Relationship banking

Relationship banking is a banking model in which a financial institution develops a continuing, often long-term association with a customer in order to obtain customer-specific information and provide multiple financial services. In academic literature, the related term relationship lending usually refers more narrowly to lending decisions based substantially on information accumulated through repeated interactions between a lender and borrower.

Under relationship banking, a bank may acquire proprietary information through loans, deposit accounts, payment activity, cash-management services, financial advice, and direct contact with a firm's owners, employees, suppliers, and customers. Much of this information is described as soft information because it is qualitative, difficult to verify, and not readily transferable outside the organization collecting it. Relationship banking is therefore commonly contrasted with transaction banking or arm's-length lending, which relies more heavily on standardized financial statements, collateral values, credit scores, market prices, and other quantifiable information.

Relationship banking is especially associated with small and medium-sized enterprises (SMEs), family-owned firms, agricultural borrowers, and other customers whose prospects cannot be assessed completely from audited accounts or market information. It has also been practiced in corporate banking, investment banking, consumer finance, and interbank markets.

Research has identified potential benefits including greater credit availability, more effective monitoring, flexible loan renegotiation, intertemporal smoothing of borrowing costs, and continued financing during temporary distress. Its potential disadvantages include higher operating costs, customer lock-in, informational monopoly, conflicts of interest, favoritism, reduced competition, excessive continuation of loans to unproductive firms, and vulnerability when a borrower's principal bank becomes distressed.

A widely used definition was proposed by economist Arnoud Boot, who defined relationship banking as the provision of financial services by an intermediary that:

1. invests in acquiring customer-specific information, often proprietary in nature; and
2. evaluates the profitability of that investment through repeated interactions with the same customer over time or through the provision of multiple financial products.^[1]: 7–9

The definition has two principal dimensions: the production of private information and the expectation of repeated business. A single loan can form part of a banking relationship if it is accompanied by prior interactions or an expectation of future dealings. Conversely, repeated standardized loans need not constitute relationship lending when each application is evaluated independently through a mechanical credit-scoring system.^[2][3]

Relationship banking is broader than relationship lending. A bank may develop relationships through deposits, transaction accounts, payments, foreign exchange, underwriting, asset management, insurance, trade finance, or financial advice even when it is not the customer's principal lender. Information obtained through one service may influence the pricing or availability of another.^[4][5]

The term is sometimes used commercially to describe personalized customer service. Academic usage is more specific: friendly service alone does not establish relationship banking unless the institution produces customer-specific information or considers the value of interactions across time or products.^{[1]: 7–10}

Relationship banking and transaction banking are not mutually exclusive. A bank may use credit scores and collateral valuations while also considering a loan officer's accumulated knowledge of the borrower. Large institutions may maintain relationship-based corporate-banking divisions alongside highly automated consumer and small-business platforms.^[3][6]

Comparison of relationship and transaction banking

Characteristic	Relationship banking	Transaction or arm's-length banking
Primary information	Proprietary, qualitative, and accumulated over time	Standardized, quantitative, and independently verifiable
Common information sources	Loan-officer contact, account activity, customer history, local knowledge, suppliers, and management reputation	Financial statements, credit scores, collateral appraisals, market prices, and external ratings
Evaluation	Considers the overall customer relationship	Assesses an individual transaction or product
Organizational tendency	Decentralized authority and greater discretion for local officers	Centralized underwriting and standardized rules
Typical borrowers	SMEs, young firms, family businesses, agricultural borrowers, and financially opaque companies	Large or transparent firms, consumers with extensive credit files, mortgages, and asset-backed borrowers
Principal advantage	Information production, flexibility, and financing continuity	Scale, speed, consistency, comparability, and lower processing costs
Principal risk	Hold-up, favoritism, excessive discretion, and dependence on one lender	Mechanical rejection, limited accommodation of qualitative information, and procyclical credit withdrawal

Relationship banking is usually explained as a response to information asymmetry. A borrower generally knows more than a prospective lender about its abilities, risks, and intended use of funds. Before lending, this creates a problem of adverse selection; after lending, it creates the possibility of moral hazard.^[7][8]

