News from
PRINCETON UNIVERSITY
Office of Communications, Stanhope Hall
Princeton, New Jersey 08544
Tel 609/258-3601; Fax 609/258-1301
FOR IMMEDIATE RELEASE
Contact: Mary Caffrey 609/258-5748
Date: September 29, 1997
Princeton Economist Finds Lotteries
Helped by a Good CausePrinceton, N.J. -- If you've ever watched the nightly televised drawing for the Pennsylvania Lottery, you've probably heard the announcement, "Lottery proceeds benefit older Pennsylvanians."
The statement may sound like simple public relations, but research by Assistant Professor of Economics and Public Affairs John Morgan shows why there's much more to it than that.
In a paper submitted to the Review of Economic Studies, Morgan and Martin Sefton of the University of Manchester have tested Morgan's theory that lotteries can successfully finance public services not only because they tap participants' love of gambling, but also because a good cause helps draw bettors.
The theory evolved from Morgan's interest in auctions. Auctions exists in nature, he says: for example, when there is competition for food, and everyone pays something in a "war of attrition. In a lottery, everyone pays some price to improve his or her chance of winning, just as animals do."
School raffle
Morgan's initial treatment of lotteries as auctions was modified by an observation made by his wife, who was helping conduct a school raffle. "She said that people weren't buying the tickets to win the prize, although they would enjoy winning. They were buying tickets to raise money for the school. I realized if I didn't include the public good motive, then I was missing the point," Morgan said.
Through a series of experiments at the Pennsylvania State University and the University of Iowa, Morgan showed that lotteries raised more money for public goods than voluntary contributions, that the lotteries with the largest prizes raised the most money, and that ticket purchases fell if potential bettors placed little value on the public good to be funded by the lottery.
In the Penn State experiment, pairs of students were divided into groups; one group made bets in a "lottery," while the other took part in a "voluntary contribution mechanism," which roughly corresponds to a request for donations for some public good (such as pledges to support a public television station). In both groups, students were given a pot of tokens, and a point system determined how much each student would be paid at the end of the session.
Students in both groups were then asked to make a series of contributions to a pool, representing the public good. All knew they would receive 100 points for each token they kept for themselves, while each token contributed to the pool would bring a student and his or her partner 75 points apiece.
For example, if both partners gave one token each in a given round, both got 150 points; if one partner gave one token and the other gave nothing, the partner giving a token received only 75 points, while the other person would get a "free ride" and 175 points.
In addition, students taking part in the voluntary contribution mechanism knew that their pool had been augmented with 600 bonus points for each participant -- which the students would receive even if they contributed nothing. In the lottery, students knew they would receive one lottery ticket for each token they contributed to the pool and that one of them would win the lottery prize of 800 bonus points.
In both cases, students as a group stood to gain points by placing all their tokens in the pool. However, each student also had an individual incentive to withhold tokens and get a "free ride" from other contributors. In the voluntary mechanism, students gave half or fewer of their available tokens, and donations dropped off in later rounds. In the lottery, students bet slightly more than half their tokens at the start of the session, and by the end were giving more than three-quarters of their available tokens. Thus, the results highlighted how lotteries can overcome the free-rider problem.
Experiments with students at the University of Iowa produced similar results. In addition, Morgan tested the effect of a lottery prize of 1,600 bonus points. Contributions increased, and so did the amount of "public good" provided, even after netting out the cost of the prize. By contrast, when students were told there would be no group account of bonus points -- and, thus, no "public good" -- bets fell below 15 percent of available tokens. Morgan and Sefton write that this last test shows the importance of the link between lotteries and public goods in understanding betting.
Morgan acknowledges that the increased use of gambling as a way to raise funds for education or the elderly has come under fire. Critics charge that low-income bettors do not understand the level of risk they face and thus give a disproportionate share of their income to lotteries.
"When we look only at the equity issues, lotteries are viewed as regressive," Morgan says. "My view is that we need to step back and ask, 'What are the efficiency properties?'"
He argues that in public good settings, and especially when financed privately (such as church bingo), lotteries can bring in funds for needed goods and services that would otherwise not be raised. In this context, he said, "Lotteries can be a good thing."
Lotteries complement taxes
In their paper, Morgan and Sefton point to a political reality faced by those who must provide social services. "Governments are increasingly turning to lotteries to complement tax-based forms of revenue generation. In the United States, under the auspices of state governments, lotteries are frequently proposed in response to de facto or de jure restriction on increases in taxation. In short, lotteries are frequently employed as a voluntary mechanism in lieu of or in addition to other confiscatory mechanisms."
Morgan said his findings have several implications, including the fact that states and charities can make more money by offering a fixed prize rather than conducting a pari-mutuel or "share of the pie" event.
He is also investigating whether long-term dependence on lotteries can have a detrimental effect on the services that receive proceeds. With most states earmarking lotteries for a popular purpose, such as education, Morgan wants to examine whether lotteries lead to a withdrawal of the tax dollars that initially supplemented the proceeds.
0929-lotteries.html