(This was originally written as a college essay. For this reason it is 3,000 words long and contains a lot of economics terms references. It is meant as a help for anyone who wants an in-depth study of the topic with supporting facts and figures. For a short version in plain English see here. The essay was marked the best in my class of 80 students. It won the Undergraduate Award 2012 in Economics along with someone from Yale. I’m not trying to boast but I’m over the moon.)
This paper discusses the conventional wisdom that unemployment benefits create a disincentive to work, the so-called “welfare trap”. It examines if higher benefit levels or longer benefit durations lead to higher unemployment rates. Surprisingly, it finds that the disincentive effect is negligible to non-existent. A wide range of different studies have found little or no disincentive effect. The orthodox economic view which states that workers will quit their jobs if they will receive more money from unemployment benefits has been found to be naive and unrealistic. Despite the large number of people who believe it to be true, it simply is not supported by evidence.
The reason for this is that a job is not simply a way to make money, it is also part of an individual’s identity. Unemployed workers suffer psychological damage from their lack of a job. They suffer from a higher rate of mental illness than those working and gain a boost in mental wellness when re-employed. The orthodox theory does not take into account the fact that many workers gain self-fulfilment from their work. There is no evidence that they suffer from a lack of work ethic either.
The traditional theory ignores inconvenient facts such as the fact that many unemployed workers are not eligible for unemployment benefits or the full range of allowances. Many allowances apply only to families, whereas most of the unemployed are childless. These findings apply both in Ireland and abroad. In fact, a study found that 97% of those unemployed in Ireland in 2011 would improve their financial situation by getting a job. In most countries it is necessary to have worked a certain length of time before it is possible to receive benefits, this would actually cause benefits to have a positive rather than negative impact on participation in the labour force.
Comparisons across countries find no link between generous benefits and high unemployment. In fact, the countries with the most generous benefits (Sweden, Denmark, Norway, and the Netherlands) have some of the lowest levels of unemployment.
This paper concludes that, after taking the wealth of studies available on the topic, generous welfare benefits do not lead to high unemployment.
Do Generous Welfare Benefits Lead To High Unemployment?
Conventional wisdom holds that if welfare rates are too high, people will simply stop working and go on the dole. In fact it has become a cliché to speak of people too lazy to work receiving unemployment benefits or single mothers having children solely to receive an allowance. It is seen as such an obvious statement that few question it. This has spread to economics where it is stated as fact that people will go on welfare if the rates are high enough. But when you get out of the political rhetoric and stereotypes, there is surprisingly little evidence to support such claims. Few studies have found a link between welfare rates and unemployment and those that have, found correlations much smaller than classical economics would have us believe.
The classical or orthodox view is quite straight forward. If unemployment benefits are too high people will not have an incentive to get a job. It commonly stated that “higher benefit duration and replacement rates tend to raise unemployment.” (Siebert 1997:52) “High unemployment is associated with . . . generous unemployment benefits that are allowed to run on indefinitely, combined with little or no pressure on the unemployed to obtain work.” (Nickell 1997:72) Ljungqvist and Sargent too argue that “unemployment insurance tends to increase the unemployment rate”. (Ljungqvist & Sargent 1995:143) It has been argued that “higher levels of state welfare benefits are associated with longer spell duration. Except for the dilemma of needing both to create appropriate incentives and protect the “truly needy”, the solution would be simple. Reduce benefits and duration will decline.” (O’Neill etc. 1987:248)
These studies have been criticized for being based on “a narrow set of studies that rely on data many decades old.” (Howell & Azizgolu 2011:14) The classical theory has been derided as “A naive model inappropriate for the assessment of the real world incentive effects of labour market policies.” (Howell & Azizgolu 2011:23)(Atkinson & Micklewright 1991:1688) They have been described as “very sensitive to rather strong assumptions. Generally they reveal what they assume, not what the data reveal.” (Blundell 1994:35) The robustness of the data has been challenged, with claims that its assumptions are flawed. When previous work is adjusted to be more realistic, “its effect ceased to be significantly different from zero.” (Atkinson & Micklewright 1991:1711)