Banks can reduce these problems by screening borrowers and monitoring them after credit has been extended. Deposit-account activity may reveal sales patterns, cash flows, overdrafts, seasonal fluctuations, and payment difficulties before those developments appear in formal financial statements. Repeated lending enables the bank to compare the borrower's promises with its subsequent performance.^{[5][9]: 13–19}

Bank-loan announcements can affect a publicly traded firm's share price, which has been interpreted as evidence that bank approval conveys information not already fully incorporated into securities prices.^[10][11] The effect has generally been found to be stronger for loan renewals and other decisions indicating that a bank has monitored a borrower over time.^[11]

Soft information includes judgments about management ability, reliability, business strategy, succession plans, local market conditions, or the reasons behind an unusual financial result. It is commonly produced through personal contact and is difficult to communicate using numerical reports. Hard information, by contrast, can be recorded, verified, and transmitted with relatively little loss of meaning.^[12][13]

The distinction has organizational consequences. A local loan officer may possess valuable information that senior managers cannot independently verify. A decentralized bank can give that officer decision-making authority, but doing so creates agency and control problems. A centralized bank can exercise more consistent oversight but may discount information that cannot be documented easily.^[14][15]

Research has therefore associated relationship lending with smaller banks, community banks, and relatively flat management structures. Large banks can engage in relationship lending, but hierarchical distance between information producers and decision-makers may reduce the influence of soft information.^{[16][6][17]: 97–108}

A relationship may permit the bank and borrower to enter an implicit contract that cannot be reproduced in a one-time transaction. For example, a bank may charge a young firm a relatively low rate while it is establishing itself, expecting to recover the cost through later lending, deposits, or additional services. The bank may also maintain a credit line during temporary distress in return for profitable business during normal periods.^[18][19]

This ability to shift costs and revenues across time or products is called intertemporal or interproduct smoothing. Core deposits can support such arrangements because stable deposit funding gives banks some insulation from changes in wholesale interest rates.^[4] The arrangement depends on both parties expecting the relationship to continue; otherwise, borrowers could accept early subsidies and then move to another lender before the bank recovered its investment.^{[1]: 14–18}

Relationship banks may act as delegated monitors for depositors and other creditors who would find it costly to evaluate borrowers individually.^[7] Monitoring can include enforcing covenants, reviewing account activity, limiting additional borrowing, requiring collateral, assisting in restructuring, or influencing management decisions.

A bank's ability to intervene may be strengthened by the priority of its claims in bankruptcy. Theoretical work suggests that seniority can improve the bank's incentive to collect information and provide additional financing when a viable borrower encounters temporary distress.^{[20]: 57–64}

Long-term banker–merchant relationships predate the expression relationship banking. Early commercial banks relied heavily on personal reputation, local knowledge, correspondent networks, and recurring dealings because audited statements, credit bureaus, and standardized accounting information were limited.^{[21]: 1–30 [22]}

In nineteenth-century New England, banks frequently lent to directors, shareholders, and firms within local business networks. Such insider lending could reduce information costs and help finance industrial development, but it also created opportunities for self-dealing and concentration of credit.^{[21]: 62–115} Studies of early American lending records have found that repeated dealings could sustain long-term relationships even when individual loans remained short in maturity.^[22]

Relationship banking became particularly associated with universal bank systems. German banks combined commercial lending, securities underwriting, payments, and sometimes equity ownership or representation on supervisory boards. A firm's principal relationship institution became known as its Hausbank, or house bank.^{[23]: 5–30 [24]: 1–32 [25]}

Economic historians have debated how decisive German universal banks were in industrialization. Traditional accounts emphasized their ability to mobilize capital, monitor industrial firms, and support investment over long periods. Later research found greater variation among industries, regions, and periods and questioned claims that universal banking alone explained German growth.^{[24]: 4–32 [26]: 5–18}

