Credit Analysis

Credit analysis is the craft of deciding, before money is lent, whether it will come back.

13 frameworks, from 1300 to today · Banking · Business, Management & Finance

still a working view ended · each leads to its entry below

Overview

Credit analysis is the craft of deciding, before money is lent, whether it will come back1. A bank's whole business rests on this judgment: it borrows short from depositors and lends long to borrowers, and every loan that fails to repay eats into the thin cushion of capital that stands between the bank and insolvency. Credit analysis is therefore not a back-office formality but the discipline that decides where a bank's money goes. Its history is a long movement from personal knowledge of the borrower, to standardized financial tests, to statistical and then computational models—each step taken because the previous one could not cope with the scale or complexity of the lending it faced.

Lending on Character: Merchants, Reputation, and the Five Cs

For most of commercial history, credit was personal. A merchant in a trading city—medieval Bruges, Renaissance Florence, or the Atlantic ports of the eighteenth century—lent to people he knew or to people vouched for by those he knew. Reputation was the collateral. A merchant's standing in his guild or trading network, his record of honoring bills of exchange, and the word of other merchants carried the weight that a credit file carries today. The great Italian banking houses of the fourteenth and fifteenth centuries, such as the Bardi and Peruzzi of Florence, lent to monarchs on the strength of political relationships—and were ruined when Edward III of England defaulted in the 1340s, a lesson in how reputation can fail when the borrower is a sovereign.

Out of this world came the durable folk taxonomy known as the Five Cs of credit: character, capacity, capital, collateral, and conditions1. The formulation is hard to date precisely and was codified in twentieth-century banking textbooks, but its logic is as old as merchant lending. Character meant the borrower's honesty and willingness to pay; capacity meant the income from which repayment would come; capital meant the borrower's own stake in the venture; collateral meant assets pledged against default; and conditions meant the economic and industry setting. The Five Cs were never a formula. They were a checklist for judgment, and they remain the vocabulary in which loan officers describe what they are weighing.

The Credit Department and the Loan Committee

As banks grew into joint-stock institutions with branch networks in the nineteenth century, lending could no longer rest on the proprietor's personal acquaintance with every borrower. The response was organizational: the separation of the lending decision from the relationship, embodied in the credit department and the loan committee. A credit department gathered information on borrowers, prepared written analyses, and recommended or declined; a loan committee of senior officers reviewed larger exposures collectively. This division of labor was a governance innovation as much as an analytical one—it put a second pair of eyes between a persuasive borrower and the bank's money, and it created a written record of why a loan was made.

The practice spread with the growth of large commercial banks in the late nineteenth and early twentieth centuries, especially in the United States and Britain. Its logic—that credit judgment should be documented, reviewed, and made by committee rather than by a single officer with a relationship to protect—is the ancestor of every modern credit approval process, including the independent risk functions that regulators now require.

Financial Statements and Ratio Analysis

The rise of the corporation gave credit analysts a new kind of evidence: audited financial statements. If a borrower was a company rather than a person, its balance sheet and income statement could be read for signs of strength or weakness. From the late nineteenth century onward, analysts developed ratio analysis—the comparison of one financial figure to another to reveal what the raw numbers conceal. A current ratio (current assets divided by current liabilities) tests whether short-term assets can cover short-term obligations; a debt-to-equity ratio measures how much of the business is financed by borrowing rather than owners' money; profitability and turnover ratios reveal how efficiently capital is used.

The systematic study of ratios is associated above all with the American analyst Alexander Wall, whose work in the early twentieth century through the Federal Reserve system and later the Robert Morris Associates (a bankers' credit association founded in 1914) promoted ratio comparison across firms and industries. The insight was comparative: a ratio means little in isolation but much against industry peers or against the same firm over time1. Ratio analysis made credit judgment auditable and teachable, and it remains the backbone of corporate credit work.