The classical view is countered by the claim that unemployment insurance may actually increase the incentive to work. This is because unemployment insurance requires a person to have worked a certain amount of time before they can claim benefits. (Howell & Rehm 2009:68) “This illustrates the fact that UI may have positive as well as negative effects on the transition from unemployment to employment.” (Atkinson & Micklewright 1991:1699) (Nickell 1997:68) Even one of the most famous conservative economists Milton Friedman stated this point in his Nobel Lecture, when he said that “unemployment insurance makes it more attractive to enter the labor force.” (Friedman 1970:7)
A fact that is often overlooked is that only a fraction of those who are unemployed actually receive benefits, One study showed that only 30% of those classed as unemployed were receiving benefits in America. The figure was 26% for the UK. (Atkinson & Micklewright 1688-9) (Howell & Rehm 2009:64) This means that although welfare rates may appear to be high, in practice they may be much lower. (NESC 2011:119) When discussing the orthodox theory that high benefits cause high unemployment, Howell and Rehm state “there is scant empirical support for such demand side effects . . . This should not be surprising, since, as noted below, in many countries, substantially less than half of the unemployed actually get benefits, and among those that do, most get income for limited durations that is far below previous earning levels. “ (Howell & Rehm 2009:63)
Many theoretical studies ignore features about actual benefits systems, such as the fact those who voluntarily quit their job, often are not entitled to benefits. Similarly benefits are denied to those who are not available for work or who refuse suitable job offers. (Atkinson & Micklewright 1991:1721) (Howell & Azizgolu 2011:9) Similarly, although some benefits may be generous, many claimants may not qualify for it. While there is a lot of commentary on the benefits families receive for each child they have, they are not representative of the unemployed as a whole. In fact, 80% of people on the live register in Ireland do not have any children, while 56% of them are single. (NESC 2011:119) This leads the NESC to conclude that “the degree of attention given to the potential disincentive effects of social welfare on ‘large families’ in the 1980s would be quite disproportionate today.” (NESC 2011:120) Similarly while rent supplement is often seen as having a large disincentive effect, only 11% of those receiving unemployment were receiving rent supplement. (NESC 2011:123)
They challenge the conventional belief that higher benefits lead to longer duration spells. In their review of the literature, they find that most estimation calculate a very small effect. Most studies estimate that 10% rise in the replacement rate would only reduce the average unemployment duration by one week. (Atkinson & Micklewright 1991:1711) They note that “this sort of evidence has important implications for policy suggesting, in the case of the U.K., that income support for most of the long-term unemployed may be increased to a higher level without concern for incentives.” (Atkinson & Micklewright 1991:1713) In fact the level of benefits has almost no effect on unemployment. Instead it is the maximum amount of time a claimant can draw benefits. It has been estimated a cut in the maximum length is twice as effective as a cut in benefit rates. Even still the effects are quite small. An increase in maximum benefit duration lengths of one week only increases average duration by between 0.10-0.15 weeks. (1991:1717) After reviewing previous studies, they conclude that there is “no strong evidence that benefits have much effect on inflows,” and that the evidence suggests that “benefits affect inflows into unemployment less than outflows.” (1991:1715) Interestingly a rise in unemployment is usually caused by fewer people finding jobs rather than more people losing theirs. (Howell & Azizgolu 2011:5-6)
The correlation between the maximum allowable duration of benefits and average duration is “extremely weak.” (Howell & Rehm 2009:69) Examination of a study found that while there was a positive link between replacement rates and unemployment, the effect was slight to negligible. Increasing the replacement rate by 10% would increase unemployment by between 0.1 and 1.3 points. (Howell & Rehm 2009:70) Figure 2 shows no correlation between the change in benefit level and the change in unemployment. In response to the OECD claim that there is a lag of 5-10 or even 10-20 years, figures 4 and 5 dispute this and show that there is no correlation. Other studies “show little evidence of causality running in the orthodox direction, from benefits to unemployment.” In some cases, “the results clearly indicate that the predicted effect run from unemployment to benefit generosity.” (Howell & Rehm 2009:79-80) A study of each individual OECD country over several decades found no correlation between its replacement rate and its unemployment level. (Howell & Rehm 2009:80) Figure 6, also shows there is little or no link between a country’s replacement rate and its unemployment level. Long potential duration of benefits, also fails to explain high unemployment and in fact the correlation works in the opposite direction. (Howell & Rehm 2009:85) In 2001 the OECD introduced a new data measurement the NRR. This too is negatively correlated to unemployment in figure 7, but in the short and long run. In fact it is also positively correlated with employment-population ratios. “Countries with increasingly generous replacement rates and benefits durations appear to be associated with increasing employment rates.” (Howell & Rehm 2009:86-7)
