Insurance by BogdanovaThe Encyclopedia of Insurance and Risk Science

RESEARCH DOSSIER

Credit Surety Insurance Bank Lending and Additionality

Abstract. Credit insurance and surety arrangements alter the allocation of loss, liquidity and enforcement risk between borrowers, lenders, beneficiaries and protection providers. Their effect on bank lending depends on which obligation is protected, the enforceability and timing of payment, the response of bank balance sheets and the availability of alternative financing. This monograph compares public small-business loan guarantees, receivables insurance, surety bonds, independent undertakings, financial guaranty insurance and development-finance guarantees. It examines financial and economic additionality against explicit counterfactuals, with particular attention to loan substitution, bank capital, contractual conditions, claim administration and fiscal exposure. Evidence from Chile, Italy and the euro area illustrates the coexistence of increased credit access, heterogeneous borrower outcomes and replacement of previously unprotected lending. The analysis distinguishes gross supported financing from causal net credit growth and from changes in loan terms. Contractual examples from the United States and a proposed Sierra Leone infrastructure transaction establish how beneficiary rights and payment sequencing mediate credit enhancement. Illustrative calculations expose the sensitivity of expected loss, effective protection and fiscal cost to coverage, collection delay and correlated default. The central conclusion is that additionality is an empirical property of a financing intervention at a specified level and horizon; contractual risk transfer is a necessary element of some interventions but does not establish that property on its own.

Keywords: Credit insurance; surety insurance; loan guarantees; bank lending; financial additionality; economic additionality; credit risk mitigation; project finance; credit substitution; contingent liabilities; claims and recoveries.

Introduction

A loan finances activity by advancing purchasing power against a promise of future repayment. A credit enhancement changes the allocation of the consequences of a failure to perform that promise or of another obligation on which repayment depends. Its relevance to finance is established by the relation between the protected obligation and the cash flows available to the lender. The same nominal amount of cover can therefore support a marginal loan, reduce the cost of an existing loan, replace a bank's retained exposure, or contribute to the completion of an infrastructure procurement without increasing aggregate borrowing.

This subject belongs to the science of credit, surety and financial guarantee insurance because the insurance obligation, the original contract and the financing transaction represent distinct objects of study. A supplier's insured invoice is an asset of the supplier; the supplier's borrowing is a liability of the supplier. A contractor's performance bond addresses completion risk; a project loan concerns debt service. A public guarantee of that loan assigns a defined part of default loss to a public institution. Treating these as identical merely because they enhance credit would destroy the information needed to evaluate the protection.

The investigation addresses three connected questions. What mechanism converts a promise of contingent payment into an improvement in access to credit? Which improvements represent additional finance or an additional economic outcome relative to a credible alternative? Where does credit risk migrate when the original bank, insurer, surety or sovereign changes its position? These questions govern the selection and interpretation of the evidence. Legal examples are jurisdiction-specific. Empirical studies are interpreted within their populations and designs. Hypothetical numerical cases serve as transparent sensitivity exercises, without representing actual portfolio data.

The method is a structured narrative synthesis of selected institutional texts, legal provisions, a dated contractual specimen, a development-project appraisal and original empirical research. It is not a statistical meta-analysis or an exhaustive jurisdictional survey. The literature search emphasized primary institutional publications and original working papers; where only an author abstract or official summary was accessible, conclusions remain confined to that material. No transaction-level loan or claims microdata were independently obtained. International comparison is functional, not a declaration that the same contractual doctrine applies everywhere.

Conceptual and Historical Foundations

The protected obligation

In credit insurance, a policyholder ordinarily obtains cover for a loss arising from the failure of an obligor to pay an insured debt. The identity of the insured, the debtor, the insurer and the ultimate recipient of proceeds must be recorded separately. When a bank lends against insured receivables, the bank may benefit through an assignment or other contractual arrangement, but its claim against the exporting firm remains legally distinct from the insured buyer's debt.

A conventional accessory surety arrangement joins a principal, an obligee and a surety. The surety's obligation is related to the principal's undertaking and the agreed bond wording. Remedies may involve financing completion, arranging another contractor, making payment or responding to a monetary claim, depending on the instrument and applicable law. In US federal procurement the bond requirements in FAR 28.102-1 distinguish performance and payment protection. The public guarantee of a surety's exposure adds another relationship: under 13 CFR section 115.15, the participating surety retains responsibilities for monitoring its principal and bonded work. Neither the existence of that public backing nor the naming of a bank in a separate financing agreement makes the bank the bond obligee.

