How Important is Intertemporal Shifting of Wage Income? Claus Thustrup Kreiner1 , Søren Leth-Petersen1 and Peer Ebbesen Skov1,2 1 Department

of Economics, University of Copenhagen and 2 Rockwool Foundation Research Unit

May 25th, 2012

2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

1 / 22

Introduction Monthly Wage (100.000 DKK approx. 17.800 USD)

j-10

450.000

400.000

350.000

300.000

j-11

d-10

n-10

o-10

a-10

s-10

j-10

j-10

a-10

m-10

m-10

f-10

o-09

s-09

a-09

j-09

a-09

j-09

m-09

j-09

m-09

f-09

d-08

a-08

150.000

o-08

s-08

n-08

j-08

j-08

m-08

a-08

m-08

f-08

200.000

j-08

DKK

250.000

100.000

d-09

n-09

50.000

0

(source: Own calculations on eIncome register)

2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

2 / 22

Introduction Gain from Shifting December Wage 12.000

10.000

Gain (DKK)

8.000

6.000

4.000

2.000

20 .0 24 00 .0 28 00 .0 0 30 0 .0 32 00 .4 36 92 .0 40 00 .0 44 00 .0 48 00 .0 0 52 0 .0 56 00 .0 60 00 .0 64 00 .0 68 00 .0 72 00 .0 0 76 0 .0 80 00 .0 84 00 .0 88 00 .0 0 92 0 .0 96 00 .0 10 00 0. 00 10 0 4. 0 10 00 6. 0 11 00 0. 0 11 00 4. 0 11 00 8. 0 12 00 2. 0 12 00 6. 13 000 0. 0 13 00 4. 0 13 00 8. 0 14 00 2. 0 14 00 6. 0 15 00 0. 00 0

0

Monthly Wage (DKK)

(source: Own calculations)

Income Year 2009 2010

Table: Income limit and marginal tax rates Income limit for Marginal tax rate top tax rate (DKK) below limit (%) 347,200 42.08 389,900 40.90

Marginal tax rate above limit (%) 62.86 56.01

2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

3 / 22

Purpose Review of some standard Public Finance results

Standard PF insight: the e¢ ciency loss from a tax is equal to the e¤ect on government revenue from all the behavioral changes generated by the tax

) All responses may be captured by looking at changes in taxable income (su¢ cient statistics) obtained from individual tax returns ) The modern empirical PF literature focus on measuring the elasticity of taxable income (ETI) rather than the hours elasticity The ETI is typically estimated using di¤-in-di¤ types of methodologies using tax reforms as exogenous variation Example: The tax rate is reduced for a treatment group and stays unchanged for a control group. Compute the (percentage) change in income from pre-reform to post-reform of the treatment group relative to the control group (counterfactural is a common trend) 2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

4 / 22

Purpose Review of some standard Public Finance results

The ETI, de…ned as ε

dz 1 τ , d (1 τ ) z

is crucial for tax policy

Consider increasing the tax on high incomes, de…ned as z τ la¤er = and z e E [z j z income.

1 , where α 1+ε α

z . Then

ze ze

z

z ] is the average income for individuals with high

For high incomes in Denmark, we have α ε: τ la¤er (%) :

0.05 85.5

0.1 74.6

3.4 implying 0.15 66.2

0.2 59.5

This should be compared to the actual tax rate in DK on high incomes: 66.2% 2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

5 / 22

Purpose Why interested in intertemporal shifting of income?

