Return of the Senior Death Discount?: Heinzerling Takes On Mannix

Is the senior death discount back? It may be, if a recent speech by an Environmental Protection Agency official is any indication.

In a keynote address for an EPA workshop on economic analysis last month, Brian Mannix (associate administrator of the agency's Office of Policy, Economics, and Innovation) called for the return of a controversial way to measure the number of human lives saved by environmental regulations.

Mannix called for EPA to abandon the practice of assessing the benefits of a proposed regulation by plugging in the number of lives saved, and to replace it with a method of counting the Value of Statistical Life Years (VSLY) saved. Because VSLY focuses on life years saved rather than lives saved, benefits from saving the lives of the elderly count for less than benefits from saving the lives of the young, because the elderly have fewer life years remaining.

Accordingly, the practice is better known as the senior death discount.

EPA proposed but then withdrew the practice of using VSLY over VSL three years ago, but the threat reemerged in Mannix’s speech.

Point - Counterpoint

The following is the text of Mannix's speech, with a point-by-point rebuttal from Georgetown law professor Lisa Heinzerling:

Point - Mannix

What I want to do today is step back and take a look at the metrics we use to describe the benefits of mortality reductions that we attribute to environmental regulations. In particular, I want to raise questions about the statistical robustness of the "lives saved" metric that is now commonplace.

Counterpoint - Heinzerling

First off, note the use of euphemistic jargon. "Benefits of mortality reductions" are, in plain English, human lives saved -- or, even more pointedly, they are the people who will be killed by pollution unless EPA acts. When you hear "benefits of mortality reductions," think: he's talking about people who are being killed by pollution. As you read the rest of the speech, try to spot the euphemistic jargon, and try to figure out what it means.

Second, see how the relevant issues are narrowed from the start: the issue is the "statistical robustness" of the "'lives saved' metric" (more euphemistic jargon), not the moral appropriateness of cost-benefit analysis.

I should say that, years ago, I was an advocate for VSL (Value of a Statistical Life) analysis, and encouraged EPA to focus its efforts on measuring lives saved. Now that I am back at EPA, I am surprised at how much progress has been made in incorporating VSL into agency analyses and decisions. I am surprised, too, to find that I am not very comfortable with where that progress has left us. And I am most surprised to find that the most serious difficulty, in my mind, turns out not to be with the V, but with the SL. That is, the economic valuation of mortality benefits is a tractable problem analytically and politically. But figuring out the right underlying metric for mortality benefits is much more problematical. For years, economists have hounded EPA to come up with a monetary valuation for human life. Now that EPA has done this, they don't like it anymore. Why might this be? Note that one of the biggest decisions facing EPA today is setting a new air quality standard for fine particulate matter. Note, too, that regulation of fine particulate matter is one regulation that overwhelmingly, in analysis after analysis, has passed a cost-benefit test. In fact, in OMB's annual reviews of the costs and benefits of federal regulation, regulation of fine particulate matter puts environmental regulation "in the black" in terms of net benefits; it produces so many benefits that a lot of programs can produce more costs than benefits and still be "paid for," if you will, by the regulations for fine particulate matter. Finally, note that many of the people saved by stricter regulation of fine particulate matter are elderly.

Now you can understand why it is no longer enough to translate human life into dollars. That analysis justifies air pollution regulation that is much too stringent for the current agency's tastes. So the "V" -- the value of life -- is all right, but the focus on life -- "SL" -- must go.

Mannix also claims that setting a dollar value for life is "politically tractable." Is that so? Is the public really aware that EPA has translated human life into dollar terms, that the going value for life is between 5 and 6 million dollars, and that it's gone down in the last few years? If the issue is so "politically tractable," why is the actual value EPA uses for human life always buried deep in dense tables, in government documents almost no one reads? Why doesn't the government put out a press release: "the value of life this year: 5.5 million dollars; stay tuned for next year's (new, improved, lower) value"?

