How much work can the long-term disabled do?

The number of people who receive long-term disability payments has grown rapidly in many countries, raising the question of whether it is possible for some of them to return to gainful employment. Analysing the impact of a major disability reform in the Netherlands, this research finds that long-term recipients of disability insurance were on average able to replace about 60% of lost benefits with labour income. This is evidence of substantial remaining work capacity among recipients and it suggests that there is scope for disability insurance reform to tap into this capacity.

The number of recipients of disability insurance (DI) in the Netherlands rose tremendously in the 1980s, to the extent that one in every eight people of working age was receiving disability payments by 1990 (Koning and Lindeboom, 2015). This led the Dutch government not only to explore ways to reduce the inflow into DI but also to review the work capacity of existing DI recipients. In 1993, they passed a major reform that entailed re-examining the work capacity of all DI recipients under the age of 50.

The Dutch experience provides evidence of substantial remaining work capacity among long-term disability insurance recipients

Disability payments make up 70% of the last earned wage for fully disabled individuals in the Netherlands, but this percentage is adjusted downward for those who are partially disabled and therefore deemed to have remaining work capacity. In addition to re-examining the work capacity of existing DI recipients, the 1993 reform tightened the rules determining work capacity. The re-examinations often increased the work capacity that recipients were deemed to have and disability payments were reduced accordingly.

The re-examinations were done by age cohort, starting with those under the age of 35 in 1994. As recipients aged 44 and younger were being examined in early 1996, there was a political backlash against disability benefit cuts, especially for those considered ‘too old’ to be expected to return to work. In response to this pressure, the Dutch government decided in November 1996 that those aged 45 and older as of 1 August 1993 would be re-examined under the more lenient pre-reform rules for the determination of remaining work capacity.

Investigating the effects of a disability benefit cut


The 1996 decision creates an opportunity to investigate the effects of a disability benefit cut, making it possible to compare the outcomes for those who were just a bit younger than 45 and who therefore faced the stricter re-examination rules with the outcomes for those who were just old enough to escape the stricter rules.

Earlier studies provide evidence of labour supply responses to DI (for example, Autor and Duggan, 2003; Chen and van der Klaauw, 2008; Maestas et al., 2013). More recent studies begin to explore the degree to which existing long-term DI recipients can be induced to work (for example, Campolieti and Riddell, 2012; Marie and Vall Castello, 2012; Kostøl and Mogstad, 2014).

Our study adds to this emerging body of evidence using data from Dutch administrative records, in which we observe the labour earnings, DI payments and other social assistance payments of all disability recipients and their partners for many years following the reform.

Because none of the later disability reforms nor of any other policies were dependent on whether or not a recipient had turned 45 by 1 August 1993, we can attribute changes in outcomes that occur exactly at this age cut-off as the effect of applying the stricter re-examination criteria.

The impact of stricter re-examination rules on average benefits


Figure 1 shows that the stricter re-examination rules substantially lowered average benefits. Those just to the left of the cut-off (indicated by the vertical blue line) were subject to the stricter re-examination and received on average about €9,500 in disability benefits per year. Those who just escaped the stricter rules (just to the right of the blue line) received annual benefits averaging €10,600.

Figure 1: Stricter re-examination of those under the age of 45 reduced their DI benefits by €1,100 per year

borghans11

Notes: Each dot shows the average disability benefit in 1999 of all people with the same age (measured in years and months). Everyone to the left of the solid blue line was subject to the stricter re-examination. The solid red line is the line of best fit for those born 30 months before and after the age cut-off for the stricter re-examination. The dashed red lines represent 95% confidence intervals.

A major disability reform in the Netherlands led recipients born after a certain date to receive 10% lower benefits on average than those born earlier

Hence, the stricter rules reduced benefits on average by €1,100 per year or about 10%. About 70% of this reduction was due to lower benefits for people that remained eligible for some disability benefits with the remainder being due to those who became ineligible.

It is important to note that only a sub-group of DI beneficiaries experienced a benefit cut following a re-examination. Of those who were just under the age of 45 (and hence subject to the stricter re-examination), only about 29% experienced a cut in benefits. Still, this fraction is almost twice as high as the 16% that experienced a benefit cut among those who just escaped the more stringent re-examination. This means that for 13% of beneficiaries, the stricter re-examination led to a cut in benefits they would not have experienced under the more lenient re-examination. This group was 80% male and had on average been on disability for a little over a decade (and at least for three years) when the re-examinations took place. About 80% of them were married and just under half were considered fully disabled.

