Insurance by BogdanovaThe Encyclopedia of Insurance and Risk Science

RESEARCH DOSSIER

Insurance and Household Consumption Smoothing

Abstract. Household consumption smoothing is the preservation of essential and valued expenditure when earnings, wealth, health and unavoidable costs vary across states of the world and over time. Insurance can support smoothing by exchanging predictable financing for contingent benefits in unfavorable states. Actual protection, however, depends on the insured loss, affordability, exclusions, claims timing, access to liquidity and credit, informal transfers, and the use to which compensation is put. This monograph integrates established expected-utility insurance theory, the complete-risk-sharing benchmark, buffer-stock saving, consumption econometrics and international empirical evidence on health insurance, unemployment protection, agricultural index insurance, and property catastrophes. It analyzes the ex ante effect of protection on production and savings, the ex post allocation of cash and debt, the identification of causal insurance benefits, and heterogeneity across household wealth positions. The primary studies show that effective protection can manifest as a reduction in medical expenditure tails, lower debt, continued food purchases, avoidance of distress asset sales, or more productive investment. These outcomes measure different economic margins. A large insurance payment need not translate into a comparably large increase in current consumption; an observed stable spending path need not imply that insurance has little welfare value. The monograph develops a source-critical comparative account of how resources become available when needed, identifies the assumptions required for welfare comparisons, and concludes that household insurance performance must be evaluated across loss states, payment time, financial position and the future economic consequences of coping decisions. Its synthesis introduces no new proprietary theory or doctrine.

Keywords: Insurance economics; household consumption smoothing; consumption insurance; risk sharing; precautionary saving; liquidity constraints; insurance payouts; medical expenditure risk; unemployment insurance; index insurance; basis risk; flood insurance; financial resilience; household welfare; distributional effects; causal identification; catastrophe risk; consumption response; asset protection.

Introduction

Insurance changes the timing and state dependence of the resources available to a household. A premium normally precedes the insured loss; an indemnity, medical service, public benefit or contingent transfer follows the specified eligibility event. The economic objective of consumption smoothing is to preserve the household's achievable consumption path when shocks would otherwise force reductions in necessary expenditure, borrowing at costly terms or liquidation of productive assets.

The foundational rational insurance analysis of Mossin (1968) places coverage choice inside expected-utility reasoning. Its household-level logic becomes richer when actual insurance is considered alongside savings, credit, labor income, family support and social programs. A key distinction comes from Morduch's study of two smoothing mechanisms: income smoothing changes exposure through earlier occupational and production choices, whereas consumption smoothing reallocates resources after income is realized. The first can make observed consumption stable by sacrificing expected income; the second can preserve expenditure by spending down capital or incurring debt.

The scope of this Research Dossier includes commercial coverage, public insurance, informal risk sharing and selected index-based mechanisms. These institutions have different financing rules and beneficiary structures. It compares established studies across medical, employment, agricultural and catastrophe contexts without treating their effect sizes as directly interchangeable. The argument proceeds from a household cash-flow identity to contractual payment timing, source-specific empirical findings, welfare measurement and distributional incidence.

Uncertain Household Consumption

The reference consumption path

Household consumption denotes the goods and services used through time. Expenditure surveys, bank transaction records and claims files observe parts of that object. Food purchases are relatively close to current use, while a vehicle or appliance yields services beyond its purchase date. A medical bill may increase spending precisely when health and well-being deteriorate. Maintaining total monetary expenditure is consequently neither a necessary nor a sufficient condition for protecting welfare.

The Global Findex 2025 report, based on interviews with approximately 148,000 adults in 141 economies during 2024, measures financial access and reported shock coping. It supplies comparative context for savings and borrowing, but it cannot by itself identify an insurance product's causal effect on post-shock consumption. Households differ in asset holdings, dependency ratios, credit access, health needs and the ability to claim benefits.

