Austin Contrarian

Give FDA the right incentives

The emerging consensus is that FDA unfairly blamed tomatoes for the salmonella "outbreak."  I say this confidently because members of Congress have begun pushing for hearings.

The tomato growers, not surprisingly, want the U.S. government to compensate them for the havoc it wreaked.  To the tune of $100 million.  Note that this probably understates the total cost. $100 million is the tomato growers and shippers' estimate of their own damages.  The total cost including losses to downstream users, particularly restaurants, is much higher.  (I can personally attest that Chipotle lost several lunch sales while it was out of salsa.)

Should the government compensate tomato growers?

I don't think tomato growers have a compelling equitable case.  Over the long run, the price of produce ought to reflect the risk of random, FDA-induced scares.  Think of FDA as a freak hailstorm or 100-year flood;  a portion of the price of produce reflects the cost to growers of self-insuring against government-induced catastrophe.  Compensating tomato growers may give them a windfall. 

The real case for compensation is that the government needs proper incentives.  If it doesn't pay for the damage caused by bogus scares, it has an incentive to over-warn.  FDA officials face political punishment for acting too slowly to quell a genuine outbreak, but little sanction other than embarrassment for acting too aggressively.  And since FDA has the experts, it is difficult for the public to assess whether it indeed acted too aggressively.  

The solution is to have Congress pay compensation on a case-by-case basis.  Presumably, FDA officials want to protect their budget, and it is a bad budget-preservation strategy to force your paymasters to shell out hundreds of millions of dollars for your blunders.

Yes, there are problems with this.  It is difficult even for Congress to determine when FDA has blundered.  And making growers unconditionally responsible for outbreaks gives them an incentive to police their industry.  But we don't want to encourage over-policing -- we don't want to pay for too much safety -- and FDA overreactions don't give anyone the right incentives.

Related posts:

• Don't mess with my pico.
• After the salmonella rush.

Now it's the jalapeños and salsa

For those of you who left snarky comments to my salmonella post, you should know that FDA is massing for a frontal assault on jalapeños and salsa.

Give it time.  Sooner or later, FDA will find something you do care about.

After the salmonella rush

It may not have been the tomatoes after all.  After a fruitless (!) search for contaminated tomatoes, FDA is beginning to look elsewhere for the source of the salmonella "outbreak":

As sicknesses continue beyond the shelf life of tomatoes — it’s been almost three months since the first Salmonella case — officials are also starting to question whether another vegetable is to blame, according to this morning’s USA Today..

“We’re broadening the investigation to be sure it encompasses food items that are commonly consumed with tomatoes,” Robert Tauxe of the Centers for Disease Control and Prevention said.

I think the good doctor is just trying to reassure us.  "Food items that are commonly consumed with tomatoes" sounds suspiciously like "the general food supply"; after all, what's not "commonly consumed with tomatoes"?  This means that any one of us could be exposed to salmonella at any time.  Why, you could be eating a big bowl of salmonella-tainted spaghetti right now.

We're screwed.

Right.

I originally planned to write this up as a possible instance of observation bias:

• FDA detected a random cluster of reported salmonella cases and put out the alert.  It didn't know whether the cluster was due to chance or to a genuine outbreak, but it gets in trouble for not warning, so it figured it was safer to over-warn.
• Doctors everywhere, alerted to a possible outbreak of salmonella poisoning, began looking for salmonella poisoning.
• The symptoms of salmonella poisoning are headache, diarrhea, nausea, fever, and abdominal cramps, symptoms that doctors ordinarily would have been content to diagnose as "stomach virus."  But doctors instead screened these patients for salmonella poisoning.  They asked them if they'd recently eaten tomatoes -- and many had, since many of them had recently eaten food.  Rather than send them straight home with some Tylenol and Phenergan, the doctors sent them to the lab to be tested for salmonella.
• Presto.  Lots of unremarkable and (epidemiologically) invisible cases of stomach virus became confirmed cases of salmonella poisoning.

As I said, that was my original take.  But I now think this story is too charitable to FDA.  It doesn't sound like there was an outbreak at all. 

FDA has counted 810 cases since April 23, a period of two months.  810 cases every two months is equivalent to a rate of 4,860 cases per year.  4,860 cases per year is not an epidemic.  It's not even the background rate:  

It is estimated that from 2 to 4 million cases of salmonellosis occur in the U.S. annually. (Source: FDA Bad Bug Book) ... Every year, approximately 40,000 cases of salmonellosis are reported in the United States. Because many milder cases are not diagnosed or reported, the actual number of infections may be twenty or more times greater. (Source: excerpt from Salmonellosis (General): DBMD) 

(I like that FDA gives its epidemiology publications cute, nursery-rhymey names.)

