VoxEU Column Labour Markets Poverty and Income Inequality Welfare state and social Europe

Marital status and redistributive aspects of the US Social Security system

At the individual level, social security is a strong source of redistribution from rich to the poor in the US, due to the concavity of the pension formula. But this column argues that spousal and survivor benefits, which are important sources of retirement income for women, introduce regressive redistributive elements to social security and also provide incentives for even highly educated women to stay at home if they are married to a high earner. A means-tested minimum benefit would simultaneously increase overall labour supply and reduce inequality, compared to the current system.

Social security is credited as the main driver of the dramatic decrease in old-age poverty during the 20th century (Engelhardt and Gruber 2006). Incomes vary widely across older households, however, with dispersion equal to or greater than that of younger households. This result is surprising given the concavity of the benefit formula, which implies redistribution from richer to poorer households. With social security providing about half of cash income – and up to 70% of retirement income for women – we might have expected a more dramatic compression of the distribution.

In the US, it is possible to also claim social security benefits based on a spouse’s earnings record. Therefore, social security benefits favour married households over singles. With the rise in divorce and decline in marriage, there are increasing numbers of single older single people in the economy. These demographic changes call for an evaluation of the current social security system.

Previous research has shown that marriage has a role of marriage in insuring against risk over the life cycle (e.g. Attanasio et al. 2005). Kaygusuz (2015) studied the role of the family in social security and Fernandez and Wong (2014) studied the effect of divorce risk on female employment. 

In our recent work (Groneck and Wallenius 2017a), we study how social security contributes to inequality in old age. We analyse reforms that decrease this inequality, without increasing the fiscal burden of social security, by developing a structural quantitative model looking at intra-generational redistribution between current retirees. We also analyse efficiency losses from the social security system by quantifying the negative incentives for female employment that are inherent in the current system. Our framework allows us to assess the budgetary implications of policy reforms, important considering the financial strain placed on public pension systems by an aging population.

Our analysis is based on auxiliary benefits, where eligibility is linked to marital status irrespective of education or income. Auxiliary benefits make it possible to claim benefits based not on one's own earnings record, but rather that of the spouse. They essentially top-up benefits for married and widowed individuals with low lifetime earnings, and are generous, with spousal benefits of 50% of a retired worker's Primary Insurance Amount (PIA) and survivor benefits equal to 100% of the deceased worker's PIA.

Auxiliary benefits are an important source of retirement income for women: about 54% of women collecting social security in the US are collecting spousal or survivor benefits (SSA 2014).

Eligibility rules potentially induce an increase in inequality at old age. The system implies redistribution from singles to married households even though never-married and divorced people are at greatest risk of poverty in old age. Poverty rates among divorced and never-married elderly women exceed 20%, whereas less than 6% of married women are below the official poverty line.

Auxiliary benefits also distort labour supply decisions. They disincentivise the labour supply of married women by awarding them larger benefits than they would be entitled to if their benefits were based on their own earnings record. Note that the auxiliary benefit system creates incentives for even highly educated women – potentially with high labour market productivity – to stay at home. This creates efficiency losses. In contrast, auxiliary benefits encourage men to postpone retirement, since by working longer the husband increases not only his benefit, but that of his wife.  

Poorer individuals also have less stable marriages, and are therefore less likely to be eligible for auxiliary benefits than their richer people. We find that the positive correlation between marital stability and income is even more pronounced than the correlation between marital stability and education found by Stevenson and Wolfers (2007). The share of married individuals is 31 percentage points lower in the lowest income quintile than in the highest quintile. In addition, poorer individuals die earlier and so receive social security for a shorter time. According to our estimates, women who drop out of high school can expect to live, on average, 6.4 years less than their peers who graduate from college. Both features introduce regressive redistributive elements to social security.

To study the employment effects and the consequences of social security for redistribution, we develop a dynamic, structural life-cycle model of singles and couples with marriage and divorce risk, and uncertain survival, that we calibrate using US data. By doing so, we take the socio-economic gradient of marital transitions into account.