The recipiency rate is measure of how many of the unemployed actually receive benefits. It has been argued that the more ‘generous’ or ‘lax’ a welfare system is, the higher the rate of unemployment. However, data finds a strong negative correlation between it and unemployment and a strong positive correlation with overall employment. “These results are not consistent with the orthodox prediction that such generosity substantially undermines employment incentives.” (Howell & Rehm 2009:86) Figure 9 shows that “increasing access to benefits by the unemployed, at least as measured by the OCED’s recipiency rate, is not associated with increasing unemployment over the last two decades.” (Howell & Rehm 2009:88)
A fact about welfare that is ignored by orthodox theorists is that there is a stigma attached it. Many people refuse a hand out on principle, regardless of their financial situation. While this is often seen as a positive thing as it discourages people from going onto benefit rolls, it has been shown that it can have negative effects on unemployment as employers can hold biases against unemployed people and may be less willing to hire them. The unemployed themselves get more discouraged the higher the stigma levels are. (Contini & Richiardi 2009:25-6) Surprisingly they find that “welfare stigma is positively related to poverty and unemployment.” (Contini & Richiardi 2009:8)
It has been calculated that the replacement rate in Ireland in 2007 was 60%. (Callan etc 2011:8) “About three-quarters of the recipients of Jobseeker’s payments face a replacement rate of less than 60 per cent and over half face a replacement rate of 50 per cent. At the other end of the scale, just over 3 per cent face a replacement rate of more than 100% (i.e., would receive more net income when unemployed than when in work).” (Callan etc 2011:11-2) In other words, 97% of the unemployed would increase their income by getting a job. This completely contradicts the common notion that people receive more money with unemployment benefits that they would by working. The authors note that there often claims made that there is a weak “financial incentive to move from unemployment into employment , and selective examples have been used to support this argument. We showed how such examples can be misleading, failing to take into account the range of factors affecting both benefit entitlements and potential earnings in work.” (Callan etc 2011:17)In an Irish context, Layte and Callan did find a link between the replacement rate and the rate of people leaving welfare, though it was extremely small at –0.005. They admit that the disincentive effect they found is “very small in comparison to those found in other national contexts.” (Layte & Callan 2001:125) They conclude that: “it is interesting that most media, government and academic attention given to the question of disincentive effects is directed at the more disadvantaged portion of the unemployed who tend to receive means tested benefits and who show no sign of disincentive behaviour in this data.” (Layte & Callan 2001:125)
In 1987 the General Accounting Office (the agency responsible for auditing government spending in America), issued a report which examined over 100 separate studies on welfare since 1975. (GAO 1987:11) It concluded that “research does not clearly support the contention that welfare creates a disincentive to work.” (GAO 1987:24) O’Neill, Bassis and Wolf state that “The common notion that welfare typically becomes a state of dependency is not supported by these statistics.” (O’Neill, Bassis & Wolf 1987:244) “The size of the disincentive effects that have been established in empirical research for replacement rates is quite modest and suggests that only a quite large reduction in social welfare, with major social consequences, would make a significant contribution to reducing high unemployment.” (NESC 2011:127) They dismiss the idea that people quit working in order to receive unemployment benefits as in both the UK and the USA people who quit their jobs are a minority of the unemployed. (Atkinson & Micklewright 1991:1715) “It has to be concluded that across a wide range of studies the effect of replacement rates on return to work probabilities has been found to be quite small. . . Given the results above, it is difficult to believe that entry into unemployment for prime age men could be significantly affected by replacement rates.” A study found “no significant effect of benefit income on the length of unemployment spells.” (Blundell 1994:33) “Most benefit systems offer income replacement at levels well below the average income, so none but the lowest wage workers receive anything close to their previous earnings level. In addition, eligibility rules exclude many of the unemployed in most countries.” (Howell & Rehm 2009:63)