I'll illustrate this with an extreme example. Suppose, on Monday, a hospital in a small town publishes a press release, announcing that, over the busy weekend, it had managed to save a dozen lives. The local TV station sends a camera crew, and asks if it can interview a few of the lucky survivors. The ER nurse tells them: "I'm sorry, that won't be possible. He died." "What do you mean, who died?" the reporter asks. "The man who was having the heart attacks," the nurse replies. "We managed to save him 12 times ... in 13 attempts." No one really talks this way about saving lives. Mannix has created an analytical problem only by acting like a visitor from another planet who doesn't understand how humans talk about life and death.
The point of the story is that, while we can easily count lives or deaths, we cannot easily count "lives saved." It is not well defined, and it is inherently unbounded. The airbag may save your life in the event that your brakes fail. But how many times did the brakes save your life when they didn't fail? The number of times my life was saved during my commute this morning is already beyond my ability to reckon. It is hard to figure out what Mannix is trying to do here. Yes, our lives are protected in dozens or hundreds of ways every day. Does this mean we can't say something special has happened if a police officer rescues us from a kidnapper? Or a doctor gives us life-saving medicine? Here, too, it is as if Mannix has just dropped in from outer space.
In some narrow contexts we might be able to come up with a workable definition of a life saved. As a lifeguard Ronald Reagan would put a notch in a log every time he saved a life, and I don't doubt that it was accurate and meaningful. If he had kept a notched log during his Presidency, however, I can't imagine how we would come up with an accurate count, or interpret it if we had one. I don't believe it is possible to come up with a definition of lives saved that is robust, that can be applied to a wide variety of situations, and that can be aggregated in a statistically meaningful way. Here is why Reagan was President, and Mannix is a policy analyst within a large government bureaucracy: Reagan understood that something special had happened when he saved someone from drowning.

It is not clear why Mannix is so skeptical about our ability to count how many lives are saved through government action. If we really can't do this, then cost-benefit analysis -- the decision-making framework addressed, and advocated, throughout this speech -- cannot be done for life-saving regulation, since cost-benefit analysis depends in the first instance on counting up the benefits regulation will produce. If we can't do this, we can't do cost-benefit analysis.

The underlying difficulty is that "lives saved" lacks a time dimension. We know that all lives are temporary, and, while the valuation problem can be quite complex, we are generally in agreement that longer life is better than a shorter one. If we don't capture this time dimension, we are unlikely to come up with a metric for mortality that is versatile and behaves well in statistical usage. But wait! It turns out Mannix doesn't really believe we can't count how many lives are saved by regulation. It turns out he just doesn't think saving lives is what matters. Lives are "temporary," he reminds us; as other fans of cost-benefit analysis have often said, lives are never saved, but only prolonged. (Here the rhetoric turns dark rather than euphemistic; it doesn't sound so nice to "prolong" a life.) We have to figure out a way to account for this fact in our analysis; otherwise it won't be "well behaved." (One can only guess at the standards for "good behavior" on the part of "metrics for mortality.")

Notice the giant non sequitur. Mannix has gone from claiming that we can't come close to figuring out how many people are saved by regulation (an empirical claim) to saying that what really matters is how long we live (a normative claim). There is no connection between Mannix's confusion about how many times his car brakes have saved his life and his "solution" of taking into account the "temporary" feature of life by changing the "metric for mortality." One is an empirical conundrum; the other is a moral judgment of the highest order.

Note, too, that if we really can't figure out how many lives are saved by environmental protection, we can't figure out how many life-years are saved, either.

Now, there is a standard statistic for measuring longevity that everyone is familiar with: the expected value of the length of life, or life expectancy. It has several advantages in communicating with the public. Everyone has a pretty good idea of what it measures. People also have a good sense of what the units mean. They may have a great deal of difficulty picturing what a 10 -6 risk of death is, but they know how long a minute is and how long ten years is. That covers more than six orders of magnitude. This also solves the problem of divisibility: some find it difficult to think about a fraction of a "life saved" or about the same life being saved multiple times, but they have no trouble dividing time into units of arbitrary size. The public is also likely to have less difficulty in attaching a monetary value to changes in life expectancy-- even those who cannot imagine attaching a finite value to a life saved. There is a lot here, none of it sensible. But let us begin.

