RESEARCH DOSSIER
Farmer Insurance Program Procurement and Competition
Abstract. The procurement of farmer insurance is an institutional decision about who will supply financial protection, which risks that supplier will bear, how service obligations will be enforced, and how the public contribution will be determined. This monograph examines insurer selection, transparency, and competition in agricultural insurance programs through comparative institutional and economic analysis. It distinguishes competition for geographically defined contracts from competition among accredited insurers for individual policyholders and from coordinated delivery through coinsurance. Primary documents from India's Pradhan Mantri Fasal Bima Yojana (PMFBY), a dated Andhra Pradesh tender, the United States Federal Crop Insurance Program, Spanish public-private agricultural insurance arrangements, and European Union risk-management policy provide the institutional evidence. Public audit findings and agricultural insurance and auction research delimit the empirical implications. Mathematical illustrations establish how historical exposure weights, unequal risk allocation, claim delays, and concentration can change the apparent value of a tender. The analysis finds that premium competition can be meaningful only within a comparable contractual specification; the procurement authority must evaluate risk-bearing capacity, service continuity, public contingent exposure, and future entry alongside current price. Agricultural insurance competition is consequently a multidimensional question of access to risk-transfer capacity and the delivery of contractually defined protection. The study proposes operational evaluation criteria derived from established insurance economics and procurement principles. It does not estimate an international causal effect of auctioning insurance contracts or attribute individual misconduct without evidence.
Keywords: Agricultural insurance; public procurement; insurer selection; crop insurance; competition; public-private partnerships; tender design; actuarial premium; insurance service quality; market concentration; fiscal risk; insurer accountability; PMFBY; Federal Crop Insurance Program.
Introduction
A public agricultural insurance program purchases or enables a promise to pay after a specified uncertain event. The authority making the purchasing decision frequently differs from the producer whose livelihood depends upon the promise. Farmers may contribute a share of premiums, choose particular coverage, submit production records, and make claims. An agricultural ministry or other purchasing authority may define the product and select the insurer. A treasury supplies funds, a regulator establishes financial and conduct requirements, and a public or commercial reinsurer absorbs a portion of the insurer's loss. The distribution of those responsibilities determines the nature of competition.
The central question of this study is how the selection of insurance suppliers affects the price, reliability, accessibility, and sustainability of financial protection for farmers. The inquiry concerns the design and outcomes of procurement as an economic institution. The calculation of crop-specific premium rates, the complete structure of a particular national scheme, and the agronomic validation of individual insurance instruments constitute neighboring inquiries. They enter the present analysis insofar as they affect supplier comparability, service delivery, or the ability to contest an award.
Three research questions organize the monograph. First, what constitutes a comparable insurance procurement object, and what evidence is necessary to distinguish genuine capacity from nominal eligibility? Second, how do allocation, disclosure, and contract-management rules govern rivalry for the program? Third, under what conditions can observed rates, service performance, and fiscal outcomes support conclusions about the consequences of competition?
The research uses purposive comparative documentary analysis. The Indian national operational rules and a dated state tender provide direct evidence of territorial award mechanisms. United States agency documentation, audit summaries, and a scholarly study provide evidence about a different delivery arrangement. Spanish administrative materials and European audit evidence establish an alternative institutional boundary. Auction literature supplies mechanisms that are applied analytically rather than transferred as measured agricultural treatment effects. All numerical examples expressly identified as illustrative are constructed for this monograph; their assumptions are stated and they are reproducible. The source cutoff is November 18, 2026. A historical document establishes the rules of its own edition, and a tender establishes announced contractual terms, not necessarily what actually occurred.
The comparison therefore examines functional equivalents across legal environments rather than treating different countries as independent repetitions of a single experiment. No tender-level microdata, insurer financial returns, individual farm records, or realized award histories were assembled into a new empirical panel. The scope of conclusions is comparative institutional reasoning, accounting consistency, and conditional economic implications. The status of current local procurement law and tender addenda must be established separately for each actual award.
Conceptual and Historical Foundations
The object of farmer insurance procurement
Farmer insurance program procurement is the organized selection and contracting of entities responsible for delivering insured protection to an agricultural population under a public, collective, or publicly supported arrangement. The object may comprise the assumption of underwriting risk, the administration of claims, policy distribution, data collection, or several of these functions under one agreement. Procurement can operate through a competitive tender, an accreditation arrangement followed by decentralized choice, or a negotiated collective delivery system. These arrangements differ in their formal legal classification; the use of procurement as an analytical category here does not classify every financial assistance agreement as a public tender.
The primary insurance object is a contingent financial obligation. An indemnity seller accepts liability for losses under specified conditions, while an index-based contract promises payment according to a predetermined observable trigger. The financial promise a farmer receives depends on the schedule of insured perils, benefit limits, exclusions, coverage dates, deductible or trigger, and settlement provisions. An administrator can be paid to execute these duties without bearing the underlying risk. A reinsurer can absorb catastrophic losses without maintaining direct contractual relations with farmers. Each supplier market must be identified separately.
A contract to handle enrollment and process claims has a different economic structure from a contract that requires the supplier to pay indemnities from its own capital. A tender that combines both functions must identify the price assigned to the risk and the price assigned to administration as far as feasible. The procurement authority cannot establish the value of an outsourced function by comparing its fee with a bundled premium that includes expected losses.
Agricultural exposure also has a specific temporal and physical structure. For crops, identify the insured crop variety or legally recognized commodity, unit of insurance, planted area, yield measure, sum insured, production season, and the precise period in which perils are covered. The crop year must be reconciled with the underwriting year, financial year, payment year, and weather observation period. For livestock, the exposure may be defined by insured animals, animal-days, feed or output prices, or a spatial index governing forage conditions. Any aggregate comparison must retain these distinctions.
The difference between the individual farmer's experience and the contractual trigger is economically consequential. Suppose a farmer loses 40 percent of a crop while the relevant area retains its normal yield. An area-yield policy might generate no payment despite substantial individual distress. Conversely, an area payment might be due even where a particular insured farm escapes loss. This difference affects the value of the product in addition to its procurement price. A measure of service reliability must therefore distinguish failure to honor a valid contractual claim from a correctly implemented contract whose trigger provides incomplete economic protection.
Public involvement and the structure of rivalry
The historical movement toward publicly supported agricultural insurance has created several distinct relationships between government and insurers. The United States Economic Research Service's crop insurance program history (2026) traces federal involvement to 1938 and describes the subsequent transition to private delivery and broader coverage. The historical changes show that public support and private distribution can coexist within a durable institutional arrangement.