The re-examination increased benefits for those whose work capacity had declined substantially since their initial examination. Such benefit increases occurred for 12% of those who just escaped the stricter re-examination but for only 5% of those who were just subject to the more stringent re-examination.

Because not everyone was affected by the stricter re-examination, any consequences of the stricter re-examination must have been due to the sub-group that was affected – that is, those for whom a stricter re-examination led to lower benefits than the benefits that they would have received under the more lenient re-examination. We find that the stricter re-examination made a difference to the benefits of 20% of beneficiaries. Among this sub-group of affected beneficiaries, the stricter re-examination reduced benefits on average by €5,500 (€5,500/0.20) per year.

The impact of benefit cuts on recipients’ employment


Having found that the stricter re-examinations substantially reduced benefits, we want to know the beneficiaries’ response to this change. Did these cuts do nothing but reduce their standard of living? Or were beneficiaries able to offset these cuts by increasing their own labour income or by finding other sources of social assistance?

Figure 2 shows that the benefit cut for those under the age of 45 increased their employment rate by about three percentage points in 1999, which is roughly two years after the re-examinations took place. The employment rate of those just under the age of 45 was 38% whereas only 35% of those who just escaped the stricter re-examination were employed.

Figure 2: Stricter re-examination of those under the age of 45 increased their employment rate

borghans2

Notes: Each dot shows the average fraction employed in 1999 of all people with the same age (measured in years and months). Everyone to the left of the solid blue line was subject to the stricter re-examination. The solid red line is the line of best fit for those born 30 months before and after the age cut-off for the stricter re-examination. The dashed red lines represent 95% confidence intervals.

Given that the stricter re-examination only made a difference for 20% of beneficiaries, it must be this sub-group that was responsible for the three percentage point increase in employment. This means that among this sub-group of affected beneficiaries, the stricter re-examination increased employment by about 15 (3/0.20) percentage points. Another way of interpreting the magnitude of the employment response is by noting that the three percentage point increase in employment was brought about by an average benefit cut of 10%.

The impact on earnings


The increase in the employment rate shows that some beneficiaries were able to return to paid employment. But it does not reveal anything about the level of earnings in these jobs nor does it capture increases in the earnings of beneficiaries who were already working.

Figure 3 shows that the stricter re-examination increased annual earnings by about €625 or about 10%. This means that disability recipients increased their earnings by about €0.60 for each €1 that their benefits were cut. Given that the increase in average earnings of €625 was brought about only by the 20% for whom the stricter re-examination led to lower benefits than a regular re-examination, the earnings increase per affected beneficiary was much larger – about €3,125 (625/0.20) per year.

Figure 3: Stricter re-examination of those under the age of 45 increased their annual earnings by €625

borghans3

Notes: Each dot shows average annual earnings in 1999 of all people with the same age (measured in years and months). Everyone to the left of the solid blue line was subject to the stricter re-examination. The solid red line is the line of best fit for those born 30 months before and after the age cut-off for the stricter re-examination. The dashed red lines represent 95% confidence intervals.

This implies that many of the affected beneficiaries had substantial remaining work capacity and that the reform induced them to harness this capacity. We also examine whether partners of the affected beneficiaries changed their labour market behaviour in response to the benefit cuts, but do not detect significant responses.

Increased reliance on social assistance programmes made up for about 30% of lost disability benefits


Effects on other social assistance programmes


While lower disability benefits induced some recipients to increase their earnings, it also led some to increase their income from other social assistance programmes. As Figure 4 shows, those just to the left of the age cut-off and subject to the stricter re-examination collected about €300 per year more in benefits from other social assistance programmes than those just to the right of the cut-off. In other words, disability recipients on average were able to offset about 30% of lost disability benefits by collecting more from other social assistance programmes.

Figure 4: Stricter re-examination of those under the age of 45 increased their income from other social assistance programmes

borghans4

Notes: Each dot shows average annual income from social assistance programmes other than disability of all people with the same age (measured in years and months). Everyone to the left of the solid blue line was subject to the stricter re-examination. The solid red line is the line of best fit for those born 30 months before and after the age cut-off for the stricter re-examination. The dashed red lines represent 95% confidence intervals.

This implies that about 30% of the money that the Dutch government intended to save through stricter disability re-examinations instead ended up being spent on higher outlays on other social assistance programmes. These findings are in line with earlier studies showing the existence of ‘spillover effects’ between different social assistance programmes in other contexts (for example, Karlström et al., 2008; Staubli, 2011; Inderbitzin et al., 2016).

