Tuesday, July 17, 2012

Regulatory thrombosis

Charles Perrow is a leading researcher on the sociology of organizations, and he is a singular expert on accidents and system failures. Several of his books are classics in their field -- Normal Accidents: Living with High-Risk Technologies, The Next Catastrophe: Reducing Our Vulnerabilities to Natural, Industrial, and Terrorist Disasters, Organizing America: Wealth, Power, and the Origins of Corporate Capitalism. So it is very striking to find that Perrow is highly skeptical about the ability of governmental organizations in the United States to protect the public from large failures and disasters of various kinds -- hurricanes, floods, chemical plant fires, software failures, terrorism. His assessment of organizations such as the Federal Emergency Management Agency, the Department of Homeland Security, or the Nuclear Regulatory Commission  is dismal. Here is his summary assessment of the Department of Homeland Security:
We should not expect too much of organizations, but the DHS is extreme in its dysfunctions. As with all organizations, the DHS has been used by its masters and outsiders for purposes that are beyond its mandate, and the usage of the DHS has been extreme. One major user of the DHS is Congress. While Congress is the arm of the government that is closest to the people, it is also the one that is most influenced by corporations and local interest groups that do not have the interests of the larger community in mind. (The Next Catastrophe, kl 205)
The most alarming chapters of The Next Catastrophe concern the failures of US agencies to effectively and intelligently organize preparations that will genuinely make us safer. Perrow provides extended analyses of the Department of Homeland Security, FEMA, and the Nuclear Regulatory Commission in their respective functions -- securing the country against the consequences of terrorist attack, preparing for and responding to major environmental disasters like Katrina, and securing nuclear power plants and spent fuel storage dumps against accident and attack. In chapter after chapter he documents the most egregious and frightening failures of each of these agencies.

The level of organizational ineptitude that he documents in the performance of these agencies is staggering -- one has the impression that a particularly gifted group of high school seniors could have done a better job of responding to the Katrina disaster. ("You're doing a great job, Brownie!") And the disarray that he documents in these organizations is genuinely frightening. He walks through plausible scenarios through which a group of a dozen determined attackers could disable the cooling systems for spent fuel rods at existing nuclear power plants, with catastrophic release of radiation affecting millions of people within 50 miles of the accident. (These scenarios are all the more believable now that we've seen what happened to the cooling ponds at Fukishima.)

What this all suggests is that the U.S. government and our political culture do a particularly bad job of creating organizational intelligence in response to crucial national challenges. By this I mean an effective group of bureaus with a clear mission, committed executive leadership, and consistent communication and collaboration among agencies and a demonstrated ability to formulate and carry out rational plans in addressing identified risks. (Perrow's general assessment of the French nuclear power system seems to be that it is more effective in maintaining safe operations and protecting nuclear materials against attack.)  And the US government's ability to provide this kind of intelligent risk abatement seems particularly weak.

Perrow doesn't endorse the general view that organizations can never succeed in accomplishing the functions we assign to them -- hospitals, police departments, even labor unions. Instead, there seem to be particular reasons why large regulatory agencies in the United States have proven particularly inept, in his assessment. The most faulty organizations are those that are designed to regulate risky activities and those that are charged to create prudent longterm plans for the future that seem particularly suspect, in his account. So what are those reasons for failure in these kinds of organizations?

One major part of his assessment focuses on the role that economic and political power plays in deforming the operations of major organizations to serve the interests of the powerful. Regulatory agencies are "captured" by the powerful industries they are supposed to oversee, whether through influence on the executive branch or through merciless lobbying of the legislative branch. Energy companies pressure the Congress and the NRC to privatize security at nuclear power plants -- with what would otherwise be comical results when it comes to testing the resulting level of security at numerous plants. Private security forces are given advance notice of the time and nature of the simulated attack -- and even so half the attacks are successful.

Another major source of dysfunction that Perrow identifies in the case of the Department of Homeland Security is the workings of Congressional politics. Committee chairs resist losing scope for their committees, so the oversight process remains disjointed and disruptive to the functioning of the agencies. Senators from low-population states block the distribution of DHS funds to enhance the ability of first-responders to be effective in the first hours of an incident, in order to get higher levels of funding for their low-risk populations. So California receives only roughly 13% the per-capita level of funding for anti-terrorism functions that Vermont or Wyoming receive. And of course the funds available through Homeland Security become a major prize for lobbyists, corporations, and other interested parties -- with resulting congressional pressure on DHS strategies and priorities.

Another culprit in this story of failure is the conservative penchant for leaving everything to private enterprise. As Michael Brown put the point during his tenure as director of FEMA, “The general idea—that the business of government is not to provide services, but to make sure that they are provided—seems self-evident to me” (kl 1867). The sustained ideological war against government regulation that has been underway since the Reagan administration has had disastrous consequences when it comes to safety. Activities like nuclear power generation, chemical plants, air travel, drug safety, and residential development in hurricane or forest fire zones are all too risky to be left to private initiative and self-regulation. We need strong, well-resourced, well-staffed, and independent regulatory systems for these activities, and increasingly our scorecard on each of these dimensions is in the failing range.