We are a nation of 300 million people who occasionally consume undercooked poultry and bad eggs and other food that tasted a little funny at the time.  Some of us will get salmonellosis.  "Some" of 300 million is "a lot" -- on average, 10 times as many confirmed cases as the Great Salmonella Poisoning Outbreak of 2008.  And that's before subtracting the extra cases doctors have detected this year because they were specifically looking for them.   

Thus, to save the general population from a risk of salmonella poisoning well below the background risk, FDA has given restaurants a $100 million butt-kicking and forced the rest of us to go for weeks without a decent salad or fajita.   

They know they've screwed up:

“It’s bad, and I think everyone will be very apologetic” if it turns out tomatoes weren’t the source, said Tim Jones, Tennessee’s state epidemiologist, describing himself as “increasingly concerned” about whether tomatoes are to blame.

That's how an epidemiologist says "D'oh!"

Don't mess with my pico.

Is the salmonella-tainted tomato "scare" over?  The New York Times reports:

Six new cases of an unusual form of salmonella linked to eating raw tomatoes have been confirmed in New York City, in addition to a previously known case, the Department of Health and Mental Hygiene said on Wednesday. In addition, Suffolk County announced that a 26-year-old West Babylon man also contracted the strain of salmonella from eating tomatoes at New York City on May 27.

Maybe I have an undue attachment to pico de gallo, but I was upset when my restaurants pulled tomatoes off the menu.  I was cheerfully willing to accept the tiny risk of encountering Mr. Salmonella.

There have been 383 cases of tomato-related salmonella reported since April.  Even if all the cases occurred in April, that's just 12-13 cases a day.  By contrast, 6 billion pounds of fresh tomatoes* are consumed each year by 300 million Americans spread over a market area of 3 million square miles.  I would be flabbergasted if only a couple handfuls (literally) of the tomatoes consumed each day were contaminated with some bacteria.  Fresh produce always has a chance of being contaminated with something.  It's an extremely small chance, though -- I imagine it is zero, to five significant digits, even when there is a confirmed "outbreak."  My chance of ending up with a contaminated tomato is still practically indistinguishable from zero.   

When the FDA detects an outbreak of salmonella, it ought to investigate.  We ought to work hard to get tainted food out of the system.  But there will always be some bad stuff out there no matter how hard we try.  Don't mess with my pico unless there's a real epidemic.

*20 pounds of fresh tomatoes per American per year times 300 million.

The City wants to get into the parking business

Condo builders are apparently reluctant to build or operate public parking downtown, so Councilmember McCracken wants the City to take up the slack.  He claims it will be "cheaper" for the City to build public parking:

McCracken believes they could be financially beneficial for the city because it could lower construction costs by issuing low-interest debt and wouldn't have to pay property taxes on the finished product. The city also would be willing to accept lower profit margins than private developers.

This is pretty shaky accounting.  The City's loss of property taxes is part of the cost of operating a City-owned garage, regardless of whether it has its own budget entry.  This also ignores the opportunity cost:  if the City could sell a lot for a gazillion dollars, but instead builds a public garage on it, the lost loss of that gazillion dollars is a cost, too.

I'm also skeptical that these will generate enough revenue to cover the construction and operating costs.  Developers' land costs and property taxes are more or less fixed costs.  The cost to them of adding public parking is just the marginal construction cost.  That they don't think public parking will provide a profitable return on this capital investment is a reliable indicator that it is not.  This shouldn't really be a surprise, since much of the return from parking is captured by merchants, who benefit from the traffic.

Parking is a problem downtown, and I'm not trying to suggest that the City sit on its hands.  But we shouldn't go into this thinking that this will be anything other than a subsidy for downtown property owners and merchants, possibly a deep one.  We also need to recognize that this subsidy will fall most heavily on those without cars.

You can think of downtown as a complex ecosystem, with the different inhabitants in a mutually dependent relationship.  The condo builders need plenty of street-level retail to make their condos attractive to buyers.  The retailers need plenty of condo dwellers.  But they also need traffic from outside downtown, because the condo residents alone won't keep them in business.  It is a fantasy to think that buses will bring enough traffic downtown, at least anytime soon, so the downtown retailers need parking.  The problem is that the condo builders who could build the parking don't want to take the hit for the benefit of everyone else.  It's a classic free-rider problem.

The City can do some things fairly easily.  It can start by charging what the market will bear for existing parking, particularly street parking.  (The price should be set high enough so there is always some free curb space.)  This will maximize revenue and encourage those traveling downtown to consider alternative transportation.

