One new and uniquely American source of information appeared in the wake of the financial crisis of 1837: credit reporting. Instead of continuing to rely on their own informal sources of information about the creditworthiness of others, merchants could now turn to a third party, a “mercantile agency,” whose business it was to provide independent and encompassing assessments of the creditworthiness of other firms.
The Mercantile Agency (later known as R.G. Dun) was founded in 1841 by Lewis Tappan, a prominent textile merchant and ardent abolitionist who had failed in the panic of 1837. According to historian Rowena Olegario, the idea of a credit agency was already “in the air,” but it was Tappan who brought this idea firmly into reality. Organizationally, Tappan established a central office (in New York City, close to his clients) and attached to this a network of independent confidential correspondents, distributed across the country, who provided information about the firms in their local vicinity. Scholar William Armstrong gave a thumbnail summary: “Lewis Tappan…applied himself to the establishment of a commercial agency, the object of which is to ascertain, by means of agents throughout the country, the character and standing of the merchants in the different towns, so that when the New York dealers receive applications for goods from traders at a distance, they have only to refer to Mr. Tappan to ascertain their degree of trustworthiness.” Using the well-functioning national postal system, correspondents mailed information about firms to the head office. This information was transcribed and edited, and then turned into credit reports that could be sold to anyone with an interest in a particular firm. As of August 1841 Tappan had acquired 133 subscribers to his new service, mostly in the dry-goods business.
Although many organizations could be interested in someone’s ability to pay their debts (e.g., banks, insurance companies, employers, etc.), rating agencies served the mercantile community first and concentrated on trade credit. In an early advertisement, Tappan attracted customers by giving them the chance to assess for free the information his firm would provide: “Any merchant who wishes to test the value of the information can do so gratuitously. No better way has been thought of than for such to bring a list of the bad debts they have made since the establishment of the agency in 1841, and ascertain how the debtors stood on the books of the agency when the goods were sold to them.”
The Merchants’ Magazine and Commercial Review of January 1851 stated that the overall purpose of the new credit rating system was to “uphold, extend, and render safe and profitable to all concerned the great credit system, on which our country had thriven.” More specifically, however:
It was mainly intended as an aid to the jobber. His customers, scattered over many states, were periodically visiting him for the purpose of renewing their stock of goods; generally canceling, in whole or in part, previous obligations, while they contracted new ones. The intelligent jobber would necessarily need to be informed, on the opening of a new account, respecting the then circumstances of his customer. From year to year, he would desire to be freshly advised of the good or ill success attending him. Information of this character can, in general, be satisfactorily obtained only at the home of the trader. Hence, the main object with the agency is to furnish the home standing of a merchant obtained from intelligent and reliable sources there.
The agency believed that secrecy was important for the accuracy of the information they provided. For example, correspondents’ letters were destroyed after being transcribed into the agency’s ledgers. But the agency had to rely on people who were knowledgeable about local businesses and business dealings, and this meant mostly using local attorneys (it also used postmasters, sheriffs, bank cashiers, and merchants). According to the Bankers’ Magazine and Statistical Register of January 1858, correspondents were selected “for their integrity, long residence in the county, general acquaintance, business experience, and judgment. Their duties are to advise the agency promptly, by letter or telegraph, of every change affecting the standing or responsibility of traders; to notify it of suits, protests, mortgages, losses by fire, endorsements, or otherwise; to answer all special inquiries addressed to them by any of the associate offices; and to revise before each trade season, or oftener if required, the previous reports of every trader in the county, noting any change for the better or worse.”
Building up the network of correspondents was key to enlarging the scope of the firm and enhancing the value of the information it provided to subscribers. In 1846 the agency had 679 correspondents around the country, and by 1851 there were some 2,000 (correspondents were less numerous in Southern states because of Tappan’s strong and public support for abolition). Correspondents were not directly employed by the agency, but it struck an informal deal with them: in exchange for information, the agency would steer collection work to correspondents. Eventually, agencies employed their own reporters, who would operate from a growing network of branch offices. The ability of these networks to gather information on a national scale was such that later in the nineteenth century the life insurance companies also sought out credit reporters as a source to obtain individual data about health risks.