Japan developed a related but distinct main bank system. A large firm typically borrowed from several institutions but maintained a particularly close association with one main bank. That bank could coordinate creditors, monitor management, organize rescue finance, and intervene during distress.^{[27]: 1–27 [28]} The system was especially prominent after the Second World War, although historians dispute how completely it characterized Japanese finance and how beneficial it was to borrowers.^[29][30]

United States banking has often been described as relatively market-oriented, but relationship banking remained important among local banks, agricultural lenders, trust companies, and institutions serving small businesses. Restrictions on interstate and branch banking historically encouraged geographically bounded banking relationships.^{[31]: 1–38 [32]}

In large-corporate finance, bankers associated with institutions such as J.P. Morgan & Co. and the Guaranty Trust Company of New York cultivated long-term relationships, sat on client boards, organized securities offerings, and assisted in reorganizations.^{[33]: 137–174 [34]: 52–88} Guaranty Trust executive Eugene Stetson, for example, maintained multi-decade relationships with The Coca-Cola Company and the Illinois Central Railroad, combining banking services with board membership and financial advice.^{[35]: 95–174, 290–334}

From the late twentieth century, consolidation, credit scoring, securitization, and information technology expanded transaction-based lending. Nevertheless, relationship lending continued to be important for borrowers without extensive public financial information.^[14][36]

A relationship bank may collect information from:

• loan applications and repayment histories;
• current and deposit accounts;
• cash-management and payment services;
• credit-card or merchant-processing activity;
• personal meetings and visits to the borrower's premises;
• the history and reputation of owners and managers;
• contacts with customers, suppliers, accountants, and local institutions;
• trade finance, foreign exchange, payroll, and investment services; and
• the borrower's conduct during earlier periods of financial difficulty.^{[1]: 7–14 [5]}

The breadth of the relationship may be as important as its duration. A bank that provides deposits, payments, and loans can observe information that would be unavailable to a lender providing only a single term loan.^[37][38]

Loan officers are central to many relationship-lending systems. They solicit information, interpret local conditions, negotiate terms, monitor performance, and transmit recommendations to the bank. Studies have found that loan officers obtain soft information through repeated contact and that their influence depends on the bank's organizational structure.^{[17]: 97–108 [39]}

Reliance on individual officers can create continuity risk. If an officer leaves the bank, some undocumented knowledge may leave with them. Personal discretion can also produce inconsistent decisions, favoritism, or discrimination. Banks respond through rotation policies, dual approval, documentation requirements, internal ratings, and limits on individual lending authority.^{[40]: 795–814 [13]: 20–31}

Physical proximity historically facilitated relationship lending by lowering the cost of meetings, site visits, and gathering local information. Research on American small-business loans found that the average distance between borrowers and lenders increased as communication and information-processing technologies improved.^[32][41]

Distance nevertheless remains relevant for loans dependent on soft information. Studies have associated greater borrower–lender distance with more standardized underwriting and, in some contexts, different default outcomes.^[38][42][43]

A major empirical finding is that relationships can increase credit availability for small or informationally opaque firms. In research using United States small-business data, longer and broader relationships were associated more strongly with the availability of credit than with large reductions in stated interest rates.^[44][37]

Relationships may allow a bank to lend when financial statements alone would produce a rejection. They may also reduce demands for collateral as the lender learns about the borrower's behavior, although empirical findings on collateral and relationship duration are mixed.^[45][46]

The effect on interest rates varies across studies and institutional settings. Research on United States lines of credit found that longer relationships were associated with lower rates and reduced collateral requirements.^[45] Belgian evidence found that interest rates could rise with relationship duration while collateral requirements declined, consistent with both information benefits and the bank's increasing market power.^[47]

Because lenders can alter interest rates, fees, collateral, covenants, and the availability of credit simultaneously, evaluating only the stated interest rate may provide an incomplete measure of relationship value.^[46]

A relationship bank may be better positioned to distinguish temporary illiquidity from long-term insolvency. Its private information can support covenant waivers, maturity extensions, emergency credit, or a coordinated restructuring that would be difficult among dispersed creditors.^[28][48]

Relationship banks may also provide liquidity insurance: borrowers pay somewhat more in ordinary periods in exchange for more favorable continuation lending during a crisis.^[49] This benefit is conditional on the relationship bank remaining financially capable of lending.

