Essays on Savings Behavior, Inflation Measurement, and Growth

Abstract: The Impact of Pension Reform on Household Income and Savings: A Quantitative Analysis of the Swedish CaseIncreases in life expectancy cause challenges for defined benefit pension systems. Sweden is one of few countries having undertaken a major reform aimed at creating a financially sustainable pension system. Since the pension reform, the household savings rate has increased significantly. We investigate to which extent the pension reform can explain this increase. We construct a life-cycle model of heterogeneous agents with rational expectations, which enables us to quantify the impact of the reform on household income and savings. The key feature of our framework is the explicit modeling of all aspects of the pension system including public pensions, occupational pensions, and the minimum guaranteed pension. Given perfectly rational and fully informed agents, the Swedish pension reform implies a major shift of income from retirement to working age, and explains about half of the observed increase in the private savings rate.Savings Responses to Pension Reform and Implications for Models of Savings BehaviorHow do individual savings respond to pension reform? What are the implications for models of savings behavior? We answer these questions by comparing the behavior observed in detailed administrative data on asset holdings to a life cycle model constructed to quantitatively account for the dynamics of pension benefits and contributions. Exploiting the transition rules across cohorts of a structural reform of the Swedish public pension system, we find that individuals do not respond to the reform: Despite a reduction in future pension income, net wealth and savings rates remain unchanged. The particular setting under study enables an evaluation of competing models of savings behavior. We find that inaction is due to inattention, and not due to inconsistent time preferences in the form of hyperbolic discounting. A model in which 71 percent of individuals are inattentive to the reform can quantitatively account for the lack of response observed in data.Micro PPI-Based Real Output ForensicsWe study the producer price index micro data on total private goods and services production in Sweden to quantify the implications of different methods of price index construction on the aggregate inflation rate.  Compared to an arithmetic index, moving to a geometric averaging of items decreases annual goods and services inflation by 0.5 and 0.4 percentage points, respectively. An index based on economic theory and estimated elasticities of substitution decreases the annual inflation rate by 3.9 percentage points for goods and 3.1 percentage points for services. These results pose a challenge for the comparability of inflation rates and real output growth rates across countries as well as a tension between (economic) theory and (statistical) measurement. A practical solution to overcome these issues is to assume a joint log-normal distribution of price growth factors and weights. Under this assumption, the true index is well approximated by only three moments.Structural Change in Production Networks: A PuzzleIn the US, the services share of intermediate inputs increased by more than 1/4 in goods production and by more than 1/3 in services production between 1947–2019. Due to virtually no trend in relative input prices, a neoclassical growth model with sector-specific total factor productivity and elasticity of substitution cannot account for these structural changes. When allowing for directed technical change in intermediate inputs, the model is still unable to account for structural change in intermediates because directed technical changes in some inputs must feature long-run declines. The paper ends with a discussion on directions for future research by suggesting either task-based production or expanding the model's nest structure.