For household i in period t and realized state s, a transparent cash-flow accounting identity is

C(i,t,s) + K(i,t,s) = Y(i,t,s) + A(i,t,s) + T(i,t,s) + I(i,t,s) + D(i,t,s) - P(i,t,s) - L(i,t,s) - Q(i,t,s),

where C is consumption spending; K is net acquisition of liquid or productive assets; Y is disposable income; A is a drawdown of existing assets; T is accessible family or public transfer income; I is insurance compensation actually spendable in that period; D is net new borrowing; P is premiums and contributions then due; L is uninsured loss-related expenditure; and Q is debt repayment. The identity is an organizing accounting statement, not an estimated behavioral model. Counting insurance as I before a claim is paid would misrepresent the present cash constraint.

The shortfall in a low-income state can be financed by lower C, lower K, more D, larger A or increased transfers. These adjustments have different consequences. Food cuts may impair nutrition; a business asset sale may reduce future production; additional expensive borrowing preserves consumption today but burdens later periods. A household can appear protected in current expenditure while accumulating severe financial fragility.

Shock dimensions

Income losses vary by duration, persistence, predictability and correlation with other households. Temporary job interruption differs from long-term disability. Illness often combines earnings loss and medical obligations. A regional flood damages property, shifts housing expenses, interrupts work and weakens local lenders at the same time. A crop failure can affect both farm revenue and food purchases. Contract design must identify which losses and times are covered.

The review of consumption responses to income by Jappelli and Pistaferri (2010) explains why anticipated and unanticipated income changes have different implications for inference and why estimating a marginal propensity to consume requires careful observation of expectations and income dynamics. Persistent shocks are harder to absorb with finite liquid reserves. Aggregate disasters can impair savings and informal assistance simultaneously.

A household may prepare by choosing a stable but less remunerative occupation or planting less risky crops. Such decisions can reduce measured income volatility at a genuine opportunity cost. The baseline for welfare analysis is therefore a fully specified alternative resource and consumption path, rather than the simple ratio of an insurance payment to a reported financial loss.

Risk Sharing and Precautionary Saving

A complete-insurance benchmark

Under enforceable complete contingent markets, appropriate Pareto weights and suitable preferences, idiosyncratic household income shocks can be pooled so that individual consumption growth does not depend on those shocks after aggregate resources and the allocation rule are accounted for. Common losses remain relevant: a community cannot share away a decline in its own aggregate food or productive resources without transfers from outside.

Cochrane's consumption-insurance test (1991) evaluates this benchmark by examining whether exogenous household-specific events predict consumption growth. The reported results reject full insurance against long illness and involuntary job loss but yield other conclusions for several alternative shocks. This is an investigation of total effective risk sharing, not a randomized test of an identifiable insurance policy.

Townsend's village insurance study (1994) analyzes three Indian villages. The full-insurance model is statistically rejected, yet village-level consumption is informative for individual consumption, and some household-specific shocks have limited additional explanatory power; protection is uneven among landless villagers. The finding is neither proof of complete risk sharing nor evidence that formal insurance alone generated the outcome.

The complete-market benchmark remains valuable because it distinguishes the risk that can in principle be pooled from the risk households still bear. Its practical translation depends on enforceability, verification costs, capital and solvency, correlated claims, information about risk and credible settlement. A nominally complete contract that cannot be performed is weaker protection than its formal coverage description suggests.

Precautionary accumulation

Self-insurance transfers purchasing power through prior saving. Deaton's liquidity-constrained saving analysis (1991) shows how uncertainty and limited borrowing generate a buffer-stock motive. The ability to finance expenditure from assets depends on their amount, accessibility and the shock's persistence. A household facing a sudden liability larger than liquid wealth cannot fully stabilize consumption by self-insurance alone.

The Browning–Lusardi savings review (1996) explains the variety of saving motives and mixed empirical findings about precautionary accumulation. Insurance can reduce desired emergency savings by covering a large hazard; it can also protect productive assets and thereby make saving and investment more attractive. Its net impact on recorded saving cannot be assigned a universal sign.

Liquid wealth may be especially important in the weeks before an indemnity arrives. The same ultimate coverage can offer sharply different protection to a household that can borrow temporarily at low cost and one that must sell a productive asset immediately. The economic analysis must therefore track both the final balance sheet and the cash path leading to it.

Informal networks and risk pooling

Philippine household risk-sharing evidence by Fafchamps and Lund (2003) shows how gifts and informal loans respond to shocks through kin and friendship networks rather than a perfectly insured village. Such arrangements offer flexible transfers but depend on social relations and the economic capacity of the network. They can contract when many members face a common disaster.

