On Large Organizations and the Bush Administration

Toyota is a large organization: it employs more than 36,000 people in the United States alone. IBM is a large organization: it employs 23,000 people in India alone. WalMart is a a large organization: it employs 1,500,000 people worldwide.

Nobody I know has any complaints about WalMart's efficiency.

Thus my jaw drops when I come across statements like Arnold Kling's:

TCS: Tech Central Station - A Challenge for Brad DeLong: Large organizations, in the private sector and the public sector alike, are inherently dehumanizing to employees, clumsy, inflexible, and unable to handle sudden new challenges.... [S]urviving [private sector] large organizations tend to be slightly less dysfunctional than those that go out of business.... I wish that we could somehow re-run the history of the last three years with Bill Clinton as President and an administrator DeLong admires... as FEMA Director. My guess is that New Orleans after Katrina would have turned out approximately the same -- maybe a little better, perhaps a little worse...

[M]arket outcomes emerge from decentralized interactions among people.... [B]elievers in Intelligent Design.... The more government fails, the harder they want to try.... If pork-barrel public works projects contributed to the catastrophe in New Orleans, then appropriate billions for pork-barrel public works projects as "relief." Is President Bush particularly bad at administration? I do not think that DeLong is in a position to judge.... I do not believe that one can mold a spectacularly effective government out of the clay of imperfect human beings. I do not believe it is possible for large organizations to behave as if they were the products of Intelligent Design...

"Inherently.

"Inherently" rather than "can easily become."

Shouldn't the fact that WalMart finds it more efficient to be a bureaucracy of 1.5 million people--rather than to split itself up into 15,000 companies of a hundred employees each--make Arnold Kling a little hesitant in his declarations that FEMA was bound to foul up this badly no matter what? Serious thoughts about when wants to use market and when one wants to use command-and-large-organizations--and how one then controls command-and-bureaucracy--would be very welcome here.

Mark Schmitt provides some:

TPMCafe: There are some good points in Matt Yglesias's article noting that plenty of Bush appointees aren't all that well qualified. His point that Michael Chertoff's background as a prosecutor doesn't make him well suited to manage most of the functions of DHS was borne out by the revelation that Chertoff bears responsibility for the delay in FEMA response to Katrina, and may not have understood what he needed to do to trigger the response. But I think this misses a bit of what's going on....

The head of any large agency is inevitably going to lack expertise in many key functions of the agency. Homeland Security, among its other problems, may be just too large and diverse for anyone to manage, but the same will be true of many agencies -- how likely is it that an Interior Secretary who knows a lot about open space and wetlands issues would also be expert in the Bureau of Indian Affairs issues, for example?

I don't really expect Chertoff to know exactly what form he needs to sign to trigger an "Incident of National Significance" declaration. But I expect him to have a chief of staff who pulls together the meeting that brings that information up to the secretary's level promptly.... [W]hat matters most about what's happened in the last few years is that career public servants don't have that same attitude anymore. As Paul Krugman pointed out, in places like FEMA, the FDA, Interior, EPA, top career people are leaving in droves.

There was a terrifying quote in Mike Allen's story about the administration: "Katrina has shown the incredible weakness of the notion that you can have weak players in key spots because the only people who matter are in the White House" -- quoting a Republican lobbyist. It would make sense to say, "you can have weak players in key spots because the people who matter are the operational bureaucrats." That's a familiar concept of government, it's how you survive an Ed Meese. But the idea that it's White House staff who would compensate for the weakness of individual cabinet officers -- that is really something new. And it's absolutely crazy. It shows a total disdain and disregard for what government does. White House staff can sometimes do the broad-brush development of a policy initiative. But even the most seriously qualified White House staff -- let's say the Program Associate Directors at the Office of Management and Budget -- can't manage an agency or implement an initiative or help it survive.

That's why it's so important to forget about Michael Brown or Chertoff or the individuals involved and focus some attention on the system that made it all possible -- a radical, unprecedented system of centralized, politicized control that is guaranteed to fail.

And yes, George W. Bush is bad at administration. He is bad at spotting administrative talent. He is bad at telling who among his subordinates is telling whom the truth and who is telling him pleasing lies.

Impeach George W. Bush. Impeach him now.

Comments

[Anderson]

After DeLong's stunning success with Kinsley, Bush should be quaking in his Hush Puppies right about now ...

[Jonathan W. King]

Brad DeLong writes:
>
> Impeach George W. Bush. Impeach him now.

