Does Unemployment Damage Your Health?

There were 3,340,958 unemployed in the UK, according to government statistics released on 31 January. There is little prospect of a dramatic fall in the number of unemployed in the near future. What effect does unemployment- especially long-term unemployment have on those who experience it? On 1 February CEPR held a one-day workshop to explore the relationships between unemployment and health. The workshop, organized as part of the CEPR's programme in Human Resources since 1900, was chaired by the programme's director, Roderick Floud  (Birkbeck College). The DHSS provided the financial support which enabled CEPR to bring together sociologists and epidemiologists, historians and economists, civil servants and journalists to discuss research in this field.

In his Introduction, Roderick Floud outlined what he hoped might be achieved in the workshop. One purpose was simply to inform participants of the variety of approaches to the study of health and unemployment. What lessons could be learned from past experience? If unemployment did lead to an increase in morbidity (ill-health), what were the short-term and medium-term costs of this increase? What were the long-term effects of unemployment on health? What action should be taken to remedy the effects of unemployment, and finally, what further research was needed?

As a first step towards answering some of these questions, Dr Derek Cook (Royal Free Hospital) presented the workshop's first paper 'Unemployment and Health', written jointly with A G Shaper. It discussed the findings of a number of recent British studies and also the methodological implications of their different approaches. Individual studies of the health of unemployed people showed, Cook suggested, that ill-health was as much a cause of unemployment as a consequence of it, and that for both employed and unemployed the greatest health problems were related to social class. Only by looking at changes in aggregate measures of health over time would it be possible to say whether unemployment was itself a cause of ill-health. One approach to this question would be to compare changes in mortality rates with changes in the level of unemployment. Such a comparison should not ask whether the rate of mortality could be correlated with unemployment, but .whether the rate of mortality decline could be correlated with unemployment. Finally, Cook warned against attempts to predict the current effects of unemployment on health from studies of its effects in the past. It was not reasonable, in his opinion, to assume that the effects of unemployment on health would be the same from one decade to the next.

Cook's view that the effects of unemployment on health were still not known and required further intensive research occasioned a lively discussion. While it was difficult to distinguish the effects of unemployment from the effects of other kinds of deprivation, nevertheless some participants believed that it had been clearly demonstrated that unemployment did cause ill-health. They argued that the need now was to identify precisely which aspects of unemployment affected health, so that remedial action could be taken against them.

Benefits did not meet basic needs

Current concerns with the effects of unemployment on health are not new. They were also voiced in Britain during the Depression of the 1930s. In his paper ‘Health, Welfare and Unemployment During the Depression', Charles Webster (Wellcome Unit for the History of Medicine and CEPR) asked whether the substantial extensions of welfare legislation introduced in that decade had benefited those most in need. At the time, it was thought that the State's 'safety net' had prevented the overwhelming majority of those affected by the Depression from experiencing acute deprivation. Webster argued that this official view was based on inaccurate statistics and was not credible. For most unemployed people and their dependants, state benefits were insufficient to meet basic needs, and In the areas of greatest need, welfare services were least developed, If there was any lesson to be ,drawn from the 1930s, Webster argued, it was that in a period of depression, welfare services organized on a local basis, without effective central planning, were of little value.

Historians who stress the harmful effects of the Depression are often described as ‘pessimists’, in contrast to the 'optimists' who prefer to emphasise the more progressive aspects of the period. In the lively discussion which followed Charles Webster's paper, attention was drawn to his heavy emphasis on the experience in Wales. The Welsh Board of Health might have been particularly 'backward' in its arrangements, and health and welfare in Wales may not have been representative of the country. As Webster had noted, the 1930s may well have witnessed a slower decline in the rate of mortality. Nevertheless, the influences on mortality are very long-term in nature, and some participants found it difficult to attribute fluctuations in mortality in the 1930s to events such as the Depression.

Webster's attempt to draw lessons from past experiences in the UK provoked much discussion; in his own paper Cook questioned the effects of unemployment on health. Are there lessons to be learned from contemporary developments in other countries? The two sessions which followed were devoted to reports of research in progress in West Germany and Sweden.

Unemployment in West Germany declined steadily in the 1950s, and apart from a brief rise in 1967 it remained low up until the mid-1970s. Following the oil crisis of 1973, unemployment began to rise again. As it has continued to rise throughout the present decade, long-term unemployment has become a more significant problem, This has prompted researchers to look at ill-health both as a cause and as a consequence of unemployment, and also as a barrier to re-employment. at both individual and aggregate levels. Jurgen John (Institute of Medical Information, FRG) discussed the current state of this research, At the individual level, it has been shown that the unemployed are indeed more likely to experience ill-health than those in employment. This is often highly dependent, however, upon the personal characteristics and circumstances of the individuals concerned, and it can be misleading to generalise about 'the unemployed' from these individual studies.