The purchaser's objective is more complex than the acquisition of an ordinary commodity. Protection must exist before disaster, yet its usefulness is most visible afterward. Farmers may have little opportunity to evaluate providers on the quality of the promise before joining a program. The supplier's solvency and reporting capability are partly unobservable. The ministry may be evaluated on immediate enrollment or budget expenditure while the economic benefit depends on settlement after a future shock. Risk selection, service effort, and the timing of subsidy payments matter even when the contract's premium is fixed.
A tender awards an enforceable delivery assignment for a defined period or population. This is competition for the market. The selected insurer may have exclusive responsibility in a territorial cluster while the contract remains in force. Accreditation followed by a producer's ability to choose among competing providers introduces competition within the market. A coordinated pool allocates risk and shares infrastructure; the competitive margin may instead involve admission, distribution, reinsurance, or procurement of supporting services. These categories can coexist within one national program.
The appropriate unit of rivalry is consequently determined by actual substitutability. An insurer that is licensed nationally but lacks capacity to cover a particular crop in a remote district is not necessarily a viable competitor for that award. An independent agent who can change carrier may create distribution rivalry without giving farmers meaningful access to another product. An insurer participating through a pool contributes underwriting capacity even where claims service is centrally organized. Counting corporate names as if each were an independent supplier of the entire bundle misrepresents competition.
The economics of uncertain future performance
Agricultural insurance creates correlated liabilities. A drought or flood can generate claims across many farms over the same period. Increasing the number of enrolled producers can reduce the relative importance of idiosyncratic fluctuations but may leave a severe shared shock essentially undiversified. If a portfolio consists of n identically distributed insured losses with variance v and pairwise covariance c, its total loss variance is n v + n(n - 1)c. With positive persistent covariance, the second term becomes increasingly important as the portfolio expands. A territorial award therefore transfers an exposure distribution, not merely a larger number of standardized policies.
This distinction shapes supplier selection. An insurer whose low unit handling cost arises from processing thousands of homogeneous contracts may still require more capital or reinsurance to accept a geographically concentrated book. A provider that already holds correlated liabilities elsewhere may be unable to diversify the additional award. Procurement design affects the portfolio that the insurer can obtain, and thus its cost of supplying the promise.
The authority must distinguish actuarial adequacy from strategic bid behavior. A quoted premium reflects a supplier's expectation of claims, expenses, the cost of financing and risk transfer, and the desired economic margin under its information set. It may also reflect the bidder's desire to enter a market, obtain a distribution network, or continue its existing operations. The quotation alone does not reveal which of these components dominates. A low bid can be an efficient price, a strategic entry price, or an optimistic valuation of the risk. A high bid can reflect prudence, expensive capital, limited rivalry, or inefficient administration. Valid judgment requires evidence on the underlying liabilities and the competitive environment.
Insurer Selection
Defining a comparable procurement specification
A sound tender starts from an exposure and benefit specification. A prospective insurer needs a consistent enumeration of farms or parcels, agricultural production units, insured values, historical yields and losses, the contractual hazard trigger, data availability, and the duties of other program participants. The commissioning authority should supply the same material data and contractual conditions to every eligible bidder. Revisions to a material assumption must be versioned, dated, and circulated under the same access conditions.
Expected participation is an important input because a program can define a large eligible population without requiring every farmer to enroll. A bidder prices the expected portfolio and uncertainty about that portfolio. If the authority forecasts participation based on the previous season, new subsidy rules or a change in crop profitability can alter both total exposure and its composition. The expected number of insurance units and the estimated insured value should accompany the forecast of acreage. The contract should define the handling of errors, late corrections, and material changes in exposure.
For a contract with multiple categories of coverage, the premium quoted for each category may depend on per-acre insured values. Consequently, a percentage-of-value comparison and a total premium amount comparison can give different rankings. Any weighted scoring rule must be announced before the bids, and its reference exposure must be recoverable. Separate consideration of extreme loss exposure is essential because increasing the expected total premium does not necessarily increase the insurer's capacity to pay claims in a severe year.
The specification should identify the duties and remedies for underwriting errors, delayed issuance, failed data matching, contested yield estimates, inaccessible support, and unpaid claims. The selected insurer needs a measurable set of performance obligations. A farmer-facing coverage document should remain clear even where the agreement between insurer and public purchaser contains complex financial provisions.
Territorial bidding in India
India's 2023 PMFBY Operational Guidelines, sections 7 and 31, document a cluster-based selection mechanism. The state or union territory invites empaneled insurers to compete over grouped districts, generally using a common coverage specification and crop-district premium quotations. The lowest evaluated weighted premium identifies the L1 implementing agency. The rules envisage contracts of at least three years and establish restrictions on withdrawal after selection. Historical insured areas enter the weighting procedure, with specified substitutes for missing historical coverage and newly notified crops.
The economic logic of the cluster is to package different areas and crops into an obligation that cannot be selectively accepted after prices are disclosed. This can help combine attractive and difficult risks and support the fixed investment needed for local staff. Bundling also changes the entry threshold and may exclude suppliers with insufficient capital or a narrowly specialized operating footprint.
The national scheme creates a particularly instructive separation between the farmer's price and the full premium. Under the 2023 guideline's section 13, for ordinary listed foodgrain and oilseed crops, the farmer's maximum premium is 2 percent of insured value for the kharif season and 1.5 percent for the rabi season, or the actuarial rate if lower. This edition also specifies normal public subsidy arrangements. The financial implication is that reducing an evaluated rate may predominantly reduce the public payment when the producer already pays the capped amount. The purchaser thus has an independent interest in effective competition.
The state award can be understood only in relation to its announced risk-sharing model and public obligations. The primary document for a concrete implementation is the Andhra Pradesh 2023–24 tender (Government of Andhra Pradesh, 2023; reference AGC02-21021/7/2023). Issued July 29, the yield-based tender specifies nine clusters across 26 districts. It requests quotations under three designated risk arrangements, comprising the traditional transfer arrangement and two cup-and-cap alternatives. Section II contains a three-cluster allocation limit for each insurer separately in yield- and weather-based business, with an exception if subsequent bidders will not match the leading rate. Section VI provides for ranking weighted premiums under each model. The state retains the decision about the applicable model.
This document establishes an announced procurement mechanism in a specific crop year. It does not establish that all bids were submitted, that any named provider won a particular district, or that the resulting rates reduced costs. Those outcomes require the bid register, evaluation sheets, award notifications, acceptance documents, subsequent amendments, and actual experience records.