By my count, you're up to a 56,320 Hz high A by now. So shrill that only dogs can hear you. :-)

But meanwhile, here's something really...weird:

http://photos.reuters.com/Pictures/ViewImage.aspx?type=News¤tPicture=23&photoName=galleries/newspictures/2005-09-14T201816Z_01_UNS93D_RTRIDSP_2_SUMMIT-UN.jpg

And if that's not weird enough, why is our president winking at the President of China?

http://photos.reuters.com/Pictures/ViewImage.aspx?type=News¤tPicture=33&photoName=galleries/newspictures/2005-09-14T182926Z_01_UNS74D_RTRIDSP_2_UN-SUMMIT.jpg

Sorry in advance for the long URLs. But I don't think you can make stuff like this up.

[howard]

if i am reading arnold kling correctly, i assume he doesn't vote, because what difference does it make who is in office? the sluggish large organization overcomes all before it.

since he does post here, i'll ask: does he vote? and if so, on what basis?

[Grumpy]

I do not believe it is possible for large organizations to behave as if they were the products of Intelligent Design...

Oddly, whenever artificial systems are used as analogies for natural selection, anti-evolutionists often object that the artificial system is necessarily the product of a guiding intelligence. The analogy proves the intelligence that underlies nature, they say. Here, we see artificial systems cited in contrast to guided intelligence.

Of course, a strict anti-materialist might believe that a deity presides over all such systems, natural and artificial. Thus, Wal-Mart's prosperity is evidence of divine favor. God, it seems, is inordinately fond of bargains.

[Altoid]

The pattern Mark Schmitt describes-- I think accurately-- is without any precedent I know of in America's national political system. It's far more like one of the following alternatives: self-aggrandizing CEO (who ends up often enough looting the company and leaving an empty shell, but gets good publicity); Stalinist or Maoist party-based system; tsarist system (the one that gave us Potemkin villages); Ottoman system.

The latter two cases are especially ironic given our current position in Iraq.

[ken melvin]

Katrina required lining up ducks and making decisions. You could put an experienced manager or executive in almost any such situation and they would probably handle it very well. It's those with no management experience who are most likely to fail. There are hundreds of project managers, ex-military officers, business executives, and job superintendents who could have handled Katrina, but there's not one in ten thousand lawyers who could have handled it. Lawyers are terrible decision makers. They just sit around arguing.

[marquer]

Further confirmation -- as though any were required -- that sensible individuals need not waste so much as a second of their daily reading time at FlackCentralStation.

[Daniel Barnes]

When it comes to government, Arnold clearly favours Unintelligent Design.

[cb]

Even with WalMart, the store managers can manage their own P&Ls and it is pretty clear, day to day, that they are supposed to be maximizing shareholder value. Such is not the case in government. You *could* make the stores semi-autonomous like gigantic 7-11's...but you can't run most government agencies/programs by looking at P&L's. You could NOT, for example, turn the armed forces into semi-autonomous units without some kind of centralized control.

But you do need an effective bureaucracy...you just hope that the disillusioned don't all leave.

[Jim Dandy]

Yes, the managerial incompetence revealed in the Katrina debacle is symptomatic of the Bush regime’s contempt for government. And, as Krugman argued, this debilitating condition extends to other agencies and departments as well.

However, this argument should not be allowed to obscure the other side of the story, i.e. their truly impressive competence when it comes to the management of the political side of their operation.

Begin with their handling of the Florida vote-recount debacle in 2000.

Then, consider the massive mobilization, first on the propaganda front (domestic and international), second on the military front, for the invasion of Afghanistan and Iraq, nowithstanding the subsequent bungling of the Iraq occupation.

Consider too the extraordinary accomplishments (financial, organizational, marketing, and Kerry-bashing) in waging a successful campaign for the 2004 election.

It is unclear how much credit for all this is to be given to Bush himself and his vaunted MBA and CEO background. I suspect very little, in view of his actual record of business failure. Some (greater) weight must obviously be given to Rove (“Bush’s Brain”) and to Cheney.

[SqueakyRat]

The US military is ruthlessly efficient at one thing above all: budget politicking. War-fighting? Not so much.

[kharris]

I get around to this thought more and more often lately - if you can guess what somebody is going to say about a situation before they say it, you don't really need to pay them any mind. At least not till they try to influence policy. Kling's views on big organizations may have a grain of truth in the purely technical sense that it takes more effort to organize large groups than small ones. That is, however, a pretty trivial observation. If a large group is what is needed to get the job done, then you figure out how to manage large groups. You don't make it an excuse for bad performance unless you are hunting for excuse. Or in Kling's case, unless you have a religious zeal for condemning public efforts and praising private efforts. This is where Kling is so predictable that one need not read him. He is not offering analysis, but rather restating his own dearly held view.