Credit Reporting: Making Reputation Portable

Personal and small-business lending faced a different problem: how to know a stranger's repayment record. The answer was the credit reporting agency, an institution that collects and sells information about borrowers' histories. Local merchant credit bureaus appeared in the United States in the nineteenth century, pooling information among tradesmen about customers who paid and those who did not2. In 1860 the Brooklyn merchant Lewis Tappan founded what became the Mercantile Agency, which grew into R.G. Dun & Company; its rival, the Bradstreet Company, was founded in 1849. These firms sent correspondents to assess businesses and sold subscription reports—an early commercial answer to the problem of assessing distant borrowers. Dun and Bradstreet merged in 1933.

Consumer credit reporting developed along a parallel track in the twentieth century, becoming nationally consolidated and, in the United States, regulated by the Fair Credit Reporting Act of 1970, which gave consumers rights to see and dispute their files. Credit reporting turned reputation into a portable, standardized record—the precondition for lending at a distance and at scale.

Statistical Credit Scoring

The decisive break came when analysts stopped judging borrowers one at a time and started predicting default statistically. The founding work is usually credited to the statistician David Durand, whose 1941 study for the National Bureau of Economic Research applied discriminant analysis—a statistical method for separating two groups—to distinguish good from bad installment loans. The idea was to weight borrower characteristics so that a score, not an opinion, sorted applicants.

Credit scoring became a mass practice with the spread of consumer credit after the Second World War. In 1956 the engineer Bill Fair and the mathematician Earl Isaac founded Fair, Isaac and Company to sell statistical scoring models to lenders; their FICO score, introduced commercially in 1958 and later widely adopted, became the standard measure of American consumer creditworthiness3. Scoring's appeal was consistency and speed: it removed some of the discretion, and some of the bias, from loan officers' judgments, and it made high-volume consumer lending economically possible. Its critics noted that it also encoded whatever patterns the historical data contained.

Cash-Flow and Debt-Service Lending

Corporate lending underwent its own shift in the later twentieth century, from lending against balance-sheet assets to lending against future cash flow. In cash-flow lending, the analyst asks not what the borrower owns but what it will earn—whether projected operating cash flow will cover interest and principal. The key measure is the debt-service coverage ratio (cash available for debt service divided by the payments due), often paired with leverage ratios such as debt to EBITDA (earnings before interest, taxes, depreciation, and amortization). This approach, associated with the growth of leveraged buyouts and project finance from the 1970s and 1980s, allowed banks to lend to firms with few tangible assets but strong earnings—and exposed them to the risk that those earnings projections were wrong.

Predicting Corporate Distress

If default could be predicted statistically for consumers, could it be predicted for companies? Edward Altman's 1968 Z-score was the landmark: a weighted combination of five financial ratios, derived from a sample of bankrupt and surviving manufacturers, that produced a single number indicating how close a firm was to distress. The Z-score demonstrated that a small set of accounting ratios carried real predictive power, and it spawned a family of successors—logit and probit models that estimate the probability of default directly, and later hazard models that estimate default timing. These corporate distress prediction models gave lenders and investors a quantitative early-warning tool and became standard in credit departments and, later, in the rating process.

Rating Agencies and the Rating of Securities

A parallel institution grew up to judge bonds rather than loans. John Moody began publishing ratings of railroad bonds in 1909; Poor's, Standard, and Fitch followed in the following decades, and the agencies became central to securities markets. Their method combined financial analysis with judgment about an issuer's ability to meet obligations, expressed in letter grades. Rating agencies acquired quasi-regulatory status when regulators began using their ratings to set capital requirements and investment rules—a delegation that came under severe scrutiny after the 2008 financial crisis, when highly rated mortgage securities defaulted en masse. The agencies' history shows both the power and the fragility of third-party credit judgment.

Portfolio Risk and the Basel Capital Framework

Through most of the twentieth century, credit analysis was loan by loan. The late twentieth century brought the recognition that a bank's risk lies in the portfolio, not only in individual exposures: many loans can default together when a common shock hits, and concentration in one borrower, industry, or region magnifies loss. Portfolio credit risk modeling—pioneered in the 1990s by firms such as J.P. Morgan (with its CreditMetrics framework, published in 1997) and by academic work on default correlation—estimated the distribution of losses across a whole book of loans.

This thinking entered regulation through the Basel Accords4. Basel I (1988) set minimum capital against credit risk with crude borrower categories; Basel II (2004) allowed large banks to use internal ratings and models to set capital, tying regulatory capital to the bank's own credit analysis. The framework made credit analysis a matter of regulatory consequence: how a bank measured risk determined how much capital it had to hold.

