Prescription Drugs

“Is the FDA Too Conservative or Too Aggressive? A Bayesian Decision Analysis of Clinical Trial Design,” by Vahid Montazerhodjat and Andrew W. Lo. August 2015. SSRN #2641547.

Conservative and libertarian critics of the U.S. Food and Drug Administration often argue for “compassionate access” to experimental drugs. Under such access, severely ill patients can receive experimental drugs that have shown no deleterious effects in their initial Phase I “safety” trials, but that haven’t completed further trials of safety and efficacy. (See “Breaking the FDA Monopoly,” Summer 2004, and “Regulation Overdose” [book review], Summer 2010.)

Critics of this access, such as prominent bioethicist and oncologist Ezekiel Emanuel, point out that many safety issues are not discovered in Phase I trials. (See “Cancer in the Courts,” New Republic, July 3, 2006.) Many forget that the drug Thalidomide, which caused the birth defects that led to the 1962 Food and Drug Act amendments requiring not only safety but efficacy testing before FDA approval, had already undergone Phase I safety trials.

According to Emanuel, another problem with compassionate access is that it would impede broader trials of a drug’s safety and efficacy. For instance, in the late 1980s many oncologists believed that bone marrow transplants could treat metastatic breast cancer. Because of political pressure, an allowance for compassionate access was created and 20,000 women received marrow transplants outside of the clinical trials. Because of that access, it took many years to gather 1,000 women who agreed to participate in a proper clinical trial of the treatment. The trial results demonstrated that the transplants were not effective relative to standard chemotherapy, which meant that as a result of the compassionate access, thousands of women suffered through a painful, grueling, unnecessary treatment that cost millions of dollars.

Both proponents and critics of compassionate access frame the issue in a binary way: compassionate access either should or shouldn’t be allowed. Economists often eschew such “yes/​no” policies in favor of policies that vary with costs and benefits. In the case of experimental drugs, the costs and benefits would involve false positive (Type‑I) and false negative (Type-II) inference errors about the true effects of a drug on disease and health. This paper outlines a practical policy toward experimental drug access that addresses the Type‑I/Type-II error issue.

The paper’s premise is that the costs and benefits of Type‑I and Type-II error avoidance vary with the disease context. For example, the five-year survival rate for pancreatic cancer is 1 percent, so patients with the disease are worried much less about Type‑I efficacy errors with the experimental drug than they are about the effects of the illness. Thus, the current FDA tendency to avoid Type‑I errors (approving drugs that don’t work or have too many side effects) creates too many Type-II errors (rejecting effective drugs).

In general, for deadly diseases, Type-II error costs are larger than Type‑I error costs—the value of lives saved offsets the costs of some false positives. For mild diseases the opposite is true: Type‑I error costs are larger than Type-II error costs. In both cases, the costs are proportional to the size of the relevant populations.

The authors use mortality rates and years lived with disability to adjust the usual sample size and confidence level requirements for trials to create their Bayesian-adjusted equivalents. The more severe the mortality and disability consequences of the disease, the less confident we need to be of a drug’s efficacy and safety in order to allow approval.

Trials currently require a difference in results between experimental and control patients that exceeds 1.96 standard deviations of the mean (that is, keep false positive errors to below 5 percent). In their framework, taking into account the deadliness of the disease, a difference of only 0.587 standard deviations (which keeps Type‑I errors to below 56 percent) should be required for approval of a pancreatic cancer drug.

The framework of this paper operationalizes the intuition behind both sides of the debate. It reduces the hurdles for drug approval for diseases that have severe mortality or disability consequences that affect large numbers of people and increases the hurdles for those diseases with the opposite characteristics.

—Peter Van Doren, Cato Institute

Curfews and Crime

“Keep the Kids Inside? Juvenile Curfews and Urban Gun Violence,” by Jillian B. Carr and Jennifer L. Doleac. December 2015. SSRN #2486903.

Washington, D.C. uses outdoor audio sensor technology to detect gunfire incidents in four of the city’s high-crime police precincts. Washington also has a curfew for people under age 17, which begins at 11 p.m. on weeknights from September to June, and at midnight on all other nights of the year. In this paper, the researchers use the sensor data and the exogenous one-hour difference in curfew times on weeknights to determine if the curfew is having a beneficial effect.

Curfews are premised on the idea that preventing nighttime interactions between youths will prevent violence. But there is a concern that curfews may have a perverse effect: by taking law-abiding youths off the streets, lawmakers may also take away “eyes” and “ears” that would discourage violence, as Jane Jacobs noted in her 1961 book, The Death and Life of Great American Cities (Random House).

The study’s authors conclude that gunfire increases in the weeknight 11 p.m. to midnight hour from September through June relative to the same hour during the summer. Thus the curfew costs lives rather than saves them.

