Tim Harford’s thought-provoking book argues that complex systems cannot be created—they must evolve—and that evolution is necessarily a messy, trial-and-error process. In nature, error often results in death. Failure in the marketplace, while harsh, is rarely fatal and often yields lessons from which success springs.
If so, failure should be expected and planned for. Lessons learned should be identified and incorporated into future efforts. Attempts to eliminate failure will themselves fail because of the myriad interconnections that exist in complex systems and because independent actors respond to such attempts in unpredictable ways.
Rather than trying to construct failsafe systems, Harford suggests constructing resilient structures and institutions that can fail yet continue to work. To do this, Harford recommends “decoupling” a system’s components so that failures do not snowball. Using the example of falling domino exhibitions in which thousands of dominoes are artfully toppled, Harford relates how experience led exhibitors to place “safety gates” between dominoes at key points during setup so that, in the event of a mishap, only a few would fall.
Harford takes the concept of decoupling and applies it to, among other things, financial systems. Surprisingly, the author, an economist generally favorable to free markets, offers few market-oriented solutions even when they fairly leap off the page. For example, he notes that the legal structure of Lehman Brothers, which failed when the housing bubble burst, had been made intentionally complex to avoid taxes. Its “Byzantine” structure made bankruptcy chaotic and resulted in assets being frozen far longer than would otherwise have been necessary. Despite this, Harford does not suggest simplifying the nation’s tax structure to reduce incentives to create such tangled webs.
He does mention the moral hazard of guaranteeing bank debts, which reduces bankers’ incentives to invest wisely and all but eliminates depositors’ incentives to ensure that the banks to which they entrust their money are well managed. However, he doesn’t propose that government end such guarantees even though relying on a single insurer is an extreme form of coupling.
Adapt also fails to note that regulations requiring institutions to do the same things, in the same ways, at the same times are another way to couple complex financial systems and thus risk falling dominoes. For example, the Securities and Exchange Commission’s mark-to-market accounting rules required financial instruments to be valued at current market prices. Those rules amplified the effects of both the boom and the bust. Mortgage-based securities, overvalued when housing prices soared, became undervalued as the panic grew, and financial institutions saw their assets become worthless. Banks, required by law to maintain specified levels of AAA-rated securities, were forced to sell devalued assets at fire-sale prices.
Another critical safety gate that Harford neglects is a diversified investment portfolio. Americans learned the value of diversification during the Depression. Branch banking laws tied the solvency of thousands of the country’s banks to insular economies—sometimes to single commodities such as a locally grown crop. When, after the stock market crash of 1929, the Federal Reserve failed to increase the money supply, a deflationary spiral caused thousands of small, isolated banks to go under. By contrast, Canada, unencumbered by branch-banking laws, suffered no failures.
Another form of regulatory coupling has been introduced by security-rating agencies, members of a cartel created by the Securities and Exchange Commission. During the boom the agencies uniformly gave unrealistically high ratings to packaged debt containing subprime loans. Basel II, an international banking accord, encouraged banks to hold highly rated mortgage-backed securities by allowing them to keep smaller cash reserves to back them than it required for traditional loans. Were financial institutions allowed to assess and manage asset risks themselves, it’s unlikely they all would have made the same mistakes at the same time. When governments act as guarantors of last resort, however, they not unreasonably attempt to limit their risks through regulation. Problems arise when one-size-fits-all rules cause institutions to act as a herd, making booms and their subsequent busts bigger and far more damaging.
Adapt is delightful, well written, and packed with fascinating examples drawn from everything from offshore oil rigs to Iraqi battlefields. Harford’s diagnoses are right on target, but his prescriptions could use a bigger dose of deregulation and free markets.