At least two reasons to be interested in intertemporal shifting of wage income: 1. Intertemporal shifting of income is a behavioral responses ) loss of economic e¢ ciency 2. The long run, structural elasticity of ETI is what really matters when considering tax reforms BUT the empirical estimates may be upward biased due to income shifting (Saez et al, 2012) Research questions How common is intertemporal shifting of "normal" wage income? How much of the ETI can be attributed to intertemporal shifting behavior? Importance of incentives, information and employer cooperation for shifting behavior? 2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

6 / 22

Previous Literature Limited knowledge on the extent of intertemporal shifting of wage income

Evidence of shifting between tax bases (Auerbach and Slemrod 1997) Timing of capital gains realizations appears to react strongly to changes in capital gains tax rates (Auerbach, 1988) Signi…cant part of the increase in taxable income of executives is due to timing in the realization of stock options (Goolsbee, 2000b) Much more di¢ cult to change the timing of regular labor income than the timing of capital income and stock-options: I

Higher transaction costs

I

Need corporation with your employer

I

More likely to become credit constrained if not receiving labor income

2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

7 / 22

Data (eIndkomst) Before: Companies are obligated to report their employees’wage income, labor market contribution, tax payments etc. on a yearly basis. (For SKL § 7 in Danish, see https://www.retsinformation.dk/Forms/R0710.aspx?id=17108)

– E¤ectively January 2008 – After: Companies are obligated to report their employees’wage income, labor market contribution, tax payments etc. on a monthly basis. In other words, the eIncome register captures all the information that is printed on a typical Danish payslip. (For eIndkomst see http://www.skat.dk/skat.aspx?oId=401651&vId=0)

Why: The eIncome register has two major innovations compared to the earlier system. First the montly reporting of income replace yearly reports thus SKAT always have up-to-date information for all tax payers. Second all public bodies share the information reducing companies’administrative burdens. (see http://www.skat.dk/SKAT.aspx?oID=416058&vID=0) 2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

8 / 22

Tax Incentives, Employer Cooperation and Information I

Tax Incentives:

The Danish 2010 tax reform reduced the marginal tax rate on earnings from 63 percent to 56 percent for the 1/3 of employees with the highest income. I

Employer Cooperation:

The shifting of income requires cooperation from the employer, who reports earnings to the tax authorities. I

Information:

The employee needs knowledge about the potential economic gain and the tax law. 2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

9 / 22

Tax Incentives, Employer Cooperation and Information Average Monthly Wage Income

Control Group: Individuals with an avg. 08 monthly wage between 25.000 -35.000 DKK. Treatment Group: Individuals with an avg. 08 monthly wage 35.000 DKK. Observations: 33.016.432 and Individuals: 892.336. 2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

10 / 22

Tax Incentives, Employer Cooperation and Information For each month we regress

h

yi ,y ,m yi ,2008,m y i ,2008

i

on a treatment dummy.

Difference-in-differences. Treatment: avg. monthly ‘08 wage is lt. 35.000 dkk. Control: avg. monthly wage in ‘08 is betwn 25.000 dkk. and 35.000 dkk. Observations: 33.016.432 and Individuals: 892.336. 15

10

5

ja n0 fe 8 bm 08 ar -0 ap 8 r-0 m 8 aj -0 ju 8 n08 ju l-0 au 8 g08 se p0 ok 8 t-0 no 8 v0 de 8 c0 ja 8 n0 fe 9 b0 m 9 ar -0 ap 9 r-0 m 9 aj -0 ju 9 n09 ju l-0 au 9 g09 se p0 ok 9 t-0 no 9 v0 de 9 c0 ja 9 n1 fe 0 b1 m 0 ar -1 ap 0 r-1 m 0 aj -1 ju 0 n10 ju l-1 au 0 g10 se p1 ok 0 t-1 no 0 v1 de 0 c1 ja 0 n11

0

-5

-10

-15

(source: Own calculations on eIncome register) 2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

11 / 22

Tax Incentives, Employer Cooperation and Information A Feldstien type approach gives an elasticity estimate of 0.1 log (yi )2009 = α + βs [log (1

log (yi )2010

m i ,2010 )

log (1

m i ,2009 )] + ei ,2010 ,m

ei ,2010 ,m

where βs is the elasticity of taxable income (where the s superscript relates to the simple approach taken here), m i ,2009 2009 is the pre-reform tax rate, which is treated as exogenous, while m i ,2010 is the predicted marginal tax rate after the reform.