How many lives are lost? Approximately seven additional gunfire incidents per week occur in that hour across the four monitored police districts. The literature suggests that one additional gunfire incident results in 0.0048 additional homicides. If one statistical life is valued at $9 million, each gunfire incident results in around $43,000 in costs. Thus, seven additional incidents per week for 45 weeks a year result in around $13.5 million dollars in social costs accounting only for homicides.

—Peter Van Doren, Cato Institute

Auto Safety

“The Transformation of Automobile Safety Regulation: Bureaucratic Adaptation to Legal Culture,” by Jerry L. Mashaw and David L. Harfst. December 2015. SSRN #2703370.

Concise narrative histories of regulatory agencies are rare and invaluable. In 1990, Jerry Mashaw and David Harfst wrote The Struggle for Auto Safety (Harvard University Press) about the first 15 years of federal auto safety regulation following the 1966 unanimous approval by Congress of the National Traffic and Motor Vehicle Safety Act. This paper extends that history to the present.

The law’s passage followed the publication of Ralph Nader’s book Unsafe at Any Speed (Grossman Publishers) the previous fall. The premise of both the book and the law was that the market for vehicle safety failed and government regulation was necessary to force manufacturers to provide technical modifications to cars that reduce the injury and fatality rate from accidents.

In the first years of its existence, the federal auto safety agency (now called the National Highway Traffic Safety Administration [NHTSA]) attempted to live up to that premise. But the statute required that the agency’s regulations be “reasonable,” “practicable,” and “objective.” Citing those requirements, car manufacturers sued to block the regulations, and between 1968 and 1978 they won six of the 10 cases litigated over the rules.

One of those cases (won by the manufacturers in 1972) was the passive restraint rule requiring manufacturers to install airbags. One provision of that rule prevented engine start unless the driver and front-passenger seatbelts were fastened. The courts left that provision intact, and NHTSA implemented it in 1974. Immediately, the agency and lawmakers were beset with public backlash. Acting with unusual haste, Congress repealed the interlock requirement that November, and even forbid NHTSA from requiring a warning light or sound lasting longer than 8 seconds indicating seatbelt non-use.

The agency responded to those setbacks and the subsequent regulation-skeptical Reagan administration by, first, passing no new rules at all, and then by issuing rules that mandated safety devices that the industry was adopting anyway. In a 2015 report on lives saved by NHTSA rules issued after the early 2000s, the agency concluded that four of those eight major rules had effective dates that were after the median new car already was equipped with the mandated device.

The lack of regulatory activity didn’t mean NHTSA was doing nothing about vehicle safety. Instead, it began issuing recalls for vehicles it deemed to have safety defects. NHTSA had realized that, though the evidence required for the issuance of defensible safety rules is high, the evidence required for recalls is light. Car makers are loath to publicly fight recalls because such fights hurt brand image and market share.

But recalls do little to improve aggregate auto safety because most accidents are the result of driver error rather than vehicle defects. A 2008 U.S. Department of Transportation report to Congress found that vehicle defects or failure accounted for only 2.4 percent of accidents, while driver error accounted for over 95 percent. According to the authors,

The efficacy of the recall program is mysterious and dubious. Yet, if anything, Congress and the public seem to want more aggressive recall activity, not less. NHTSA is happy to oblige.

Mashaw and Harfst conclude that the premise of the 1966 legislation (and industry opposition to it) was that safety does not sell and that Congress must compel auto manufacturers to provide it. But, they decide, “there does now appear to be something resembling a private market for safety, and industry is supplying it.” And NHTSA codifies what they are supplying.

—Peter Van Doren, Cato Institute

Fiscal Rules

“Can Fiscal Rules Constrain the Size of Government? An Analysis of the ‘Crown Jewel’ of Tax and Expenditure Limitations,” by Paul Eliason and Byron Lutz. Federal Reserve working paper, February 2016.

In the battle to rein in government spending, the Colorado Taxpayer Bill of Rights (TABOR) seemed to be a rare panacea. The rule, passed in the early 1990s, limited the state government’s spending growth to the combined rate of inflation and population growth. If revenue increased beyond that rate—which could occur, for instance, if economic growth were to concomitantly boost incomes and tax revenue—then the surplus funds would be returned to taxpayers. The state’s apparently salutary budget health in the early 2000s was attributed to TABOR, and then-governor Bill Owens briefly became the poster child for the libertarian small-government crowd.

However, a closer look at Colorado’s budget since TABOR’s passage reveals that it has not caused the state to behave any differently than similar states, according to a new working paper by Paul Eliason of Duke University and Byron Lutz of the Federal Reserve Board. Instead, the salutary view of TABOR stumbles on two serious problems.

