Research
Research
Disability Benefit Reform, Employment and Welfare: Evidence from Australia’s Welfare-to-Work Reforms
with Chung Tran
Presented at: Australian Conference of Economists (ACE), University of Canberra, 2026
Upcoming: 40th PhD Conference in Economics and Business, The University of Sydney, 2026; 11th Workshop of the Australasian Macroeconomics Society (WAMS), University of Tokyo
Abstract. Tighter disability-benefit eligibility can generate opposing labour-supply responses: recipients may reduce hours to retain support or increase work when additional earnings are more valuable than continued benefits. We study Australia’s Disability Support Pension (DSP), a non-contributory, means-tested payment whose eligibility turns on an hours-based work capacity test rather than on earnings history, and whose recipients have one of the lowest employment rates in the OECD. The 2005 Welfare-to-Work reform lowered the work-capacity threshold from 30 to 15 hours per week from July 2006. Using linked administrative records and a staggered difference-in-differences design, we find that earnings fall upon exposure and employment declines subsequently. Continuing recipients experience persistent losses, while those who exit initially earn more but their gains diminish over time. A calibrated lifecycle model with disability risk and means-tested benefits clarifies these competing incentives. At benchmark prices and unchanged taxes, the tighter hours rule reduces labour supply, lifetime consumption, and entrant welfare. Removing the income test encourages work among severely disabled households and raises welfare in every group, but at a fiscal cost, whereas reducing the maximum payment encourages employment and lowers expenditure but reduces welfare in every group. Our findings show that tighter eligibility need not promote employment and can reduce income protection against disability-related earnings losses.
Disability, Labour Productivity and Lifetime Earnings: New Evidence from Australia
with Chung Tran · PDF
Award: Best Paper Prize, HILDA Survey Research Conference, University of Melbourne, 2026
Presented at: Tax and Transfer Policy Institute (TTPI) Seminar, ANU, 2025; Asia-Pacific Productivity Conference, The University of Queensland, 2025; 2nd Workshop of the East and South-East Asian Macroeconomic Society (WESEAMS), Asia School of Business, Kuala Lumpur, 2026; HILDA Survey Research Conference, University of Melbourne, 2026
Upcoming: 2026 Econometric Society Australasia Meeting (ESAM), Adelaide University
Abstract. Disability can lower wages by reducing the value of an hour of work (labour efficiency) or the return to accumulated experience (experience efficiency). Using 24 waves of Australian panel data, we find that moderate disability lowers labour efficiency by 12% and severe disability by 17%. Only severe disability lowers experience efficiency, by 23% on average and 32% for workers without a university degree. Moderate disability therefore shifts the wage profile down, while severe disability also flattens it. In a life-cycle model with learning by doing, these losses compound because less work today means less experience and lower wages later. Severely disabled workers earn 59% to 86% less over a lifetime than non-disabled workers with the same education, several times the wage penalty in any single year. Learning by doing also makes the earnings response to removing disability income support about 40% larger for severely disabled workers without a degree.
Disability, Social Insurance and Employment: Facts and Correlates from Australia
with Chung Tran
Presented at: 5th Annual RSE PhD Workshop, ANU, 2023; Tax and Transfer Policy Institute (TTPI) Seminar, ANU, 2024; Women in Macroeconomics Workshop IV, UNSW Sydney, 2025
Abstract. The Disability Support Pension (DSP) is Australia’s main income support payment for working-age people with disability. This paper assembles two decades of household survey data—the Household, Income and Labour Dynamics in Australia (HILDA) survey and the Survey of Income and Housing (SIH)—together with administrative records from the Person Level Integrated Data Asset (PLIDA), covering 2000 to 2025, to document how DSP receipt, benefit payments and labour market outcomes evolve over time and over the life cycle. Two empirical exercises complement the descriptive facts: a logistic regression identifying demographic predictors of DSP entry, and a difference-in-differences analysis of working hours around DSP entry and exit. Four findings stand out. First, the inflow of new recipients has fallen sharply since the eligibility reforms of the early 2010s, while average benefits among continuing recipients have grown. Second, non-employment is pervasive among beneficiaries, whose income relies overwhelmingly on public transfers—88% of aggregate income on average. Third, health, labour force status and marital status are the demographic characteristics that most sharply distinguish recipients from non-recipients. Fourth, working hours drop upon DSP entry and rebound after exit, and partners of entering recipients also reduce their hours, consistent with the program’s means tests and work-capacity rules shaping labour supply.