A Goolsbee Type Decomposition Approach reveals that shifting accounts for 60% of the simple elasticity estimate. log (yi )yr ,m

log (yi )2008,m = α + βT 09 d Tx 09 + β10 d 10 + βT 10 d Tx 10 + ei ,2010 ,m

Where βs = βT 10

ei ,2010 ,m

βT 09

Note that shifting implies that βT 09 will be negative and therefore that βs overestimates the structural elasticity. The structural elasticity will equal η = βT 10 + βT 09 which take into account that some of the increase in observed income in 2010 is due to shifting of income from 2009 to 2010. 2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

12 / 22

Tax Incentives Difference-in-differences (A/S). Treatment: avg. monthly ‘08 wage is lt. 35.000 dkk. Control: avg. monthly wage in ‘08 is betwn 25.000 dkk. and 35.000 dkk. Observations: 18.486.680 and Individuals: 499.640 15

.

10

5

ja

fe

n08 bm 08 ar -0 ap 8 r-0 m 8 aj -0 ju 8 n08 ju l-0 au 8 g08 se p0 ok 8 t-0 no 8 v0 de 8 c0 ja 8 n0 fe 9 b0 m 9 ar -0 ap 9 r-0 m 9 aj -0 ju 9 n09 ju l-0 au 9 g09 se p0 ok 9 t-0 no 9 v0 de 9 c0 ja 9 n1 fe 0 b1 m 0 ar -1 ap 0 r-1 m 0 aj -1 ju 0 n10 ju l-1 au 0 g10 se p1 ok 0 t-1 no 0 v1 de 0 c1 ja 0 n11

0

-5

-10

-15

(source: Own calculations on eIncome register)

2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

13 / 22

Tax Incentives For each month we regress

Panel A

hy

yi ,2008,m y i ,2008

i ,y ,m

i

on a treatment dummy and sector dummies.

Panel B

37 Consecutive company speci…c wage observations

Unconditional

Difference-in-differences (A/S). Treatment: avg. monthly ‘08 wage is lt. 35.000 dkk. Control: avg. monthly wage in ‘08 is betwn 25.000 dkk. and 35.000 dkk. Observations: 18.486.680 and Individuals: 499.640 15

.

Difference-in-differences (A/S). Treatment: avg. monthly ‘08 wage is lt. 35.000 dkk. Control: avg. monthly wage in ‘08 is betwn 25.000 dkk. and 35.000 dkk. Observations: 8.110.807 and Individuals: 219.211 15,0

5

5,0

0

0,0 jan -0 feb 8 -0 ma 8 rap 08 r-0 ma 8 j-0 jun 8 -0 jul- 8 au 08 g-0 se 8 p-0 ok 8 tno 08 v-0 de 8 c-0 jan 8 -0 feb 9 -0 ma 9 rap 09 r-0 ma 9 j-0 jun 9 -0 jul- 9 au 09 g-0 se 9 p-0 ok 9 tno 09 v-0 de 9 c-0 jan 9 -1 feb 0 -1 ma 0 rap 10 r-1 ma 0 j-1 jun 0 -1 jul- 0 au 10 g-1 se 0 p-1 ok 0 tno 10 v-1 de 0 c-1 jan 0 -11

10,0

jan -0 feb 8 -0 ma 8 r-0 ap 8 r-0 ma 8 j-0 jun 8 -0 jul- 8 au 08 g-0 se 8 p-0 ok 8 tno 08 v-0 de 8 c-0 jan 8 -0 feb 9 -0 ma 9 r-0 ap 9 rma 09 j-0 jun 9 -0 jul- 9 au 09 g-0 se 9 p-0 ok 9 tno 09 v-0 de 9 c-0 jan 9 -1 feb 0 -1 ma 0 r-1 ap 0 r-1 ma 0 j-1 jun 0 -1 jul- 0 au 10 g-1 se 0 p-1 ok 0 tno 10 v-1 de 0 c-1 0

10

-5

-5,0

-10

-10,0

-15

-15,0

(source: Own calculations on eIncome register)

.