They argue that strict tightening of unemployment can lead to workers accepting lower paid work, or jobs that do not utilise their talents. This leads to a less efficient economy in the long run. (Howell & Azizgolu 2011:13) Higher benefits can insure that workers such longer and find a job they are more suited to, rather than accepting the first offer. This can increase labour productivity. (NESC 2011:129) “Respondents are using the resources provided by benefits for more effective job search and thus a better more stable job.” (Layte & Callan 2001)
Financial considerations are not the only reasons why people work. “Many people at work value the social contract their job brings while many unemployed experience isolation and a lack of structure to their lives. . . Positive well being effects are associated with being in work, while strong ill-effects unambiguously attend being unemployed.” (NESC 2011:126) People are often defined by their job, so that unemployment robs them of part of their social and personal identity. (Jackson 1995:119) “Employment offers the best prospects for meaningful activity which individuals find inherently satisfying: without it most of us feel deprived.” (Bryson 1994:123) Surveys show that most people do not see a job solely as a means of making money and would work even if they did not need the money. For them financial incentives are irrelevant. (Howell & Azizgolu 2011:11) (Bryson 1994:126) “The evidence overwhelmingly shows that in the real world holding a job is typically highly valued independently of the income it generates, and if this is so, changes in the UI generosity may have nothing to do with changes in the disutility to work.” (Howell & Azizgolu 2011:9)
Classical theory states that the unemployed should have an increase in life satisfaction as they have more lesure time and are free from the stress of work. Howver, this is contradicted by a wide field of studies which demonstrate that unemployment causes a lack of self-esteem and feelings of helplessness. (Carroll 2007:288-9) Classical theory ignores the fact that not working is breaking a social norm, which causes a loss of reputation. (Carroll 2007:289) A cross section of 52 seperate studies showed that the unemployed suffered from poorer mental health, lower life satisfaction, lower martial or family satisfaction and poorer physical health.” (Kinicki 2005:61) A survey of 15 studies shows that reemployment caused “significant improvements” in mental health, life satisfaction and physical health. (Kinicki 2005:61) (Carroll 2007:293) While there are questions regarding correlation and casaulity, there is a large amount of consistenty among studies finding links between unemployment and lower well being. (Kinicki 2005:67) “Because individuals with high work-role centrality find the work role as providing meaning and fulfilment, the absence of work for those individuals has been proposed by many authors to at lead to lower psychological and psychical well being. (Kinicki 2005:56) Interestingly, it found that the actual level of benefits has no effect on mental health. This implies it is the absence of work rather than the loss of income that is most damaging. (Kinicki 2005:67) An unemployed man would have to be compensated with 42,100 Australian dollars (almost 33,500 euro) in order to bring his life satisfaction to the same level of an employed man. This difference is even larger for women who would have to receive 86,300 Australian dollars (68,500 euro). (Carroll 2007:298) Similar findings regards life satisfaction and income compensation have been found in studies in other countries, with unemployed men in the UK particularly low levels of life satisfaction 69% lower than those of the unemployed. (Carroll 2007:300)
Despite the conventional wisdom, studies have found that the unemployed do not suffer from a lack of work ethic. No evidence was found supporting the idea that it is lack of motivation or work ethic that prevented the unemployed from getting a job. (Jackson 1994:114) In “most cases the probability of accepting a job offer is close to unity.” (Atkinson & Micklewright 1991:1712) While economists focus on the large disincentives having children are supposed to have on working. However surveys show that it is the unemployed that have children that are the most motivated to get a job. (Jackson 1994:112)
After drawing together different studies, viewpoints and opinions it is clear that the disincentive effect that is supposed to affect the unemployed is greatly exaggerated. The conventional wisdom is simply not supported by facts. In fact most available data contradicts it. The orthodox model fails to give an accurate description of the real world, instead relying on overly simplistic assumptions. It ignores features of welfare systems that prevent people from receiving assistant, such as means tests and work tests. It ignores the stigma associated with receiving hand-outs and the psychological damage unemployment does. It glosses over the non financial benefits people gain from work. Contrary the orthodox opinion, generous welfare benefits do not lead to higher unemployment, longer durations of unemployment or a disincentive to work.