Independent undertakings are structured differently. Articles 2 and 3 of the United Nations' 1995 independent guarantees convention establish the documentary character and independence of the undertaking within the convention's scope. A complying demand under an independent guarantee or standby credit can therefore give rise to a payment duty without adjudication of the underlying transaction. The precise documentary conditions, exceptions, applicable law and expiry remain essential. Independence is a legal characteristic rather than a guarantee of immediate settlement in every dispute.

Financial guaranty insurance can directly concern debt instruments and scheduled obligations. New York's Insurance Law section 6901 illustrates why regulatory classification must be tested against legislation: it identifies financial guaranty insurance by events and obligations and provides explicit exclusions for separately defined kinds of insurance. It is unsuitable to infer from a product label alone either the insurer's licensing authority or its capital obligations in another jurisdiction.

Public participation and the development of financial institutions

The financing function of third-party promises has expanded from transaction security to organized insurance and public credit support. The institutional evolution is reflected in the 1995 convention for independent undertakings, the dated 2015 international principles for public SME schemes and the 2018 multilateral additionality guidance. These documents address different problems: cross-border payment certainty, sustainable public credit intermediation and the justification of development-finance intervention. They do not establish a single historical origin of guarantees.

The OECD's export credit policy overview differentiates public direct financing from public support delivered through insurance or guarantees of privately funded credit. This distinction determines both the balance-sheet entry and the economic counterfactual. A guarantee can support private origination without furnishing the initial principal; a direct official loan supplies funding itself. Both may entail public risk. A comparative assessment must identify the source of initial cash separately from the bearer of contingent loss.

Comparative contract characteristics

Instrument Protected exposure Typical immediate beneficiary Important residual risk
Trade credit insurance Insured buyer receivable Seller or qualifying assignee Uninsured percentage, disputes, exclusions, timing
Loan guarantee Eligible borrower debt Lender Uncovered share, servicer compliance, provider default
Performance surety Contract performance Named obligee Completion cost beyond limit, bond defenses
Payment bond Specified payment obligations Qualifying claimants Claim eligibility and time limits
Independent standby or demand guarantee Stated documentary contingency Named beneficiary Expiry, presentation, issuer credit
Financial guaranty insurance Specified financial obligation Insured creditor or trustee Coverage scope, insurer solvency
Project payment guarantee Specified contractual payment chain Contractually designated beneficiary Other project risks, draw and reimbursement conditions

The table is a functional comparison, not a substitute for the actual terms. A single project may use several instruments with distinct triggers, limits and beneficiaries. In that case the protections must be mapped to separate loss events before their value is aggregated. A guarantee of a guarantee creates a further recourse relationship rather than a second independently recoverable amount for the original creditor.

Credit Access

Borrowing constraints and channel of transmission

Access to credit depends on a bank's assessment of repayment capacity, information quality, enforceable security, funding and capital. Borrowers with valuable projects may have assets of limited collateral value or a short history of operating results. From the bank's perspective, a promise by a credible third party can lower the expected shortfall if a covered borrower fails to pay. It can also reduce uncertainty surrounding recovery or relax a binding portfolio limit. The resulting change in eligibility is contingent on the economics of the proposed loan: protection fees, servicing requirements and retained losses must fit within expected revenues.

A useful decomposition of a bank's minimum acceptable loan return is funding cost plus servicing cost, expected net credit loss, required compensation for capital and liquidity, and a commercial margin. A guarantee may change several of these components simultaneously. A lender constrained by liquidity may still reject an application despite a reduction in risk-weighted assets, while a lender constrained by capital may increase lending after eligible risk mitigation. Pass-through into borrower pricing depends on the competitive position of the bank and the ability of the borrower to obtain alternatives.

The World Bank and FIRST Initiative (2015) public guarantee scheme principles place delivery, partial cover, eligibility, pricing and claims within governance and risk management. This institutional design recognizes a basic trade-off: broader protection can facilitate participation, but the financial sustainability of the intermediary and continued bank underwriting matter. Partial coverage preserves a direct financial stake in loan performance; it cannot by itself guarantee prudent behavior if losses are transferred elsewhere through collateral, side agreements or other hedges.

Receivables finance and lender rights

An exporter that sells on deferred terms may request a revolving advance secured by eligible invoices. If a lender excludes foreign receivables from its borrowing base because of collection or country risk, credit insurance can make some invoices acceptable. The EXIM multi-buyer insurance description expressly identifies insured receivables and approved assignment of proceeds as possible supports for financing. This is evidence of a contractual financing pathway, not evidence of the size of any incremental borrowing.

Consider a hypothetical exporter with USD 1 million of eligible invoices and an advance rate of 65%, giving a borrowing base of USD 650,000. If a lender accepts 85% after suitable insurance and verification, the base becomes USD 850,000: a potential USD 200,000 increase. The actual increase depends on available commitments, drawdowns, reserves, other security and demand for funding. If the exporter already had unused capacity of USD 300,000, the additional eligible base need not generate a single additional dollar of financing.