John's own research had been carried out using aggregate data for 1950-1977, data originally analysed by Brenner. Brenner had found that mortality was influenced by unemployment, but not by trends and fluctuations in economic growth. Using a different approach, John showed that on the contrary, economic growth did influence mortality. John also found that the relationship between unemployment and mortality was not positive, as one might expect, but negative. The only exception was infant mortality, which was positively related to the level of unemployment. Even this relationship disappeared, however, if the first five years were excluded from the estimation period.

John concluded that both individual and aggregate data were useful and might well be combined in the same study, if the very considerable practical problems involved could be surmounted. More work needed to be done on the concept of unemployment: the differences between official and personal definitions of unemployment could in themselves affect the results of the studies that had been made. John argued that a promising subject for further research might be the distribution not only of unemployment but of unhealthy employment among the workforce.

Unemployment did result in increased stress

Urban Janlert (Karolinska Institutet, Sweden) discussed current Swedish findings in his paper 'Unemployment and Health -Recent Swedish Experience and Research'. Sweden has historically experienced low levels of unemployment. This, together with the conviction that the existence of unemployment was an 'accident', meant that until recently the effects of rest of unemployment did not receive serious study. Even though Swedish unemployment began to rise in the 1960s, its effects upon health were largely ignored until the beginning of this decade. A number of tentative conclusions had since begun to emerge. In one study of the effects of a factory closure and  consequent redundancies In Olofstrom, unemployment did result in increased stress. A similar study in Kopmanhoimen had suggested that although differences in health may have determined who became unemployed, the relative disadvantage in the health of unemployed people did not increase during the period of their unemployment. A study of school-leavers in Lulea also found that health status played an important part in determining who became unemployed. In contrast to the Kopmanholmen study, however, those who became unemployed in Lulea then experienced a further deterioration in their health. The experiences of unemployed shipyard workers in Finnboda had led researchers to believe that age and marital status may have a strong influence on the way in which unemployment affected health. All these conclusions related to the unemployed individuals themselves.

Janlert suggested that more knowledge was needed on the effects of unemployment upon the parents and children of those unemployed and on the effects of white collar unemployment. Future research should also take account of what he called the 'anticipation effect'. In a period of industrial depression, expediency or competitive pressures may cause health and safety conditions in workplaces to deteriorate. Workers in employment may also experience stress and ill health as a result of the threat of unemployment. Both these factors related to unemployment are prejudicial to health.

Janlert was asked whether any studies had been made of the health of young workers in 'make-work' schemes. In the UK, experience had shown participants in the Youth Opportunities Programme did not experience the psychological problems otherwise associated with unemployment, provided they found employment on completion of the programme. Failure to find employment resulted in tangible disillusionment. Unfortunately, however, no similar research had been undertaken in Sweden which might have shed light on these British findings.

Thomas McKeown has argued that the most important single influence on mortality levels before the Second World War was the level of nutrition. In the last 20 years, most mortality rates have behaved in a way broadly consistent with this hypothesis. The maternal mortality rate, however, is an important exception. Comprehensive measures of the maternal mortality rate (MMR) are available from the 1840s. Despite enormous improvements in nutrition and decline in almost every other mortality rate, the MMR stayed the same. When it did eventually fall in the 1930s, it did so spectacularly.

Any explanation of maternal mortality has to account both for the spectacular decline and for the long period of stability which preceded it. It has, moreover, to be consistent with the fact that the two components of the MMR, deaths from puerperal fever and accidents of childbirth, both fell at roughly the same time. Thus the introduction in 1936 of the sulphonamide drug prontosil to combat puerperal fever can explain only half the story. In his paper 'The Historical Evidence Concerning the Effects of Social and Economic Deprivation on Maternal Mortality', Irvine Loudon (Wellcome Unit for the History of Medicine) offered a simple explanation consistent with all these facts. He argued that the availability and quality of obstetric care has been the single most powerful influence on maternal mortality during the past two hundred years. In the early 1930s Andrew Topping had introduced basic reforms in obstetric practice in Rochdale. The MMR in Rochdale fell from 8.9 to 1.75 deaths per thousand births between 1929/30 and 1935 -before the introduction of sulphonamides, and at the height of the Depression.

One Important Implication of Loudon’s argument is that social and economic deprivation had little impact on maternal mortality; Loudon himself found this difficult to accept. He suggested that the explanation may lie in the assumption that there is a simple linear relationship between maternal mortality and malnutrition. Some Third World studies seemed to indicate that the physiological processes of childbirth were little affected by quite severe degrees of deprivation. Beyond a certain threshold, however, deprivation caused maternal mortality to increase sharply. If deprivation only had effects when it reached this threshold, Loudon argued, unemployment in Britain today may not induce levels of hardship severe enough to affect the maternal mortality rate, In the discussion which followed, it was suggested that the difficulty may lie as much with the index used for maternal mortality (deaths per thousand births) as with the phenomenon itself.