Dercon's review of coping strategies documents the limits imposed by asset indivisibility, entry barriers to income diversification and incomplete informal protection. Public insurance and private contracts interact with these arrangements and can shift financial burdens between recipients and nonrecipients.

The Mobarak–Rosenzweig index-insurance study (2013) addresses interactions among village risk sharing, index protection and production incentives. Its preliminary findings warrant careful interpretation. A formal contract may complement support networks by financing shared disasters or change expectations of transfers, and the net effect requires evidence on the whole social arrangement.

Insurance Payouts and Liquidity

Contractual payment and usable cash

A policy's effective current protection depends on a qualifying loss, a payable benefit, the interval to settlement and the household's control over the funds. Deductibles, benefit limits, exclusions, verification rules and required direct payments can leave residual risk. In a stylized indemnity contract, the maximum indemnity is capped by the applicable benefit limit after subtracting the deductible from eligible loss; exact policy language can change this calculation.

For a household whose necessary monthly consumption is 1,000 monetary units, current available resources are 400 and income interruption removes another 600, a timely benefit of 600 can close the immediate deficit if there are no further liabilities. The same nominal benefit paid after three months cannot finance the first month's consumption without a bridge. The example is illustrative accounting, not an estimated insurance effect. Premiums, repayment, other transfers and future consumption must be included in the total comparison.

An advance claim payment, direct provider billing, emergency housing benefit or accessible pre-authorized credit can reduce the gap between the loss and usable resources. Conversely, delays, deductibles, dispute costs and uncertain benefit realization can leave a family exposed despite a positive expected indemnity. Under household liquidity constraints, speed and reliability belong in the measurement of effective coverage.

Premium finance and participation

Premia paid before shocks constrain current expenditure. Financially fragile households may face a large annual premium on the same day that rent or agricultural input spending falls due. Purchase decisions are influenced by both the expected value of a contract and the path of the required payments.

The randomized Indian insurance-demand experiments (2013) find that demand for rainfall insurance is sensitive to price and affected by trust, liquidity and salience. Rainfall insurance participation research by Giné, Townsend and Vickery (2008) similarly associates take-up with basis risk, household wealth and binding credit constraints. These studies illuminate barriers to purchase; they do not independently identify how much household consumption rises after an actual indemnity.

Index-based policies pay according to rainfall or another measured index. Basis risk occurs when the index differs from the household's realized loss. A household may suffer a severe loss without the trigger being met. The reliability of an index payout in the relevant adverse state matters as much as the claim speed that simplified index adjustment may offer.

Effects before a shock

Insurance changes feasible behavior before it pays a claim. Ghanaian insurance and credit experiments by Karlan and coauthors (2014) find greater agricultural investment and riskier production when farmers receive protection against catastrophic risk; payout experience and trust influence later demand. An insurance-backed investment is an ex ante behavioral effect, not itself a direct measure of ex post consumption stabilization.

Jing Cai's Chinese insurance study (2016) finds increases in insured crop production and borrowing together with changes in savings composition. These responses are important channels from insurance to future resources, but their welfare implications require the eventual income distribution, financing cost and household consumption to be assessed.

Insurance can also alter access to loans, protecting lender collateral or lowering expected defaults. When claim proceeds repay a secured obligation, the immediate increase in food purchases may be small although the family's long-term balance sheet improves. Economic measurement must distinguish contemporaneous expenditure from solvency and future productive capacity.

Health and Property Loss Effects

Health insurance as expenditure protection

Severe illness can simultaneously reduce labor income and increase medical spending. The Indonesian illness-and-consumption study (2002) by Gertler and Gruber finds imperfect consumption insurance against major illness. This evidence establishes a substantial residual financial burden while leaving the exact performance of any proposed coverage intervention to separate analysis.

Health insurance can operate through direct payment to providers, reductions in out-of-pocket medical charges, caps on catastrophic expenditure or improvements in treatment access. An increase in health service use may be beneficial even when the household's recorded medical expenses grow. Household medical outlays, total medical resource use and health outcomes must remain separate research variables.