(source: Own calculations on eIncome register)

2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

14 / 22

Tax Incentives Panel A

Panel B

Treatment group: 35.000 - (75th percentile)

Treatment group: 75th - 90th percentile

37 Consecutive company speci…c wage observations

37 Consecutive company speci…c wage observations

Difference-in-differences (A/S). Treatment: avg. monthly ‘08 wage is betwn 38.942 dkk. and 51.539 dkk. Control: avg. monthly wage in ‘08 is betwn 25.000 dkk. and 35.000 dkk. Observations: 6.025.820 and Individuals: 162.860

Difference-in-differences (A/S). Treatment: avg. monthly ‘08 wage is betwn 35.000 dkk. and 38.942 dkk. Control: avg. monthly wage in ‘08 is betwn 25.000 dkk. and 35.000 dkk. Observations: 7.105.961 and Individuals: 192.053

10

10

5

5

0

0 jan -0 feb 8 -0 ma 8 rap 08 rma 08 j-0 jun 8 -0 jul- 8 au 08 g-0 se 8 p-0 ok 8 tno 08 v-0 de 8 c-0 jan 8 -0 feb 9 -0 ma 9 r-0 ap 9 rma 09 j-0 jun 9 -0 jul- 9 au 09 g-0 se 9 p-0 ok 9 tno 09 v-0 de 9 c-0 jan 9 -1 feb 0 -1 ma 0 r-1 ap 0 r-1 ma 0 j-1 jun 0 -1 jul- 0 au 10 g-1 se 0 p-1 ok 0 tno 10 v-1 de 0 c-1 jan 0 -11

15

jan -0 feb 8 ma -08 rap 08 r ma -08 j-0 jun 8 -0 jul- 8 au 08 gse 08 p-0 ok 8 tno 08 vde 08 cjan 08 -0 feb 9 ma -09 rap 09 r ma -09 jjun 09 -0 jul- 9 au 09 g se -09 p-0 ok 9 tno 09 vde 09 cjan 09 -1 feb 0 ma -10 rap 10 r-1 ma 0 jjun 10 -1 jul- 0 au 10 g-1 se 0 p-1 ok 0 tno 10 vde 10 cjan 10 -11

15

-5

-5

-10

-10

-15

-15

(source: Own calculations on eIncome register)

(source: Own calculations on eIncome register)

2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

15 / 22

Tax Incentives Panel C

Panel D

Treatment group: 90th - 95th percentile

Treatment group: 95th+-percentile

37 Consecutive company speci…c wage observations

37 Consecutive company speci…c wage observations

Difference-in-differences (A/S). Treatment: avg. monthly ‘08 wage is betwn 51.539 dkk. and 62.434 dkk. Control: avg. monthly wage in ‘08 is betwn 25.000 dkk. and 35.000 dkk. Observations: 4.945.753 and Individuals: 133.669

Difference-in-differences (A/S). Treatment: avg. monthly ‘08 wage is lt. 62.434 dkk. Control: avg. monthly wage in ‘08 is betwn 25.000 dkk. and 35.000 dkk. Observations: 4.945.716 and Individuals: 133.668

10

10

5

5

0

0 jan -0 feb 8 -0 ma 8 rap 08 rma 08 j-0 jun 8 -0 jul- 8 au 08 g-0 se 8 p-0 ok 8 tno 08 v-0 de 8 c-0 jan 8 -0 feb 9 -0 ma 9 r-0 ap 9 rma 09 j-0 jun 9 -0 jul- 9 au 09 g-0 se 9 p-0 ok 9 tno 09 v-0 de 9 c-0 jan 9 -1 feb 0 -1 ma 0 r-1 ap 0 r-1 ma 0 j-1 jun 0 -1 jul- 0 au 10 g-1 se 0 p-1 ok 0 tno 10 v-1 de 0 c-1 jan 0 -11