The distinction between nominal insurance and enforceable lender rights is particularly clear in EXIM's March 2013 enhanced assignment agreement. The specimen links favorable payment rights to shipment-related documentation, timely premiums, notice, claim filing and the assignee's own compliance; it also sets limits and recourse following payment. The specific conditions are historical and cannot be assumed to govern a contemporary facility. They show why a credit officer must distinguish an ordinary assignment of proceeds from an undertaking that protects a compliant assignee against certain acts of the original insured.

Invoice fraud, credit notes, commercial disputes and double financing can compromise the value of receivables even when the buyer appears creditworthy. An insured invoice that overstates actual goods supplied is not equivalent to an enforceable debt. The bank should therefore analyze assignment priority, cash control, reconciliation of invoice and shipment records, coverage limits by buyer, the treatment of rebates and dilution, and the applicable legal regime. These are analytical underwriting considerations, not an assertion that every policy contains the same provisions.

Prudential recognition and capital capacity

Basel Committee (2020) CRE22 credit risk mitigation sets out conditions for recognizing protection in the standardized capital approach, including documented, direct and timely enforceable rights, relevant restrictions on cancellation and conditions, and adjustments to eligible coverage. National implementation and the lender's regulatory approach determine the actual calculation. An insurance contract that reduces expected economic loss may fail regulatory eligibility; eligibility itself does not establish real lending additionality.

Suppose a hypothetical USD 100 million exposure carries a 100% risk weight; qualifying proportional protection of 80% has an assumed provider risk weight of 20%. With no applicable mismatch or other adjustment, risk-weighted assets become 80 x 0.20 + 20 x 1.00 = USD 36 million, rather than USD 100 million. At an illustrative capital ratio of 10%, the associated requirement falls from USD 10 million to USD 3.6 million, a reduction of USD 6.4 million. These risk weights and capital ratio are assumed, not a determination for an actual provider, borrower or jurisdiction. The example excludes buffers, leverage constraints, expected loss treatment and other supervisory requirements.

The bank could use the released capacity to finance a previously excluded borrower, maintain a portfolio that would otherwise contract, acquire safer assets, or reduce the capital it allocates to the business. The change in risk-weighted assets identifies a channel through which lending may increase. Only subsequent portfolio and counterfactual evidence establishes how that capacity was used.

Project financing

A project company generally repays debt through cash flows from the project, with the exact recourse defined in the financing documents. A guarantee of an offtaker's payments can stabilize an input into that cash-flow chain. A construction completion bond can reduce the chance of losing the investment before operation; a loan guarantee protects specified creditor payments. The World Bank guarantee description distinguishes project loan guarantees from guarantees of government payment obligations outside loan contracts. The intercreditor arrangements guidance describes counterindemnity and the coordination of claims over project assets.

The distinction is concrete in the World Bank Group's 2016 Western Area project appraisal. For the proposed Sierra Leone power project, IDA contemplated up to USD 40 million of cover on repayment to a commercial bank of draws under a standby letter of credit. The letter of credit was intended to respond to specified payment-security events for the project company, and a draw would create a reimbursement loan to the public counterparties with a 12-month reimbursement period. The guarantee beneficiary and project company's rights under the letter of credit were therefore different relationships. The appraisal is evidence of proposed structure, not proof that the proposed documents were signed, a draw occurred or additional electricity was produced.

For financial analysis, let annual cash flow available for debt service be USD 12 million and scheduled debt service USD 10 million, giving a debt service coverage ratio of 1.20. If an offtaker delays USD 3 million, available cash may temporarily fall to USD 9 million, below scheduled service. Timely liquidity support of USD 3 million can restore the ratio in this simplified period. If repayment of the support subsequently creates another senior obligation, the next period's debt-service burden must be modeled as well. An isolated period of restored coverage should not be treated as permanently improved project solvency.

Additionality

Counterfactual definition and measurement

Financial additionality is the difference between a financing outcome with an intervention and the financing outcome that would have occurred without it, for an explicitly defined unit and horizon. Examples are access, net borrowing, interest cost, collateral burden, maturity, committed capacity and ability to finance an otherwise infeasible transaction. Economic additionality concerns outcomes such as productive investment, output, employment, productivity, survival or welfare. The two outcomes require separate tests.

The 2018 MDB additionality framework locates development-finance contribution beyond market availability and identifies crowding out as an issue. The OECD (2017) evaluation of guarantee programs concludes from its review and survey that evidence is generally more favorable for financing access than for economic outcomes, with limitations in evaluations of counterfactuals and sustainability. These different criteria explain why a volume report by an administering institution cannot serve by itself as an impact evaluation.