Mannix is right that everyone understands the concept of life expectancy. But everyone also understands the concepts of life and death. One might even be so bold as to suggest that life and death are more "tractable," "analytically," for the average person, than life expectancy is.

Mannix doesn't want to talk about life and death, though; he wants to talk about a 10-6 (one in a million) risk of death. This is what cost-benefit folks always do: they insist that we talk about risk of death, not death itself. This is because their analysis doesn't work if we talk about death. Cost-benefit analysis today asks how much money people are willing to pay for regulatory benefits, or alternatively, how much money they will accept to give up those benefits. If we tried to figure out these values for life and death, we'd end up either just measuring how much people were able to pay (because presumably most people would give up everything they have to avoid certain death), or we'd end up with no number at all (because most people won't accept certain death in exchange for money). But when pollution kills, it kills real people. Pretending that we're just talking about risk, not death, doesn't change this simple fact.

Even on its own terms, Mannix's analysis is nonsensical. He thinks people can understand life expectancy, but not risk. How, then, does he propose conducting cost-benefit analysis of changes in life expectancy? Are we to presume that the changes in life expectancy are certain to occur (since people won't be able to understand risk rather than certainty)? If so, how can he be so sure economic analysis here won't be doomed by the same dilemma noted above -- that is, if people are given a choice between living five more years and taking a big wad of cash, what if just won't sell?

And how much comfort are we to derive from the fact that people can divide time into units of "arbitrary size"?

The real advantage of life expectancy, however, is that it is a well defined and well behaved summary statistic that reflects the mortality risks across an entire population, including risks of all kind and at all ages, without discriminating against any particular subgroup. Mannix's notion of non-discrimination is disingenuous. Although tersely stated here, the idea is, I think, that valuing people according to their life expectancy (with the elderly faring worse than the young) is a way of avoiding discrimination because it's a way of treating everyone's "life-years" the same. If I have 40 years left to live, and you have 4, then the only way to treat us equally is to value me more; that gives equal respect to each of the life-years each of us has left. No one I know of thinks this way other than the political appointees who have dreamed this up as a way of rescuing their beloved methodology. Officials at OMB started to talk this way, too, when the original "senior death discount" came under fire.
Let's suppose we are evaluating a range of policy options, all of which have small marginal effects on mortality risks. If we take as our mandate to maximize life expectancy, using limited resources, we can easily solve the problem, at least on paper. We know that the solution will give us a cost-effectiveness criterion--a fixed dollar amount per incremental year of life expectancy. And the decision rule would be to adopt those measures that met the cost-effectiveness criterion, and to avoid committing resources to those that didn't. Here are the things we have to go along with in order to accept Mannix's hypothetical scenario:
    • The mortality risks EPA addresses are "small." This surely isn't true for the big policy issue now on EPA's plate: the revision of the fine particulate matter NAAQS.

    • Our mandate is to maximize life expectancy. Only if Mannix's view of the world holds; that is, only if we think we shouldn't protect the elderly as much as we protect the young.

      Note, too, that even if we're interested in saving lives, period, no matter how old the people who will be saved are, it may be that maximizing this value alone won't be very satisfactory. EPA is also in the business of protecting ecosystems, nonhuman species, watersheds, etc. A rigid "maximize life expectancy" mandate ignores every one of these other values.

    • We have limited resources. How could anyone argue with this idea? The counterpoint is not that we have unlimited resources, but that the people who talk about "limited resources" make two huge errors.

      First, they act as though money that we refrain from spending on one program will be diverted to another program. If we set a more relaxed standard for particulate matter, for example, we'll spend the money we saved on health care programs for children. This never happens. There is not even a legal mechanism by which it could happen. The money not spent on cleaning the air stays in the pockets of the industries that pollute; the children don't get the money.

