The Basel Capital Accord: How 8% Became a Global Banking Rule, 1988

On 8 January 1987 the Federal Reserve Board and the Bank of England published a joint proposal for measuring bank capital. It was a technical document, full of conversion factors and asset categories, and it was also an ultimatum. Two of the three jurisdictions that mattered in international banking had agreed a common standard between themselves and invited everyone else to match it. The third jurisdiction, Japan, whose banks were then expanding across the Eurodollar market faster than any in history, had not been consulted.

Eighteen months later, on 11 July 1988, the Basel Committee on Banking Supervision released a twenty-eight-page paper titled International Convergence of Capital Measurement and Capital Standards. It contained a single number that would govern lending decisions from Frankfurt to Seoul for the next two decades: eight per cent. Banks operating internationally were to hold capital equal to at least 8% of their assets, with each asset weighted according to how risky the Committee judged it to be. No treaty was signed. No parliament ratified anything. Within a decade more than a hundred countries had written the rule into law.

A Committee Born in Cologne

Herstatt is where this begins. When German regulators withdrew the banking licence of Bankhaus I.D. Herstatt on the afternoon of 26 June 1974, counterparties in New York had already paid Deutschmarks into Cologne and were waiting for dollars that never came. The episode gave the world the phrase settlement risk and a practical lesson in what cross-border banking had become. Franklin National Bank failed in New York four months later. At the end of 1974 the central-bank governors of the Group of Ten established a standing Committee on Banking Regulations and Supervisory Practices, which first met at the Bank for International Settlements in Basel in February 1975 and was later renamed after the city.

Its early work was jurisdictional rather than quantitative. The 1975 Concordat divided responsibility for supervising banks' foreign establishments between host and parent authorities, and its governing principle was stated as a negative: no foreign banking establishment should escape supervision. That principle turned out to be easier to write than to apply. Banco Ambrosiano collapsed in 1982 with a Luxembourg holding company that neither Italian nor Luxembourg supervisors would claim. In May 1983 the Committee issued a Revised Concordat introducing consolidated supervision — the idea that somebody must look at a banking group as a single balance sheet.

Peter Cooke of the Bank of England chaired the Committee from 1977 to 1988, long enough that practitioners simply called it the Cooke Committee. Charles Goodhart's archival history of these years describes a body with no legal powers, no enforcement mechanism and no secretariat to speak of, whose influence rested entirely on the fact that the people around the table went home and wrote their countries' rules (Goodhart, 2011).

Capital Nobody Could Compare

Two pressures turned a coordination body into a rule-setter.

The first was Mexico. When Jesús Silva-Herzog told Washington in August 1982 that Mexico could not service its debt, American regulators discovered how thin the cushion under their largest banks was. Claims on Argentina, Brazil, Mexico and Venezuela held by the nine largest US banks came to roughly 180% of their combined capital. Writing those loans down to anything like market value would have rendered the money-centre banks insolvent on paper, which is why the Latin American debt crisis was managed for seven years as a liquidity problem rather than a solvency one. Congress responded with the International Lending Supervision Act of November 1983, which for the first time instructed the three federal banking agencies to set and enforce minimum capital levels. By 1985 they had settled on a uniform floor of 5.5% primary capital to total assets.

A flat leverage ratio has an obvious defect: it charges the same capital against a Treasury bill and a loan to a shipping company. Banks facing it did the arithmetic and moved exposure into forms the ratio did not count — standby letters of credit, note issuance facilities, loan commitments, interest-rate swaps. American supervisors watched the denominator of their own rule hollow out.

The second pressure was competitive. Japanese banks had gone from a modest presence in international lending at the start of the 1980s to roughly a fifth of cross-border bank claims by 1988, and by that year most of the world's ten largest banks by assets were Japanese. They operated on capital ratios that looked alarming by American measures, in some cases near 3%, while holding enormous unrealised gains on equity portfolios accumulated over decades of cross-shareholding. Whether Japanese banks were undercapitalised or merely accounted for differently was a genuine analytical question. It was also a commercial grievance, and Congress treated it as one.

The Bilateral That Forced a Multilateral

Paul Volcker's Federal Reserve and Robin Leigh-Pemberton's Bank of England had converged on the same diagnosis by 1986: a risk-weighted ratio that captured off-balance-sheet exposure was the only measure worth having, and neither country would impose it alone while the other's banks were unconstrained. Their January 1987 announcement was designed to be joined. London and New York between them cleared most of the world's wholesale banking, so a Japanese or German bank that wanted to operate in both centres would eventually face the standard whether or not its home supervisor endorsed it.

Ethan Kapstein's account of the negotiation reads it as a straightforward exercise of market power by the two states that controlled access to the deepest capital markets (Kapstein, 1992). Thomas Oatley and Robert Nabors went further, arguing that the Accord was less a public-goods bargain than a redistributive one, engineered by American legislators to raise foreign banks' costs and paid for by borrowers everywhere (Oatley and Nabors, 1998). Daniel Tarullo, who later sat on the Federal Reserve Board and supervised the implementation of its successors, described the bilateral as the moment the Committee lost the option of moving slowly (Tarullo, 2008).

Tokyo joined in September 1987. A consultative draft went out that December, and the final text appeared the following July. Its opening pages are unusually candid about serving two masters:

"Two fundamental objectives lie at the heart of the Committee's work on regulatory convergence. These are, firstly, that the new framework should serve to strengthen the soundness and stability of the international banking system; and secondly that the framework should be fair and have a high degree of consistency in its application to banks in different countries with a view to diminishing an existing source of competitive inequality among international banks."