Define total bank debt B for a borrower as guaranteed debt G plus other bank debt U. At a fixed date, relative to the same no-intervention counterfactual, the identity is:

Net change in B = change in G + change in U.

If the intervention introduces USD 100 of guaranteed credit and unprotected credit is USD 80 below its counterfactual, net debt increases by USD 20. A ratio of net credit change to the change in guaranteed credit is then 0.20. This arithmetic defines a descriptive or causal ratio according to how the component differences were estimated; it does not automatically identify causation. Gross origination flows cannot be substituted for debt stocks without reconciling repayments, cancellations and exchange-rate changes.

The relevant counterfactual may differ by time and level. A new loan can be additional in a crisis because without support a viable firm would have faced a withdrawal of its existing facility. A stable observed credit stock may hide an economically important positive effect relative to a declining alternative. Conversely, rising aggregate debt may reflect unrelated policy easing or recovery. The evaluator must specify what would have happened in the absence of the guarantee, including concurrent monetary and fiscal measures.

Evidence from eligibility thresholds

Mullins and Toro (2018), in Credit Guarantees and New Bank Relationships, examine a Chilean SME program using discontinuity in eligibility. The authors report additional borrowing and subsequent formation of other banking relationships within eighteen months, as well as growth in firm activity. The design supports an interpretation local to the eligibility margin under its identifying assumptions. It does not imply that all firms in Chile or guarantees elsewhere would have the same response.

An increase in relationships provides a plausible durable access channel. A bank's willingness to accept a new client can be enhanced by the initial protection, and subsequent repayment may generate a credit history. Competing explanations include borrower switching in response to favorable pricing and concurrent improvements in firm demand. Data on the dates of first borrowing, previous bank relationships, contract terms and loan termination provide a way to separate such mechanisms.

de Blasio and colleagues (2017), in Public Guarantees on Loans to SMEs, exploit the Italian fund's score-based eligibility discontinuity. Near the boundary they report increased borrowing, no detected improvement in interest cost and increased repayment difficulty. This result is compatible with a guarantee relaxing a rationing constraint while admitting riskier borrowers at the margin. It is not sufficient to establish that banks reduced their screening effort. That competing hypothesis requires observing lender behavior or exploiting a change in retained risk independent of borrower composition.

The two studies establish why heterogeneous effects matter. Additional volume, pricing and default can move in different directions. Comparing an effect estimated at one country's eligibility threshold with another country's national average would mix estimands. Pooling these results without harmonized exposure definitions, outcomes and sampling designs would create a number that lacks a coherent empirical meaning.

Financial access and use of proceeds

An additional dollar of debt may finance inventories, substitute for equity, refinance expensive obligations, pay wages or permit a firm to retain cash. Economic additionality depends on the outcome the intervention is intended to produce. A loan used to refinance debt can avert liquidation without increasing gross investment. Such a result may be valuable when the policy objective is preservation of viable businesses in a crisis. For an investment-expansion program, the same use of funds may indicate weak targeting.

A useful evaluation records whether the firm is a first-time bank borrower, the reason the prior application was constrained, the use of disbursed funds and changes in other financing. Economic outcomes should be assessed over horizons compatible with the intervention. Employment effects in six months cannot resolve the lifetime productivity consequences of a fifteen-year project. The appropriate horizon follows the expected economic mechanism, rather than the administrative reporting cycle.

Substitution Risks

Replacement of unguaranteed lending

The working-paper abstract of Altavilla and colleagues (2023), Loan Guarantees and Bank Lending, reports that euro-area banks extending guaranteed loans during pandemic programs reduced their unguaranteed credit by more than 30% relative to other banks serving the same firm. Recipients also obtained more favorable interest rates and longer maturities. The percentage concerns a relative estimate for unguaranteed lending; it is not the fraction of all guaranteed euros displaced across the economy.

This coexistence of substitution and improved terms is economically coherent. A lender may protect exposure previously borne on its balance sheet and share part of the benefit through lower pricing or a longer tenor. The borrower obtains a contractual improvement even if total quantity changes little. The government or insurer, however, has assumed a contingent obligation. A full evaluation asks who captures the value of the new protection and whether the same improvement could have been achieved through a less costly instrument.

Consider a bank extending a USD 100 guarantee-backed loan that repays an existing USD 100 facility at that bank. Gross supported origination equals USD 100, but the customer's loan stock is unchanged. If the interest rate drops from 8% to 5% on the same outstanding balance and fees add 1%, the illustrative annual financing-cost saving is USD 2, before amortization or other changes. If the guarantee fee is paid by the borrower, it must be included when measuring pass-through. The public or private value of that benefit depends on the guarantee's cost and the counterfactual risk of credit withdrawal.

