Wage Elasticity of Labor Supply: A Survey-Based Experimental Approach

Transcription

1 Wage Elasticity of Labor Supply: A Survey-Based Experimental Approach 15 February 2006 Fumio Ohtake (Osaka University) Shinji Takenaka (Osaka University) Kengo Yasui (Osaka University) 0

2 Purpose of This Research 1.To Estimate i) Marshallian Elasticity of Labor Supply ii) Hicksian Elasticity of Labor Supply using the survey responses to hypothetical wage rises To Compare a) All Workers (JPN) and High-Income Earners (JPN) b) All Workers (JPN) and All Workers (USA) 2. To analyze the elasticity of retirement as well as the elasticity of work hours 1

3 Why is the Marshallian Elasticity Crucial? Argument: A progressive taxation inhibits the highincome earners labor supply based on Assumption: The high-income earners wage elasticity is greater than that of the other income classes. Examination on the Assumption through... Comparison of the Marshallian elasticity between high-income earners and all workers facing hypothetical, permanent wage shocks 2

4 Why is the Hicksian Elasticity Crucial? If the Hicksian elasticity of labor supply gets greater for the increase in marginal income tax rate, the welfare loss is greater for the high-income group. Examination on the underlined part through Comparison of the Hicksian elasticity between high-income earners and all workers 3

5 Why is the International Comparison Crucial? International difference in labor supply elasticities (concerning work hours and retirement) Different responses to a common taxation policy in the other countries Comparison of the labor supply elasticities between Japan and the United States as an example 4

6 Questionnaire Surveys in Japan 1) OPSJ: Osaka University 21st Century Center of Excellence Panel Survey [in Japan] (Feb. 2005) Sample: 6,000 households randomly sampled in 2004 # of respondents: 2,987 ( all workers in the following) (Questions yielding critical info. are reproduced in Appendix 1.) 2) JHIES: Japanese High-Income Earners Survey (Feb. 2005) Sample: Randomly-sampled 10 thousand high-income earners taxed more than US$ 2,320.4 of their income in 2004 # of respondents: 921 ( High-Income in the following) 5

7 Questionnaire Survey in the USA 3) OPSUS: Osaka University 21st Century Center of Excellence Panel Survey [in USA] (Jan.-Feb. 2005) Sample: 12,338 individuals # of respondents: 4,979 ( all workers in the following) (All the questions were the same as those in the OPSJ.) 6

8 Samples The following analyses are based on the samples of male respondents... up to 64 years old working for pay. 7

9 Marshallian Elasticity of Labor Supply Definition Percentage change in the optimal labor supply for a 1% rise in his wage Measurement of Marshallian elasticity By asking how many hours he would change his work hours if his wage were permanently doubled 8

10 Hicksian Elasticity of Labor Supply Measurement with the Slutsky Equation Hicksian elasticity = Marshallian elasticity - (wh / R 0 ) x Income elasticity w : real hourly wage h : work hours a week R 0 : potential income (= nonlabor income+ real wage x time endowment) Measurement of Income elasticity By asking how many hours he would change his work hours if he won a lottery that would pay the amount equal to last year s family income every year 9

11 Elasticity of Retirement: Introduction Change in labor supply encountering a wage shock 1) Change in work hours and 2) Change in the retirement 10

12 Marshallian Elasticity of Retirement Measurement of Marshallian Elasticity of Retirement By asking the expected age of retirement in case his wage were doubled permanently 11

13 Hicksian Elasticity of Retirement Measurement of Hicksian Elasticity of Retirement 1. The income elasticity of retirement was measured by asking the expected age of retirement in case of winning the same lottery projected in the question on work hours. 2. Using the income elasticity and the Marshallian, the Hicksian was derived through the Slutsky equation. 12

14 Marshallian Elasticity of Work Hours Density kdensity marshall All Japanese Workers Density kdensity marshall All US Workers Density kdensity marshall Japanese High-Income Earners Mean Japan High-income US

15 Hicksian Elasticity of Work Hours Density kdensity hicks40 All Japanese Workers Density kdensity hicks40 All US Workers Density kdensity hicks40 Japanese High-Income Earners Mean Japan High-income US

16 Marshallian Elasticity of Retirement Density kdensity retire_40 All Japanese Workers Density kdensity retire_40 All US Workers Density kdensity retire_40 Japanese High-Income Earners Mean Japan High-income US

17 Hicksian Elasticity of Retirement Density kdensity retire_hicks_40_1 All Japanese Workers Density kdensity retire_hicks_40_1 Japanese High-Income Earners All US Workers Density kdensity retire_hicks_40_1 Japan High-income US Mean