Banks may benefit from greater customer retention, cross-selling, improved screening, lower monitoring costs after the initial investment, and the ability to price services across the entire relationship. Proprietary information can reduce losses by allowing earlier intervention when a borrower deteriorates.^{[1]: 10–18 [4]}

Relationship banking is costly, however. It requires skilled employees, repeated contact, decentralized judgment, and the maintenance of information that cannot always be reused with another customer. Relationship banks may therefore have higher operating costs than transaction lenders.^[49]

The value of a lending relationship may be lost when a bank fails or is forced to contract. Even if the borrower ultimately finds another lender, it may face higher rates, additional collateral, or a delay while the new lender reproduces information accumulated by the former bank.^[50][51]

The effect of banking competition is theoretically ambiguous. Competition can prevent an incumbent bank from exploiting its information advantage. At the same time, if customers can leave immediately, banks may be unwilling to bear the initial expense of learning about them.^[18][19]

Borrowers may maintain relationships with multiple banks to obtain greater credit capacity, diversify funding, compare prices, and reduce dependence on one institution. Multiple relationships can also weaken each bank's incentive to monitor and create coordination difficulties during distress.^[52][2]

Empirical studies have linked the transition from a single bank to multiple banks with firm growth, borrowing needs, informational opacity, and concern about becoming locked into an incumbent lender.^[53][54] The typical number of banking relationships differs substantially between countries because of variations in creditor rights, bank concentration, disclosure, and financial-system structure.^{[55]: 27–45}

Small banks are commonly associated with relationship lending because local officers can exercise discretion and communicate directly with senior decision-makers. Large banks may have advantages in credit scoring, diversified portfolios, syndicated loans, asset-based lending, and other technologies based on hard information.^[16][6]

The distinction is not absolute. Large banks can create specialized relationship units, while small banks can employ automated scoring. Research therefore treats relationship lending as one of several lending technologies rather than an activity exclusive to one class of institution.^[3][56]

Foreign-owned banks may face disadvantages in processing local soft information because decisions are made at greater organizational or geographic distance. Studies in Argentina and other developing economies found that large and foreign-owned banks tended to have greater difficulty lending to informationally opaque small firms, although foreign banks may serve such customers through acquired domestic institutions or standardized technologies.^[57][58][59]

Relationship banking can stabilize credit during an external shock because the lender possesses information enabling it to identify viable borrowers. Italian credit-register evidence following the collapse of Lehman Brothers found that relationship banks offered relatively favorable continuation credit to established borrowers compared with transaction lenders.^[49][60]

Studies have also associated stronger lending relationships with firm investment, employment, and survival after credit shocks.^[61][62]

The relationship can instead transmit distress from banks to borrowers. When a principal lender suffers capital or liquidity losses, dependent firms may be unable to replace its credit quickly. Studies of Pakistan, Peru, Portugal, and cross-border lending have used bank-specific shocks to demonstrate that reductions in bank funding can produce declines in lending and real economic activity.^{[63][64][65][66]}

The effect may be especially severe for borrowers whose information is difficult to transfer. The same exclusivity that makes a relationship valuable in normal periods can make the firm vulnerable to the financial health of one institution.^[2][67]

Banking relationships affected the distribution of government-supported credit during the COVID-19 pandemic. Under the United States Paycheck Protection Program (PPP), established borrowers often obtained loans more quickly through banks with which they already conducted business. Banks with stronger existing small-business relationships supplied more PPP credit, and local differences in bank participation affected the timing of access to funds.^[68][69]

The episode also illustrated limitations. Firms without established banking relationships sometimes encountered delays, while financial-technology lenders served borrowers outside traditional branch networks.^[70][71] Research has therefore treated the PPP as evidence both of the value of established relationships and of the unequal access that can result when emergency programs are administered through incumbent lenders.