The randomized first-year Oregon Medicaid study (2012) reports higher insurance coverage and utilization with lower medical out-of-pocket spending and debt among households offered the opportunity to apply. Random assignment applies to the offer, not automatically to the subsequent choice to enroll; interpretation of the effect of coverage itself requires further instrumental assumptions.

The Oregon two-year clinical study (2013) found near-elimination of catastrophic out-of-pocket expenditures, decreased financial strain and improved depression screening results, alongside no statistically significant improvements in several specified measures of physical health. The correct inference is outcome-specific. A null estimate for a particular measured condition is neither proof of no general health benefit nor evidence against the observed financial protection.

Medicare age-eligibility evidence (2015) reports a substantial decline in average and upper-tail out-of-pocket expenses at age 65 together with reduced financial strain. The historical Medicare introduction study (2008) likewise identifies a reduction in medical spending risk, without a discernible mortality improvement over its initial observation horizon. These results reinforce the importance of the upper tail of loss and direct financial measures.

Unemployment protection and benefit exhaustion

Unemployment benefits finance consumption during a disruption of earnings. Eligibility requirements, replacement rates, waiting periods, coverage of part-time work and maximum duration determine the effective household benefit path. Gruber's consumption-smoothing work (1997) is a landmark in this research line; the available bibliographic record is used for its intellectual placement rather than unverified numerical estimates.

Ganong and Noel's bank-account evidence (2019) shows a sharp reduction in spending when unemployment benefits expire, despite the predictability of expiration. In their setting and under the paper's normative assumptions, extension of benefit duration produced larger estimated smoothing gains than raising benefit levels. The relative magnitude is specific to the U.S. program and the study's model.

Unemployment insurance also influences job search and the time to reemployment. Baily's optimal insurance analysis (1978) studies consumption insurance together with search responses, leisure and imperfect capital markets. A well-defined welfare evaluation must therefore include the household's insurance gain and any changes in work incentives, taxation and program finances.

Flooding, property insurance and debt

Floods can cause immediate displacement, costly temporary shelter, destruction of housing wealth, interruption of employment and collateral losses. Property indemnity may be contractually or practically directed to repairs and lenders rather than to unrestricted consumption. Thus a reported large insurance payment is not equal to a large rise in ordinary household expenditure.

The Hurricane Katrina household-finance analysis (2017) finds that flood insurance proceeds were often used to reduce outstanding mortgages rather than rebuild damaged homes. Debt repayment can improve solvency while leaving housing or current consumption impaired. The destination of insurance cash must be observed before labeling the outcome consumption smoothing.

The hurricane fiscal-transfer analysis (2017) demonstrates long-run increases in ordinary social benefits in hurricane-affected U.S. counties beyond direct disaster programs. This is evidence about the broader fiscal risk-sharing response; aggregate county transfers cannot identify each household's consumption trajectory.

A catastrophe evaluation requires a dated sequence of loss, displacement, claim filing, first payment, final settlement, repairs, debt change and return to work. A household may be protected against foreclosure but receive insufficient cash for relocation. Both outcomes matter, and neither can be substituted for the other.

Agricultural shocks and subsistence expenditure

Agricultural insurance may stabilize farm revenue, maintain necessary food consumption and prevent the distress sale of productive animals or equipment. Index products face basis risk, especially when the selected weather observation differs from a particular farm's harvest loss. The incidence of protection depends on landholding, assets, liquidity and the coverage trigger.

The Kenyan post-drought insurance experiment (2019) by Janzen and Carter reports reductions in costly coping behavior. Among wealthier households, insurance chiefly reduced asset sales; among poorer households, it chiefly reduced food consumption cuts. The findings concern distinct responses and cannot be merged into one undifferentiated percentage.

Ex ante investment effects in Ghana and China complement these ex post Kenyan results. Together, these studies support a multi-period analysis of household welfare: insurance can change the risks households accept today, the losses they experience later and the coping choices they make after a shock. Each stage requires a distinct measurement strategy.

Distribution of Smoothing Benefits

Household position and the use of proceeds

The marginal welfare value of a transfer depends on resources and needs at the time of the loss. A payment that prevents a food shortage in a low-asset household may preserve very different welfare components from one that finances repairs or debt service for a higher-asset household. The Kenyan microinsurance evidence directly supports this separation between consumption protection and productive-asset preservation.

