Difficulties in the U.S. mortgage market played a key role in spawning a widening financial crisis, which is now producing real economic distress on a global scale. Discussion of the causes has focused on lending practices in the “subprime” market, where large numbers of poorly qualified borrowers apparently obtained mortgage financing with great ease. These subprime loans were “securitized,” i.e., packaged with other loans in “mortgage-backed securities” (MBS). The securities attracted investors because they appeared to diversify away much of the risk ordinarily faced by a traditional mortgage lender, an assessment that is as fine an example of a seductive half -truth as economic life has to offer. Rating agencies and supposedly sophisticated investors seem to have missed an “obvious” qualification to the genuine diversification advantages of MBS: The general level of housing prices is a source of “systematic risk” for these securities.
The weakness produced by incautious lending inevitably surfaced. When it did, the consequences were amplified because the bad loans were hidden away in MBS and in even more complex instruments. The resulting financial uncertainty, coupled with a sharp decline in housing prices, produced a depressive contagion that imperiled large banks and is still spreading.
What lessons should be drawn from this episode, for business strategy and public policy, for the short run and the long? While the foregoing sketch captures much of the structure and short-term dynamics of the crisis, it misses key contributing factors found at deeper levels and at longer term. It is therefore inadequate as a source of lessons for the future. As is always the case when a complex system fails in a disastrous fashion, the list of contributing factors and mechanisms is a long one. For every explanation that identifies a “mistake” that seems to have mattered, there are (or should be) inescapable follow-on questions to be confronted: Why was the system not organized in such as way as to make that sort of mistake impossible? Was there, on the contrary, some factor actually tending to make that sort of mistake more likely?
In this case, reasonable answers to such questions might focus on the institutional changes that have transformed the structure of mortgage lending in recent years. If, for example, one inquires into the origins of the peculiar business practice of making mortgage loans without documenting the capacity of the borrower to repay (so-called “no doc” loans), the answer probably has to do with the fact that the arranger of the loan is far removed institutionally from the investor whose capital is placed at risk. That “arranger,” typically a mortgage broker, will likely consider the transaction successfully completed when that loan is handed off to another party, ultimately to become an anonymous member of the group of mortgage loans bundled in a mortgage-backed security. This is not your friendly local bank, placing its shareholders money at risk, that has decided to forego the traditional assurances about the possibility of repayment.
A similar institutional change occurred at the other end of the process, and is now playing a key role in complicating efforts to deal with the developing foreclosure crisis. Just as it was not the ultimate investor who arranged the loan, it is also not the ultimate investor who is collecting the payments on the loan. Rather, the payments are often being processed by a separate mortgage servicing entity with specialized capabilities in large scale, computer-based transaction processing, akin to credit card operations. Like the broker who arranged the loan, the mortgage servicer has no capital at stake and is being paid a fee or commission for its work. When the borrower misses payments, therefore, the institutions make it very difficult to strike a deal that efficiently balances the interests of the borrower with those of the many MBS investors who own a small piece of that particular mortgage. In this picture, there simply is no “mortgage lender” who could cut a deal that is appropriately and necessarily responsive to the current circumstances of the individual borrower and the key asset, the house. A traditional lender, such as a local bank with strong incentives to protect the principal value of the loan, might find it both feasible and advantageous to offer an adjustment of terms to the delinquent borrower.
These illustrations from the two ends of the process reflect only two among many places in the institutional structure where perverse incentives promoted the prolific generation of risky loans, speculative increases in home prices, erroneous assessments of investment risk, and incapacity for dealing with difficulties as they emerged.
The point of these observations is not to suggest that the institutional changes themselves were a terrible mistake, perhaps one that sound economic policy could have prevented. The idea of an unbundled structure for providing mortgage finance has clear merits – or at least the promise of real merits, if institutions can be created that are robust even in times of contraction and credit restriction. The fundamental institution of the current system is the mortgage-backed security itself, and it offers a fundamental improvement in the availability of mortgage credit, which has enormously positive social consequences. There are also efficiency arguments for other institutional changes that together accomplished the “vertical dis-integration” of mortgage banking. For example, it makes sense that mortgage servicing should at least under normal circumstances be done by an organization with the appropriate specialized skills.
