The impact of social security compensation inequality on earnings distribution due to sickness and disability

Abstract: One of the main purposes of social security systems is to compensate for loss of earnings during periods of sickness and disability. Of interest is therefore the extent to which such compensation reduces earnings inequality. The aim of the thesis is two-fold. First, it intends to identify the factors that influence the distribution of compensations from social security due to sickness and disability (Papers I, II, III). Second, it intends to analyse the redistributive outcomes of social security insurance programmes applying to sickness and disability (Papers II, IV). In Paper I socioeconomic differences and gender imbalance in compensation from occupational injury insurance are analyzed. It is tested whether occupational risk explains differences in compensation from Swedish occupational injury insurance schemes in relation to socioeconomic differences in all municipalities in Stockholm County, Sweden. An occupational risk level is calculated for each municipality, which considers the proportions of workers in various industrial branches and the probability of a worker being injured according to reported occupational accidents and diseases. A regression analysis was performed, treating socioeconomic conditions and risk level as predictors of occupational injury insurance compensation. After controlling for variation in socioeconomic factors, the occupational risk level explains the pattern of compensation to males but not to females. From a gender perspective, we observed that women, as a group, are not fully compensated for their occupational risks. It is concluded that socioeconomic variations in compensations from occupational injury insurance are observed and female workers are not compensated according to their risk at work. Paper II investigates how compensation from nine social security schemes for sickness and disability are distributed across 25 municipalities in Stockholm County. Concentration indexes and their decompositions showed how low income earning municipalities are both absolutely and relatively favoured by the compensations. Four principles are used to distribute payments via the Swedish social insurance system for sickness and disability. The analysis is based on aggregate social insurance data from the 25 municipalities in Stockholm County. Compensation inequality was measured on a concentration index. The redistributive effects of compensations on earnings inequality were measured by the differences between Gini coefficients with social insurance payments excluded from and included with earnings. The results showed that the concentration index for compensations from the nine social insurance schemes in total was -0.0469. The Gini coefficient fell from 0.0437 to 0.0379 when compensations were included with earnings. Decomposition by scheme showed that the largest redistribution effect on earnings inequality was made by disability pension. It is concluded that municipalities with low average income are favoured by the Swedish socialinsurance system when considering sickness and disability compensations. Principles of compensations can he ranked according to their redistributive capacity: mix of compensating-lost-income and flat-rate, compensating-lost-income, means-testing, flat-rate, and need-based respectively. The nine social-insurance schemes contribute very differently to income redistribution. The schemes that cover a greater proportion of the population and compensates in relation to previous earnings have a higher redistributive effect. Paper III deals with cross-country experiences of the magnitude of compensations due to sickness and disability. Labour market conditions have often been considered as a determinant of sick leave. The relationships between macroeconomic conditions, measured by unemployment rate and social security spending from four social security schemes and total spending due to sickness and disability were analyzed. Aggregated panel data from 12 OECD countries for the period 1980-1996 were employed. Regression analysis with fixed effect models were used to examine spending on sickness benefits, disability pensions, occupational injury benefits, survivor's pensions, and their total. The results showed that a decline in unemployment increased sickness benefit spending and reduced disability pension spending. These effects reversed direction after four and two years of unemployment respectively. Inclusion of the mortality rate as an additional variable in the analysis did not influence the findings. It was concluded that higher unemployment reduces sickness benefit spending and increases the spending on disability pension benefits. Paper IV focused on the redistributive outcomes of sickness benefits in relation to a social insurance institutional typology and sickness risk by using data from nine OECD countries for a number of years. The typology comprises five categories: encompassing, corporatist, basic security, voluntary state subsidised and targeting systems. The redistributive outcomes of sickness benefits, using this typology of social insurance institutions in comparison with external factors for sickness risk, were analyzed. Gini coefficients were used for measuring earnings inequality. Relative changes in earnings inequality for sickness benefits were explained by social insurance institutional dummies using multiple regression analyses. Sickness benefits as a percentage of total social expenditure, labour force demography and educational attainment were included in extended models. The typology shows expected pattern in redistributive outcomes. However, after adding external variables, the observed pattern of the typology was not repeated. Our results indicate that encompassing system redistributes most followed by basic Security and targeting system among the current social security systems that can be influenced by political interventions. When considering demographic structure and educational attainment as a proxy for sickness risk, the observed redistributive pattern could not the verified. Socioeconomic difference in compensations from occupational injury insurance is observed. However, compensations to females do not correspond to the risks involved. Inequality in compensations due to sickness and disability can be explained largely by risk factors for sickness and disability and by institutional factors (i.e. social insurance compensation regulations). Further, different factors may not influence each type of compensation in the same way. For example, the macroeconomic condition, measured as the unemployment rate, influences sickness benefits and disability benefits differently. Redistributive outcomes can be explained by the social insurance typology only when no other factors are considered. But, after considering sickness risk factors, the redistributive effect of social insurance cannot be confirmed.