Differences in credit access, household dependency, eligibility, prior wealth, insured assets and exposure to correlated risks shape the distribution of benefit. An average payout amount, take-up fraction or consumption coefficient can conceal these patterns. Distributional evaluation requires outcomes for groups defined in advance and reports both effect estimates and uncertainty.

Informal transfers can redistribute insurance benefits within kin networks. Dercon's safety-net analysis identifies limits and potential indirect effects extending beyond direct beneficiaries. A program may relieve the burden on relatives who would otherwise provide support, or alter the incentive to transfer funds. These general mechanisms demand empirical estimation in the specific institutional setting.

Coverage barriers and learning

A policy's availability does not guarantee affordable and reliable use. The Indian randomized insurance purchases study and the Indian rainfall participation evidence show demand affected by price, liquidity, trust and index basis risk. Enrollment alone is therefore an intermediate measure of access, not proof of sufficient risk protection.

The U.S. flood-insurance learning study (2014) finds take-up rises after recent floods before declining toward baseline. Post-disaster demand comparisons can be confounded by changed risk beliefs, disaster experience and altered exposure. An empirical design should observe when coverage was purchased relative to risk realization.

Contracts may be attached to credit, employment or social benefit systems. In each case the claimant and eventual recipient of payment must be identified. A bank receiving a property claim, a provider receiving a medical payment and a household receiving a cash transfer have distinct implications for spendable liquidity and debt.

Heterogeneity as a causal question

Wealth, credit constraints, baseline food insecurity, loss severity, location and program eligibility are plausible effect modifiers. Subgroup tests should use pre-specified interactions, consistent denominators and uncertainty intervals. A statistically significant estimate in one subgroup and nonsignificant estimate in another does not by itself establish a statistically significant difference between them.

Measurement also confronts selection into coverage, nonrandom claims and attrition following disasters. Households with stronger finances may purchase insurance more frequently and recover faster independent of that purchase. Any unadjusted comparison of covered and uncovered households can therefore attribute pre-existing advantages to the contract.

The effect of insurance on households indirectly affected by a policy can also matter. Public insurance may finance care that reduces caregiver work disruption; farm protection may change labor demand; flood payments may influence local credit and construction markets. These secondary effects should be investigated where evidence supports them.

Empirical Welfare and Household Outcomes

What the principal research designs identify

A randomized offer identifies the causal effect of the offer when assignment and follow-up are valid. When the offer changes enrollment incompletely, the intention-to-treat estimate differs from an estimate of coverage itself. The Oregon Medicaid design illustrates the distinction; instrumental-variables interpretation of coverage requires assumptions such as relevance, exclusion, monotonicity and stable treatment assignment.

Regression discontinuity around Medicare eligibility identifies a local change under continuity assumptions. Difference-in-differences for agricultural insurance depends on credible trends and treatment comparisons. Disaster exposures are not always randomly assigned to otherwise comparable households, and migration, rebuilding decisions or correlated public relief may alter the observed population. Data on bank spending can provide high-frequency responses but may omit cash and other accounts.

A useful insurance-impact study observes policy premiums, deductible and limits, loss size, claim eligibility, claim filing and settlement dates, benefit actually available, subsequent spending, debt, liquid assets, income, health and production. Financial protection from the upper tail of medical spending, reduction in food cuts, lower mortgage balances and greater crop investment are distinct estimands.

Jappelli and Pistaferri's methodological synthesis highlights the need to distinguish anticipated income movements from innovations. Claims are larger when losses are larger; simply regressing household spending on claim receipts confounds financial compensation with the adverse shock giving rise to payment.

A consistent causal contrast

Let C_t(1) be period-t household consumption under a specified insurance arrangement and C_t(0) under a specified alternative, including corresponding premia, expectations, behavior and other transfers. The mean causal effect in the defined population is E[C_t(1) - C_t(0)]. Both potential outcomes cannot normally be observed for the same household. Causal identification requires a design that estimates the missing counterfactual.

Conditioning on the same realized loss can be misleading when insurance changes prevention, production or risk taking and thereby changes the distribution of losses. A total policy effect compares complete treated and untreated paths; a fixed-loss comparison examines a narrower controlled mechanism. Neither should be substituted for the other.