The most prominent criticism is the hold-up problem. Once a bank has acquired private information, competing lenders may be unable to determine whether the customer is seeking a new bank because it is dissatisfied or because the incumbent has discovered an undisclosed weakness. The incumbent can use this informational advantage to charge higher rates, impose fees, demand collateral, or restrict the customer's ability to switch.^[72][73][74]

Borrowers may respond by maintaining multiple banking relationships, issuing public debt, cultivating deposit accounts with alternative banks, or periodically inviting competing credit proposals. These measures reduce dependency but can also diminish the informational advantages of exclusivity.^[53][52]

A close bank may continue lending to an unproductive or insolvent borrower to avoid recognizing a loss on existing loans. This practice is variously called evergreening, loan forbearance, or zombie lending. It can preserve employment and organizational value during temporary distress, but it can also trap capital and labor in low-productivity firms.^[75][76]

Research on Japan's prolonged banking difficulties found that weak banks sometimes extended credit to distressed borrowers to prevent defaults from revealing damage to bank balance sheets. The resulting support of weak firms impeded restructuring and reduced the competitiveness of healthier companies.^[75][76]

Relationship banking can blur the separation between lender, adviser, shareholder, and corporate director. A bank may recommend transactions that generate fees, protect an existing loan, or benefit another client rather than maximize the borrower's value. Equity ownership and board representation can improve monitoring but also create conflicts between the bank's interests and those of minority shareholders or other creditors.^[25][73]

Related lending to a bank's owners, directors, political allies, or affiliated businesses is a more severe form of relationship-based credit allocation. International evidence has associated related lending with unusually favorable terms, higher default rates, and weak legal enforcement.^[77][78][79]

Because soft information requires judgment, relationship banking can incorporate knowledge overlooked by automated systems, but discretion can also reflect conscious or unconscious bias. Customers excluded from established business networks may have less access to relationship credit even when their underlying projects are sound.^[39][36]

Relationship-based delivery of emergency credit can favor firms with existing accounts, experienced advisers, or stronger connections to participating banks. The experience of the PPP renewed debate over whether speed and fraud control justified reliance on existing banking networks when that reliance disadvantaged unbanked or weakly banked businesses.^[70][71]

Soft information is difficult to audit, communicate, or transfer. A borrower may consequently receive different decisions from different officers, and outsiders may be unable to determine whether a loan was approved because of legitimate private knowledge or favoritism.^[13]

A relationship can also discourage the borrower from developing transparent accounts or accessing capital markets. Japanese firms closely tied to banks were found in some studies to face higher borrowing costs and slower profitability growth, suggesting that close relationships could benefit banks more than their clients.^[29]

The German Hausbank traditionally acts as a firm's principal lender and may coordinate additional banks. Empirical studies have found that German house banks provide liquidity insurance and play a distinctive role during financial distress, although evidence that they always provide lower interest rates is limited.^[48][80][81]

Germany's universal-banking tradition permits institutions to offer a broad range of commercial and investment services. This scope can strengthen information production across products, while also raising questions about conflicts of interest and market power.^{[25][24]: 1–32}

The Japanese main-bank system has included monitoring, creditor coordination, equity ownership, personnel exchanges, and the expectation that a main bank may organize assistance during distress.^{[27]: 1–27 [28]}

Evidence from the postwar period suggested that firms with close bank ties experienced fewer liquidity constraints.^[82] Later work identified possible costs, including rent extraction and support for inefficient borrowers, while some economic historians questioned whether the main-bank model was ever as universal as conventionally portrayed.^[29][30]

In the United States, relationship lending is closely associated with community banks and small-business finance. Community banks may possess local knowledge and flatter organizational structures, while nationwide banks have advantages in automation, diversification, and standardized products.^[16][39]

American firms often combine relationship loans with credit cards, leases, asset-based finance, trade credit, and capital-market funding. The importance of a bank relationship generally decreases as a firm becomes larger, develops audited public information, and obtains access to bond or equity markets.^[54][36]