The cup-and-cap alternatives illustrate the importance of treating public and private risk allocation as part of the object of competition. A buyer may select a lower posted premium under an agreement that returns a portion of favorable experience to the state and shifts part of unfavorable experience back to the public budget. A full transfer rate and a shared-risk rate need to be compared by examining the distribution of cash flows under the actual terms. The labels of the models alone do not establish either their net fiscal cost or the amount of risk retained by an insurer.
Accreditation and decentralized competition in the United States
The United States uses a different contractual architecture. The Risk Management Agency's public-private partnership description (n.d.-a) identifies the Standard Reinsurance Agreement and Livestock Price Reinsurance Agreement as cooperative financial assistance agreements between the Federal Crop Insurance Corporation and participating insurance companies. Approved Insurance Providers deliver coverage through private agents. Accordingly, the relationship is not equivalent to an annual exclusive district premium auction.
The Economic Research Service (2026) describes the federal program as sharing underwriting experience and subsidizing both producer premiums and delivery expenses. This multiple-channel support means that the public price of the program is not observable in the farmer's premium alone. The procurement function involves maintaining an adequate and willing population of participating carriers, arranging risk sharing, and supporting distribution. Comparative evaluation should examine those dimensions alongside farmers' actual access to agents and covered products.
The agreement also changes over time. RMA's March 9, 2026 agreement revision memorandum records updates to the 2026 and 2027 Standard Reinsurance Agreements, including administrative and operating reimbursement provisions. Historical audit findings predate these changes, so an evaluation of present compensation cannot simply apply an earlier formula.
The United States arrangement demonstrates a structural limit to the use of the term procurement competition. An increase in the number of participating providers may create additional competition for agents and producer business without creating a conventional tender price for the same federally specified coverage. The evaluator must identify which prices are administratively determined, which fees are contractually reimbursed, and which margins are negotiated or subject to rivalry.
Coordinated supply in Spain and European risk management
Spain provides a further institutional contrast. The Spanish Ministry of Agriculture's 2024 ENESA–Agroseguro agreement announcement records a cooperation agreement for implementing the forty-fifth combined agricultural insurance plan, with EUR 284.5 million budgeted for premium subsidies for 2024. The announcement assigns ENESA the subsidy function and Agroseguro the management function on behalf of participating coinsurers.
Agroseguro's own system description distinguishes the management entity from the insurers that constitute the risk-bearing pool. This is evidence from an operational participant about its structure, rather than an independent evaluation of comparative efficiency. Coordinated administration can reduce duplicate infrastructure and support common processes. By concentrating administrative responsibility, however, it places particular weight on pool governance, performance accountability, entry terms, and the availability of alternative suppliers of contracted services.
At the wider European level, the EU CAP Network's insurance and mutual funds analysis (2025) describes varying national choices about agricultural risk management. Subsidized contracts and mutual arrangements differ in how farmers participate and how liabilities are borne. A uniform European insurer award mechanism cannot be inferred from a shared policy framework.
The comparative finding is that the number of insurers and the apparent concentration of delivery do not measure the same institutional feature in these three settings. India may allocate regional delivery through tenders; the United States can sustain multiple accredited carriers; Spain has shared management of coinsured risk. Each arrangement needs examination at the margin on which a farmer, authority, or potential provider can actually make a choice.
Eligibility and financial capacity
Prudential authorization is the beginning of insurer qualification. The contracting authority should establish that the provider has an enforceable ability to issue the specified policies, funds and liquid assets adequate for its expected retained risk, and access to suitable risk transfer where necessary. Reinsurance commitments must be assessed for contractual certainty, counterparties, exclusions, attachment points, limits, currencies, and the timing of recoveries. A letter indicating interest in supplying reinsurance is not equivalent to an executed treaty.
A capacity assessment should evaluate combined risk across all awards rather than treating each tender separately. Several individually affordable awards may together create a large regional accumulation. The analysis should distinguish ultimate net loss from the immediate peak payment requirement. An insurer can be solvent on an ultimate-loss basis and still face a liquidity deficit if reinsurance recovery is delayed. Stress testing should therefore represent both losses and payment sequences.
Qualification thresholds should correspond to the obligations of the contract. A requirement for a nationwide office network may unnecessarily exclude a qualified regional carrier if equivalent service can be supplied using a demonstrable local arrangement. Conversely, a cheap promise to establish service after winning is inadequate when claims must be handled from the first day of coverage. The authority should require evidence of readiness at a clearly defined milestone.
The choice between technical qualification followed by price ranking and a combined price-quality score depends on the precision with which performance obligations can be specified and monitored. Where the desired service is sufficiently measurable, binding minimum standards may support transparent price-based ranking. Where relevant quality differs in ways a minimum cannot capture, a documented quality assessment may be justified, subject to applicable procurement law and enforceability. Discretion should be structured so that a favorable narrative about service cannot compensate for absent basic financial capacity.
Procurement Transparency
Information symmetry and the tender record
Transparency is the capacity to reconstruct and verify the procurement decision. A complete event record links the initial specifications, common data release, bidder questions, authorized responses, submission deadlines, eligibility checks, financial bids, evaluation weights, award decision, conflict disclosures, and subsequent amendments. The record should make the evaluation reproducible without relying on the recollection of its participants.
Equal access to credible historic data is particularly important in insurance procurement. An incumbent may possess years of individual claim and enrollment records that are not available to entrants. Differences in information can produce differences in pricing that appear to be differences in operational efficiency. A public data custodian can reduce this advantage by establishing lawful common access to appropriately protected exposure and experience data, while respecting privacy, commercial confidentiality, and agricultural data rights.
A reliable dataset must retain metadata defining units, coverage conditions, and the circumstances of measurement. A yield series assembled under changing crop nomenclature or geographic boundaries can provide an unreliable basis for a new tender. Missing observations must be documented; substitution rules must work identically for all bidders. A version identifier and checksum can establish which dataset was used, but cannot certify its agronomic accuracy.
In a multiyear contract, transparency extends to the revisions that occur between awards. Changes in covered perils, valuation, territorial boundaries, expense allowances, and public guarantees should be identified and quantified. The authority should record which changes arise from preannounced adjustment provisions and which require separate authorization. Confidential negotiations after award may alter the substantive bargain and weaken the benefit of the original competition.
Balanced disclosure and competition protection
The OECD's 2025a tender design checklist identifies market research, careful lot design, appropriate award conditions, and control of commercially sensitive information as important procurement safeguards. Transparency needs to permit public accountability without unnecessarily facilitating communication among competitors.
This balance is significant in repeated agricultural awards, where participants may encounter the same geography and rival firms over several crop years. The publication of every firm's granular forward-looking rate assumptions could allow suppliers to learn rivals' strategies more precisely. Conversely, suppression of award criteria and reasons prevents meaningful review. The appropriate disclosure regime should therefore distinguish information needed by bidders before submission, information owed to the public after award, and commercially sensitive evidence retained for authorized oversight.