Lender behavior and contractual administration

Risk transfer can influence screening, monitoring and recovery because the bank's retained stake changes. Increasing the nominal guarantee percentage may encourage banks to approve previously excluded applicants; it may also alter the benefit of maintaining careful underwriting. A partial guarantee should be assessed together with collateral allocation, pricing, seniority, lender compensation and the entitlement to subsequent recoveries. A lender with an ostensibly substantial uncovered portion may nevertheless retain little actual downside if it enjoys priority access to collateral.

US SBA loan-guarantee administration demonstrates why operational risk matters. Section 120.520 requires documentation sufficient to justify purchase and addresses continuing servicing obligations. Section 120.524 permits release of liability for specified failures, including material noncompliance, and establishes recovery rights. These provisions illustrate a contingent relationship between loan administration and guarantee realization; they should not be conflated with Basel eligibility or foreign banking rules.

A model that assumes all nominally covered losses are paid immediately will overstate protection whenever there are material claims disputes, documentary deficiencies, settlement delays or provider solvency constraints. Conversely, a surety's intervention before formal default can preserve project value through completion and thereby reduce the final loss. Assessment of such an intervention needs claims development records, costs incurred and actual recoveries, rather than a simple default indicator.

Private capacity and systemic transmission

A guarantee transfers risk to the institution that promises payment. If that institution diversifies exposure across borrowers with weakly related losses, it may be able to absorb risks more efficiently than an individual bank. If its book is concentrated in a country, sector, common buyer or reinsurer, diversification can fail in stressed conditions. The bank may then replace many heterogeneous borrower exposures with a common contingent exposure to one insurer.

van der Veer (2015), in Loss Shocks and Insurance Supply, reports using a single global insurer's data for 1992-2006 that greater claims accompanied reduced availability and higher pricing of private export credit insurance; the study also reports evidence of transmission across countries. The result supplies a mechanism for possible insurance-driven tightening of trade and receivables credit. It does not directly estimate the causal response of a bank's loan supply or establish that every insurer behaves similarly.

The Berne Union's June 2026 2025 industry report announcement reports member claims exceeding USD 11 billion and attributes a material part of the change to significant political events. The announcement measures reported industry developments, and is not a complete portfolio-level dataset. Claims in one year may derive from earlier commitments. Annual claims divided by annual new business is therefore not a cohort loss ratio, and aggregate industry capacity cannot be equated with net new bank lending.

Fiscal substitution and the sovereign-bank link

A public guarantee can convert contingent private credit losses into a future government cash requirement. The IMF's debt guarantees assessment tool emphasizes disclosure, monitoring, risk pricing and stress analysis of those liabilities. Effects on government debt and cash depend on when the guarantee is called, how it is funded and the applicable reporting system. The absence of a current cash outflow does not imply absence of an economic cost.

A downturn can simultaneously raise corporate defaults, weaken bank capital, reduce insurer capacity and strain public revenue. Government cover that is most needed in a downturn may therefore become expensive precisely when fiscal capacity is weakest. If lenders hold debt of the supporting government and rely on its guarantees, the same macroeconomic shock may impair both assets. This does not negate the usefulness of countercyclical support, but it explains the need to evaluate bank resilience and fiscal exposure jointly.

Evidence Verification and Model Sensitivity

An illustrative expected-loss calculation

The following arithmetic is a stylized one-period exercise, not a newly proposed actuarial theory. Let E be the loan exposure, p its probability of default over the period, l the loss fraction after ordinary recoveries, c the proportional coverage of that loss, q the conditional probability of full valid settlement by the protection provider, d the discount factor for payment delay, and f the guarantee fee paid by the bank in money units. Assume identical conditional loss severity when protection is paid and when it fails, no partial settlement in the latter event, and no behavioral response.

Expected unprotected loss = E x p x l.

Present value of expected protection proceeds = E x p x l x c x q x d.

Expected bank cost = E x p x l - present value of expected protection proceeds + f.

Assume E = USD 10 million, p = 5%, l = 60%, c = 80%, q = 95%, a one-year delay discounted at 5% (d = 1/1.05), and f = USD 60,000 paid immediately. Expected unprotected loss is USD 300,000; expected nominal proceeds are USD 228,000; their present value is approximately USD 217,143. The expected bank cost is approximately USD 142,857 including the fee. The calculation discounts protection proceeds only, assuming unprotected loss is valued at the base date; it deliberately does not model dated collateral or premium streams.