Perfect example -- "My guess is that New Orleans after Katrina would have turned out approximately the same -- maybe a little better, perhaps a little worse.." if Clinton's FEMA had been responding rather than Bush's. Who cares? I don't give a tinkers toot what Kling's "guess" is. My guess is that Clinton would have single-handedly held back the waters. In a world where self-confirming guesses count for anything, Mr. Kling has been neutralized. Wasn't that fun? Now, let's get back to a real discussion.

[Arttu]

I think a lot of right wing conservatives in the US like Bill Oreilly fail to see that bureucracies have become more prominent throughout the last two centuries when capitalism has become to flourish. The irony is the prominence of the 'visible hand' along with the invisible hand. Many people discount bureucracies, but they are extremely effective if you have competent managers. The problem with public adminstration is that there is no exact way to measure and evaluate performance... you effectively depend on qualified ellected adminstrators. That is why it is crucial that appointments are not made arbitrarily and that corruption stays at a low. Ask any sociologist, economic historian or business historian - and they will tell you, capitalism depends on bureaucracies.

[Barry]

The efficiency of the US military is pretty much an object of faith among right-wingers, but it doesn't take competancy to look good when you have massive funding, a large tech base to draw on *and* generally the luxury of picking its fights.

[Jonathan Goldberg]

"I do not believe it is possible for large organizations to behave as if they were the products of Intelligent Design..."

Where does he think large organizations come from? They ARE the products of itelligent design. At least, they are if you think human beings are intelligent. And large, effectve organizations exist; besides those already mentioned I'd cite the VA Health Administration and GE. Not to mention Prussia.

Of course, when your primary design OBJECTIVE is paying off your political supporters, OTHER goals may suffer...

[Uncle Bruno]

And they're very efficient in paying off political supporters.

[Macneil]

Even if he once was, Arnold Kling is not an economist.

[kharris]

SD,

I assume the writing in question is Krugman's writing on the NYT Opinion page. And, yes, I pretty much ignore Krugman unless somebody bothers to point a column out to me. To the extent that "Bush is a lying screw-up" is the "gist" of his writing, yes, you can pretty much predict what he is going to say. However, if you are willing to define "gist" somewhat differently, then he does reach conclusions that I would not have guessed at before hand. There is also a considerable difference between fact-based and faith-based opinion. Krugman does have a habit of relying on facts. Kling's assertions, in the present case and in others I have seen, are no more than assertions. It is his view that "surviving [private sector] large organizations tend to be slightly less dysfunctional than those that go out of business.." We have not actually learned from him that there disfunctionality is endemic in large organizations. Only that he believes it to be so. I am not aware of an objective measure of corporate disfuntion on which he could base such an observation. He cannot have worked in a representative sample of large organizations. His own experience with such organizations cannot, by definition, be objective. So we are dealing with mere assetion, as a starting point for his analysis.

So yes, I read Krugman sparingly. I do not, however, see much comparability between Kling's assertions of fact when no fact is in evidence, and Krugman's use of actual fact in support of writing clearly marked "opinion".

[lawrence franko]

Wal Mart developed their organizational skills in a competitive market place. They are not hamstrung by unions or by civil service rules.

[Or by having George W. Bush and company select their top executives. That looks to be the real killer...]

[Ari]

http://www.nytimes.com/2005/09/14/opinion/14dowd.html

A Fatal Incuriosity
By MAUREEN DOWD

How many places will be in shambles by the time the Bush crew leaves office?

Given that the Bush team has dealt with both gulf crises, Iraq and Katrina, with the same deadly mixture of arrogance and incompetence, and a refusal to face reality, it's frightening to think how it will handle the most demanding act of government domestic investment since the New Deal.

Even though we know W. likes to be in his bubble with his feather pillow, the stories this week are breathtaking about the lengths the White House staff had to go to in order to capture Incurious George's attention.

Newsweek reported that the reality of Katrina did not sink in for the president until days after the levees broke, turning New Orleans into a watery grave. It took a virtual intervention of his top aides to make W. watch the news about the worst natural disaster in a century. Dan Bartlett made a DVD of newscasts on the hurricane to show the president on Friday morning as he flew down to the Gulf Coast.

The aides were scared to tell the isolated president that he should cut short his vacation by a couple of days, Newsweek said, because he can be "cold and snappish in private." Mike Allen wrote in Time about one "youngish aide" who was so terrified about telling Mr. Bush he was wrong about something during the first term, he "had dry heaves" afterward.

The president had to be truly zoned out not to jump at the word "hurricane," given that he has always used his father's term as a reverse playbook and his father almost lost Florida in 1992 because of his slow-footed response to Hurricane Andrew. And W.'s chief of staff, Andy Card, was the White House transportation secretary the senior President Bush sent to the rescue after FEMA bungled that one.

W. has said he prefers to get his information straight up from aides, rather than filtered through newspapers or newscasts. But he surrounds himself with weak sisters who don't have the nerve to break bad news to him, or ideologues with agendas that require warping reality or chuckleheaded cronies like Brownie....