Binding published specifications, an anonymized common question-and-answer mechanism where lawful, auditable electronic submission, and recorded conflicts of interest can reduce the opportunity for selective access. An audit trail for late corrections preserves both the original and corrected material. Independent observers or reviewers should be able to determine whether a change affected the ranking.
Control of beneficial ownership and common interests is also relevant because nominally different bidders may belong to the same economic group or share resources. A relevant competition count must distinguish independent decision makers from legal entities whose bids cannot reasonably be treated as independent. Joint bidding requires careful scrutiny of both the parties' capacity contributions and the reason for the arrangement. A genuine capacity-sharing consortium may make a large award feasible where its members could not reasonably bid alone; an arrangement among independently capable rivals can raise different competition concerns.
The weighted-premium problem
A bid ranking is conditional on the weights used in its evaluation. Let two crop-district categories have area weights of 80 and 20 hectares. Assume equal insured value per hectare. Insurer A offers 5 percent and 12 percent, while insurer B offers 6 percent and 8 percent. The weighted rate for A is (80 × 5 + 20 × 12)/100 = 6.4 percent. The weighted rate for B is (80 × 6 + 20 × 8)/100 = 6.4 percent. The announced reference composition yields a tie.
If actual exposure becomes 50 hectares in each category with unchanged insured value, the actual average rates become 8.5 percent for A and 7 percent for B. On a total insured value of 10 million currency units, the corresponding premiums are 850,000 and 700,000. The actual difference is 150,000 currency units, although the reference scores were identical. If the second category also has a higher insured value per hectare, the monetary divergence can be greater than an area-weighted comparison suggests.
This illustration demonstrates composition sensitivity. It does not imply that either bidder manipulated its rates. A bidder with a legitimately higher expected loss cost in the second category could produce the same rate profile. Transparent procurement requires publication of the weights and their eligibility basis, reproducible scoring at the announced reference composition, and a separate analysis of exposure change for fiscal planning.
Replacing the announced calculation after the bids would create another problem: the purchasing authority could unintentionally or deliberately alter the winner. Sensitivity analysis should therefore be part of preaward design and authorized risk assessment. Once bids have been received, the declared procedure must govern ranking except where applicable law and the stated rules permit a properly documented correction, cancellation, or new tender.
Competition monitoring and evidence of misconduct
Klemperer's auction design analysis (2002) emphasizes entry and antitrust concerns in auction design. Agricultural insurance shares these concerns with other procurement markets, but the structure of correlated claims and the need for licensed risk capacity make simple transplantation of auction rules inappropriate.
Repeated territorial allocation may produce suspicious patterns such as rotation of successful bidders, systematic absence from particular areas, identical unusual bid components, or unexpected withdrawal. The OECD's bid-rigging detection list (2025b) stresses that such indicators require investigation rather than automatic classification as collusion. Cost differences, risk limits, independent capacity constraints, and shared information can produce similar patterns.
Consider an insurer that consistently avoids a drought-exposed district. Its conduct could reflect elevated expected claims or a treaty exclusion that prevents it from accepting the risk. A statistical pattern of absence is relevant evidence for inquiry, but neither an econometric residual nor a visual bidding pattern alone establishes intent. Investigators need contemporaneous records, ownership and capacity information, the relevant legal standards, and evidence of coordination where required. The tender authority should preserve evidence without making public accusations on the basis of an initial screen.
A useful monitoring database tracks eligible firms, actual bidders, disqualifications, bid vectors for each coverage class, geographic and crop exposures, bidder ownership, capacity limits, subcontractor relationships, withdrawals, award modifications, and the dates of each observation. This record allows the authority to distinguish entry difficulties from unusually coordinated behavior and makes independent assessment feasible.
Competition Outcomes
Rates and the cost of protection
The gross insurance premium must be related to the form of risk transfer actually purchased. As an accounting illustration, a gross premium can be decomposed into the expected retained claim cost after reinsurance recoveries, the ceded reinsurance premium, administrative expense, capital and financing expense, and margin. If ceded claim recoveries have already been deducted from gross expected losses, the corresponding reinsurance premium must be added consistently. Otherwise the comparison double counts or omits a component.
Public expenditures form a different accounting boundary. They include premium subsidies, delivery payments made outside the premium, public administration costs, and the net cost of retained or guaranteed risks. Actual recoveries and repayments must be deducted when assessing net expenditure. Producer premium payments should be separately identified, and indemnities funded from those premiums must not be counted again as additional public expenditure unless the state independently funds that payment.
The distinction between program cost and social value matters because a farmer receives a transfer after loss, while society bears the resource costs of administration and the opportunity costs associated with taxation and capital. A fiscal analysis focuses on the budget and contingent liabilities. A welfare analysis would also require displaced activity, risk reduction, equity, household liquidity, and the value of timely protection. Adding all cash movements as if they were independent social resource costs gives an incoherent result.
Suppose full private risk transfer is offered for 9 million currency units and a shared-risk contract for 7 million. In the second arrangement, the public expected retained loss payment is 1 million in an ordinary scenario model and 8 million in a severe scenario. Assume for illustration probabilities of 90 and 10 percent respectively, no public recoveries, and otherwise identical administrative and farmer terms. The expected additional public payment equals 0.9 × 1 + 0.1 × 8 = 1.7 million. Expected cost under the shared arrangement is 8.7 million, slightly below the 9 million transfer price; its severe-scenario cost is 15 million. If the severe-state probability is 20 percent, the shared arrangement's expected cost becomes 9.4 million. A procurement ranking based solely on premiums hides the difference in both expected and stressed public expenditure.
This example uses assumed probabilities, not calibrated agriculture data. It demonstrates a sensitivity relationship that can be evaluated with a real risk-sharing contract by replacing the hypothetical cash flows with its exact payment schedule, incorporating reinsurance, discounting, and the public authority's tolerance of fiscal shocks.
Empirical evidence from the United States
Public audit records reveal why private delivery payments deserve independent analysis. The United States Government Accountability Office's 2017 program delivery audit recommended evaluating the relationship between company compensation, market conditions, risk sharing, and expense reimbursements. The policy environment has subsequently changed, so the recommendations constitute a dated evaluation rather than a current legal statement.
The GAO's 2023 program cost update reported that federal expenditure in 2022 totaled approximately USD 17.3 billion, with about USD 3.7 billion attributed to compensation for private delivery, comprising approximately USD 2.2 billion in administrative and operating subsidies and USD 1.5 billion in the companies' share of underwriting gains. Its analysis placed the average annual return on retained premiums in 2011–2022 at 16.8 percent, above its 10.2 percent estimated market-based benchmark.