18 Marshallian Elasticity of Work Hours (All Japanese workers and Japanese High-Income earners) Constant *** *** *** *** High-income dummy * Work hours & its square, Marital status, Education, Types of Employment Hourly wage Annual income Almost no significant difference between all workers and high-income earners 17

19 Summary 1: Comparison between All Japanese Workers and High-Income earners On both work hours and retirement, almost no significant difference in the Marshallian and Hicksian elasticities was found between all Japanese workers and high-income earners. 18

20 Marshallian Elasticity of Work Hours (All Japanese Workers and All US workers) Constant *** *** ** *** US sample dummy *** *** *** *** Work hours & its square, Marital status, Education, Types of Employment Hourly wage Annual income Significantly greater in the United States 19

21 Summary 2: Comparison between Japan & USA (1) Elasticity of Work hours (1) Elasticity of Work Hours a) On both Marshallian & Hicksian terms, the US elasticities were found significantly larger. b) Average Marshallian of Japanese Average Marshallian of the U.S c) Average Hicksian of Japanese Average Hicksian of the U.S (under the assumption of 40 work hours maximum) 20

22 Summary 2: Comparison between Japan & USA (2) Elasticity of Retirement (2) Elasticity of Retirement a) On both Marshallian & Hicksian terms, the US elasticities were found smaller. ( w/o control variables, the differences were significant) b) Average Marshallian of Japanese Average Marshallian of the U.S

23 Summary 2: Comparison between Japan & USA (3) Features of Labor Supply (1) Elasticity of Work Hours b) Average Marshallian of Japanese Average Marshallian of the U.S (2) Elasticity of Retirement b) Average Marshallian of Japanese Average Marshallian of the U.S (1b) & (2b) In case of a wage rise, on average, JPN: a small decrease in work hours & later retirement USA: a tiny increase in work hours & earlier retirement 22

24 Summary 3: Comparison between Japan & USA (3) Elasticity and aging Elasticity and aging a) With regard to work hours, Both Japanese and US workers become less elastic by aging. b) With regard to retirement, Japanese become more elastic by aging (i.e., delay their retirement). 23

25 Thank you 24

26 Hicksian Elasticity of Work Hours (All Japanese workers and Japanese High-Income earners) Constant *** *** *** *** High-income dummy Work hours & its square, Marital status, Education, Types of Employment Hourly wage Annual income No significant difference between all workers and high-income earners 25

27 Marshallian Elasticity of Retirement (All Japanese Workers and Japanese High-Income Earners) Constant High-income dummy Work hours & its square, Marital status, Education, Types of Employment Hourly wage Annual income No significant difference between all workers and high-income earners 26

28 Hicksian Elasticity of Retirement (All Japanese Workers and Japanese High-Income Earners) Constant *** High-income dummy ** Work hours & its square, Marital status, Education, Types of Employment Hourly wage Annual income No significant difference between all workers and high-income earners 27

29 Hicksian Elasticity of Work Hours (All Japanese Workers and All US Workers) Constant *** *** ** *** US sample dummy *** *** *** *** Work hours & its square, Marital status, Education, Types of Employment Hourly wage Annual income Significantly greater in the United States 28

30 Marshallian Elasticity of Retirement (All Japanese Workers and All US Workers) Constant * US sample dummy *** Work hours & its square, Marital status, Education, Types of Employment Hourly wage Annual income Without control variables, the US elasticity of retirement is significantly smaller. 29

31 Hicksian Elasticity of Retirement (All Japanese Workers and All US Workers) Constant *** US sample dummy *** Work hours & its square, Marital status, Education, Types of Employment Hourly wage Annual income Without control variables, the US elasticity of retirement is significantly smaller. 30

32 Survey-based measurement of Labor Supply, compared to Contingent Valuation Method (CVM) A major criticism on Contingent Valuation Method: Dubious existence of the preference for unfamiliar public goods in question like a remote natural environment (Diamond and Hausman, 1994) Existence of preference for leisure and consumption The balance between work hours and leisure are based on one s preference. The response to the questions should represent their thought on the balance. 31

33 Survey-based measurement of labor supply: what the questionnaire measured Actual data of labor supply can be affected by complexity and frictions of workschedule changes. Survey-based data of labor supply are obtained in hypothetical environments but would reveal one s optimal work hours. The survey responses can at least provide complementary information on their labor supply. Reference: Kimball and Shapiro (2003) 32