Consider a hypothetical uninsured household with a 25% decline in spending during a shock and an otherwise comparable insured household with a 10% decline. The difference of declines is 15 percentage points, not automatically a causal insurance benefit. Differences in baseline wealth, shock severity and coverage choice may explain some or all of the gap. This is an illustrative identification example without estimated data.

Welfare and the insurance value of low-consumption states

Expected utility assigns especially high marginal value to resources in adverse low-consumption states under risk aversion. Welfare evaluation considers the entire state distribution, timing, premium costs and effects on work, saving and risk taking. An insurance contract may have little measured average consumption effect and still reduce costly tail risks or the need to avoid valuable productive opportunities.

The Chetty–Looney social-insurance welfare analysis (2006) demonstrates that smooth consumption does not by itself imply small benefits from social insurance, since high risk aversion can also generate a smooth observed path. The study provides a theoretical qualification to naive welfare inference, not a single measured global social-insurance benefit.

Financial strain, indebtedness and preservation of schooling or treatment access can be consequential even where a short-run expenditure measure is stable. Evidence about each additional effect must be assessed on its own. Public insurance evaluation also requires financing costs, coverage incidence, eligibility incentives and fiscal sustainability.

The optimal generosity of unemployment protection depends jointly on consumption protection and labor-market incentives, as Baily's search-and-insurance analysis establishes. The positive and normative conclusions of one country-specific study cannot simply be imported into jurisdictions with different employment dynamics or liquidity distribution.

Comparative evidence and limits

Insurance setting Principal observation Supported inference Important limitation
Indian villages, Townsend (1994) Consumption partly protected from individual shocks Informal and composite risk sharing can matter No identification of a specific policy's causal effect
Indonesia, Gertler and Gruber (2002) Major illness disrupts consumption Significant uninsured risk remains Observational illness variation
Oregon, Finkelstein et al. (2012) Lower out-of-pocket costs and debt following coverage offer Financial protection from insurance access Offer effect differs from enrollment effect
United States, Ganong and Noel (2019) Spending decreases at benefit exhaustion Timing and duration of income protection matter Program and data-specific normative assumptions
Ghana, Karlan et al. (2014) Higher agricultural investment after insurance Protection changes ex ante decisions Investment is not a direct food-consumption outcome
Kenya, Janzen and Carter (2019) Food cuts and asset sales reduced for different wealth groups Benefits operate on heterogeneous coping margins Context and subgroup-specific results
Katrina, Gallagher and Hartley (2017) Insurance pays down mortgages Protection may alter household balance sheets Debt repayment is not proven rebuilding or consumption

The table synthesizes published study findings and their inferential boundaries. It is a comparative critical reading, not a quantitative meta-analysis. Harmonization of treatment definitions, spending measures, geography, price indexes and horizons would be essential before numerical pooling.

Evaluation implications

First, specify the risk insured and the population entitled to payment. Second, reconstruct the dates of premium payment, loss, eligibility, initial household cash shortfall, first usable compensation and final settlement. Third, trace the disposition of funds into current goods, assets, debt or essential services. Fourth, report causal effects for consumption categories and balance-sheet outcomes separately, with explicit heterogeneity and uncertainty. Finally, express welfare results under transparent risk-preference, financing and incentive assumptions, checking how conclusions change when these assumptions vary.

These are research synthesis and evaluation implications derived from established economic reasoning and the reviewed studies. They constitute no separately claimed scientific theory, new branded model or universal quantitative rule.

Conclusion

Insurance can smooth household consumption by reallocating purchasing power toward low-resource states and by preserving options when losses would otherwise cause expensive borrowing, involuntary asset sales or interruption of essential expenditure. Its practical performance depends on usable payment timing, covered loss, claim reliability, limits, liquidity, credit, production incentives and other risk-sharing resources.

Complete-risk-sharing research shows why individual shocks can be partially insured through a combination of institutions; the empirical literature shows why protection is often incomplete. Health insurance can principally reduce medical expenditure tails and debt. Unemployment benefits can stabilize expenditures until entitlement expires. Agricultural insurance can affect both pre-shock investment and post-shock food or asset protection. Property insurance can repair household balance sheets through debt repayment without establishing that housing was rebuilt or ordinary consumption maintained.