Long-term disability insurance recipients offset benefit cuts mostly by increased labour earnings


Offsetting disability benefit cuts with income from other sources


A remarkable implication of our findings is that disability recipients were able to offset around 90% of their reduction in benefits by increasing other sources of income, two thirds of which was labour income. Thus, on average, the reform did not reduce beneficiaries’ income by much, though this average masks some people experiencing declines and others experiencing increases in total income.

We find that this general pattern not only holds in 1999 (about two years after the re-examinations), but for all years until 2005 (the last year examined). If anything, the fraction of lost disability benefits replaced by labour income increased somewhat over time and the fraction replaced by other social assistance decreased somewhat over time.

These findings are generally similar across demographic sub-groups of DI recipients. For example, even those who were deemed fully disabled were able to offset just over half of their lost benefits with increased labour income and make up the rest with increased income from other social assistance programmes. The same holds when we limit our analysis to the half of the sample that had the longest duration of disability benefit receipt, namely at least eight years at the time of the re-examinations.

Our exact estimates of the degree to which people were able to offset disability benefit cuts by increasing income from other sources are obviously specific to this particular reform. In particular, the reform cut benefits only for those for whom the re-examination procedure indicated that they should be able to work more. In addition, it might have been easier in the Netherlands than elsewhere to increase social assistance income, as the country has a relatively generous system of alternative programmes.

General lessons


We believe that our research offers two general lessons that are widely applicable. First, it provides strong evidence that spillover effects between social assistance programmes can be substantial, including for prime-aged individuals. Thus, any analysis of social assistance reforms would be well advised to consider the possibility of benefit substitution between programmes.

Second, we show that long-term DI recipients may still have substantial capacity to change their labour income in response to changes in DI generosity. In other words, labour supply among existing DI recipients is not just determined by limitations from disability, but also by economic incentives.

Policies that affect the behaviour of the large stock of existing beneficiaries have a more immediate impact than policies that affect the inflow of DI recipients over time. Hence, policies that focus on harnessing the remaining work capacity of existing DI recipients are a promising avenue for a relatively rapid reduction in disability expenditure.

This article summarizes ‘Social Support Substitution and the Earnings Rebound: Evidence from a Regression Discontinuity in Disability Insurance Reform’ by Lex Borghans (Maastricht University), Anne C. Gielen (Erasmus University Rotterdam) and Erzo F.P. Luttmer (Dartmouth College), published in the American Economic Journal: Economic Policy in 2014.

Further reading

Autor, David H., and Duggan, Mark G (2003) ‘The Rise in the Disability Rolls and the Decline in Unemployment’, Quarterly Journal of Economics 118(1): 157-206.

Campolieti, Michele, and Chris Riddell (2012) ‘Disability Policy and the Labor Market: Evidence from a Natural Experiment in Canada, 1998-2006’, Journal of Public Economics 96(3-4): 306-16.

Chen, Susan, and Wilbert van der Klaauw (2008) ‘The Work Disincentive Effects of the Disability Insurance Program in the 1990s’, Journal of Econometrics 142(2): 757-84.

Inderbitzin, Lukas, Stefan Staubli, and Josef Zweimüller (2016) ‘Extended Unemployment Benefits and Early Retirement: Program Complementarity and Program Substitution’, American Economic Journal: Economic Policy 8(1): 253-88.

Karlström, Anders, Mårten Palme, and Ingemar Svensson (2008) ‘The Employment Effect of Stricter Rules for Eligibility for DI: Evidence from a Natural Experiment in Sweden’, Journal of Public Economics 92(10-11): 2071-82.

Koning, Pierre, and Maarten Lindeboom (2015) ‘The Rise and Fall of Disability Insurance Enrollment in the Netherlands’, Journal of Economic Perspectives 29(2): 151-72.

Kostøl, Andreas R., and Magne Mogstad (2013) ‘How Financial Incentives Induce Disability Insurance Recipients to Return to Work’, American Economic Review 104(2): 624-55.

Maestas, Nicole, Kathleen Mullen, and Alexander Strand (2013) ‘Does Disability Insurance Receipt Discourage Work? Using Examiner Assignment to Estimate Causal Effects of SSDI Receipt’, American Economic Review 103(5): 1797-1829.

Marie, Olivier, and Judit Vall Castello (2012) ‘Measuring the (Income) Effect of Disability Insurance Generosity on Labour Market Participation’, Journal of Public Economics 96(1-2): 198-210.

Staubli, Stefan (2011) ‘The Impact of Stricter Criteria for Disability Insurance on Labor Force Participation’, Journal of Public Economics 95(9-10): 1223-35.