[Richard Cownie]

The US military is *effective*, not *efficient*. Two very different concepts. In fact, any military organization has to be designed to keep functioning when a large proportion of its personnel and equipment have been shot/blown-up/disabled, so it has to have a great deal of redundancy, which in normal circumstances is far from efficient.

[Gideon S]

"Whose bright idea was it to put volunteer firefighters for FEMA in sensitivity-training classes rather than sending them to New Orleans? Hint: not Republicans nor Libertarians"

Not sure I follow. We have a single party in charge of everything right now. And it's not the libertarians or democrats. So who couldn't be bothered to work around the rules to get volunteer firefighters to NO for the presidential photo ops?

Not sure if you've been paying attention since 1980 Lawrence but you're favored party has been the outstanding cut taxes and spend party. The Dems are the sane ones that have to clean up the Repub mess. The welfare state gag was old news in 1979.

[KennyBabes]

Lawrence Franko, Professor of Finance, University of Massachusetts Boston

"Whose bright idea was it to put volunteer firefighters for FEMA in sensitivity-training classes rather than sending them to New Orleans? Hint: not Republicans nor Libertarians.)"

Unfreaking believable
Who is in charge of FEMA
oh right you forgot: Republicans (a god damned failed horse lawyer whose only qualification is being the butt buddy of the head of DHS) and speaking of the Guy responsible for the guy responsible for FEMA he is nothing but a god damn fundraiser for W.

Nice set of qualifications, I am surprised they didn't have the volunteer firefighters landscaping their yards.

Can the state of Massachusetts get a refund on your salary you ignorant fucktard.

[Arttu]

Personally, the idea of competition within public adminstration is a little scary! Look at what happened in Nazi Germany when pure competition was implemented. You had an arbitrary head who rewarded and punished.

It seems to me that a certain level of competition does exist between different government departments. It works kind of like a mini-capital market.
Again, the problem is that measuring performance is difficult, and the government has to decide which is the best way in allocating tax revenue. Companies like Walmart and McDonalds can work the way they do because performance = profit...

[bakho]

The SNAFU that was the Katrina response was a mismanagement from the top. Emergency MGMT or any large organization is subdivided into increasingly smaller subunits with a chain of command and control. When you have Natl Guard troops ready to deploy, but lack orders for days after Katrina, when you have a large hospital ship ready to move but lacking orders, when you have thousands of trained rescue volunteers ready to pitch in, but someone decides they need to go to Atlanta for sensitivity training then you have an dysfunctional organization. The organization is dysfunctional not because it lacks resources or that command and control at the lower levels is missing, it is dysfunction because the leadership is dysfunctional. Inexperienced leadership that does not know the resources that are available or know how to command and control those resources will naturally be ineffective and incompetent. It would be the equivalent of pulling a native out of the jungles and allowing them to coach an NFL football team. (First and goal at the 5 yard line and the coach thinks we only need 4 men on the field?)

[Arttu]

this is why it is vital for the public sector to attain some kind of credibility... Structure matters, but I think strict Parsonian functionalism is wrong. As De Long and Schmitt point out - the players do matter.

[anne]

http://www.nytimes.com/2005/09/14/opinion/14friedman.html

Singapore and Katrina
By THOMAS L. FRIEDMAN

Singapore

There is something troublingly self-indulgent and slothful about America today - something that Katrina highlighted and that people who live in countries where the laws of gravity still apply really noticed. It has rattled them - like watching a parent melt down.

That is certainly the sense I got after observing the Katrina debacle from half a world away here in Singapore - a city-state that, if it believes in anything, believes in good governance. It may roll up the sidewalks pretty early here, and it may even fine you if you spit out your gum, but if you had to choose anywhere in Asia you would want to be caught in a typhoon, it would be Singapore. Trust me, the head of Civil Defense here is not simply someone's college roommate.

Indeed, Singapore believes so strongly that you have to get the best-qualified and least-corruptible people you can into senior positions in the government, judiciary and civil service that its pays its prime minister a salary of $1.1 million a year. It pays its cabinet ministers and Supreme Court justices just under $1 million a year, and pays judges and senior civil servants handsomely down the line.

From Singapore's early years, good governance mattered because the ruling party was in a struggle for the people's hearts and minds with the Communists, who were perceived to be both noncorrupt and caring - so the state had to be the same and more.

Even after the Communists faded, Singapore maintained a tradition of good governance because as a country of only four million people with no natural resources, it had to live by its wits. It needed to run its economy and schools in a way that would extract the maximum from each citizen, which is how four million people built reserves of $100 billion....