Ideally, it would figure out some way of shifting the cost of new parking to downtown property owners.  Spending some of the density bonuses on parking would be a better use of the money than some of the others uses that have been proposed.  Developers would probably be more receptive to that use, since ultimately it will enhance the value of their property. 

Could the City create a public improvement district to fund parking?  The district would build the off-street parking and assess the owners. That would eliminate the free-riding problem, and eliminate the subsidy.  And if the City put the decision-making in the hands of the downtown merchants and property owners, it could reduce the risk of oversupplying parking -- presumably, downtown owners won't pay to build more parking than they need. 

I'm not a local government finance expert.  McCracken may have already considered and rejected this possibility.  But I'd like to hear that all options have been considered before the City launches into a risky, property-subsidizing enterprise.

P.S.  Shilli at Austinist has a similar take (although he appears to be more skeptical of the need for new parking).

Law school chum

My law school roommate has been sued by convicted terror plotter Jose Padilla:

Jose Padilla, the American citizen who was held in military detention for more than three years as an enemy combatant, filed a lawsuit Friday against a former Justice Department lawyer who helped provide the legal justifications for what the suit says was Mr. Padilla’s unconstitutional confinement and “gross physical and psychological abuse.”

The lawyer, John C. Yoo, now a law professor at the University of California, Berkeley, wrote or helped prepare a series of legal memorandums on interrogations and the treatment of detainees after the Sept. 11 attacks.

A lawyer for Mr. Yoo, Eric M. George, called Mr. Padilla’s suit “a political diatribe” that “belongs, at best, in a journal, not before a federal court.”

Mr. Padilla, 37, was transferred from military custody to the criminal justice system in 2006, and in August he was convicted of terrorism-related charges in Miami. He awaits sentencing.

The new lawsuit, filed in federal court in San Francisco, seeks only one dollar in damages. “That’s what Padilla directed us to ask for,” said Jonathan M. Freiman, one of Mr. Padilla’s lawyers. “At bottom, this isn’t about money. It’s about right and wrong.”

John was well to the right of me in law school, at least on stuff like this.  We did not talk about politics or even law much, though.  Our interests didn't mesh.  He was interested in esoteric issues of federal jurisdiction and executive powers and I was interested in property law and land-use controls.  The one significant exception occurred when he got assigned a regulatory-takings case for moot court.  I got stuck with some Establishment Clause case about prayer at a high-school graduation that I couldn't have cared less about.  I spent more time working through the fine points of the takings issues with John than I spent thinking about my own case.

We were a real odd couple -- an urbane Philadelphian and a snuff-dipping redneck from Mississippi.  We rented a house on the shore of the Long Island Sound in a New Haven suburb.  It was a long way from the law school, too far really for me to justify going to class, so I spent my time skipping Bankruptcy and watching the Movie Channel.  John spent his time editing articles for the Law Journal, hobnobbing with professors, and doing the other things that overachievers do. 

John is a very smart guy (he was tenured at Berkeley by 30 or some such ridiculous age) and a thoughtful guy, media portrayals notwithstanding.  He knows his stuff, and has a calm, reasonable style that can be very persuasive.  (After all, he convinced a slew of seasoned government lawyers to sign on to his legal analyses.) 

I always thought John was wildly off the mark with Padilla, though.  How can a US citizen apprehended on US soil be held without trial merely because the executive has classified him as an "enemy combatant"?  John has some precedent involving saboteurs apprehended on a beach during World War II, but it never seemed that persuasive to me.

Still, I can't believe this case is worth the filing fee:

The suit is based in part on a recent book by Jack Goldsmith, a Harvard law professor who, while serving in the Justice Department in 2003 and 2004, disavowed some of Mr. Yoo’s work. In the book, “The Terror Presidency” (W.W. Norton), Mr. Goldsmith wrote that two of Mr. Yoo’s memorandums were “legally flawed” and “tendentious in substance and tone.”

I freely confess my bias in favor of my friend and my lack of familiarity with the case law governing this type of civil-rights claim.  But I am skeptical that a court will allow a government lawyer to be exposed to civil-rights liability for a legal analysis he prepared for his client.  That would, quite predictably, chill candid legal advice. (I'm sure one of the ulterior motives of this suit is to do just that).  In any event, mere evidence of "tendentiousness" won't be enough -- they hand out tendentiousness with the law degree.   Even if the court allows the claim to proceed, I imagine it will require proof the analysis was prepared in bad faith or without a colorable basis or something similar.  Padilla won't be able to prove that.

Anyway, I hope this thing gets thrown out for my friend's sake.

Now back to our regularly scheduled programming . . .