases by 0.13%p after the income tax system change, if the effective tax rate decreases by 1%p by tax system such as EITC, their working hours can be expected to increase by 753 hours per year. By the same reasoning, if the progressive tax system is strengthened to increase the effective tax rate of high-income male workers by 1%p, the annual working hours of them can be expected to increase by 66 hours.

20 Discussion (Cont d) In the case of female workers, both low- and high-income workers do not statistically significant change in working hours compared to the control group. It can be that there is no big difference in the magnitude of the substitution effect or the income effect, However, since the number of samples in the treatment group is not large enough, the possibility of sampling error can not be ignored.

21 Conclusion The control group that is not influenced by the income tax system change is clearly classified and analyzed in this study. The estimation results show that the government s tax policy to improve taxation equity by changing the income tax system has resulted in inspiring willingness to work and raising real income for low-income male workers. In the case of high-income male workers, although the after-tax wage rate is reduced due to the increase in the tax burden, their labor supply does not decrease because their labor supply curve is backward bending.

22 Conclusion (Cont d) For female workers, we do not find a tax system change effect on their labor supply. However, sample size many not be big enough to generalize this conclusion. The decision of labor supply of female workers is more complicated than that of male workers. Therefore, we suggest follow-up studies reflecting these characteristics well. We hope that the results of this study can be used as a reference for predicting the effect on the labor supply of the workers in the process of redesigning the tax policy in the future.