Relationship lending can compensate for limited credit reporting, weak accounting standards, and incomplete contract enforcement in developing economies. Domestic private banks may have an informational advantage over foreign institutions, while state-owned banks may possess broader geographic networks but face greater political influence.^[57][59]

Cross-country research indicates that the effectiveness of any lending technology depends on creditor rights, contract enforcement, competition, bank regulation, information-sharing systems, and the broader institutional environment.^[83][84]

Large corporations frequently maintain designated relationship banks even when loans are syndicated among multiple institutions. A lead arranger may accumulate information through repeated syndications, treasury services, underwriting, and advisory mandates. Prior lending relationships influence the selection of arrangers and the allocation of future investment-banking business.^[85][86]

Consumer relationships can include current accounts, savings, mortgages, credit cards, and personal loans. Research using German savings-bank data found that prior relationships and the breadth of a customer's accounts contained information about subsequent retail-loan defaults.^[87]

Retail relationship banking raises additional privacy and consumer-protection issues because banks can combine transaction data with other personal information. Automated analysis may make relationship information more portable within a banking group while increasing the scale of profiling.

Banks themselves form lending relationships. Repeated transactions in unsecured interbank markets can reduce search and information costs and provide more reliable access to liquidity. Such networks may supply informal insurance but can also transmit shocks among institutions.^{[88]: 33–56 [89]}

Credit bureaus, online banking, electronic payments, cloud accounting, and machine-learning systems have reduced the cost of producing hard information. These developments allow distant and nonbank lenders to evaluate borrowers without maintaining a conventional branch-based relationship.^[90][91]

Technology does not necessarily eliminate relationship banking. Digital transaction data can deepen a bank's knowledge of customers, and communications technology can support continued contact over greater distances. Some lending models combine automated scoring with human judgment, using algorithms to organize information while leaving exceptional decisions to relationship managers.^[3][43]

Fintech lenders can compete with incumbent banks by using real-time sales, payment, and platform data. During the PPP, fintech institutions reached borrowers outside conventional branch networks, although banks retained advantages arising from existing customer identification and account relationships.^[71][69]

Research on Italian banks' adoption of artificial intelligence found that AI-based screening could complement information accumulated through relationships and mitigate some rent extraction in normal periods. The technology did not, however, reproduce all of the credit-protection effects associated with relationships during the COVID-19 shock.^[92]

Relationship banking cannot be observed through a single universally accepted measure. Researchers have used:

• duration of the bank–borrower relationship;
• number of financial products obtained from the bank;
• share of the borrower's total debt supplied by one institution;
• whether the institution is identified as the principal or main bank;
• exclusivity or number of concurrent banking relationships;
• geographic distance;
• frequency of personal contact;
• use of deposit and payment accounts;
• prior loans or loan renewals;
• the lender's position as lead arranger; and
• survey responses from borrowers, loan officers, or banks.^[2][3]

These measures are not interchangeable. A long relationship may be inactive, while a short but intensive relationship may produce substantial information. A borrower may also concentrate deposits at one bank but obtain loans from another.^[2]

Causal identification is difficult because strong firms may attract long-term lenders, and banks may selectively maintain relationships with better borrowers. Researchers have addressed this problem by using bank failures, liquidity shocks, regulatory changes, lender mergers, credit-register data, and multiple-bank borrowers that allow changes in supply to be separated from changes in demand.^[63][62][64]

The economic literature does not treat relationship banking as universally superior to transaction banking. Relationship lending is most valuable where borrower information is costly to produce, difficult to verify, and capable of becoming more accurate through repeated interaction. Transaction lending tends to be more efficient when relevant information can be standardized cheaply or when collateral and market prices provide reliable measures of risk.^[3][56]

The two models increasingly coexist. Banks use customer relationships to interpret data and manage exceptions while relying on automated systems for initial screening, regulatory compliance, pricing, and portfolio monitoring. The balance depends on the borrower, product, institution, legal system, competitive environment, and stage of the economic cycle.^{[1]: 20–23 [91]}

• Community bank
• Relationship marketing
• Transaction banking