Atkinson, A. and Micklewright, J., (1991) “Unemployment Compensations and Labor Market Transitions: A Critical Review”, Journal of Economic Literature, Vol. 29, No.3, 1679-1727
Blundell, R., “Work Incentives and Labour Supply in the UK”, in Is It Worth Working, (1994), Bryson, A., & McKay, S. (eds), 19-38, Policy Studies Institute, London
Bryson, A., “Is It Worth Working?” in Is It Worth Working, (1994), Bryson, A., & McKay, S. (eds), 122-127, Policy Studies Institute, London
Callan, T., (ed) Crilly, N., Keane, C., Walsh, J., (2011) Tax, Welfare and Work Incentives, Budget Perspectives 2012, ESRI, Dublin
Carroll, N., (2007) “Unemployment and Psychological Well-Being” in The Economic Record, Vol.83, No.262, 287-302
Contini, D. and Richiardi, M., (2009) Reconsidering the Effects of Stigma on Unemployment
Friedman, M., (1970), Nobel Lecture: Inflation and Unemployment
Howell, D. And Azizoglu, B., (2011) “Unemployment Benefits and Work Incentives: The US Labor Market in the Great Recession.” Political Economy Research Institute, Working Paper No.257
Howell, D. and Rehm, M., (2009) “Unemployment Compensation and High European Unemployment: A Reassessment With New Benefit Indicators”, Oxford Review of Economic Policy, Vol.25, No.1, 60-93
General Accounting Office, Welfare: Issues to Consider in Assessing Proposals for Reform, February 1987, Washington
Jackson, P., “Influences on Commitment to Employment and Commitment to Work”, in Is It Worth Working, (1994), Bryson, A., & McKay, S. (eds), 110-121, Policy Studies Institute, London
Kinicki, A., McKee-Ryan, F., Song, Z., Wanberg, C., (2005) “Psychological and Physical Well-Being During Unemployment: A Meta-Analytic Study”, in Journal of Applied Psychology, Vol. 90, No.1, 53-76
Layate, R. And Callan,T.,(2001) Unemployment, Welfare Benefits and the Financial Incentive to Work, in Economic and Social Review, Vol. 32, No.2, 103-129
Ljungqvist, L. and Sargent, T., (1995) “Welfare States and Unemployment”, in Economic Theory, Vol. 6, 143-160
Ljungqvist, L. and Sargent,T., (1998) “European Unemployment Problem”, in journal of Political Economy, Vol. 106, No.3, 514-550
National Economic & Social Council, Supports and Services for Unemployed Jobseekers: Challenges and Opportunities in a Time of Recession, No.123, August 2011
Nickell, S., (1997) “Unemployment and Labor Market Rigidities: Europe versus North America”, in Journal of Economic Perspectives, Vol. 11, No.3, 55-74
O’Neill, J., Bassi, L. and Wolf, D., (1987) “Duration of Welfare Spells”, in The Review of Economic and Statistics, Vol.69 No.2, 241-8
Siebert, H., (1997) “Labour Market Rigidities: At the Root of Unemployment in Europe”, in Journal of Economic Perspectives, Vol.11, No.3, 37-54