[piotr]

Walmart and the military can be both dehumanizing and efficient (I do not know that they are).

Organization may excel at some tasks and fail at others. Walmart could fail miserably marketting upscale apparel. Military that can demolish organized units equipped with decent -- but not outstanding weapons -- may be rather miserable as a force of occupation (as almost all militaries are).

It seems that FEMA tried to perfect swift ditribution of benefits to folks affected by some calamities like hurricanes or earthquakes. There is also not trace, in report and discussions, that FEMA ever made a coherent plant how to quickly evacuate 1,000,000 people of whom 10% may require assistance etc.

The glowing report of FEMA readiness cited from FT was quite probably the way high FEMA officials have seen their mission. First of all, funds were prepare and 1000 FEMA workers were position to distribute checks to stricken homeowners etc. But resquing folks? C'mon. It is a surprise that they could not do it on their own.

It would be nice to learn WHY FEMA made no appropriate plans. Walmart and Toyota may choose their market segments. A wise politician can choose wars for the military (e.g. high altitude bombing campaigns and conquest of small islands with English-speaking population). But FEMA cannot choose.

[antimba]

i guess W having an mba from hbs did not help.

[Auros]

Jack Welch for Secretary Homeland Security!

[robert waldmann]

I think an anonymous "talking points memo" reader makes a strong case that Chertoff's declaration was not necessary, since national emergencies are, by definition incidents of national signficance.

see
http://www.talkingpointsmemo.com/archives/week_2005_09_11.php#006536

no apologies for double post as the comment is relevant to two threads.

[Dave]

I remember reading early on that Brownie was gonna take the fall. Could it be that is what he is paid for? He can always go back to f*cking horses or whatever. His mistake was not having UPS/Fedex supply the logistic talent and support necessary to at least get the job started.

[Sebastian Holsclaw]

"Shouldn't the fact that WalMart finds it more efficient to be a bureaucracy of 1.5 million people--rather than to split itself up into 15,000 companies of a hundred employees each--make Arnold Kling a little hesitant in his declarations that FEMA was bound to foul up this badly no matter what? Serious thoughts about when wants to use market and when one wants to use command-and-large-organizations--and how one then controls command-and-bureaucracy--would be very welcome here."

Large company bureaucracy can sometimes minimize the inherent problems of bureaucracy because they are constantly disciplined by losing money when they don't act properly. Government bureaucracy almost never has that kind of month-to-month accountability. Small government bureaucracy can be disciplined by other means--usually social connections with the community. These types of things don't scale up into a large bureaucracy. The military sometimes avoids this trap by repeatedly focusing the minds of their soldiers on the fact that their own personal survival depends on working efficiently--and even then the military often has the kinds of problems that are endemic in a large government bureaucracy. An organization like FEMA, with tests to its efficiency often coming years apart, is highly likely to be deeply inefficient.

Also, the large non-government bureaucracies which become inefficient often do so because they have become insulated from market disciplining forces.

[BillCross]

a. isn't this just a rehash of the arguments about Taylorism in the 90s?

b. The people went that went to Atlanta for sensitivity training did so because the responded to a request for responders to go door to door giving out information for FEMA.

c, It's sad that the US has no plans for evacuation but Cuba can evacuate up to 1.5 million people successfully and the Netherlands can do 250,000 people and over 1 million livestock

[Brad DeLong]

Sebastian Holsclaw writes: "Large company bureaucracy can sometimes minimize the inherent problems of bureaucracy because they are constantly disciplined by losing money when they don't act properly. Government bureaucracy almost never has that kind of month-to-month accountability."

Sebastian Holsclaw is an extremely insightful writer and thinker from whom I have learned an enormous amount: that is why it is worth noting the times he goes awry, and I think he goes awry here.

There is no magic "market dust" that the market sprinkles on organizations to make them efficient. Managers *make* organizations efficient. Managers of organizations embedded in markets have strong incentives--the fear of losing your job and the desire for more commodious living--to work hard at making their organizations more efficient. But there are other possible motivators as well: the pleasant feeling of a job well done, the pleasant feeling of having helped someone, the shame of public humiliation when you have messed up, loyalty to the person who put you in the job, and so forth.

I think we would all agree that these alternative motivating factors were not sufficient to induce appropriate performance and effort on the part of George W. Bush, Michael Chertoff, and Michael Brown. But this is at least as much a statement about who they are as a statement about whether FEMA could have been expected to do a reasonable job in New Orleans.

[sd]

Brad:

You make an excellet point about the "other" motivations besides desire for survival and money that can motivate people to do a good job. But what I think makes government and non-profit organizations often less efficient than their private sector counterparts is that it is much harder to measure and evaluate performance - even for the employees themselves. I agree that people can be highly motivated by pride in their work and the desire to accomplish great things, but the same structural forces that make it difficult to evaluate government and non-profit employees to determine whether they will keep their jobs and whether they will be promoted, get raises, etc. make it difficult for the employees themselves to tell whether or not they are doing a good job.