Credit rating was an idea whose time had clearly come, and soon the agency attracted competitors like Bradstreet’s (established in 1849). The absence of nationally branched banks undoubtedly played a role in creating an opening for the mercantile agencies: had they existed, such banks could have used their own internal systems to acquire and circulate information about creditworthiness. As it was, many states only allowed unit banks or severely restricted the number of branches. Competition prompted innovation and emulation, and credit raters rivaled each other in breadth of coverage and the type of information provided. Bradstreet’s began to issue reference books, that is, compilations of summary assessments of the creditworthiness of thousands of firms published in book format, and its success with this format encouraged the Mercantile Agency to follow suit in 1859.
A key that explained the meaning of the rating categories accompanied the reference book, as part of the introductory material. The category systems changed over time and varied from agency to agency. In 1860 Bradstreet’s used a single dimension that ranked firms from “AA” at the top to “E” at the bottom. Rating systems later became two-dimensional, and in 1880, for example, the Mercantile Agency measured pecuniary strength (a measure of net worth) using an ordinal scale that estimated the capital in the firm, ranging from A+ ($1,000,000 or more) down to K (less than $2,000). General credit was also measured with ordered categories, from A1 (“unlimited”) through 1, 1½ (“high”), 2 and 2½ (“good”), down to 3 and 3½ (“fair”). Sometimes firms were listed but given no rating (either for pecuniary strength or general credit, or both). In this case, the accompanying key stated: “The absence of a rating indicates those whose business and investments render it difficult to rate satisfactorily to ourselves. We therefore prefer, in justice to these, to give the detailed reports on record at our offices.”
Although ratings were intended to add transparency to credit relations, the process of evaluating a particular firm and classifying it into the agency’s category system was itself opaque. The agency gathered large amounts of information about firms from multiple sources possessing varying degrees of credibility. Some of what it learned verged on gossip and hearsay, much of it was conflicting or ambiguous, and most of it was qualitative. Across different firms, input information was unsystematic and arrived at the headquarters at uneven intervals: sometimes the agency knew a lot (as measured by the size of the entry in the ledgers and how often information was updated), and sometimes it knew only a little. It might seek information directly from a firm (by having personnel try to obtain a statement from the firm’s proprietor), but it also frequently used indirect sources (industry “insiders,” local notables, agency subscribers, the local business community, etc.). And for a long time, whatever financial information an agency could obtain was undisciplined by credible accounting standards or rules. No regulations required firms to construct a balance sheet, disclose financial records, or even calculate income in a systematic fashion.
There are indications of what rating agencies wanted from their correspondents and reporters, but for a long time the agencies were not very specific about what informational “inputs” were to be used in the rating process. One letter of instruction sent out to Bradstreet’s correspondents in 1869 states:
If there are any dealers who have recently come into your midst of whom little is known, please state where they were previously located (after giving us such information as you can), that we may obtain the full particulars from that place. 1) Give length of time in business. 2) Amount of own capital in business. 3) Amount of net worth, after deducting all liabilities of every nature. 4) Of what is estimated wealth composed? (Viz.: Real estate less incumbrances, capital in business, personal property, which includes bonds and mortgages, stocks, notes, etc., etc.) 5) Character? Good, fair, poor. 6) Habits? Good, medium, poor. 7) Business qualifications? Very good, good, medium, poor. 8) Prospects of success? Good, fair, medium, poor. 9) Succeeded whom? If any person or firm, state whom. 10) Give individual names of partners, with age.
Overall, the creation of ratings in the nineteenth century did not involve a systematic process through which specified pieces of information were gathered and combined, and then transformed into standardized ratings in an algorithmic fashion. The process was both irregular and opaque.