The reported return denominator is retained premiums, not necessarily shareholder equity or regulatory capital. It must not be interpreted as a universal corporate return on investment. The audit's estimated benchmark is specific to its methods and time period. Its figures identify the importance of the delivery-compensation channel without establishing that an exclusive territorial tender would improve the American program.
Using the rounded 2022 amounts, delivery compensation divided by total government cost is approximately 3.7/17.3 = 21.4 percent. This calculation is an original arithmetic restatement of the reported aggregates. It is neither the share of all social program costs nor a measure of excessive compensation. The risk actually retained and the level of service secured would have to be incorporated into such an evaluative judgment.
Competition among agents can also affect the selection of coverage. Delay, Chouinard, Walters, and Wandschneider's agent competition study (2020) reports evidence from insurance-unit data in five US states linking agent market concentration and agent market share to producer coverage choices. The authors characterize the economic magnitude as relatively small and the relationships as heterogeneous across states. This result supports examining local distribution incentives but does not demonstrate the effect of changing a national procurement system. The accessible abstract establishes the reported findings; no unexamined regression result is used here.
In a standard-rate or centrally specified environment, insurers may compete by recruiting agents rather than reducing producer premiums. Whether that competition improves farmer welfare depends on actual advice, availability, and allocation of program compensation. A provider entering high-return territories may increase rivalry there without expanding access in less profitable regions. Evaluation requires examining where entry occurs and who is served.
Claims service and the timing of payment
The usefulness of an insurance contract for a farm depends on the temporal relationship between damage, assessment, payment, and agricultural investment. A payment arriving after a producer has lost the opportunity to purchase seed or maintain a herd may have less economic value than an equal payment delivered before that deadline. The procurement authority therefore purchases a service process as well as a contingent liability.
Claim processing involves dates for notification, completion of required records, availability of weather or yield data, assessment, decision, authorization, transfer, and confirmation of receipt. A single measure from the date of the hazard to the date of payment conflates delays under the control of the insurer with delays arising from official data, enrollment errors, or the payment system. Both total farmer waiting time and attributable administrative delay are relevant, but they answer different questions.
A hypothetical 100,000-unit indemnity received six months later has a present value of approximately 93,900 units at an annual effective discount rate of 13.5 percent, calculated as 100,000/(1.135)^0.5. The calculation describes financial time value only. If delayed liquidity forces the sale of productive assets or prevents timely planting, the economic loss may exceed the ordinary discount. A procurement score that ignores delay can favor a supplier whose nominal premium advantage is outweighed by poor service.
Service metrics should separate the time to an initial decision from the time to final payment; reopened disputes from unresolved first decisions; and valid rejection from administrative abandonment. A favorable closure rate may result from inappropriate denial. Similarly, rapid processing of easy claims can mask delays among smallholders or in high-loss districts. Report distributions and upper quantiles, not only the mean. For unsettled claims, use appropriately defined censoring rather than treating pending cases as zero-delay observations.
Enrollment, access, and effective protection
Competition may be credited with expanding the number of policies when the additional records represent repeat crops or units rather than newly protected households. For this reason, the denominator of coverage must be explicit. The number of insured farm households should be distinguished from policies, individual crop units, area, livestock exposure, and total insured value. Multiple insurance contracts covering one producer are valid separate contracts without representing multiple farmers.
The European Court of Auditors' 2019 income stabilization audit examined earlier Common Agricultural Policy instruments and found limited and uneven use of supported insurance. Its discussion of the aggregate participation indicator illustrates how grouping different risk-management measures can obscure actual insurance participation. This historical finding supports maintaining separate denominators; it does not measure procurement performance under the later CAP period.
The effective protection of an insured population can be studied using independent estimates of physical or economic damage and the contractually promised indemnity. Paid indemnities can then be compared with the amounts due under the contract and with the broader economic damage. These two comparisons distinguish service performance from product adequacy. A loss ratio, defined as incurred claims divided by earned premium for a corresponding exposure period, supplies different information again. It is generally not the ratio of farmer losses to public subsidies.
A strong program should be capable of identifying cohorts for whom formal participation is high but the realized protection is weak. Examples include insured farms exposed to localized losses outside an area trigger, producers unable to complete a digital notice process, and tenant farmers whose economic interest is not sufficiently documented. These examples indicate mechanisms requiring local evidence; no incidence rate is asserted here.
Entry, concentration, and dynamic rivalry
Concentration measures can summarize award distribution if the relevant market and exposure basis are defined. For illustration, four independent insurers with 40, 30, 20, and 10 percent of awarded insured value produce a Herfindahl-Hirschman Index of 3,000 on the percentage-share scale: 40 squared + 30 squared + 20 squared + 10 squared. Four equal shares would yield 2,500. The reciprocal of concentration on the fractional scale gives an effective equal-size supplier count of approximately 3.33 for the first example.
These calculations describe the distribution of award value, not independently established market power. A national award concentration measure may conceal a local exclusive territory, while a local measure based on insurer licenses may overstate meaningful alternatives. Agent concentration, independent bidding concentration, retained-risk concentration, and reinsurance concentration may all differ. No universal inference about agricultural insurance misconduct follows from a single threshold.
Lot limits can preserve an independent supplier population and constrain accumulation. They can also prevent an efficient insurer from combining territories into a cheaper portfolio and induce otherwise unnecessary fragmentation. Seeking a second bidder's acceptance of the lowest price may preserve the initial nominal rate, but that bidder must be able to honor the required terms. If suitable alternatives cannot accept the rate, the award authority may face a trade-off between the announced diversity objective and continuity of insurance coverage.
A multi-year award strengthens incentives to invest in local knowledge, systems, and personnel if the supplier expects to recover fixed costs over several seasons. It also lengthens the interval before another firm can compete for the territory. The consequence depends on data portability, term-end obligations, service enforcement, and the process for replacing or removing a deficient provider. Competitive procurement must be evaluated over a sequence of contracts, not solely at the moment an award is announced.
Common information and the winner's curse
Agricultural insurers encounter private information about internal expenses and common uncertainty about the underlying regional loss distribution. If several bidders estimate the same future loss with error, the most optimistic estimate may generate the lowest price. Winning can thus provide information about the relation between one's estimate and rivals' estimates. A prudent bidder may adjust for that conditional information.
Hong and Shum's procurement competition study (2002) reports a winner's-curse mechanism in a model estimated using New Jersey procurement bids; for a substantial subset, modeled median procurement cost increased with competition. Its setting was not agricultural insurance, and its magnitude cannot be applied to crop tenders. The relevant theoretical implication is that rivalry under common uncertainty need not generate uniformly lower viable bids.