This is certainly not to make excuses for the "executive" leadership of FEMA. Clearly FEMA's mission is to make sure disasters are mitigated and in this instance they have failed. We can fire them and take away their livelihood, but its also relatively easy for they themselves to figure out whether or not they did a good job. They should now feel shame. Similarly, the CEO of Wal-Mart can be evaluated on his performance and can know himself whether he has done good work.

But where the two organizations differ is further down the org chart. Just as the Wal-Mart CEO's job is to make money, the individual store manager's job is to make money. But what is the job of the low-level FEMA manager? Without the clear, unambiguous and transparent yardstick of a P&L, it is very difficult to to tell whether a middle manager at FEMA is doing a good job. And for the FEMA employees themselves it is difficult to tell whether they are doing a good job.

This problem can be gotten around in two ways: 1) A well defined professional culture that reinforces norms of excellence; 2) A rigorous effort from the outside to map out the underlying drivers of excellence coupled with a plan to measure those drivers so that employees can be rewarded (and can assess themselves) accordingly. The first option works well for jobs in which the individual is expected to to be a largely self-regulated generalist: college professor, policeman, doctor. The second option works (and is probably the only thing that works) for jobs in which the individual is expected to operate in an interdependent team as a focused specialist. In any event, jobs of the second type abound in government (as they surely must in any large complex organization), and the only way we'll ever get organizational excellence is to make a real effort at defining what the objectives of those jobs are, measuring performance against those objectives, rewarding excellent performance, and making the results of the measurement fully transparent to the employees themselves.

[Glen Raphael]

Brad, you seem to have missed the point of the Intelligent Design analogy. Walmart wasn't just created and managed, it also *evolved* into what it is now, in a competitive marketplace. Sure, the managers tried to make it successful, but the reason Walmart is big is that the strategies those managers came up with *worked*. It might have been good sense or it might have been sheer blind luck that led to the structure Walmart now has, but in a competitive marketplace companies that are well-adapted for their market niche grow and those that aren't shrink. So we have reason to believe, simply due to the fact that Walmart is big, that it is relatively efficient. We have no similar reason to believe the same of FEMA.

In short, the market *does* sprinkle "magic market dust" on organizations to make them more efficient. Just like evolution sprinkles magic evolution dust on species to make them more efficient. It's a weeding-out process and a growth dynamic that can't possibly apply to government programs unless government programs are regularly allowed to fail, go broke, and be dismantled.

[sd]

Glen,

I'm sympathetic to your argument, but part of the problem with FEMA is the long feedback loop. So Brown failed? Great, fire him. Now we get a new guy. We'll know whether he's any good when lots of people do or do not die after the next big disaster.

Private sector enterprises that have similarly long feedback loops suffer some of the same problems. Want to know whether the CEO of Wal-Mart is any good? Wait till next quarter. Want to know if the CEO of Boeing is any good? Wait till 2018.

[Sebastian Holsclaw]

I agree that there isn't any magic market dust, but I think that sd and Glen both articulate better than I would have what the response to that is. Maybe I've been in Sarbanes-Oxely mode too long (in fact I know I have, damn that law) but if your only real testable events come years apart, and if you don't have good non-critical tests to measure things along the way, it is very difficult to get a good outcome.

An analogous (and in my opinion much more important situation) can be found in the testing problems of the education system. The teacher's unions pretty much say it is impossible to accurately test how teachers do (in terms of outcomes) in order to separate good teachers from bad ones. We don't in fact test in such a way. I would argue that as a result of this lack of testing, we don't have good outcomes in teaching.

The problem of "what is a good job" is far more regularly testable in school situations than it is in disaster prep situations. Disaster simulations are not particularly realistic in that you never involve millions of people in a test and it isn't easy to simulate the nasty effect of tens of millions of dollars in property damage. Disaster prep is not as frequently testable as school performance.

Also, like the classic farming example of why command economies don't work, there are many localized conditions for which localized knowledge is valuable and impossible to capture far up the chain in a bureaucracy. This is why the old model of FEMA as a backup to local agencies (who are supposed to be responsible for a deep understanding of their own locales) strikes me as better than a centralized system. On the other hand, the testing events for an individual locale come even less frequently for each locale than they do for FEMA as a whole. This contributes to local problems which do not get exposed until the disaster strikes. I admit that this is problematic, and probably contributed greatly to the levee problem in New Orleans--locals had talked about disaster but didn't do too much about it for decades. After escaping the consequences of that for almost an entire lifetime, it is very human to believe that the danger is overblown.

