Jonna Olsson

I am a PhD Candidate at the Institute for International Economic Studies (IIES), Stockholm University. I do research in macroeconomics, with an emphasis on quantitative models and labor supply questions in the short and long run.

In September I will join the Macro and International (MInt) research group at the University of Amsterdam as an assistant professor.

Link to CV

email jonna.olsson@iies.su.se | tel +46(0)70 765 5237 | address IIES, Stockholm University, Universitetsvägen 10, SE-10691 Stockholm, Sweden |


Structural transformation of the labor market and the aggregate economy [JMP]

Women's increased involvement in the economy has been the most significant change in labor markets during the past century. In this paper, I account for this period of structural change of the labor market in a macroeconomic model, and study how the increase in female labor force participation has affected the economy's response to aggregate shocks. I explicitly model heterogeneity in gender and household composition as well as the historical decrease of the gender wage gap. The model captures the salient features of historical data, including a strong increase in employment among married women, low crowding-out of married men, and relatively stable employment over time for single women. I then study how the changing labor force composition affects the economy's aggregate employment dynamics. The underlying trend in employment, driven by growth in female labor force participation, contributed to the perceived quick employment recovery after recessions before 1990, and the absence of growth thereafter consequently explains the more recent slower employment recoveries. In general, incorporating both one- and two-person households matters for employment dynamics, with single households reacting more strongly to shocks and employment responses by subgroups changing over time. Despite relatively large changes by subgroup, the aggregate effect is unchanged between the 1970s and the present time due to multiple counteracting forces.

Research papers in progress

"Labor Supply under Heterogeneous Agents: The Case of Complete Markets" (with Timo Boppart and Per Krusell)

Most of applied macro -- e.g., newkeynesian models, but also long-run modelling -- uses a representative agent or abstracts from labor supply, and often both. Sometimes the models study frictional labor markets, but rarely with active labor supply and heterogeneity, and then there is usually no frictionless benchmark to refer to. This is the gap we fill in this paper: we consider what we believe to be highly realistic and a priori relevant heterogeneity and its role for aggregate labor supply and for who will/should work. The findings have bearing both on short-run macroeconomic analysis -- in particular the sensitivity of ``the'' Frisch elasticity to heterogeneity -- and on long-run growth, where the growth rate of the economy can depend nontrivially on working behavior. The class of potential utility functions to consider is large so we restrict attention to those that are consistent with balanced growth (BK/KPR). Moreover, we focus especially on the functions most commonly used in the applied macroeconomic literatures. We conclude that, although a realistic model of the economy would feature incomplete markets and other frictions, the results we derive here will reflect a benchmark first-best allocation also in those economies and will, likely to an important extent, be driving the positive and normative features of those frictional economies as well.


"Can Stable Preferences Explain Postwar U.S. Hours worked?" (with Timo Boppart and Per Krusell)

Cross-country and time-series evidence suggest that, along a path of approximately constant real output and productivity growth, income effects on labor supply slightly outweigh substitution effect. Yet in the postwar U.S., where these approximate growth features are satisfied, hours have not had a trend, unlike in the average of OECD economies. In this paper we try to account for the U.S. facts with a standard neoclassical -- and, for convenience, frictionless -- model, using preferences that do mean that income effects exceed substitution effects. It turns out, first, that it is possible to account for the data if one defines ``hours worked'' to include hours worked at home. Second, to account for hours at home and in the market separately, one (i) needs to consider women explicitly and (ii) needs trends in drivers of women's labor supply. We briefly discuss possible such drivers, some of which can be quantified and are shown to contribute part way toward the total.


"Subjective Life Expectancy, Health, and Wealth” (with Richard Foltyn)

Evidence from the US suggests that individuals hold asymmetric subjective beliefs about their own life expectancy: At ages below 75 they on average underestimate survival probabilities compared to population averages (as measured by the life tables), while at higher ages they are overoptimistic. Moreover, within a cohort the expectation errors are not random but highly correlated with self-reported health status. Healthy individuals on average overestimate their expected life-span, while individuals in poor health underestimate it, compared to expected life-span conditioning on self-reported health. On the other hand, households have substantially higher and more dispersed financial asset holdings in higher age compared to what standard life-cycle models would predict. Traditionally, the literature has suggested bequest motives and precautionary savings to insure against medical costs as possible ways to rationalize observed asset levels at high age, while some recent contributions started to explore to what extent overoptimism with respect to survival to high age can help explain the slow decumulation of wealth. We embed the latter approach into a full-featured life-cycle model to assess the quantitative importance of overoptimism and systematic error along the health gradient on asset holdings and wealth distribution.


"Self-selection into Retirement and Social Security Reform" (with Richard Foltyn)

We investigate optimal retirement behavior in a life-cycle model of agents who are heterogeneous with respect to age, assets, productivity, health and social security entitlements. Our model replicates stylized facts observed in the US, such as the delayed retirement of high earners. Furthermore, unlike in earlier work, we carefully model health transitions estimated from a panel of the elderly which allows us to quantify the effects of (expected) longevity on retirement decisions. We use our model to evaluate reforms to the US Social Security system, such as changes to the full retirement age or the earnings test for early retirees, and the heterogeneous effects on high vs low earners. Preliminary results indicate that the earnings test as currently implemented in the US is a major factor in delaying retirement for high earners.