Evidence on the Negative Income Tax

Ben Eisen, Blog, Canada, Taxation, Uncategorized, Welfare

Carol Goar’s op-ed in The Toronto Star today discusses the negative income tax. The negative income tax is basically a policy proposal to create a guaranteed annual income for all households. The government would set a minimum threshold that it wants everybody to have, and then “top up” the earnings of everybody below the line. So if the guaranteed income was set at $15, 000 and you earned $12, 000 in a year, the government would give you a $3,000 cheque. It’s a compelling idea that has won adherents on the left and right over the decades. It appeals to many libertarians who feel that it would be better to put money currently spent on social programs directly in the hands of poor people who could better decide for themselves how to spend the money to improve their own lives. Goar writes that a pilot program along these lines was initiated in Dauphin, Manitoba in 1974 before being terminated in 1978. The author notes that the program apparently had a number of positive effects, but that a detailed evaluation can’t be undertaken until the data is digitized. Most importantly, Goar writes: “contrary to policy-makers’ fears, people in Dauphin did not stop working or reduce their hours to get “free” money from the government.” Well, OK, maybe, but readers of Goar’s column should be aware that previous experiments conducted in the United States have shown evidence of work disincentives from negative income taxes. There’s a reason the idea went out of vogue after enjoying some popularity in the 1970s – most of the experiments and pilot projects showed significant effects on work effort. Carefully designed experimental research (the gold standard of social science research) showed in several different instances that workers in areas in which an NIT was implemented worked, on average, fewer hours per week than members of control groups. By all means, we should look into the Dauphin experiment pointed out by Goar – but even if the experiment provides no evidence of work disincentives, we shouldn’t conclude that the matter is settled. There is enough evidence for the common-sense proposition that NITs discourage work effort that we will need more than one study (even a good one) to overturn the conventional wisdom.