In any case, the dangers of a large government bureaucracy are real. If you want to get around them you have to make it possible for medium and low-level workers to be tested on their performance so they can make positive changes. I can imagine what a teacher's test would look like because it intersects with the real world on a frequent basis in student performance. I'm not really sure what a good FEMA test would look like.

The good news is that if you can figure out how to discipline an agency like FEMA apart from their very few and very infrequent real-world tests, your method is likely to be applicable to a wide range of agencies with much more frequent testable intersections with their own real world problems.

Unfortunately I have no good suggestions for how that would actually work because all of the good methods I know involve actually testing against the real-world results as a check against the theory.

I guess when I talk about the market approach I'm really talking about a scientific approach where you don't get to just theorize. You have to test your hypothesis in the real world. Unfortunately in many political situations there is enough ambiguity to avoid a real test or fudge the answer to 'validate' your own hypothesis.

None of which excuses Bush for apparently not trying to solve these problems (or at least minimize their ramifications), by the way.

[Sebastian Holsclaw]

Hmm, wish I had thought of it earlier.

There is no magic market dust to increase efficiency. The market is one of the best known methods for testing hypotheses about how things work in a social/economic world. As with the scientific method, testing hypotheses about how things function in the real world is an important aspect of making sure your theories is correct. Unfortunately the best method is rarely available in large government bureaucracy situations.

[Arttu Ulmanen]

Sebastian: Measurement and testing perfomance is problematic, but you can still route some of these problems simply by reducing corruption and building an administration based on public professionals. This is crucial for any sized system to work. I guess Brad De Long's point is that its not size per se that matters... it is also the people involved. If you have an administration which disdains merit and promotes loyality - it ceases to be an effective.

What about political credibility from the public? Surely government adminstrators should respond a little to what people think - that's exactly why Brown resigned. It was Bush who famously said that I've earned political capital and now I intend to spend it. In a liberal democracy performance is measured simply by the public. And currently, the public ain't too happy - no surprise really.

[Sebastian Holsclaw]

"What about political credibility from the public? Surely government adminstrators should respond a little to what people think - that's exactly why Brown resigned. It was Bush who famously said that I've earned political capital and now I intend to spend it. In a liberal democracy performance is measured simply by the public. And currently, the public ain't too happy - no surprise really."

The problem with public accountability is that it only becomes crucial when there is a disaster--which is by definition too late in a disaster prep agency.

"If you have an administration which disdains merit and promotes loyality - it ceases to be an effective."

I think a serious problem with agencies like FEMA is that a good definition of merit is hard to find on a regular basis. So even if you had a truly virtuous administration, you have trouble testing your agency's effectiveness until disaster actually strikes.

[burritoboy]

"Large company bureaucracy can sometimes minimize the inherent problems of bureaucracy because they are constantly disciplined by losing money when they don't act properly."

No, this old chesnut from the socialist planning debate does not hold up. First, there is often no clear connection between action and overall firm performance in the large company setting. A firm may be doing well largely due to positional advantages it gained decades (sometimes centuries) ago. While it's possible that a competitor will come along and reveal the firm's recent errors, this can take decades or more to arrive. The big 3 US automakers were likely making large, irreversible errors in the 1950s, but because no one has been able to start a new major US automaker since the 1920s, it took until the mid-1970s and later (and the Japanese invading from their own protected markets) to fully reveal those errors.

Second, Holsclaw's theory underestimates the truly irreducible chaos of markets. Causes do not lead to predictable effects. The firm will build an internal narrative about what's going on, but this narrative is.....a narrative, and suffers from the same problems that any other narrative does. Conversely, capital markets will build their own narratives as well. Literally, investment analysts (like myself for most of my career) will ask each other: "what's the story on company Z?". There is a common narrative structure on how such stories should be structured (and again, these narratives suffer from the inherent faults of all narratives).

Of course, there are a few cases where clearly the firm is performing ineffectively (they're performing far more poorly than other comparable firms) or clearly effectively - but those instances are comparatively rare and are often more ambigious than they initially seem as well.

So, it's never clear that large private for-profit organizations do not face the exact same (or, at minimum, very similar) problems as do governmental organizations. Indeed, in this particular instance, since FEMA encounters disasters on a regular basis, there is lpossibly more, not fewer, opportunities for performance checks than there may be in many large private for-profit firms.

[Sebastian Holsclaw]

"A firm may be doing well largely due to positional advantages it gained decades (sometimes centuries) ago. While it's possible that a competitor will come along and reveal the firm's recent errors, this can take decades or more to arrive. The big 3 US automakers were likely making large, irreversible errors in the 1950s, but because no one has been able to start a new major US automaker since the 1920s, it took until the mid-1970s and later (and the Japanese invading from their own protected markets) to fully reveal those errors."