15

jan -0 feb 8 -0 ma 8 rap 08 r-0 ma 8 j-0 jun 8 -0 jul- 8 au 08 g-0 se 8 p-0 ok 8 tno 08 v-0 de 8 c-0 jan 8 -0 feb 9 -0 ma 9 r-0 ap 9 rma 09 j-0 jun 9 -0 jul- 9 au 09 g-0 se 9 p-0 ok 9 tno 09 v-0 de 9 c-0 jan 9 -1 feb 0 -1 ma 0 r-1 ap 0 rma 10 j-1 jun 0 -1 jul- 0 au 10 g-1 se 0 p-1 ok 0 tno 10 v-1 de 0 c-1 jan 0 -11

15

-5

-5

-10

-10

-15

-15

(source: Own calculations on eIncome register)

(source: Own calculations on eIncome register)

2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

16 / 22

Employer Cooperation Difference-in-differences (Gov't). Treatment: avg. monthly ‘08 wage is lt. 35.000 dkk. Control: avg. monthly wage in ‘08 is betwn 25.000 dkk. and 35.000 dkk. Observations: 637.066 and Individuals: 17.218 15

10

5

ja

fe

n08 bm 08 ar -0 ap 8 r-0 m 8 aj -0 ju 8 n08 ju l-0 au 8 g08 se p0 ok 8 t-0 no 8 v0 de 8 c0 ja 8 n0 fe 9 b0 m 9 ar -0 ap 9 r-0 m 9 aj -0 ju 9 n09 ju l-0 au 9 g09 se p0 ok 9 t-0 no 9 v0 de 9 c0 ja 9 n1 fe 0 b1 m 0 ar -1 ap 0 r-1 m 0 aj -1 ju 0 n10 ju l-1 au 0 g10 se p1 ok 0 t-1 no 0 v1 de 0 c1 ja 0 n11

0

-5

-10

-15

(source: Own calculations on eIncome register)

2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

17 / 22

Employer Cooperation Triple differences (A/S vs. Gov't). Treatment: avg. monthly ‘08 wage is lt. 35.000 dkk. Control: avg. monthly wage in ‘08 is betwn 25.000 dkk. and 35.000 dkk. Observations: 8.747.873 and Individuals: 236.429 15

10

5

ja n0 fe 8 bm 08 ar -0 ap 8 r-0 m 8 aj -0 ju 8 n08 ju l-0 au 8 g0 se 8 p0 ok 8 t-0 no 8 v0 de 8 c0 ja 8 n0 fe 9 b0 m 9 ar -0 ap 9 r-0 m 9 aj -0 ju 9 n09 ju l-0 au 9 g09 se p0 ok 9 t-0 no 9 v0 de 9 c0 ja 9 n1 fe 0 bm 10 ar -1 ap 0 r-1 m 0 aj -1 ju 0 n10 ju l-1 au 0 g10 se p1 ok 0 t-1 no 0 v1 de 0 c1 ja 0 n11

0

-5

-10

-15

(source: Own calculations on eIncome register)

2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

18 / 22

Information Survey Questionarie in February 2010: Q1 (Brief) Imagine that you were entitled to a little extra wage income in December ‘09: when would you prefer to receive the extra income payment? December ‘09 or January ‘10. Q2 (Brief) Is it legal or illegal to postpone the wage income you earned in 2009 to 2010? Timing December ‘09 January ‘10 Doesn’t matter Total (%)

Legal (%) 3.44 8.78 21.17 33.39

Illegal (%) 6.06 14.77 45.78 66.61

Total (%) 9.50 23.56 66.94 100.00

Fairly intense newspaper scrutinization: For example: www.epn.dk and www.borsen.dk (date: October 12th and 15th) outlines the scheme, but questions if it is legal. A spokes person from SKAT says YES IT IS! The popular press also picks it up: DR (November 6th), TV2 (October 15th), BT (November 29th) 2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