Suppose three equally capable bidders independently estimate the same expected aggregate insured loss as 80, 100, and 120 monetary units, while the common underlying expectation is 100. If each adds a fixed 15 units for all other costs, the lowest bid is 95. Awarding that bid does not make the underlying expected loss 80. If its provider maintains the promised risk allocation and actual common expectation is 100, the expected underwriting result before the stipulated non-loss costs is already negative by 5, and the entire expected result after those costs is negative by 20. This stylized example does not describe an equilibrium bidding strategy. It isolates the risk of selecting an optimistic estimate.
Real insurers may possess distinct expertise or reinsurance arrangements, so an apparently optimistic estimate may instead be well founded. The authority should require an adequate explanation of anomalous prices and demonstrable ability to fulfill the contract, using preannounced legal and actuarial standards. This avoids substituting a subjective official premium for legitimate actuarial competition.
Evidence Validity and Sensitivity
Source hierarchy and temporal comparability
The present study gives priority to primary operational rules, tender documents, public records and official audits. A tender clause establishes a procedure or obligation for its specified contract. A public agency's description establishes its documented institutional responsibilities. An audit establishes findings and recommendations under its own observation period. Research abstracts establish only the findings expressly reported in them. These classes of evidence answer different questions.
The national Indian operational document was examined through retrievable text covering the relevant provisions; the dated Andhra Pradesh tender was examined through searchable extracted text. Visual PDF retrieval was unavailable in the research environment. The conclusions use accessible prose rather than an unverified figure, detailed model settlement formula, or handwritten annotation. The tender is analyzed as a 2023 document rather than asserted to represent every subsequent state contract. No claim is made that its posted terms establish the signed agreement or subsequent performance.
The Spanish ministry document establishes the announced 2024 arrangement, and the management entity describes its institutional role. The European Court of Auditors report concerns an earlier CAP period. The 2026 American agency memorandum establishes revisions, while the 2023 audit supplies historical financial figures. These dates are integral parts of the observations.
The research uses selected studies in economics to identify mechanisms, not a systematic pooled effect size. The article records from the American Economic Association, the University of Nebraska repository, and Caltech provide bibliographic and abstract-level evidence. The comparison makes no claims about unexamined coefficients, sample construction details, or robustness tests.
Direct legal application would require the legal status of the procuring authority, applicable procurement and insurance regulations, the operative contract and amendments, and any binding decisions on that procurement. The present study supplies general comparative analysis; its proposed evaluation elements need adjustment to jurisdiction-specific legal requirements.
Constructing an observation dataset
A rigorous evaluation begins with a linked set of procurement, exposure, policy, claim, and fiscal tables. The primary tender identifier should be joined to cluster, award period, product, and legal version. Bid tables preserve both successful and unsuccessful observations, disqualification reasons, and all financial components. The award table identifies the contracting party, associated entities, contractual model, assigned territory, and actual start and end dates.
An exposure table identifies the insured crop or livestock unit, policy period, expected and realized acreage or other exposure, insured value, peril, and trigger. Claims must distinguish reported, admitted, denied, withdrawn, outstanding, and paid amounts. Payment dates, insurer reserves, subsidy receivables, and reinsurance recoveries require separate fields. Changes in the reporting system, boundaries, or definitions must remain observable rather than disappearing through overwriting.
For the United States, the RMA's Summary of Business provides access to aggregate historical reports and data across several dimensions, including geography and product-related categories (RMA, n.d.-b). The existence of those data is evidence of an available source for further research, not proof that every desired provider-level bid or agency identifier is publicly present. An empirical project would inspect the actual layouts, access conditions, and missing fields before choosing an estimand.
For any linked panel, the relevant date is the vintage at which records were frozen. Claim and premium data can be revised as late submissions are processed. An observation comparing paid claims from yesterday's extract with final premiums from a later extract may produce an invalid ratio. Cohorts should be defined by a common evaluation date or adjusted through an explicit development method.
Causal identification and rival explanations
A correlation between the number of bidders and the premium rate is particularly difficult to interpret. More insurers may select low-risk districts, making those districts cheaper even without a causal effect of additional rivalry. High-risk districts may also generate profitable opportunities under generous public risk sharing and attract a greater number of bidders. Both mechanisms produce dependence between market entry and risk.
Conceptually, a credible causal analysis compares the observed outcome under one procurement arrangement with the outcome that would have occurred for the same eligible exposure under an alternative arrangement. That counterfactual is unobserved. An observational design must therefore identify a source of variation that plausibly changes the procurement conditions without introducing unrelated changes in expected loss or farmer eligibility.
A difference-in-differences design may be possible when some regions alter their procurement rules at different dates. The critical question is whether, absent the rule change, the relevant outcome trends would have been comparable. The evaluation should inspect pretreatment trajectories, changes in subsidies and product offerings, anticipation, regional weather patterns, crop composition, and possible spillovers into neighboring markets. Where treatments occur at different times and effects vary, the estimation procedure must address treatment timing and heterogeneous effects.
A near-threshold design might be considered where an administrative eligibility condition creates a discontinuous change in bidding or award rights, but it requires credible evidence that the threshold is applied consistently and cannot be strategically manipulated. An instrumental-variable strategy similarly needs a defensible exclusion restriction. Mere availability of a regulatory change or a historical variable does not establish causal validity.
An evaluation of farmer service should consider selection into both the insurance program and observed claims. A scheme serving more high-risk or disadvantaged producers may record greater claim complexity, more disputes, and higher average loss costs. Comparing service outcomes without controlling for the types of cases handled could punish an insurer that expands access. Yet adjusting for postaward variables affected by the procurement rule may remove part of the effect the researcher seeks to measure. The outcome, treatment, mediators, and baseline covariates must be defined before estimation.
Climate and agronomic sensitivity
Climate-related nonstationarity can change future loss distributions relative to past averages. Historical loss rates may be less useful when heat exposure, water availability, planting dates, crop mix, varieties, or land use have shifted. The direction of the change is exposure-specific; no universal rate of worsening across all farms follows from the general concept.
An actuarial tender evaluation should examine multiple plausible weather and exposure scenarios rather than treating a single historical mean as the expected cost. Long sequences of favorable yields should not be interpreted as proof that future catastrophe risk is small. Correlation across districts matters to both capital requirements and the distribution of aggregate public retained loss.