Sd covered this objection with "Private sector enterprises that have similarly long feedback loops suffer some of the same problems. Want to know whether the CEO of Wal-Mart is any good? Wait till next quarter. Want to know if the CEO of Boeing is any good? Wait till 2018."

Yes, not all companies feel market discipline. Yes this tends to mean that they make undisciplined errors which later cause serious problems. Yes this is analogous to FEMA. Your argument has the tone of disagreement, but all of this is agreement. My suggestion is that market discipline tends to out efficiency problems. My argument was not that market discipline effects all corporations equally (especially when they can get the government to regulate their competitors but that is a different argument.)Suggesting that some corporations are able to avoid market discipline for one reason or another and that this causes large inefficiencies is evidence supporting my argument that avoiding market discipline is likely to allow large inefficiencies to survive long enough to cause large disasters.

The problem is that while some companies avoid market discipline, nearly all government agencies do. So pointing out the ill effects of that avoidance causes in some corporations tells us nothing about avoiding such failures in FEMA.

[Sebastian Holsclaw]

What a poorly formed last paragraph. Let's try:

The problem is that while some companies avoid market discipline, nearly all government agencies do. So pointing out the ill effects that avoidance causes in some corporations tells us nothing about avoiding such failures in FEMA.

Funny how one word can make a sentence almost incomprehensible.

[burritoboy]

Nope, you don't get it at all. All organizations try to evolve toward stability. Business organizations not only try to construct themselves, insofar as possible, to avoid market discipline but also to change their very environments to avoid market discipline. On the highest level, organizations often suceed in constructing mental frames such that very few people (within or without the organization) can even concieve of the organizations' failure.

An example is the Big 3 automakers. They managed to convince both themselves and everyone else in the world that the Big 3 were competitive amongst themselves, produced the best and only really desirable cars and that, if you wanted to be driving the best car, you would drive the cars that the Big 3 were best at building. The other automakers around the world only survived post-WWII because their national governments protected them. Small cars, for example, were bad because Americans didn't drive small cars. If you had to drive a small car, it was because you were poor and couldn't afford a large one.

None of this was an error on the Big 3's part. They were vastly profitable companies. The mental frames were conversely a massive success. Japanese auto-makers didn't even actually make cars, as defined by the mental frames. They made sardine cans on wheels, things that were non-cars.

The breaking of the mental frames came only when political change (higher oil prices jacked up by OPEC)broke part of the frames. This was not a failure by the Big 3 - small cars were not cars at all until the mental frames about what a "car" was shifted. The only actual thing one can point to as "failure", was that the Big 3 were not nimble enough to quickly switch to producing higher-quality smaller cars. But remember, that lack of that particular type of agility was part of the system built under the decades of successful adaptation.

Such shifts happen very, very infrequently. Most industries evolve to the point that these shifts very, very rarely occur. This is not a bad thing for those industries, but a good thing. Repeated drastic market shifts usually do not benefit the economy as a whole, either. Go read Elbaum and Lazonick, or Chandler.

Government agencies are often more frequently tested, not less, than business organizations. Business organizations exist within very protected environments. There are few domestic political events that they cannot influence, or at least predict long in advance. The legal and institutional environments are usually quite stable. Competing organizations can either be purchased or allied with. They are essentially self-governing, self-perpetuating organizations that answer to very few external rulers - at base, only to creditors IF the firm is in bankruptcy (they certainly do not respond to equity shareholders). None of this is the case with government organizations, who must often inherently deal with international politics (out of our control), extreme political shifts (wars, terrorism, revolutions, etc) as well as being ruled by multiple external forces (voters, political oversight, business organizations, etc.).

[Arttu]

'The problem with public accountability is that it only becomes crucial when there is a disaster--which is by definition too late in a disaster prep agency.'

I was referring to government institutions in general - not just disaster relief, but I take your point. Still, I'm thinking more in terms of solutions to the problems posed here. In my opinion it has to be accountability.
There are also surveys which evaluate government performance to a certain degree.

For example:

http://www.transparency.org/pressreleases_archive/2004/2004.10.20.cpi.en.html

'I think a serious problem with agencies like FEMA is that a good definition of merit is hard to find on a regular basis. So even if you had a truly virtuous administration, you have trouble testing your agency's effectiveness until disaster actually strikes.'

I agree that the definition of merit is ambiguous and is probably a function of the political culture which exists. The notion of a 'truly virtuous administration' is also ambiguous in the same way. I still think there should be no obstacles for public employers to recruit on the basis of certain defined credentials. This was not the case with FEMA and the exodus of those with actual credentials from specific public institutions is worrying.