Presentation contains brief review of the analysis of return to commuting in Sweden. The research is based on highly informative administrative data set and various econometric techniques to tackle the problem of endogeneity. Results assign the monetary value for the return per kilometer of commuting.

1. RETURN TO COMMUTING
IN SWEDEN
Sergii Troshchenkov
PhD student
L.A.S.E.R

2. Return to commuting in Sweden
The main idea is to estimate the return to
commuting for the male and female subsamples
of the population:
Are employees compensated by employers for
commuting expenses?
Does gender matter?

3. Return to commuting in Sweden
The policy application:
• Decreasing unemployment
• Suburbanization
• Improvement of the search—matching process
Contribution:
• Comprehensive dataset based on administrative registers
• Addressing the endogeneity through the FE estimation
• Extensive set of variables that controls individual
characteristics and features of the labor market

4. Return to commuting in Sweden
Wage formation in Sweden
• The wage bargaining and setting minimum wages occurs between
employer cartel and labor unions:
• Setting collective agreement
• Possibility of re-bargain small share of wage
Unique feature
• Swedish labor unions are very strong (Laval case)
Conclusion
No immediate impact of unemployment rate on the wage level. The
opposite story: regions with high wages are characterized by high
unemployment rate.

5. Return to commuting in Sweden
Earnings by region.
Male sample
Earnings by region.
Female sample
(2118.116,3570.393]
(2008.6765,2118.116]
(1925.8884,2008.6765]
[1634.9245,1925.8884]

6. Return to commuting in Sweden
Commuting patterns:
Median of travel distance:
• 8 kilometers –for females
• 12 kilometers—for males.
The long-distance commuters earn less:
2300 hundreds of annual income in comparison to 3369 for
“stayers”
Evident spatial patterns of commuting

7. Return to commuting in Sweden
Commuting by region.
Male sample
Commuting by region.
Female sample
(.3636364,1.125]
(.2595403,.3636364]
(.205074,.2595403]
[.0819521,.205074]

8. Return to commuting in Sweden
The model:
lnLoneInk0it=αit+γ1distit+γ2squaredistit+β1Xit+zi+εit
i=1…T stands for cross-section units (individuals)
t=1…K indicates time period
α, γ and β are coefficients to be estimated
X is a generic set of variables that captures lifecycle events and
labor market conditions
z reflects individual time invariant fixed effects

9. Return to commuting in Sweden
Description of variables of main interest
• LnLoneInk0 indicates the annual earnings of individual
received from the employment type of activity. Earning was
considered in the interval 1500 hundreds SEK and 8518
hundreds SEK of gross annual income.
• dist is a variable that defines a commuting distance in
kilometers for every individual calculated using Pythagoras
formula on 100 meters span. Distance was considered on
the interval between 0.25 and 80 km one way journey.
• squaredist is a squared term of commuting distance
introduced in order to capture the concave profile of
commuting in earning equation

10. Return to commuting in Sweden
Description of the sample
• The data is (LOUISE) and Swedish National Tax Board.
• LOUISE: information about employment status, sources of
earning, family conditions and education.
• Data from the Swedish National Tax Board: labor and non-
labor earnings of individuals over an analyzed period.
• The sample – employed individuals in the age between 20
and 64 who experienced a job change in 2008.
• Separate analysis for males and females
• Students, people on parental leave and unemployed are
excluded from the sample

11. Return to commuting in Sweden
Initial assumptions for analysis
Male sample
• One hour of commuting time which is approximately equal to
35 kilometers
• Hourly wage of 199 SEK (22.11 EUR)
Female sample
• One hour of commuting time which is approximately equal to
35 kilometers
• Hourly wage of 174 SEK (19.33 EUR)

12. Return to commuting in Sweden
Results. Male sample:
OLS
Regression coefficients:
• Commuting distance coefficients of OLS is 0.000566
• Squared term of the commuting distance divided by 100 is ̶ 0.0000129.
Conclusions:
• individuals receive a compensation of 31 SEK (3.44 EUR) per one hour of
daily commuting. It constitutes 16 % of hourly wage.
FE
Regression coefficients:
• Commuting distance coefficients of FE is 0.000305
• Squared term of the commuting distance divided by 100 is ̶ 0.000156.
Conclusions:
• individuals receive a compensation of 17SEK (2 EUR) per one hour of
daily commuting. It constitutes 8.5 % of hourly wage.

13. Return to commuting in Sweden
Results. Female sample:
OLS
Regression coefficients:
• Commuting distance coefficient of OLS is 0.000684
• Squared term of the commuting distance divided by 100 is ̶ 0.000076 .
Conclusions:
• individuals receive a compensation of 33SEK (3.66EUR) per one hour of
daily commuting. It constitutes 17 % of hourly wage.
FE
Regression coefficients:
• Commuting distance coefficient of OLS is 0.000301
• Squared term of the commuting distance divided by 100 is ̶ 0.000217.
Conclusions:
• individuals receive a compensation of 14 SEK (1.5 EUR) per one hour of
daily commuting. It constitutes 7.3 % of hourly wage.

14. Return to commuting in Sweden
Firm FE
Male sample
Regression coefficients:
• Commuting distance coefficient of FE is 0.00036
• Squared term of the commuting distance divided by 100 is ̶ 0.000013 .
Conclusions:
• Individuals receive a compensation of 18,54SEK (1,95EUR) per one hour
of daily commuting. It constitutes 10,8 % of hourly wage.
Female sample
Regression coefficients:
• Commuting distance coefficient of FE is 0.00033
• Squared term of the commuting distance divided by 100 is ̶ 0.00007 .
Conclusions:
• Individuals receive a compensation of 17.53 SEK (1,83EUR) per one hour
of daily commuting. It constitutes 9,34 % of hourly wage.

15. Return to commuting in Sweden
Wage growth model
Male sample
Regression coefficients:
• Commuting distance coefficient is 0.000759
Conclusions:
• Individuals who commute one hour per day have 0.4% higher
earnings growth rate
Female sample
Regression coefficients:
• Commuting distance coefficient is 0.000326
Conclusions:
• Individuals who commute one hour per day have 0.2% higher
earnings growth rate

16. Return to commuting in Sweden
Conclusions
• Results demonstrates similar relative return to commuting for
men and woman
• The results are somewhat lower than in the previous studies
• In line with the theory of bargaining power that suggests this
return to be between 0 and 30%
Limitations
• Separated hourly wages and labor supply patterns would
improve the picture of efficient job-search process/bargaining
power process
• Introduction of commuting expenses would allow to estimate
Net return to commuting

17. Thank you to your attention
Author: Sergii Troshchenkov
Email: sergii.troshchenkov@unimi.it