Stress Testing

The 2008 crisis exposed how badly models could underestimate correlated loss. The regulatory response was stress testing: rather than estimating the probability of normal-times loss, stress tests ask what happens to a bank's capital under severe but plausible scenarios—a deep recession, a housing collapse, a spike in unemployment5. The United States' Supervisory Capital Assessment Program of 2009, followed by the annual Comprehensive Capital Analysis and Review, and the European Banking Authority's exercises made stress testing a central supervisory tool. It forced banks to model credit losses under adversity and to show they could survive them.

Machine Learning and Alternative Data

The most recent turn applies computational methods to credit assessment. Machine learning models—gradient boosting, neural networks—can find patterns in large datasets that traditional scorecards miss, and alternative data (utility payments, mobile-phone usage, transaction records, even social and behavioral signals) extends credit assessment to people with thin or no credit files. This has opened lending to previously excluded borrowers and raised new questions about fairness, explainability, and the risk of overfitting to the past. Regulators in many jurisdictions now require that credit models be interpretable and that adverse decisions be explainable to borrowers—a constraint that pulls against the opacity of the most powerful methods.

Where Credit Analysis Stands

Today's credit analysis is layered rather than replaced. A consumer lender may combine a statistical score with alternative-data signals; a corporate credit department still reads financial statements and ratios, still applies the Five Cs in judgment, and increasingly runs cash-flow projections and distress models; a bank's risk function aggregates exposures into portfolio models and stress tests them against regulatory scenarios. The through-line is a single question asked in ever more systematic ways: given what we can observe about this borrower and this portfolio, what is the probability of loss, and can the bank bear it? The open questions are the ones the history keeps returning to—how much to trust the model over the judgment, how to measure risk that only appears in a crisis, and how to extend credit fairly to those the records have never captured.

References

  1. Credit analysis (introduction, Classic credit analysis, Credit scoring systems), Wikipedia
  2. Credit bureau (History), Wikipedia
  3. FICO (introduction, History), Wikipedia
  4. Bank, Wikipedia
  5. Stress test (financial), Wikipedia

The frameworks of Credit Analysis

Each framework is a way of working the field has used: when it held, and the question it answers.

  1. 1300-1850
  2. 1849-Present

    Credit bureau

    How can a stranger's repayment record be known?

    Fly to its tower in Nousfarer

  3. 1880-Present
  4. 1900-Present

    Financial statement analysis

    What do a firm's financial ratios reveal?

    Fly to its tower in Nousfarer

  5. 1900-Present

    Five Cs of Credit

    What makes a borrower creditworthy?

    Fly to its tower in Nousfarer

  6. 1909-Present
  7. 1941-Present

    Statistical Credit Scoring

    Can default be predicted from borrower traits?

    Fly to its tower in Nousfarer

  8. 1968-Present
  9. 1970-Present
  10. 1988-Present
  11. 1997-Present
  12. 2009-Present

    Stress test

    Would a bank survive a severe downturn?

    Fly to its tower in Nousfarer

  13. 2010-Present

How the ideas connect

Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.

Merchant Credit and Reputation Networ…Credit bureauCredit Department and Loan CommitteeFinancial statement analysisFive Cs of CreditCredit rating agencyStatistical Credit ScoringCorporate Distress Prediction ModelsCash-Flow and Debt-Service LendingBasel Capital FrameworkPortfolio Credit Risk ModelingStress testMachine Learning and Alternative Data…

Beyond the field

Where Credit Analysis's ideas came from in other fields, and what they led to there: the strongest 12 of 61 links.

A city in Nousfarer: a red-carpeted stair climbing terraces to a crowned tower on an acropolis, more towers round it, forest and snowy mountains beyond.
A city in Nousfarer, the game played in the browser.

Credit Analysis is a planet in Nousfarer

Its city has a tower for each of these 13 frameworks. Land, walk in, and climb one concept at a time.

Land on Credit Analysis

Related fields

Fields its ideas are linked to

More in Banking

See where Credit Analysis sits on the Map of Human Knowledge