Maternity Leave in Denmark
Introduction
Brenøe et al. (2020) estimate the effect of maternity leave on firms’ outcomes (e.g., costs) and the outcomes for the coworkers of women taking maternity leave (e.g., hours worked). The authors use administrative data from Denmark spanning 2001-2013 and considers two groups: (1) small firms (fewer than 30 employees) where a female employee took maternity leave (“treated firms”) and (2) similar firms where no employee took maternity leave (“control firms”). The authors match firms from the two groups and uses the matched sample to implement a differences-in-differences identification identification strategy.
The authors find that firms respond to an employee taking maternity leave by hiring temporary workers, increasing hours of current workers, and decreasing firing. As a result, the total hours worked in treated firms remains largely unchanged. Employees in treated firms see a slight increase in earnings, but the total wage bill for the firm remains unchanged because the government reimburses firms for the wages paid to the employee on maternity leave. Treated firms’ output and likelihood of survival also do not change.
Paradisi (2021) uses these results to compute the MVPF of introducing a paid parental leave policy.
MVPF = 0.7
Net Cost
There are four components to the net cost estimate: the mechanical cost, and three fiscal externalities.
The mechanical cost of the policy is given by the total transfer paid to workers on parental leave. This is given by the product of the average amount of leave (0.96 years) and what would have been the total annual earnings of the workers who took parental leave (260,350 DKK), for a total of 0.96 x 260,350 DKK = 250,336 DKK.
The first fiscal externality comes from the fact that the policy yields increased income tax revenue. Paradisi (2021) assumes that, in the absence of the policy, workers would take six months of unpaid leave (resulting in 6 months without income tax remittances). The reduction in government revenue as a result of the unpaid leave would be equal to 35% x 0.5 years x 260,350 DKK = 45,561 DKK, where 35% is the average income tax rate among workers who take parental leave.
The second fiscal externality stems from the finding in Brenøe et al. (2020) that parental leave increases the earnings of coworkers who remain in the firm, yielding increased government income tax revenue. The impact on net costs is given by the estimated impact of leave on coworker’s earnings (1.7%) times the total earnings of coworkers (3,629,760 DKK) times the income tax rate (35%), for a total of 0.017 x 0.35 x 3,629,760 = 22,422 DKK.
Brenøe et al. (2020) also find that parental leave increases hiring of new workers and decreases firing of incumbent coworkers. This also increases income tax revenues and, consequently, lowers net government costs. The reduction in net costs is given by multiplying the average expected tax revenue by the percentage increase in employment that resulted from the policy. The average expected tax revenue is equal to the average wage coworkers (304,000 DKK) times the increase in earnings of new hires (32%) times the tax rate (35%). The estimated increase in employment was 10.2%. As a result, the change to net cost is 304,000 DKK x 0.32 x 0.35 x 0.102 = 3,472 DKK.
The net costs to the government are the sum of the four components described above: 250,336 DKK – 45,561 DKK – 22,442 DKK – 3,472 DKK = 178,879 DKK.
Willingness to Pay
Paradisi (2021) calculates the willingness to pay for three groups.
First, the willingness to pay of parents who take parental leave is equal to the net-of-tax increase in income they experience as a result of the parental leave policy. To compute this component, Paradisi (2021) assumes that parents would take 6 months of unpaid leave in the absence of the policy. Given that the policy replaces their wages 1-to-1 during leave, these individuals experience an increase in income equal to the total, after-tax annual income of this group (260,350 DKK) times the amount of time they were on leave. With a tax rate of 35%, the willingness to pay for these individuals is 0.5 x 260,350 DKK x (1 – 0.35) = 84,613 DKK.
Second, Paradisi computes the willingness to pay of coworkers of parents who take leave for the changes in wages caused by the parental leave. This is given by the percentage increase in wages of coworkers of the workers who leave (1.7%) times the total earnings of coworkers (3,629,760 DKK) times the net-of-tax rate (65%), totaling 0.017 x 3,629,760 DKK x 0.65 = 41,642 DKK.
Third, Paradisi computes the willingness to pay of outside workers for the increase in earnings they experience due to increase in employment caused by the policy. The amount is given by the percentage increase in employment due to the policy (10.2%) times the change average increase in earnings of new hires (32%) times the total earnings of coworkers (304,000 DKK) times the tax rate (35%), for a total of 0.102 x 0.32 x 304,000 DKK x 0.65 = 6,449 DKK.
The total willingness to pay for the policy is the sum of the three components, that is, 84,613 DKK + 41,642 DKK + 6,449 DKK = 132,705 DKK.
MVPF Summary
Paradisi (2021) arrives at the MVPF by dividing the net cost by the willingness to pay: 132,705 / 178,879 = 0.74.
The estimates used to calculate this MVPF may have been updated in a more recent working or published version of the paper.
References
Brenøe, Anne A., Serena P. Canaan, Nikolaj A. Harmon, and Heather N. Royer. 2020. “Is Parental Leave Costly for Firms and Coworkers?” NBER Working Paper. https://www.nber.org/system/files/working_papers/w26622/w26622.pdf
Paradisi, Matteo (2021). Firms and Policy Incidence. Working Paper.
https://uploads-ssl.webflow.com/5d8e3657fd776a7142924af1/60b107ce7b635e55ff194322_paradisi_2021_firms_policy_incidence.pdf
Policy Features
- Category
- Social Insurance
- Sub-Category
- DI/SSI
- Beneficiary Type(s)
- Mothers and pregnant women, Parents
- Country of Implementation
- Denmark
- Empirical Method
- Difference in Differences
- Research Type
- Primary
- Peer Reviewed
- No
- MVPF Publication Link
- www.nber.org/system/files/working_papers/w26622/w26622.pdf