Neither would it be plausible to suggest that the new institutions somehow suddenly emerged and then triggered a major market malfunction beginning five years ago or so. These institutions developed and became dominant over an extended period, dating back to the invention of the new investment vehicle, marketable securities backed by government-insured mortgages, in the 1970s. The current crisis has roots not only in that extended process of institutional change, but also in the specific economic circumstances of recent years, including the low interest rate environment and a number of more specific errors and omissions. In the omissions category, strong arguments for greater regulatory attention to the subprime market went unheeded.
The most important lesson from this very costly episode is of broad relevance for business strategy and public policy. Innovations in business practice and in market institutions are dangerous in ways characteristic of innovations in general: They are often introduced well before all the risks associated with them are identified and considered – and, necessarily, before hypothetical adverse scenarios have been validated by harsh lessons of experience. Financial innovations that have not been tested in a down market are too young to deserve celebration. It is no more reasonable to trust such innovations to function well in an adverse economic environment than it would be to certify a new design for a passenger aircraft on the ground that, one sunny day, it took off, flew around for a while in an uncrowded sky, and successfully landed again.
In the case of passenger aircraft, we do not collectively behave that way. We have succeeded in being highly attentive to numerous potential risks, addressing the “hypothetical adverse scenarios” before they have been validated by tragic experience. Rarely indeed, in recent decades, has an aircraft design flaw contributed to a disaster that cost lives. The arrangements that support this achievement are not, to put it mildly, characterized by great inhibitions about “regulatory interference in the marketplace.” Those inhibitions are apparently overbalanced by a great reluctance to trust “the market” when the stakes are so high. Airline operations are shaped by regulatory arrangements that are a match in complexity for the operations themselves, affecting not only aircraft design certification but a wide range of operational practices. In airline safety, we accept that regulations shape what firms do, and focus particularly on the use of regulation to reduce the downside risks.
Of course, the “high stakes” rationale for the difference in approach will not survive dispassionate analysis. The human consequences of a major recession, in lives as well as wealth, are much greater than the consequences of an airline disaster – possibly much greater than all the airline disasters in history.
Why this crisis happened is a question that academics and practitioners should study very carefully, as much for its broad theoretical significance as its current policy importance. It provides an excellent test-bed for alternative hypotheses on what Richard Nelson and I have called “the competence puzzle” – the question of why the record of organizational functioning provides such dramatic examples of both competence and incompetence. The case seems largely consistent with basic evolutionary perspectives on the puzzle: It is learning and practice, not deliberation and careful calculation, that largely account for observed competence. It is feedback, not foresight, that drives economic behavior. This means that fragile systems can easily evolve, and appear successful for extended periods, when the dynamic mechanisms are such as to permit a long postponement of encounters with adverse scenarios. Severe challenges to those systems may then be produced by mechanisms that are just as “endogenous” to the dynamics as the earlier period of apparent success – as is illustrated both by the credit crunch and by the much larger problem of climate change.
While the context dependence of well-practiced routines is a big part of the answer to the puzzle, it is not the whole story. This diagnosis does not tell us why some organizations do appear to display significant foresight – as Goldman Sachs evidently did, but Bear Stearns apparently did not. We should put aside any impression that competence is high and uniform in the organizations of the financial sector and set about the task of discovering the sources of the differences. Questions easily spring to mind about structure and systems, about governance and accountability, about organizational demographics and culture, about judgment biases and the deep sources of myopia.
When a tragic failure occurs in the domain of airline safety, a broad and penetrating technical investigation of the causes typically ensues. Academic scholarship could do the same for the credit crunch.
At this point, it already seems clear that we should not have relied so heavily on the limited protection afforded by private incentives and private foresight. Society has every right to expect a higher standard of safe design. As the airline safety example illustrates, an appropriate regulatory framework can do much to guide the evolution of our complex social systems toward a more robust form of competence.
* I am indebted to Michael Jacobides for his significant contributions to this note. For a discussion of the relevant background in the institutional evolution of the mortgage market in the U.S., see his prizewinning article, “Industry Change through Vertical Disintegration: How and Why Markets Emerged in Mortgage Banking,” Academy of Management Journal 48: 465-498 (2005).
Sidney G. Winter is the Deloitte and Touche Professor of Management at The Wharton School of the University ofPennsylvania. Before joining Wharton in 1993, he served for four years as Chief Economist of the U.S. General Accounting Office (now called the Government Accountability Office) in Washington, D.C.
© 2008 Sidney G. Winter