Information received from correspondents and reporters was transferred into the proprietary ledgers. Ledger entries were frequently narrational in recording small stories about what had reportedly happened to a firm. For example, a man called J.W. Finerty was in the furniture business in Illinois, and the Mercantile Agency ledger entry from December 20, 1855, noted that Finerty “has been doing business for one Mitchell who failed in NY City & sent his goods here. Finerty has helped him to smuggle his effects—from Easter creditors—Finerty was a journeyman cabinetmaker, without a $ in the world. They are now beginning to quarrel, the result will be a scramble & failure (unless I am mistaken).” A year and a half later, Finerty had formed a partnership with a Mr. Liebenstein, as upholsterers and cabinetmakers: “Have done little bus[iness] this past winter & spring & yet have retained their position pretty well. L a saving Hebrew. F a rough and ready Irishman who wd be v. likely to blunder if left to himself but with his partner they get along admirably & doubtless mkg a little. Means limited by shd thk them safe.” The ethnic stereotyping revealed in these passages was common. And although there were few businesses owned or operated by African Americans, when they were registered in agency ledgers, the racial identity of the proprietor was carefully noted.
The credit reporting process necessarily involved a substantial reduction of information. Ambiguities, inconsistencies, and contradictions disappeared. In the ledgers, some facts indicated greater creditworthiness while others signaled less. Some information was explicitly disbelieved by the rating agency personnel (or at least the records document their skepticism). But the published rating placed a firm into a single unambiguous category of creditworthiness and invited comparison with other classified firms. Messy complications were kept in the background and out of sight.
The Mercantile Agency’s reference books of 1859 listed about 20,000 firms. By 1868 the reference books covered some 400,000 firms, and at the end of the nineteenth century the agency was rating over one million firms on a truly panoptic scale.
As credit ratings increased in importance, they became more problematic for those who were rated and more contestable. A good rating helped firms get better access to credit, whereas a low rating made life difficult. Firms receiving low ratings had little recourse and sometimes threatened legal action. In fact, credit rating agencies were justifiably concerned about the possibility of litigation from several directions. To put it simply, credit raters could make two kinds of mistakes: they could give a good rating to a firm that wasn’t creditworthy, or they could give a poor rating to a firm that was in fact creditworthy. In the first instance, a client who extended credit to an untrustworthy debtor on the basis of the high rating, only to have the debtor default, could well blame the misleadingly high credit rating and seek satisfaction in court. This kind of mistake could produce highly visible consequences that prompted legal action. Firms that received erroneously low ratings might seek legal redress on the grounds that their business problems were actually caused by the agency’s mistakenly low rating. Firms that declined to extend credit to a truly creditworthy customer because of the low rating suffered from a lost opportunity, but such mistakes had lower salience and seldom prompted legal action.
Agencies went to some lengths to keep a suitable legal distance from the information they provided. Dun employees were instructed to mind the language they used when writing reports for clients: “When reports are cautionary, it is unsafe to say ‘We advise caution’ or ‘We believe they are in trouble’; the words ‘Caution is deemed expedient’ or ‘Trouble is reported’ should be used. Let it be borne in mind that it is never the opinion of the agency which is being given, the word ‘We’ must therefore never be used in any report.” In addition to adopting the passive voice when they wrote, employees were also instructed that they were not in the business of giving advice for which they or Dun could be held accountable: “The business of the agency is to give the facts, figures, and opinions obtained by its reporters. It is not within its province to give advice to anybody. The words ‘advise’ or ‘recommend’ are objectionable. In giving unfavorable credit comments, it is proper to say ‘the trade’ or ‘those consulted do not consider the account desirable.’ It is improper to say ‘Credit is not advised’ or ‘recommended,’ or in any way to convey the slightest inference that the language used comes direct from the agency instead of from those consulted.” And, of course, agencies added protective language to the standard contracts they signed with their clients.
As the nineteenth century progressed, and despite worrisome litigation, credit ratings became more widespread and were increasingly regarded as legitimate encapsulations of a debtor’s true creditworthiness. And by the beginning of the twentieth century, a credit expert declared that the credit agency “is the most indispensable, the oftenest used, and the most versatile tool with which the credit man is equipped.” Credit rating had become part of the fabric of American commercial life.
Excerpted from The Economy of Promises: Trust, Power, and Credit in America by Bruce G. Carruthers. Copyright © 2022 by Princeton University Press. Reprinted by permission of Princeton University Press.