Where indices enter the payout mechanism, observations should match the physical conditions of the relevant crop season. A rainfall measurement for an appropriate calendar month may still fail to represent water stress in a particular growth phase. Remote sensing observations can improve geographic coverage while introducing version, resolution, and validation problems. Independent field data are needed to evaluate false nonpayment and false payment; claims records alone are an incomplete basis for estimating physical damage.
The procurement specification should state which party bears correction and model risk. If a weather series is revised after settlement, there must be an agreed source hierarchy, correction window, and appeal process. An insurer cannot be expected to supply a stable price for a trigger that is materially altered after contracting without an appropriate contractual adjustment. The authority must preserve the original data and associated notices to make claims decisions auditable.
Reproducible illustrative sensitivity tests
The weighted-rate example yields a tie under 80/20 area weights, but a 1.5 percentage point difference under 50/50 weights. With the stated rates, if w is the first category's exposure share, the difference between A and B is 4 - 5w percentage points. Equality occurs at w = 0.8. This makes explicit the range of compositions under which the evaluated preference changes.
For the shared-risk fiscal example, if q is the severe-state probability, expected total cost is 7 + (1 - q) × 1 + q × 8 = 8 + 7q million. It equals the full transfer price of 9 million when q = 1/7, about 14.29 percent. The result assumes the specified payments remain fixed and ignores discounting. It provides a transparent break-even calculation rather than a real-world probability estimate.
For the concentration example, a transfer of 10 percentage points from the largest to the smallest insurer changes shares from 40/30/20/10 to 30/30/20/20. HHI declines from 3,000 to 2,600. If the two largest providers in the redistributed example belong to one controlling group, the economic-group shares become 60/20/20 and HHI becomes 4,400. Thus consolidation at the beneficial ownership level can change concentration much more than a redistribution among licensed entities.
These examples use arithmetic and clearly declared hypothetical assumptions. They do not demonstrate the magnitude of real procurement effects. Their role is to expose the variables to which observed rankings and budget consequences are sensitive, identify required records, and prevent an interpretation stronger than the underlying observations justify.
Scientific Synthesis and Research Agenda
Competition as a property of contractual options
The comparative evidence supports a functional interpretation of competition in publicly supported farmer insurance. The presence of multiple insurers within a national jurisdiction is insufficient to establish that producers or authorities can substitute one provider for another under equivalent conditions. Competitive pressure arises only where alternative providers can actually assume the same obligations and where selection or switching is institutionally possible.
A territorial tender, decentralized accreditation system, and shared delivery pool each embody a distinct choice about the contract boundary. Exclusive territorial awards can make responsibility clear and support investment in service infrastructure; they also create a period during which producer-level alternatives may be limited. Decentralized participation offers choice where sufficient independent providers operate, but rivalry can be directed toward intermediaries rather than farmers' net price. Coordinated supply can exploit economies of scale while emphasizing access to the pool and the accountability of common operations.
A meaningful international comparison should therefore use a consistent description of the competitive margin while retaining country-specific institutional identities. It should avoid a composite index that gives a cluster tender, agent competition, and shared coinsurance membership the same meaning simply because each involves private companies.
Assessing affordability and the incidence of procurement savings
A reduction in the quoted actuarial rate, holding contractual terms constant, reduces the premium amount for a fixed insured exposure. Its immediate recipient depends on the premium-sharing formula. When the farmer pays a fixed capped rate and the public authority covers the remaining eligible premium, the direct saving may accrue primarily to the government. When the farmer pays a percentage of the gross premium, the saving may be distributed between farmer and public budgets.
This observation has practical implications for evaluation. A farmer survey may detect no improvement in paid premiums despite successful price competition from the treasury's perspective. Conversely, a program may lower farmer contributions through increased subsidy while procurement efficiency remains unchanged. The affordability outcome should be decomposed into the effect of rate changes, the effect of public contribution rules, the change in insured amount, and the producer's ability to pay at the required time.
The ultimate incidence can differ from the immediate accounting incidence because additional enrollment, supplier entry, changes in compensation, and changes in farm behavior may follow. Estimating those responses requires further data. The arithmetic of the initial subsidy transaction establishes where the payment change occurs; it does not establish a complete general-equilibrium welfare effect.
Balancing scale, contestability, and protection
The authority should analyze alternative lot structures before publishing a tender. Candidate configurations can be compared on the basis of expected administration cost, exposure diversification, retained catastrophe liability, plausible qualified bidders, service logistics, and dependence on common reinsurers. Large lots may offer opportunities for diversification and efficient claims operations, but small or specialized firms may be unable to carry the necessary financial commitment. Smaller lots may enhance entry while generating fragmented service and additional coordination costs.
A documented sensitivity comparison is preferable to a fixed universal rule for the number or size of lots. When diversity constraints are important, the allocation process must specify what happens if alternative bidders cannot meet the announced price or service conditions. The expected consequences for future participation should form part of the initial design and should not be inferred exclusively from the number of winners.
Performance obligations need to be compatible with the prescribed risk-sharing terms. A supplier with little exposure to underwriting losses may face weaker financial incentives to control costs; excessive retained risk may encourage restrictive claims behavior or threaten solvency. The same service score can operate differently under different financial arrangements. Contract management should include credible payment verification, documented reasons for refusals, and oversight of conflict between margin incentives and farmer outcomes.
Practical evaluation across the procurement lifecycle
Before inviting bids, the procuring body should reconcile the legal product, coverage horizon, exposure file, forecast participation, and proposed public obligations. Independent actuarial review should characterize the range of reasonable expected loss and stress outcomes. Market engagement should assess whether several qualified entities are plausibly capable of bidding. The authority should fix the evaluation method and the information disclosure policy before bids are opened.
During selection, the authority should record the number of independent valid bids, eligibility decisions, any authorized corrections, rankings under the declared rules, and evidence that the selected provider can meet the combined capital and service requirements. Any alternate model award needs its own documented fiscal-risk comparison. The terms ultimately executed should be reconciled with the specifications and evaluation.
During operation, the relevant outcomes include insured exposure by eligible cohort, policy issuance accuracy, claim decisions and time to payment, complaint outcomes, actual administration expense, premium and subsidy receipts, and financial balances. Severe-event evaluation should assess whether all parties met the contractual payment obligations and whether the risk-sharing mechanism operated as specified. Farmer experience should be evaluated through accessible channels appropriate to the affected population.
Before renewal or retendering, the program administrator should prepare a documented account of realized costs, revisions, service performance, remaining claims, and data transferred to the public custodian. Existing contractual liabilities must survive the change of implementing agency. The outgoing provider's and incoming provider's duties require precise separation so that unresolved claims do not fall between them.
These steps are operational consequences of established procurement and insurance principles. They are presented as analytical recommendations within the scope of this monograph rather than as a new named theory or a legal code applicable in every jurisdiction.