Scenario Default probability Loss fraction Conditional settlement Expected bank cost
Base 5% 60% 95% USD 142,857
Lower settlement reliability 5% 60% 65% USD 211,429
Recession stress 12% 70% 80% USD 388,000
Recession plus weaker provider 12% 70% 50% USD 580,000

Coverage, fee and discount factor remain fixed across rows; all four figures follow the stated assumptions. The stress exercise makes the dependence between underlying defaults and protection performance visible through changes in conditional settlement. The base-case unconditional default probability of the guarantor is not an adequate substitute for its likelihood of paying a claim when the insured obligor is in distress.

These are expectations, not measures of the distribution of loss. Two portfolios with identical expected claims may have radically different aggregate losses at high quantiles. For a provider guaranteeing loans with common exposures, scenario analysis should include concentration by bank, borrower, sector, geography, maturity and underlying correlated economic factors. A risk-transfer instrument with a portfolio loss cap must be modeled at the aggregate level; per-loan coverage percentages alone cannot recover exhaustion behavior.

Coverage structure and liquidity

Consider a realized loan loss of USD 100. An 80% proportional guarantee allocates USD 80 to the guarantor and USD 20 to the bank before contractual limitations. A first-loss guarantee capped at USD 80 would instead pay the first USD 80; at a USD 30 loss it would pay USD 30 rather than the USD 24 of proportional coverage. If the bank has a first-loss position of USD 20 and the guarantor covers the next USD 80, a USD 30 loss generates only USD 10 of guarantee proceeds. The same maximum protection can consequently have quite different screening incentives and expected losses.

Claims administration has a separate liquidity dimension. A borrower may miss a payment today while a valid insurance claim is paid six months later. The bank still has obligations to its depositors and creditors during that interval. The protection's discounted value and the bank's financing cost during the waiting period are related but distinct quantities, and both should be considered when evaluating a delayed indemnity. Liquidity facilities and collateralized guarantees can change these effects without necessarily changing ultimate loss.

For public financial evaluation, apply separate accounts for guarantee payments, recoveries, fees and administrative costs, with dates for each stream. For example, hypothetical lifetime paid claims of USD 6 million, recoveries of USD 1.5 million, fees of USD 2 million and administration of USD 0.5 million imply USD 3 million nominal net cost. If additional borrowing attributable to the program is USD 25 million, nominal cost per additional dollar is USD 0.12. This omits discounting, state-contingent risk and indirect fiscal effects. If additional borrowing is USD 5 million, the ratio becomes USD 0.60. A positive reported guarantee volume cannot resolve the denominator.

Data design and identification

Administrative information should connect applications (including rejections), eligibility criteria, guarantee approval, loan agreements, drawdowns, repayment, default, claims, payments and recoveries by durable identifiers. A guarantee ID should not be mistaken for a loan ID or a borrower ID. Multiple guarantees can refer to one facility, while renewal can create administrative records without incremental principal. A claims payment may relate to origination years earlier. Cohort analysis should therefore preserve both underwriting and settlement dates.

For cross-border exposures, record original currencies, exchange rates and valuation dates separately. Comparisons of flows may use transaction-date exchange rates while balance-sheet exposures require a stated reference date. A rise in reported dollar exposure can reflect exchange-rate appreciation rather than new cover. Nominal and inflation-adjusted borrower outcomes also differ. Data definitions must state whether reported guarantees are signed commitments, amounts outstanding, exposure at risk, disbursements or actual claim payments.

An eligibility cutoff may identify a local intention-to-treat effect if potential outcomes vary smoothly through the threshold in the absence of assignment, manipulation is absent and other policies do not change at the same boundary. If not all eligible firms obtain protection, a fuzzy discontinuity may estimate an effect for compliers under additional assumptions. The validity checks include density near the cutoff, continuity of predetermined variables, sensitivity to bandwidth, and investigation of contemporaneous eligibility rules. Effects at the margin should not be extrapolated across the whole distribution without further identification.

An alternative is randomized assignment of scarce guarantee access among otherwise eligible borrowers, where legally and practically feasible. Difference-in-differences studies need credible untreated comparators and temporal assumptions; bank-firm fixed effects may help control for stable relationship differences but do not automatically remove shocks that vary by lender or borrower over time. Spillovers through banks, other firms and markets matter for the interpretation of each design. An evaluation of treated borrowers may fail to detect credit displacement from untreated borrowers competing for the same bank capacity.

Researchers should report confidence intervals, missing-data patterns, sample attrition, changes in outcome definitions and the sensitivity of results to alternative counterfactuals. Rejected claims, withdrawn applications, firms exiting the sample and loans with unresolved recoveries must remain visible. Loss ratios on paid claims alone systematically omit the contingent exposures still developing; causal comparisons based only on successful borrowers select on an outcome.

