A new study published in the Review of Economic Studies indicates that introducing additional welfare components to tax systems can make people worse off by changing their behavior. But certain policy interventions carry sizable work incentives and make people better off by helping them commit to long-run goals.
Researchers find that most people are time-inconsistent, meaning that they do not pursue the same goals as time changes. People also exhibit varying degrees of present-bias, giving stronger weight to immediate payoffs when making decisions.
Policymakers have long been concerned that social welfare programs carry incentives that can lead to dependence. Welfare programs can generate work disincentives, inducing people to work less. Moreover, if the work disincentive is strong, income can drop, making it difficult for recipients to ever get out of poverty.
Researchers used administrative data from a welfare reform experiment in the United States to estimate how severely people are subject to such dependence. Researchers considered how likely it was that people would pick a smaller reward because it came sooner over a larger reward that comes later.
Researchers proposed dynamic tax/sanctions and work subsidies to improve incentives. A dynamic sanction makes benefit rules stricter and saves the government money. These policies also induce sizable commitment-related work incentives and increase the amount people (especially those who are most present-biased) make for the limited work they do.
Researchers calculated that, in the baseline scenario, the average employment and welfare participation rates were 44.7 and 32.4 percent, respectively. Working welfare recipients constituted 30 percent of all welfare recipients, or 9.6 percent of all people in the control group. The average monthly earnings were $390.30 and the average monthly net government expenditure was $205.40 per recipient.
When people tend to focus on short-term payoffs they work less. There is a reduction in earnings and a small increase in government expenditure.
People have no commitment problem when they are time-consistent, and work dependably even when payoffs change. Researchers found that the average employment and welfare participation rates increased to 48.6 and 50.3 percent, respectively. Because people were no longer delaying employment and welfare use, the proportion of idle people reduced from 32.5 to 21.8 percent. As a whole, earnings increased slightly, and net government expenditure increased substantially, to $286.60 a person.
Earnings growth becomes somewhat slower when people are focused on initial payments rather than long-term goals. Researchers found that when people were time-consistent the initial utility was $61.2 lower than the baseline scenario due to the initial costs incurred by working welfare recipients. However, utility was higher than the baseline scenario in the long-run.
The results indicate that pro-work limits on welfare payments could improve employment meaningfully, and would have less radical effects on other outcomes. Researchers found that fewer people would leave welfare, and there would be a smaller effect on earnings and net government expenditure. The benefit reduction limit, which would reduce someone's benefits after a specified number of months, would have smaller effects than both the standard limit and pro-work limits assessed.
"This provides a comprehensive piece of evidence on the distribution of time preference in the welfare population," said author Marc K. Chan. "It identifies policies that are more effective in helping people achieve their long-run goals. The analysis also opens up an array of dynamic policy interventions for future research."
The paper "Welfare Dependence and Self-Control: An Empirical Analysis" is available at: https:/
Direct correspondence to:
Marc K. Chan
University of Technology Sydney, UTS Business School, Sydney, Ultimo NSW 2007, Australia
To request a copy of the study, please contact:
Daniel Luzerdaniel.email@example.com or 212-743-6113
Sharing on social media? Find Oxford Journals online at @OxfordJournals