The first is that Colorado can suspend TABOR temporarily. The state has done this several times, especially in the last decade. It remains a truism that it’s impossible for legislation to tie the hands of future lawmakers.

The second problem is that it is unclear what is the counterfactual—that is, what would Colorado have done if there were no TABOR? Simply comparing it to nearby states is deceiving: by dint of its major metropolitan area in Denver and the state’s relatively well-educated and wealthy populace, it does not make sense to compare its budgets to neighboring states. Economically speaking, Colorado has little in common with Wyoming, New Mexico, or Utah. Neither does it make sense to compare it to any of the coastal states, which have economies substantially different than Colorado’s.

To carry out their analysis, Eliason and Lutz cleverly create a synthetic state—a complex combination of states that have the most in common with Colorado—and compare Colorado to the synthetic state’s tax and spending evolution over the years immediately before and after TABOR.

What they find is that there is no discernible difference between Colorado’s actual spending patterns and its synthetic, unconstrained cousin. In short, they argue, Colorado did show a modicum of budget restraint at some point, but it didn’t last all that long and it merely reflected the preference of the populace, manifested in the government it elected.

The reality is that budget gimmicks are not the answer to controlling government. There is simply no substitute for an informed populace that uses the ballot box to show its preference for limited government.

—Ike Brannon, Cato Institute

Tax Breaks for the Elderly

“Do Tax Breaks for the Elderly Promote Economic Growth?” by Ben Brewer, Karen Smith Conway, and Jonathan Rork. University of New Hampshire working paper, April 2016.

States compete on taxes in myriad ways. Several states do not have an income tax. Others have reduced taxes on businesses large and small. (Kansas has gone so far as to have abolished small businesses taxes two years ago.) And, of course, states offer a wide variety of tax breaks with the intent of luring specific industries or businesses to locate or remain in their realm.

But many states also compete for senior citizens by offering a lower tax rate for pension income. How states do this varies greatly. Some exempt military pension income from the state income tax altogether. Others offer tax breaks for public employee pensions and Social Security benefits as well. Still others give tax breaks for all pension income, whether public or private.

The rationale for this is that senior citizens don’t place much of a demand on state services—their children are usually done with public schools, for instance—so having seniors around costs the state relatively little. The authors of this working paper acknowledge that fiscal tradeoff, but they suspect there’s little reason to believe that retirees are influenced all that much by the tax breaks. Many retirees are unwilling to trade that home state—where their friends, relatives, and grandkids are abundant—for a neighboring state just to save a few grand a year. And the retirees who do move often “snowbird” in Florida for just six months (and one day, of course, to qualify as residents) and then return to their home states.

Or so the authors thought. The reality, they found, is that lower tax rates on retiree income do boost a state’s economic growth. Why this is so remains a mystery to the authors: they do find that taxes on lower-income workers seem to matter the most to seniors, so it’s tempting to construct some sort of Keynesian aggregate-demand story to explain the higher economic growth. However, if fiscal stimulus spending can’t explain sustained higher growth rates in an economy, then it seems unlikely that lower tax rates on seniors would do so.

I have an alternative hypothesis that’s broadly consistent with their data. Hundreds of communities across the country are chasing prospective “angel investors”: wealthy folks (but not exorbitantly so; think of a developer in a small city who sold the business for $15 million when he hit 60). These folks are not quite ready to retire, but they’ve wisely cashed out on their career. Yet they still have a desire to chase one or two more “big ideas.” They won’t risk their entire nest egg, but if they find something promising, they’ll be tempted to throw some savings at it, along with their personal attention and skills. And the place they’d want to do this isn’t Florida, but their hometowns, where they’ve lived for decades.

Suppose these potential angels already snowbird, returning home in the summer to see the grandkids. They’re less inclined to try a business idea if they’re only going to be in-state for less than half of the year. But if their home states adopt senior-friendly tax cuts, the Florida retirement village may no longer seem so inviting. Sure, they’ll still go south to skip the worst of the winter, but they’ll spend more time back home, where they’ll be more inclined to try that startup and give it a decade to see if it flies. And if it does (or even if it doesn’t), they might then be inclined to try another idea.

Ohio Gov. John Kasich (R) offered this idea to support tax breaks implemented in Ohio in 2011, and there’s some evidence that it’s working. It is, of course, a controversial idea that, for what it’s worth, the AARP rejected vehemently when the breaks were first proposed, precisely because they mainly benefited upper-income seniors. (Perhaps those folks are less likely to join the organization?)

It is worth noting that the working paper authors didn’t expect to find that retiree tax rates significantly affected economic growth. But to their credit they behaved like real scientists, did yeoman’s work in discerning the robustness of their result (very), and then set forth trying to explain it—and not explain it away.

This is fertile ground for much more research.

—Ike Brannon, Cato Institute