Integrated Scientific Synthesis and Research Agenda

The chain from contractual promise to economic outcome

The literature and original institutional documents support a differentiated interpretation of credit enhancement. The first step is the creation of a legally enforceable protection relationship. The second is recognition of that protection by the lender in underwriting, pricing, liquidity management or capital allocation. The third is a change in financing relative to its alternative. The fourth is the use of finance in economic activity and the distribution of its benefits and residual losses. Each step can succeed while a subsequent step fails.

A guarantee that passes legal scrutiny may be commercially redundant. A bank may value insurance but retain the preexisting lending limit. A borrower may receive credit and undertake unproductive investment. A project may become financeable while generating substantial contingent fiscal liabilities. These observations identify separate empirical tests rather than successive elements of a new named theory.

There is no universal optimal guarantee percentage independent of the constraints at issue. A scheme designed for otherwise viable firms lacking collateral may need sufficient protection to make new underwriting feasible while retaining bank involvement. A crisis facility may emphasize prompt preservation of liquidity and must anticipate refinancing of existing exposures. Long-term infrastructure support may address a narrow public payment or termination risk rather than broad commercial failure. Targeting and performance measures should follow the intervention's declared purpose.

Evidence matrix and interpretation

Source or case Observed or documented object Supported inference Boundary
Chilean SME working paper Eligibility threshold and borrowing relationships Positive access effect in studied setting Local design; no universal effect size
Italian SME working paper Eligibility score, lending and repayment Quantity and repayment effects can coexist Threshold estimates are local
Euro-area working paper Within-firm bank comparison during pandemic Credit substitution can coexist with better terms Working-paper abstract only
EXIM assignment specimen Historical contractual lender conditions Documentary compliance affects payment Actual current policy not examined
Sierra Leone appraisal Proposed LC and reimbursement guarantee structure Payment security can support project liquidity No realized impact established
Private export credit working paper Insurer's bilateral supply and claims data Insurance availability may react to loss shocks One insurer; abstract only
Basel framework Dated regulatory requirements Capital relief requires eligible protection Local implementation varies
OECD evaluation report Literature review and program survey Evaluation of economic impact remains difficult Historical review, not current global census

The evidence gives stronger grounds for identifying mechanisms and heterogeneity than for estimating a pooled international impact. The analysis is intentionally asymmetric: causal evidence is concentrated in public SME guarantees, while private trade-credit and project-finance sources establish contractual pathways and supply behavior. Evidence of a mechanism in those markets does not authorize borrowing an effect coefficient from the public SME literature.

Directions for research and monitoring

Program evaluation should establish a baseline before introducing new protection. It should preserve application-level evidence, specify the intended outcome and record displacement across financing instruments. For private insurers, a particularly valuable research design would link buyer-limit changes, policy renewals and receivables eligibility to bank borrowing, while separating common buyer deterioration from exogenous changes in insurer capacity. For surety, the research unit could be the contractor-project-bond relationship, observing bid participation, cash collateral, working-capital utilization, completion intervention and actual recoveries.

Project-finance studies require executed contracts and dated cash-flow data. A financing-close comparison based on appraisal documents alone cannot establish additionality. Useful evidence includes commercial bids without public protection, competing term sheets, covenant schedules, draw records, reimbursement and guarantee payments, project completion and service outputs. The research should account for public subsidy, fee income, debt-service consequences and competing uses of fiscal capacity.

A common reporting protocol can preserve contractual diversity while enabling comparison: identify the obligation and beneficiary; establish the nominal and effective protected share; measure net credit quantity and terms against an explicit alternative; follow claims through recovery; document fiscal and portfolio concentration; and observe the economic outcome at a suitable horizon. These are methodological recommendations derived from existing evidence and standard accounting distinctions, not an assertion of scientific priority.

Limitations

The selected sources establish a comparative account of contractual mechanisms and several empirical patterns. The search was selective and no complete systematic-review protocol or bibliometric census was undertaken. Several sources were available only as official summaries or author abstracts; their claims are restricted accordingly. PDF evidence was assessed through extracted text of stated sections; attempted page-image retrieval failed. Jurisdiction-specific illustrations are dated, and local regulatory transposition, current signed contracts, solvency information, independently verified transaction-level claims and recovery data are outside the evidence actually obtained. No universal effect size, optimal parameter, or novel theory is claimed.

Conclusion

Credit insurance, surety and financial guarantees expand the set of methods by which banks and other financiers can manage default, payment and performance risk. Their economic role arises from the exact relationship between the covered event and the lender's exposure. Direct loan protection, insured trade receivables, completion security and an offtaker payment guarantee occupy different positions in that relationship.

Financial additionality must be estimated against an explicit counterfactual. Gross supported lending measures program activity; net financing effects require treatment of repayment, refinancing, other credit, rejected applications and displaced borrowers. Improvements in price, tenor or continuity are economically meaningful even where debt stocks change little, but they require their own valuation and comparison with alternatives. Economic additionality involves a further question: the effect of the financing change on real activity or welfare.