Priority research questions
An important empirical priority is to build longitudinal registers of agricultural insurance tenders containing complete bid vectors, published exposure data, and award outcomes. The resulting dataset would permit researchers to evaluate entry, composition-adjusted rate dispersion, persistence of incumbency, and the frequency and consequences of alternative-bidder matching. Identification of relevant financial and legal versions is essential.
A second research priority is to measure the relation between provider competition and service quality using cohort-specific claim development and independent farmer experience data. Enrollment, time to decision, time to payment, complaint outcomes, and conformity with contractual benefit calculations should be evaluated jointly. The design should investigate whether administrative changes improve service in remote and underrepresented areas rather than only reducing average processing cost.
A third research priority is to estimate the fiscal distribution of alternative risk-sharing mechanisms under jointly modeled climate, price, and participation uncertainty. A model should include exact contract payment rules, public subsidy timing, expected reinsurance recoveries, and liquidity needs. Scenario analysis can reveal trade-offs that are invisible in comparisons of the headline premium.
A fourth research priority is to distinguish gains from actual competitive entry from the apparent effect of increasing the number of licensed bidders. Corporate ownership, reinsurance availability, geographic concentration, administrative subcontracting, and the potential relationship between carrier and agent markets should be measured explicitly. A researcher should document the mechanisms that make alternatives viable in a particular setting.
Finally, future comparative research should evaluate the sustainability of supplier participation over complete procurement cycles. Lower prices at one award may be accompanied by exit, renegotiation, or diminished service at subsequent awards. Higher initial expenditure may support reliable long-term participation but can also reflect rent extraction. Only longitudinal exposure-adjusted and service-adjusted evidence can discriminate among these possibilities.
Conclusion
Farmer insurance program procurement is the institutional organization of agricultural financial protection through the selection, contracting, and oversight of insurance suppliers and related service providers. Its defining economic feature is the acquisition of a contingent promise whose performance depends on the contractual allocation of risk, the reliability of agricultural data, the provider's financial capacity, and the timeliness of service after a loss.
The comparative analysis establishes that territorial competition, accreditation-based delivery, and coordinated coinsurance generate different competitive margins. India's documented cluster mechanism shows how weighted premium evaluation and territorial allocation can organize rivalry for a program. The United States illustrates a delivery structure in which government compensation, risk sharing, and intermediary competition shape incentives. Spain demonstrates the distinct issues arising from common administration within a coinsurance arrangement. These institutional differences preclude a simple international ranking by the number of providers or the gross premium.
Premium competition is informative when the compared offers provide equivalent obligations and operate under a common evaluation specification. Historical weights, variation in future insured exposure, and alternative public risk retention can alter the economic implications of the evaluated rate. The illustrative calculations show why price ranking needs to be accompanied by risk and fiscal assessment. The analysis does not establish that a particular mechanism has universally superior outcomes.
Transparency is a property of verifiable decision-making throughout the procurement lifecycle. Shared and appropriately protected information, preannounced selection criteria, independent capacity review, traceable scoring, and documented contract performance create the basis for meaningful competition. Suspicious bidding patterns warrant evidence-based investigation under the applicable rules; concentration alone cannot determine whether misconduct occurred.
The central evaluative result is that an efficient farmer insurance program supplies reliable and accessible protection at an economically justified cost while maintaining the financial capacity and contestability necessary for future seasons. Evaluating that result requires distinct measures of rate, public net expenditure, retained risk, claims service, effective protection, and the availability of independent suppliers. The realization of agricultural risk protection is the ultimate object against which procurement arrangements should be assessed.
References
Agroseguro. (n.d.). The Spanish Agricultural Insurance System. Institutional description.
Delay, N., Chouinard, H., Walters, C., & Wandschneider, P. (2020). The influence of crop insurance agents on coverage choices: The role of agent competition. Agricultural Economics, 51, 623–638. DOI: 10.1111/agec.12576.
EU CAP Network. (2025, October 8). Insurance and mutual funds for farm resilience and adaptation in the EU. Policy Insights.
European Court of Auditors. (2019). Farmers’ income stabilisation: Comprehensive set of tools, but low uptake of instruments and overcompensation need to be tackled. Special Report 23/2019. Luxembourg.
Government of Andhra Pradesh, Department of Agriculture. (2023). Tender Document: Dr YSR Free Crop Insurance (DrYSRFCI), Pradhan Mantri Fasal Bima Yojana (PMFBY), Yield Based, 2023–24. Tender reference AGC02-21021/7/2023, issued July 29, 2023.
Government of India, Department of Agriculture and Farmers Welfare. (2023). Operational Guidelines 2023 of PMFBY. Sections 7, 13, and 31; guidelines edition identified in retrieved text.
Hong, H., & Shum, M. (2002). Increasing competition and the winner's curse: Evidence from procurement. Review of Economic Studies, 69(4), 871–898. DOI: 10.1111/1467-937X.00229. Published version repository record.
Klemperer, P. (2002). What really matters in auction design. Journal of Economic Perspectives, 16(1), 169–189. DOI: 10.1257/0895330027166.
Ministerio de Agricultura, Pesca y Alimentación. (2024, April 2). El Gobierno aprueba el convenio entre ENESA y Agroseguro para la ejecución de los planes de seguros agrarios combinados en 2024, por 284,5 millones de euros. Government announcement.
OECD. (2025a). Tender design checklist. In OECD Guidelines for Fighting Bid Rigging in Public Procurement (2025 Update). OECD Publishing.
OECD. (2025b). Bid-rigging detection list. In OECD Guidelines for Fighting Bid Rigging in Public Procurement (2025 Update). OECD Publishing.
U.S. Department of Agriculture, Economic Research Service. (2026, April 2). Farm & Commodity Policy: Title XI—Crop Insurance Program Provisions.
U.S. Department of Agriculture, Risk Management Agency. (n.d.-a). Public-Private Partnerships.
U.S. Department of Agriculture, Risk Management Agency. (n.d.-b). Summary of Business. Data access portal.
U.S. Department of Agriculture, Risk Management Agency. (2026, March 9). IS-26-001: 2026/2027 Standard Reinsurance Agreement (SRA) and 2027 Livestock Price Reinsurance Agreement (LPRA). Informational memorandum.
U.S. Government Accountability Office. (2017). Crop insurance: Opportunities exist to improve program delivery and reduce costs. GAO-17-501. Washington, DC.
U.S. Government Accountability Office. (2023, November 7). Crop insurance: Update on opportunities to reduce program costs. GAO-24-106086. Washington, DC. Publicly released December 4, 2023.