Soundness and fairness are not the same goal, and where they conflicted, fairness — which meant an agreement everyone could sign — generally won.

Four Buckets

The mechanism was deliberately crude. Every asset went into one of a handful of risk baskets, and the weights were negotiated rather than estimated.

Off-balance-sheet items were pulled back in through credit conversion factors — 100% for direct credit substitutes, 50% for transaction-related contingencies, nil for commitments of under a year — and derivatives were captured by marking to market and adding a notional charge for future exposure.

Capital itself was split in two. Tier 1 meant paid-up equity and disclosed reserves from post-tax retained earnings; Tier 2 admitted revaluation reserves, general provisions, hybrid instruments and subordinated term debt, and could not exceed Tier 1 in size. On this point the drafters were explicit: "the key element of capital on which the main emphasis should be placed is equity capital and published reserves from post-tax retained earnings."

Then came the concession that bought Japan's signature. Banks were permitted to count unrealised gains on their equity holdings as a Tier 2 revaluation reserve, discounted by 55% — so 45 yen of every 100 yen of latent gain on a cross-shareholding became regulatory capital. Japanese bank solvency was thereby wired directly to the Tokyo Stock Exchange, eighteen months before the Nikkei peaked at 38,915.87 on 29 December 1989 and began a decline that erased the concession. Joe Peek and Eric Rosengren traced how capital-impaired Japanese banks then transmitted that shock outward, cutting lending through their American branches in proportion to the damage at head office (Peek and Rosengren, 1997).

What the Rule Did to Lending

In the United States the effect on measured capital was rapid and large.

Capital drifted sideways through the deregulated 1980s and dipped in 1987, the year the money-centre banks finally took large provisions against their Latin American books. From 1990 the line turns and keeps climbing, through the phase-in deadline at the end of 1992 and past it. A ratio that had spent a decade near 6% settled above 8% and stayed there.

Getting there was not painless. Ben Bernanke and Cara Lown, writing in 1991 while it was happening, found that capital-short banks were shrinking their loan books rather than raising equity. New England, where property losses had eaten capital fastest, showed the sharpest contraction (Bernanke and Lown, 1991). Whether Basel caused the 1990–91 credit crunch or merely coincided with a property bust and a recession is still argued. The direction of the incentive is not in doubt. A bank short of capital could satisfy the ratio by lending less as easily as by raising more, and lending less was cheaper.

There was a second-order effect, visible in the weights table. Government paper carried no charge at all, so a bank rebuilding its ratio could buy Treasuries instead of making loans and improve the numerator's relationship to the denominator without raising a cent.

The Arbitrage the Rule Invited

A four-bucket system prices credit risk in steps of fifty percentage points, which means that inside each bucket the incentive runs entirely one way. A loan to a AAA-rated utility and a loan to a leveraged retailer both attracted a 100% weight and an identical 8% capital charge, so the rational bank kept the retailer and sold the utility. David Jones, working at the Federal Reserve Board, documented how securitisation structures were being built for no purpose other than to move assets whose true risk was lower than their regulatory weight off the balance sheet while retaining the economics (Jones, 2000). The same logic drove synthetic transfer: the credit default swap that J.P. Morgan's team built in the mid-1990s was, among other things, a device for buying a lower risk weight.

The OECD line drew the sharpest criticism. Membership of a Paris-based club determined whether a sovereign's debt carried a 0% or a 100% weight, and whether claims on its banks cost 20% or 100%. Mexico joined the OECD in May 1994 and its government paper was reweighted accordingly; seven months later the peso collapsed and Washington assembled a $50 billion rescue. South Korea joined in December 1996. More consequential was the maturity distinction: a claim of under one year on a bank anywhere in the world carried a 20% weight, while the same claim at thirteen months carried 100%. Lenders to Seoul and Bangkok duly lent short, rolling over interbank lines that cost them a fifth of the capital of a term loan, and when confidence broke in 1997 the entire stock of funding came due at once. Maturity mismatch was the mechanism by which the Asian financial crisis turned a current-account problem into a run.

The Committee kept patching. A 1991 amendment tightened what counted as a general provision. The 1996 Market Risk Amendment added a separate charge for trading books and let banks compute it with their own value-at-risk models — the first concession of the principle that the supervisor sets the numbers. After BCCI was shut down across more than a dozen jurisdictions in July 1991, the Committee issued Minimum Standards in 1992 requiring that every international bank have an identifiable consolidated supervisor — the Concordat principle, restated because it had failed again.

Basel's Successors

Work on a replacement began in 1999 and produced Basel II in June 2004, which let large banks use internal ratings to set their own risk weights. Banks that adopted it in Europe were, by 2008, reporting comfortable ratios on balance sheets levered more than thirty times, and the crisis that began with American mortgage securities in 2007 arrived before most of the framework was fully in force. Basel III, agreed in December 2010, raised minimum common equity to 4.5% of risk-weighted assets with a 2.5% conservation buffer on top, and reintroduced the simple leverage ratio that the whole risk-weighting project had been built to replace.

Peter Cooke's committee had no authority to make law and produced one of the most widely adopted rules in financial history, on the strength of twelve central banks agreeing that a number written in Basel should mean the same thing in Osaka as in Ohio. The number they chose was not derived from any model of bank failure. Eight per cent was roughly where the larger American and British banks already stood in 1987, rounded to something a negotiator could defend.