Giesecke and Jaeger (2021) compute the MVPF of marginal increase in the first old-age pension program to ever exist in the UK. In 1908, the Old-Age Pension Act (OPA) created a pension benefit to low-income, elderly individuals. All individuals above age 70 with income below 630 shillings/year became eligible to receive 5 shillings/week (22% of the average income at the time). The paper exploits a newly released census dataset covering the full UK population in 1911 to implement a regression-discontinuity design around the pension age threshold of 70. The paper finds that labor force participation declined by 6 percentage points (from a baseline of 46%) at age 70 just after the implementation of the OPA. The effect is concentrated in workers in low-earnings occupations, where eligibility is highest.

MVPF = 0.8

To estimate the net costs to the government the authors consider the mechanical cost of transfers, administrative costs, and costs due to behavioral responses.

The mechanical cost is simply £1 for each £1 in transfers. They then assume that administrative costs are 3% of mechanical costs.

Finally, they assume that behavioral costs (either due to higher participation in other social programs or due to early retirement lowering tax revenues) amount to 10% of the mechanical cost. The total costs are therefore £1 + £0.03 + £0.10 = £1.13.

£1.1
Net Cost

Upper Margin
Lower Margin

The paper considers two types of beneficiaries in order to estimate the willingness to pay: marginal beneficiaries and infra-marginal beneficiaries.

Marginal beneficiaries are those who dropped out of the labor force to receive pensions. The envelope theorem implies that marginal beneficiaries are indifferent between (a) receiving the benefit and not working, and (b) working and not receiving the benefit. As a result, the willingness to pay for this group is 0.

Therefore, the overall willingness to pay for a £1 increase in OPA benefits is simply the fraction of beneficiaries that were infra-marginal. The infra-marginal beneficiaries were those whose behavior did not change in response to OPA (i.e., those beneficiaries who would have retired/been retired anyway).

To estimate this fraction, the paper needs to estimate how many beneficiaries retired in response to OPA (i.e., the size of the marginal group). For individuals at 70-years-old, the paper relies on its RD estimates of a 6pp drop in labor force participation. The number of eligible 70-year-olds in the sample is 140,288, so the number of marginal individuals in this group is 0.06 x 140,228 = 8,459. For individuals in the 71+ age group, the paper extrapolates labor-force participation from what was observed in the 65-69 age group and estimates that OPA is responsible for a 5.1% drop in labor force participation among individuals aged 71 and older. There were 928,198 individuals in the 71+ age group, which yields 0.051 x 928,198 = 47,909 marginal beneficiaries. Thus, the fraction of infra-marginal beneficiaries in among the total 613,873 pension recipients they observe is given by (613,873 – (8,459 + 47,909)) / 613,873 = 0.91 and the willingness to pay for a £1 increase in OPA is £0.91.

£0.9
WTP

Upper Margin
Lower Margin

Dividing the willingness to pay by the total costs yields an MVPF of 0.91 / 1.13 = 0.8. Making more conservative assumptions on the behavioral response (100% instead of 10% of mechanical costs) and on the fraction of infra-marginal individuals yields a lower bound MVPF estimate of 0.37.

0.8
MVPF

Upper Margin
0.4 Lower Margin

Giesecke, Matthias, and Philipp Jaeger (2021). “Pension incentives and labor supply: Evidence from the introduction of universal old-age assistance in the UK.” https://ftp.iza.org/dp14469.pdf

- Category
- Social Insurance
- Sub-Category
- Retirement
- Beneficiary Type(s)
- Adults, Seniors (65+)
- Average Age
- 75
- Country of Implementation
- United Kingdom
- Year of Implementation
- 1909
- Empirical Method
- Regression Discontinuity
- Research Type
- Primary
- Peer Reviewed
- No
- MVPF Publication Link
- ftp.iza.org/dp14469.pdf