19 / 22

Pension Payments The 2010 tax reform also encouraged taxpayers to move pension payments forward. The reasoning behind moving pension payment forward is closely related to the motivation for income shifting. Deposits to an annuity pension plan (ratepension) can be deducted from the taxable income. The decrease of the marginal tax rates lowers the deduction value (in percentage terms) of pension payments. Thus, a taxpayer faced with the top tax in both 2009 and 2010 should favor deposits in 2009 and thereby take advantage of the 62.86 percent deduction value. (#) Gross Income = A-income + (1 /3 ) ATP contributions + (1 /3 ) PensionPayments

If the taxpayer increases his pension payments then gross income will be constant, A-income and ATP contributions will drop and pension payments will be increased. The approx in (#) is useful since pension payments are not reported directly to eIncome and thus have to be calculated. Notice that the 2010 tax reform invoked a max deduction of a 100.000 DKK for contributions to pension annuity and discontinuing annuities. 2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

20 / 22

Pension Payments For each month we regress

h pens

pensi ,2008,m y i ,2008

i ,y ,m

i

on a treatment dummy.

Panel A

Panel B

PrSE (A/S)

PbSE (Gov’t)

37 Consecutive company speci…c wage observations

37 Consecutive company speci…c wage observations

Difference-in-differences (A/S). Treatment: avg. monthly ‘08 wage is lt. 35.000 dkk. Control: avg. monthly wage in ‘08 is betwn 25.000 dkk. and 35.000 dkk. Observations: 8.110.807 and Individuals: 219.211 5

Difference-in-differences (Gov't). Treatment: avg. monthly ‘08 wage is lt. 35.000 dkk. Control: avg. monthly wage in ‘08 is betwn 25.000 dkk. and 35.000 dkk. Observations: 637.066 and Individuals: 17.218 5

.

4

4

3

3

2

2

1

1

-1

jan -0 feb 8 -0 ma 8 r-0 ap 8 rma 08 j-0 jun 8 -0 jul- 8 au 08 g-0 se 8 p-0 ok 8 tno 08 v-0 de 8 c-0 jan 8 -0 feb 9 -0 ma 9 r-0 ap 9 rma 09 j-0 jun 9 -0 jul- 9 au 09 g-0 se 9 p-0 ok 9 tno 09 v-0 de 9 c-0 jan 9 -1 feb 0 -1 ma 0 r-1 ap 0 r-1 ma 0 j-1 jun 0 -1 jul- 0 au 10 g-1 se 0 p-1 ok 0 tno 10 v-1 de 0 c-1 jan 0 -11

0

jan -0 feb 8 -0 ma 8 r-0 ap 8 r-0 ma 8 j-0 jun 8 -0 jul- 8 au 08 g-0 se 8 p-0 ok 8 tno 08 v-0 de 8 c-0 jan 8 -0 feb 9 -0 ma 9 r-0 ap 9 r-0 ma 9 j-0 jun 9 -0 jul- 9 au 09 g-0 se 9 p-0 ok 9 tno 09 v-0 de 9 c-0 jan 9 -1 feb 0 -1 ma 0 r-1 ap 0 rma 10 j-1 jun 0 -1 jul- 0 au 10 g-1 se 0 p-1 ok 0 tno 10 v-1 de 0 c-1 jan 0 -11

0

-1

-2

-2

-3

-3

-4

-5 (1 Department

-4

of Economics, University of Copenhagen

-5 2 Rockwool andIncome Shifting Foundation

Research Unit)

May 25th, 2012

21 / 22

The road ahead

Tax Incentives: Credit Constrained individuals Employer Cooperation: Other Public Sector employees, municipality/region etc. Large vs. small companies Information: Taxpayers with accountants who have e-access to their tax information (tast-selv)

2 Rockwool (1 Department of Economics, University of Copenhagen andIncome Shifting Foundation Research Unit)

May 25th, 2012

22 / 22