The evidence from different guarantee programs demonstrates the feasibility of positive credit effects, variation in repayment consequences, and substantial substitution of previously unprotected exposures. These findings are consistent because they concern different contractual and empirical margins. They reinforce the need to preserve precise outcome definitions and the context of identification.

Risk transfer also changes the identity of the residual loss bearer. The solvency and liquidity of insurers, sureties, banks and states, the reliability of contingent payment, and the disposition of recoveries are integral to assessing this transfer. A satisfactory judgment links the legal promise, the bank's response, the borrower's alternative, and the ultimate distribution of economic benefit and risk. Only evidence at all relevant levels supports a conclusion about the public or private value of credit enhancement.

References

  • Altavilla, C., Ellul, A., Pagano, M., Polo, A., & Vlassopoulos, T. (2023). Loan guarantees, bank lending and credit risk reallocation. ECGI Finance Working Paper No. 944/2023. European Corporate Governance Institute. Source

  • Basel Committee on Banking Supervision. (2020). Basel Framework: CRE22, Standardised approach: credit risk mitigation. Version published November 26, 2020, effective January 1, 2023. Bank for International Settlements. Source

  • Berne Union. (2026, June 24). State of the industry report 2025 published. Official release announcement. Source

  • de Blasio, G., De Mitri, S., D'Ignazio, A., Finaldi Russo, P., & Stoppani, L. (2017). Public guarantees on loans to SMEs: An RDD evaluation. Temi di discussione No. 1111. Bank of Italy, Rome. Source

  • Export-Import Bank of the United States. (2013). Enhanced assignment agreement, small business multibuyer export credit insurance policy. Form EIB-92-53, revised March 2013. Source

  • Export-Import Bank of the United States. (n.d.). Multi-buyer standard insurance. Official product description. Accessed November 18, 2026. Source

  • International Monetary Fund. (n.d.). Debt guarantees and loans assessment tool (DGLAT). Fiscal Risks Toolkit. Accessed November 18, 2026. Source

  • Mullins, W., & Toro, P. (2018). Credit guarantees and new bank relationships. Working Paper No. 820. Central Bank of Chile, Santiago. Source

  • Multilateral Development Banks. (2018). Multilateral development banks' harmonized framework for additionality in private sector operations. International Finance Corporation. Official publication overview. Source

  • New York State. (2023). Insurance Law, Article 69, section 6901: Definitions. Revision October 27, 2023. New York State Senate. Source

  • OECD. (2017). Evaluating publicly supported credit guarantee programmes for SMEs. OECD Publishing, Paris. DOI: 10.1787/0f952b4a-en. Source

  • OECD. (n.d.). Export credits. Official policy overview. Accessed November 18, 2026. Source

  • United Nations. (1995). United Nations Convention on Independent Guarantees and Stand-by Letters of Credit. New York, December 11, 1995. UNCITRAL. Source

  • United States. (2007, codified text accessed 2026). 13 CFR section 120.520: Purchase of 7(a) loan guarantees. Electronic Code of Federal Regulations, reproduced by Cornell Legal Information Institute. Source

  • United States. (n.d.). 13 CFR section 115.15: Underwriting and servicing standards. Electronic Code of Federal Regulations, reproduced by Cornell Legal Information Institute. Accessed November 18, 2026. Source

  • United States. (n.d.). 13 CFR section 120.524: When is SBA released from liability on its guarantee? Electronic Code of Federal Regulations, reproduced by Cornell Legal Information Institute. Accessed November 18, 2026. Source

  • United States. (n.d.). Federal Acquisition Regulation 28.102-1: General (performance and payment bonds and alternative payment protections). Acquisition.gov. Accessed November 18, 2026. Source

  • van der Veer, K. (2015). Loss shocks and the quantity and price of private export credit insurance: Evidence from a global insurer. Working Paper No. 462. De Nederlandsche Bank, Amsterdam, February 26. Source

  • World Bank and FIRST Initiative. (2015). Principles for public credit guarantee schemes (CGSs) for SMEs. World Bank, Washington, DC. Official summary of 16 principles. Source

  • World Bank Group. (2016, June 16). Sierra Leone: Western Area Power Generation Project (WAPGP). Project appraisal document, Report No. 103305-SL, paras. 38-43 and Annex 2. World Bank, IFC and MIGA. Source

  • World Bank Treasury. (n.d.). Credit enhancement - guarantees. Official product overview. Accessed November 18, 2026. Source

  • World Bank. (n.d.). Intercreditor arrangements. PPP Resource Center. Accessed November 18, 2026. Source