Summary
Staubli (2011) exploits Austria's 1996 Structural Adjustment Act — which raised the age threshold for relaxed Disability Insurance (DI) eligibility from 55 to 57 for men only — as a natural experiment to estimate the causal effects of stricter eligibility criteria on DI enrollment and labor market outcomes. Using a difference-in-differences (DiD) design on the Austrian Social Security Database (ASSD) (N=236,218 private-sector workers), he finds that stricter criteria reduced DI enrollment by 6–7.4 percentage points (pp) but generated large spillovers to unemployment insurance (UI) and sickness insurance (SI), with only modest net employment gains. Most displaced workers simply postponed DI application until the new threshold at age 57, rather than returning permanently to work.
Key Claims
- Main DiD results (preferred specification, excluding transition years): DI enrollment −6.0 to −7.4 pp (pre-reform base: 22.6%); employment +1.6 to +3.4 pp; unemployment +3.5 to +3.9 pp; sickness insurance +0.6 to +0.7 pp; total labor force participation (LFP) (incl. UI/SI) +6.1 to +7.5 pp.
- Postponement, not deterrence: DI enrollment drops ~10 pp at ages 55–56 after the reform but fully recovers at ages 58–59 — the dominant behavioral response is delaying application until the new threshold, not exiting the DI pipeline.
- Blue-collar vs. white-collar: Blue-collar DI −8.9 to −11.2 pp, employment +2.7 to +6.0 pp (pre-reform DI base 33.3%); white-collar DI −1.3 to −1.6 pp, employment null (pre-reform DI base 7.5%). White-collar workers already faced a stricter standard below 55 (same occupational group rather than "any reasonable occupation"), so the reform had less binding effect.
- Earnings gradient: Lowest quintile DI −12.0 pp (base 43.4%); highest quintile −0.6 pp (base 2.1%) — explained by the progressive benefit formula: lower earners have higher replacement rates and weaker incentive to remain employed.
- Health heterogeneity: Unhealthy workers DI −9.7 to −12.4 pp, employment +3.3 to +7.1 pp (base 43.5%); healthy workers DI −3.3 to −3.9 pp, employment null (base 12.0%).
- Spillovers: DI, UI, and sickness insurance act as close substitutes; UI absorbs most of the DI reduction. The reform "succeeded only partially."
- Net fiscal impact: −€12.1M/year net saving (DI savings −€18.1M, partially offset by higher UI +€9.2M, SI +€0.5M, partly recovered by tax revenues +€5.5M).
- Self-screening (Parsons 1991): Stricter criteria reduce applications as well as awards; the enrollment decline reflects both fewer applications and fewer approvals.
- International comparison: Austria shows a larger employment response than Sweden (Karlström et al. 2008) or the US (Chen and van der Klaauw 2008), attributed to Austria's reform targeting a younger, more employable margin of the DI applicant pool.
- Data: Austrian Social Security Database (ASSD) — universe of Austrian private-sector workers, quarterly observations 1991–2002, N=236,218 after exclusions (excl. public servants, self-employed, <5 contribution years).
Concepts Introduced or Extended
- Vocational Grid — Austrian two-tier eligibility standard (any occupation below 55 / similar occupation above 55) is the closest international analogue to the Social Security Administration's (SSA) vocational grid; the 1996 reform shifts Austria's age cutoff from 55 to 57, generating causal identification from the cliff
- Conditional DI Applicants — heterogeneity by health status, collar type, and earnings directly maps onto the conditional applicant typology; blue-collar/unhealthy workers at the eligibility cliff show the largest enrollment responses, confirming that marginal enrollees are workers with moderate impairments and viable employment alternatives
- DI Application Costs and Take-Up — Parsons (1991) self-screening mechanism applies: stricter criteria deter applications, not just awards; transition analysis introduces the postponement pattern as a distinct behavioral mode
- Causal Effects of DI Receipt — adds international DiD evidence on eligibility criteria effects; employment response benchmarked against Swedish and US estimates; confirms large individual-level effects for the blue-collar/unhealthy margin
- Disability Insurance Program — Austria's DI program as the reform subject; two-tier age-based eligibility standard
Entities Mentioned
Quotes
"The 1996 reform succeeded only partially. While there was a significant decline in DI enrollment and an increase in employment, more than half of the decline in DI enrollment was offset by increases in unemployment and sickness insurance take-up." (p. 1233)
"The evidence strongly suggests that most workers who stopped enrolling in DI at ages 55 and 56 simply delayed their DI enrollment until they reached the new threshold age of 57." (p. 1226)
My Take
The paper's cleanest contribution is the anatomy of spillovers: it demonstrates that DI, UI, and SI are close institutional substitutes, and that a reform reducing access to one will predictably shift caseload to the others. The "partial success" framing is honest and important — a naive reading of DI enrollment alone would overstate how much work the reform actually induced.
The transition analysis showing enrollment recovery at ages 58–59 is the most compelling piece of evidence for the postponement mechanism and has a direct policy implication: reforms that merely raise an age threshold rather than tighten medical criteria may shift costs across programs and time rather than generating permanent employment gains.
The Austrian context limits direct US generalizability: Austria's ~8% DI rate (vs. ~5% US) reflects a more generous eligibility standard and a structurally weaker social assistance alternative, which may amplify DI-as-early-retirement dynamics. The international employment response comparison (Austria > Sweden > US) rests on untested claims about marginal enrollee characteristics rather than direct verification.