The resulting scientific conclusion is precise: consumption smoothing, financial protection and welfare improvement are linked economic objects with different measurement requirements. Assessment must connect each actual payment to the household's contemporaneous liquidity, counterfactual consumption path and future resources. Evidence supports these differentiated conclusions across several well-studied settings; it does not support a universal insurance-benefit coefficient or a single globally optimal contract.

References

  • Mossin, Jan. 1968. “Aspects of Rational Insurance Purchasing.” Journal of Political Economy 76 (4, Part 1): 553–568. Mossin, Jan — source

  • Cochrane, John H. 1991. “A Simple Test of Consumption Insurance.” Journal of Political Economy 99 (5): 957–976. Cochrane, John H — source

  • Townsend, Robert M. 1994. “Risk and Insurance in Village India.” Econometrica 62 (3): 539–591. Townsend, Robert M — source

  • Deaton, Angus. 1991. “Saving and Liquidity Constraints.” Econometrica 59 (5): 1221–1248. Deaton, Angus — source

  • Morduch, Jonathan. 1995. “Income Smoothing and Consumption Smoothing.” Journal of Economic Perspectives 9 (3): 103–114. Morduch, Jonathan — source

  • Browning, Martin, and Annamaria Lusardi. 1996. “Household Saving: Micro Theories and Micro Facts.” Journal of Economic Literature 34 (4): 1797–1855. Browning, Martin, and Annamaria Lusardi — source

  • Chetty, Raj, and Adam Looney. 2006. “Consumption Smoothing and the Welfare Consequences of Social Insurance in Developing Economies.” Journal of Public Economics 90 (12): 2351–2356. DOI: 10.1016/j.jpubeco.2006.07.002. Chetty, Raj, and Adam Looney — source

  • Jappelli, Tullio, and Luigi Pistaferri. 2010. “The Consumption Response to Income Changes.” Annual Review of Economics 2: 479–506. DOI: 10.1146/annurev.economics.050708.142933. Jappelli, Tullio, and Luigi Pistaferri — source

  • Gertler, Paul, and Jonathan Gruber. 2002. “Insuring Consumption Against Illness.” American Economic Review 92 (1): 51–70. DOI: 10.1257/000282802760015603. Gertler, Paul, and Jonathan Gruber — source

  • Finkelstein, Amy, Sarah Taubman, Bill Wright, Mira Bernstein, Jonathan Gruber, Joseph P. Newhouse, Heidi Allen, Katherine Baicker, and the Oregon Health Study Group. 2012. “The Oregon Health Insurance Experiment: Evidence from the First Year.” Quarterly Journal of Economics 127 (3): 1057–1106. DOI: 10.1093/qje/qjs020. Finkelstein, Amy, Sarah Taubman, Bill Wright, Mira Bernstein, Jonathan Gruber, Joseph P — source

  • Baicker, Katherine, Sarah L. Taubman, Heidi L. Allen, Mira Bernstein, Jonathan H. Gruber, Joseph P. Newhouse, Eric C. Schneider, Bill J. Wright, Alan M. Zaslavsky, and Amy N. Finkelstein. 2013. “The Oregon Experiment—Effects of Medicaid on Clinical Outcomes.” New England Journal of Medicine 368 (18): 1713–1722. DOI: 10.1056/NEJMsa1212321. Baicker, Katherine, Sarah L — source

  • Barcellos, Silvia Helena, and Mireille Jacobson. 2015. “The Effects of Medicare on Medical Expenditure Risk and Financial Strain.” American Economic Journal: Economic Policy 7 (4): 41–70. DOI: 10.1257/pol.20140262. Barcellos, Silvia Helena, and Mireille Jacobson — source

  • Finkelstein, Amy, and Robin McKnight. 2008. “What Did Medicare Do? The Initial Impact of Medicare on Mortality and Out of Pocket Medical Spending.” Journal of Public Economics 92 (7): 1644–1668. DOI: 10.1016/j.jpubeco.2007.10.005. Finkelstein, Amy, and Robin McKnight — source