In her response to Loudon s paper, Alison Macfarlane (National Perinatal Unit, Oxford) sought to extend the analysis to other 'outcomes of pregnancy'. The clear relationship between infant mortality and social class was generally acknowledged: to what extent was that relationship due to maternal deprivation before or during pregnancy? How was infant health affected by previous parental experience of unemployment?

Unemployment and parasuicide rates related ?

Many of the workshop papers had drawn attention to the clear and unsurprising relationship between unemployment and stress. How severe can this stress become? Stephen Platt (Royal Edinburgh Hospital) pursued this question in his paper ‘Parasuicide and Unemployment Among Men in Edinburgh 1968 and 1983’. Platt found that over the entire period, unemployment and attempted suicide rates were positively related. The estimated relationship was much stronger over the first eight years only, and in the second half of the sample the relationship was actually negative, though not statistically significant. A study of the relationship between parasuicide and unemployment in Oxford had produced similar results. Unemployment and parasucide were also found to be related in cross-sectional analyses of Edinburgh in 1971 and in 1981. The relationship disappeared, however, when a measure of poverty was included in the estimated equation. Throughout the period Platt studied the parasuicide rate among unemployed men was much greater than among employed men, but the difference between the rates had been declining over most of the period. Either the kind of man now becoming unemployed was less suicidally inclined, or the increasingly common experience of unemployment carried a lesser stigma. Although the immediate impact of a loss of employment was marked, men out of work for long periods were more likely to attempt suicide than men out of work for shorter periods. This finding also was in line with the Oxford study. Overall, the incidence of parasuicide had

shown a tendency to rise in line with unemployment, Platt argued that these findings were consistent with the view that unemployment was a cause of parasuicide, but his methodology was heavily criticized in the discussion which followed. Without further research, these conclusions will not gain widespread acceptance.

Charles Webster had earlier discussed the effectiveness of State welfare provision during the 1930s. In the final workshop paper Noel Whileside (Bristol) returned to the question of the 'Hungry Thirties', In her paper 'Counting the Cost: Sickness and Disability among Working People in an Era of Industrial Recession (1920-1939)', she suggested that we may miss an important effect of the Depression if we simply regard the increase in claims for health insurance benefits as reflecting an increase in the level of unemployment. Increased take-up of health insurance in the 1920s and 19305 may indicate increasing hardship and disability benefits rose together; in the 1930s, claims for long-term disability benefits rose disproportionately. Previous occupation, rather than age, was the most significant factor in determining the type of claim made.

What could explain these results? The Depression may have made 'hidden' rates of sickness more visible, and unemployment aggravated its incidence. Another explanation is that the provision of sickness benefits fostered a kind of 'voluntary sickness' which allowed the marginally ill to languish indefinitely at public expense. The latter suggestion seemed barely credible to Whiteside: in the face of intense competition for jobs, workers were much less likely to admit to ill-health for fear of losing the jobs they had. In any event, Whiteside noted, health insurance payments remained far below the rates available under unemployment benefit. In a 1938 report the Government Actuary asked whether it would not have been more economical for the Government to have promoted health actively rather than sustain avoidable disability. The question was no less pertinent today.

In the discussion, Whiteside was asked whether it was right to concentrate on the negative aspects of the Depression: even at its height, conditions were not as bad as they had been thirty years previously. In her view, the point was that in any era poverty increases dependence on public support, and that by neglecting the consequences of poverty we increase the costs of dependence.

What conclusions emerged from the workshop? The general discussions at the end of the morning and afternoon sessions focused on two questions: how are the effects of unemployment on health to be studied, and how are the results to be translated into policy? Noel Whiteside’s paper served to remind the workshop how little was known effects of relative deprivation. Earlier it had been suggested that there was an urgent need to examine not only the extent to which the health of unemployed people today might have been affected by unemployment among their parents, but also to follow through the effects of their unemployment upon their own children" The danger with all such 'prospective' studies is that the questions which interest researchers and the techniques they employ change over time. Nevertheless, the papers presented at the workshop demonstrated that many of the fundamental concerns remained unchanged over several decades.

Such prospective studies may well help the unemployed in future generations; they will do little for the unemployed today. In his account of research in Sweden, Urban Janlert distinguished between those who believed that the way to avoid the unhealthy consequences of unemployment was to promote full employment, and those who felt it necessary to identify the particular aspects of unemployment which were unhealthy and to try to mitigate their particular effects. Here in Britain, while mass unemployment persists, research of this kind is urgently required. It is necessary to ascertain which aspects of unemployment affect health, which aspects of health are affected by unemployment, and what remedial measures can be taken to protect the health of deprivation. In the 1920s, claims for sickness and unemployed people.