We find the following:

  • Redistribution in the current system. We start by computing replacement rates to analyse the redistribution built into the current US social security system. We find that the current system is progressive for a given marital status, implying that the replacement rate is declining in education. There is strong redistribution from non-eligible never-married and divorced to eligible married and widowed females, however, with replacement rates for eligible women up to three times larger than for singles. This introduces a regressive element to social security, implying greater replacement rates for eligible high-educated women than for non-eligible low-educated women. In addition, we find that the current system became highly regressive (it redistributed from the bottom to the top) when we adjust the replacement rates. This is because less educated individuals spend fewer years in retirement. We find that differences in longevity overturn the concavity of the social security system.
  • Abolishing auxiliary benefits. Auxiliary benefits significantly dampen female labour supply. We found that eliminating auxiliary benefits would increase the employment rate of married women by 6.4 percentage points. This would translate into an increase in aggregate hours of 1.8% for the whole economy. Poorer households would experience the largest decline in social security income, however, resulting in increasing inequality in old age. The Gini of social security income would rise by 7 percentage points.
  • Replacing auxiliary benefits with a minimum benefit system. In a counterfactual, we replace auxiliary benefits with a minimum benefit, means-tested on household income. Minimum benefits are prevalent in many European countries. We consider a revenue-neutral minimum benefit, which allows us to top up average household benefits to 35% of average income in the economy (both spouses awarded the same benefit). We find that a minimum benefit like this would increase the average employment of married women by 1.8 percentage points relative to the current system. While employment would fall for those who drop out of high-school (-1.1 percentage points) since the new minimum benefit system would be more generous for them, this effect would be overturned by a positive employment effect for high-school (1.7 percentage points) and college-educated women (1.7 percentage points). As well as the positive overall employment effect, the minimum benefit would redistribute from richer to poorer households. Specifically, low-educated women who were married to low-educated men would gain. Widows would lose the most relative to the benchmark. Overall, the Gini of social security income would fall by 1 percentage point.

In our current work (Groneck and Wallenius 2017b), we draw attention to single mothers. The share of single mothers is large and rising. In 2010, 30% of all mothers lived alone with their children, and a large share of these single mothers live in poverty throughout their life. Hence, the current welfare state does not seem to provide sufficient insurance against the risk of becoming a single mother. The social security system also favours married women over singles.

To conclude, our analysis shows that the current social security system with auxiliary benefits depresses female labour supply and introduces regressive elements to social security. This points to the growing – but until now far less emphasised – problem of poor, elderly singles created by rising divorce and declining marriage rates.


Attanasio, O, H Low and V Sanchez-Marcos (2005), "Female Labor Supply as Insurance Against Idiosyncratic Risk", Journal of the European Economic Association, 3(2-3): 755-764.

Engelhardt, G and J Gruber (2006), "Social Security and the Evolution of Elderly Poverty", in A Auerbach, D Card and J Quigley (eds) Public Policy and the Income Distribution, New York: Russell Sage Foundation.

Fernandez, R and J C Wong (2014), "Divorce Risk, Wages and Working Wives: A Quantitative Life-Cycle Analysis of Female Labor Force Participation", The Economic Journal, 124(576): 319-358.

Groneck, M and J Wallenius (2017a), "It Sucks to Be Single! Marital Status and Redistribution of Social Security", SSE Working Paper Series in Economics No. 2017: 1.

Groneck, M and J Wallenius (2017b), "Single Mothers and the Welfare State", Mimeo.

Kaygusuz, R (2015), "Social Security and Two-Earner Households", Journal of Economic Dynamics and Control, 59: 163-178.

SSA (2014), "Annual statistical supplement to the social security bulletin", Social Security Administration No. 13-11700.

Stevenson, B and J Wolfers (2007), "Marriage and Divorce: Changes and their Driving Forces", The Journal of Economic Perspectives, 21(2): 27-52.

1,680 Reads