Overall it appears that Perrow believes that agencies like DHS and FEMA would function better if they were under clear authority of the executive branch rather than Congressional oversight and direction. Presidential authority would not guarantee success -- witness George W. Bush's hapless management of the first iteration of Homeland Security within the White House -- but the odds are better. With a President with a clearly stated and implemented priority for effective management of the risk of terrorism, the planning and coordination needed would have a greater likelihood of success.

It often sounds as though Perrow is faulting these organizations for defects that are inherent in all large organizations.  But it seems more fair to say that his analysis does not identify a general feature of organizations that leads to failure in these cases, but rather a situational fact having to do with the power of business to resist regulation and the susceptibility of Congress and the President to political pressures that hamstring effective regulatory organizations. Perrow does refer to specific organizational hazards -- bad executive leadership, faltering morale, inability to collaborate across agencies, excessively hierarchical architecture -- but the heart of his argument lies elsewhere. The key set of problems spiral back to the inordinate power that corporations have in the United States, and the distortions they create in Congress and the executive branch. The risks that any sober and independent assessment would identify as highest priority are ignored in pursuit of more immediate political or personal gain. It is specifics of the US political system rather than general defects of large organizations per se that lead to the bad outcomes that Perrow identifies. There are strong democracies that do a much better job of regulating risky industries and planning for disasters than we do -- for example, France and Germany.  (Here is a discussion of nuclear safety systems in France published in Nature and a discussion of nuclear safety in Germany published by Nature Conservation and Nuclear Safety.)

It is significant that even though Perrow endorses the need for strengthened regulatory agencies, he doesn't think this would be enough to prevent major catastrophes in the future. So he advocates strongly for reducing the concentration of hazards and populations. As a society, he argues, we need to come to grips with the fact that there are some kinds of activities we should simply not engage in anymore -- intensive residential building in hurricane and forest fire zones, placement of chemical and nuclear plants near cities, routing rail tankers of chlorine through cities like Baltimore and Chicago. (For that matter, a reasonable conclusion one can draw from his account of near-disasters at Indian Point in New York and Davis-Besse in Toledo, is that nuclear power is simply too high a risk to continue to tolerate.) Here is a clear statement of the gravity of culture change this would require:
But what if FEMA were given a mandate to deal with settlement density, escape routes, building codes, and concentrations of hazardous materials in vulnerable sites? We would need a change in out mindset to make basic vulnerabilities such as the size of cities in risky areas and the amounts of hazardous materials in urban areas as high a priority as rescue and relief. (kl 1141)
So who will provide the political will that is needed to reverse course on nuclear and chemical regulation? The public seems to believe (falsely, it would appear) that the NRC is a rigorous and independent agency and that nuclear plants are unlikely to melt down. There isn't much public concern about these risks, and legislators are therefore free to ignore them as well. (Here is an earlier post on "quiet politics" that is relevant.) So where will the political demand for strong regulation come from? Will we need to wait for the bad news we've managed by good fortune to have avoided up to this point?


JR Hulls said...

This post raises an interesting point as to the governmental regulation of risk across all fields. The real question is the extent to which man made accidents are subject to human control. Pellow tends to blame organizational structure, but some accidents are simply the result of perverse chance. It is useful to contrast Pellow's book, Normal Accidents:Living with High Risk Technologies (1984) with Scott D. Sagan’s book, ‘The Limits of Safety: Organizations, Accidents and Nuclear Weapons,’ 1993. discussed further in "Hedging the Apocalypse" at http://somewhatlogically.com/?p=598 looking at risk in the management of financial markets

One example where Pellew is more correct to blame corruption of the regulatory process is the existence of black box trading. Modern supercomputers monitoring these markets can greatly reduce risk. It seems apparent that if the financial community wants to play this kind of game, they should be forced to pay the costs to provide for the technology and regulatory expertise to make it as safe for the public and certainly should not be rewarded with tax incentives. For an interesting discussion on supercomputers, regulation and the flash crash see http://oecdinsights.org/2012/06/28/what-can-nascar-teach-nasdaq-about-avoiding-crashes/

For me, the underlying question is why politicians are able to ignore so much of the hard-won knowledge from other fields concerning risk management in complex systems. In spite of this, they seem to largely escape the consequences when, as in the financial markets, disaster occurs almost entirely because of their regulatory failure.

Peter T said...

Among western states the US is at the low end of the spectrum of bureaucratic efficiency, for lots of reasons (size, decentralised political structure, historical distrust of government, institutionalised penetration of government by private interests, heavy reliance on formal law and so on). The New Deal agencies (TVA, PWA and so on) seem to have been the historical high point, but then you had a strong President with majorities in both houses, plus exceptional times. I doubt the US is going to get too far up the scale any time soon.

AhabTRuler said...

I think equally as important is his work on system complexity: AF 477, Fukushima, & Deepwater Horizon were all unsurprising to those who have followed his work. Not even the initial accident (as they will happen), but instead the individual organizations' inability to understand and cope with the problem and it's follow on effects. Additionally, both AF477 and Deepwater Horizon also reflect his criticisms of the over-reliance on Emergency Safety Devices, and how the faith that such systems will (always) work erodes safety margins and contributes to disaster.

I am so a Charles Perrow fanboi.