  • Gruber, Jonathan. 1997. “The Consumption Smoothing Benefits of Unemployment Insurance.” American Economic Review 87 (1): 192–205. Gruber, Jonathan — source

  • Ganong, Peter, and Pascal Noel. 2019. “Consumer Spending during Unemployment: Positive and Normative Implications.” American Economic Review 109 (7): 2383–2424. DOI: 10.1257/aer.20170537. Ganong, Peter, and Pascal Noel — source

  • Baily, Martin Neil. 1978. “Some Aspects of Optimal Unemployment Insurance: Search, Leisure, Savings and Capital Market Imperfections.” Journal of Public Economics 10 (3): 403–425. DOI: 10.1016/0047-2727(78)90054-3. Baily, Martin Neil — source

  • Fafchamps, Marcel, and Susan Lund. 2003. “Risk-Sharing Networks in Rural Philippines.” Journal of Development Economics 71 (2): 261–287. DOI: 10.1016/S0304-3878(03)00029-4. Fafchamps, Marcel, and Susan Lund — source

  • Dercon, Stefan. 2002. “Income Risk, Coping Strategies, and Safety Nets.” World Bank Research Observer 17 (2): 141–166. DOI: 10.1093/wbro/17.2.141. Dercon, Stefan — source

  • Cole, Shawn, Xavier Giné, Jeremy Tobacman, Petia Topalova, Robert Townsend, and James Vickery. 2013. “Barriers to Household Risk Management: Evidence from India.” American Economic Journal: Applied Economics 5 (1): 104–135. DOI: 10.1257/app.5.1.104. Cole, Shawn, Xavier Giné, Jeremy Tobacman, Petia Topalova, Robert Townsend, and James Vickery — source

  • Karlan, Dean, Robert Osei, Isaac Osei-Akoto, and Christopher Udry. 2014. “Agricultural Decisions after Relaxing Credit and Risk Constraints.” Quarterly Journal of Economics 129 (2): 597–652. DOI: 10.1093/qje/qju002. Karlan, Dean, Robert Osei, Isaac Osei-Akoto, and Christopher Udry — source

  • Janzen, Sarah A., and Michael R. Carter. 2019. “After the Drought: The Impact of Microinsurance on Consumption Smoothing and Asset Protection.” American Journal of Agricultural Economics 101 (3): 651–671. DOI: 10.1093/ajae/aay061. Janzen, Sarah A — source

  • Cai, Jing. 2016. “The Impact of Insurance Provision on Household Production and Financial Decisions.” American Economic Journal: Economic Policy 8 (2): 44–88. DOI: 10.1257/pol.20130371. Cai, Jing — source

  • Gallagher, Justin, and Daniel Hartley. 2017. “Household Finance after a Natural Disaster: The Case of Hurricane Katrina.” American Economic Journal: Economic Policy 9 (3): 199–228. DOI: 10.1257/pol.20140273. Gallagher, Justin, and Daniel Hartley — source

  • Deryugina, Tatyana. 2017. “The Fiscal Cost of Hurricanes: Disaster Aid versus Social Insurance.” American Economic Journal: Economic Policy 9 (3): 168–198. DOI: 10.1257/pol.20140296. Deryugina, Tatyana — source

  • Mobarak, Ahmed Mushfiq, and Mark R. Rosenzweig. 2013. “Informal Risk Sharing, Index Insurance, and Risk Taking in Developing Countries.” American Economic Review 103 (3): 375–380. DOI: 10.1257/aer.103.3.375. Mobarak, Ahmed Mushfiq, and Mark R — source

  • Gallagher, Justin. 2014. “Learning about an Infrequent Event: Evidence from Flood Insurance Take-Up in the United States.” American Economic Journal: Applied Economics 6 (3): 206–233. DOI: 10.1257/app.6.3.206. Gallagher, Justin — source

  • Giné, Xavier, Robert Townsend, and James Vickery. 2008. “Patterns of Rainfall Insurance Participation in Rural India.” World Bank Economic Review 22 (3): 539–566. Giné, Xavier, Robert Townsend, and James Vickery — source

  • World Bank. 2025. The Global Findex Database 2025: Connectivity and Financial Inclusion in the Digital Economy. Washington, DC: World Bank. World Bank — source