Adam Solomon

Distinct risks are typically insured separately. A single `aggregate' contract that pays more when many shocks occur simultaneously, but less when positive shocks offset negative shocks, is utility-increasing absent moral hazard. However, an aggregate contract discourages diversification, leading to a novel insurance-incentive trade-off. We study the US Federal Crop Insurance Program (FCIP), where farmers can choose the `scope' of their policy - whether to insure each field separately, or all fields of the crop as an aggregate unit. Starting in 2009, the FCIP introduced a large subsidy increase for aggregate insurance. We show that farms that moved to aggregate insurance reduced crop diversity and irrigation, farmed less and conserved more land, and insured price risk --- all reducing the diversification of their risks. This increased the variability of farm yield by 14%, raising the fiscal cost of aggregate insurance by about $1.5 billion per year. We derive and estimate a `Baily-Chetty' style formula for the optimal contract scope. We find that an aggregate policy is never welfare maximizing, but that the optimal policy lies partway between separate and aggregate. More generally, we discuss scope's widespread relevance in insurance design.

This paper studies imperfectly-perceived private information in insurance markets when contracts endogenously respond. Equilibrium contracts, pooling and welfare depend on the joint distribution of risk and misperception. In the Health and Retirement Study (HRS), I show that misperceptions typically co-vary with (medical, long-term care, disability and mortality) risk type: high types under-perceive their risk, low types over-perceive. I develop a general model and algorithm to estimate the equilibrium contracts, pooling and welfare impact of misperceptions that is applicable in many settings. I offer suggestive evidence from US annuity markets that contracts are distorted due to misperceptions, with welfare likely increasing.

We study Colonel Blotto games with sequential battles and a majoritarian objective. For a large class of contest success functions, the equilibrium is unique and characterized by an even split: Each battle that is reached before one of the players wins a majority of battles is allocated the same amount of resources from the player's overall budget. As a consequence, a player's chance of winning any particular battle is independent of the battlefield and of the number of victories and losses the player has accumulated in prior battles. This result is in stark contrast to equilibrium behavior in sequential contests that do not involve either fixed budgets or a majoritarian objective. We also consider the equilibrium choice of an overall budget. For many contest success functions, if the sequence of battles is long enough the pay

Increasing climate risk has caused insurance in many locations to become unaffordable or unavailable. I study a novel policy response in Australian home insurance: government provided, mandatory, actuarially fair, reinsurance for cyclone damage. In this scheme, the government reinsures the cyclone risk, while the private market covers the remaining idiosyncratic risk. I find that public reinsurance leads to a 21% decrease in home insurance premiums and an 11% increase in the probability of insurance being offered at all. In terms of mechanisms, I rule out subsidization and show that the ambiguity of the risk has a minimal impact on premiums and insurance offerings. Instead, the entirety of the increase in insurance offered, and much of the decrease in premiums, comes from reducing the implicit costs associated with insuring spatially correlated risk. Increased competition due to insurer entry explains the remaining premium reductions. This isolates the cause of market dysfunction - correlated risk - and suggests that public reinsurance is a cost-effective policy to rehabilitate insurance markets for catastrophic climate risks.

This paper examines a classic rationale for voluntary take-up over automatic enrollment in transfer programs: Take-up may be advantageously ``self-targeted'' on characteristics that are infeasible to use as eligibility criteria. Using a correlation test, we find self-targeting in eight U.S. transfers: on average, recipients have lower consumption and lifetime incomes than similar eligible nonrecipients. Self-targeting focuses redistribution toward the lifetime poor and also increases transfers' within-lifetime insurance value. With a new sufficient-statistics result, we value the social benefits of self-targeting at approximately nine cents per transfer dollar, offsetting the social costs of take-up in some programs.

Every insurance contract bundles risks, and explicit bundling discounts are common. I show theoretically that bundling arises in a competitive market whenever correlation between risk types enables insurer "cream-skimming": willingness-to-pay for insurance against one risk must be negatively correlated with expected costs from the other risk. I analyze long-term care insurance, in which both-spouse bundles are discounted by 20-35%. I show that cream-skimming incentives are sufficient to explain these discounts, and rule out standard economies-of-scale. Counterfactually, banning bundling would raise welfare by 5% by correcting separate-market unraveling, while mandatory family bundling would reduce welfare by 5% by exacerbating advantageous selection.

Estimates the expected present discounted value (EPDV) of future payouts on both immediate and deferred annuities are sensitive to the discount rate used to value future payment streams and assumptions about future mortality rates. This paper illustrates this with respect to annuities that were available in the US retail insurance market in 2020. The spread between the interest rates on Treasury and corporate bonds was high by historical standards as a share of the riskless Treasury yield during much of 2020, making the choice of discount rate more consequential than in the past. The EPDV estimates also depend on whether the rapid but since-attenuated decline in US old-age mortality rates during the 1990s and early 2000s is extrapolated to future decades. The "money's worth" is the EPDV divided by the annuity's purchase price. Our central estimates, using discount rates drawn from the corporate BBB yield curve and future mortality rates that combine a Society of Actuaries individual annuitant mortality table with projections of future mortality improvements from the Social Security Administration, suggest money's worth values for annuities offered to 65-year-old men and women of about 92 cents per premium dollar. Recent Department of Labor rulemaking requires defined contribution plan sponsors to provide participants with estimates of the annuity income stream that their plan balance could purchase. These estimates, like EPDVs, are also sensitive to both prospective rate of return and mortality rate assumptions.

Unemployment insurance (UI) systems are either individualized (e.g. the US) or family-based (e.g. Australia and the UK). In a family-based system, benefits are means-tested against spousal income: otherwise comparable unemployed people with low-earning spouses receive a higher benefit than those with high-earning spouses. A family-based system targets payments to needier households, but levies an implicit tax against spousal earnings, potentially depressing labor supply. I examine this trade-off in Australia, exploiting variation in the implicit spousal tax rate that ranged from 60% to 25%. When the implicit tax rate fell, spousal earnings rose by 15%, implying a spousal income elasticity of ~0.25. I use the empirical estimates to estimate a model of optimal family-based UI that trades-off targeting with labor supply responses.

In many settings where households face substantial and changing climate risk, the government provides both ex-ante insurance subsidies and ex-post disaster assistance. Removing the former so that prices accurately reflect the risk faced would increase expenditures on the latter. We study the US National Flood Insurance Program, which has recently re-priced policies to be actuarially fair, and the spillovers this has on FEMA disaster assistance. We estimate that 1% flood insurance subsidy increases coverage by 0.66%. If a flood occurs, this coverage expansion reduces FEMA disaster assistance expenditures by $148 and disaster loans by $321 per house in the affected county. We explore heterogeneity by risk and region. Using these results, we estimate a model of optimal ex-ante subsidy versus ex-post assistance.

These are the materials from when I was the TA for Prof. Amy Finkelstein in Fall 2022

These are the materials and course evaluations from when I was the TA for Prof. Nathan Hendren in Spring 2024

These are the course evaluations from when I was the TA for Prof. Jon Gruber in Fall 2021