      Second, the people who talk about "limited resources" act as though there is some natural limit on the amount we can or will spend on environmental protection. There is not. The amount we are willing to spend depends on, among other things, our awareness of environmental problems and our sense that something must be done about them. Perhaps if we talked about "the pollution that kills people" rather than "benefits of mortality reductions," we would be willing to spend more for environmental protection.

Note that if we use another decision criterion in place of this one, we will get a shorter life expectancy for the same expenditure of resources. If we use a VSL rule, for example, we might "save more lives," whatever that might mean; but, on average, people will live shorter lives. In most cases I think the two criteria would lead to very similar outcomes. When they don't, however, we have to ask ourselves whether selecting a portfolio of policies that results in shorter life expectancy can really be said to be improving public health. These claims are completely dependent on the "limited resources" idea.
Similarly, if we adopt maximum life expectancy as the goal, but make adjustments to our metric for age, quality, or willingness to pay, the result will be that people live shorter lives. Better, perhaps, in some sense. But shorter. We need a translation here. What Mannix is recognizing, without saying so, is that some studies have concluded that the elderly are willing to pay more to avoid risk than younger people are. If we adjust our "metric" to account for "willingness to pay" in this context, then we'll end up back where we started: we'll protect the old as much as the young. Can't have that, surely. And just in case you thought that didn't sound so bad, please know that your life will be shorter as a result -- because, as you'll recall, the amount spent on protecting the elderly is coming right out of your pocket of limited resources.
I believe this creates a strong presumption for using life expectancy as a standard metric in evaluating regulatory decisions, using a flat VSLY as the cost-effectiveness criterion. As a first-order approximation of mortality benefits, I think this is vastly superior to the VSL approach. And I think that anyone advancing some other decision rule needs to explain how we can justify adopting policies that will lead to a shorter life expectancy. I don't rule out that such justifications may exist, but I think we should be cautious in entertaining them. Once again, Mannix's fearmongering -- we'll all live shorter lives if we protect the elderly as much as the young -- is completely dependent on his strong assumption of limited resources. If it turns out that we, the richest country in the world and in the history of the world, can afford both to clean up the air and to vaccinate our children, then we needn't give in to Mannix's hectoring.
Al McGartland has pointed out to me that there is a contradiction here. I embrace the use of willingness-to-pay data in figuring out what our cost-effectiveness criterion should be. But I shrink from looking any deeper into the data to find out how it might vary from group to group or person to person. I think this is a contradiction that I can live with. Certainly an individual-- perhaps because he is wealthy--who is willing to pay, and does pay, much more than average to reduce his own mortality risks, should be able to do so. But I am not ready to concede that that same individual is entitled to tilt public health measures in his favor simply because he would be willing to pay--but does not pay for them. When writing general rules, or spending public funds, there is an egalitarian consideration that does not apply when individuals are spending their own money. Certainly it does not seem fair for me to spend other people's money on my own health care more lavishly than they are willing to spend it on themselves. For those of you not familiar with EPA's organizational chart, Al McGartland is the Director of EPA's National Center for Environmental Economics.

Up until this point in the speech, I had assumed that Mannix's "cost-effectiveness criterion" was derived by simply figuring out how much money EPA had to spend and then identifying the regulatory approaches that would save the most life-years, given that amount of money. But now we learn that Mannix has somehow worked willingness-to-pay into his cost-effectiveness criterion. It's mysterious just how he has done this, or why. Cost-effectiveness analysis is often conceived of as a way of avoiding the moral and political complexities of cost-benefit analysis, since it doesn't require that we identify a dollar value for human life and other benefits of regulation. So it's not clear why Mannix has smuggled some form of cost-benefit analysis (through a focus on willingness-to-pay) into his cost-effectiveness analysis.

In any event, what Mannix is saying here is that he's unwilling to take willingness-to-pay to its logical conclusion, which is that the lives of the rich are worth more than the lives of the poor -- because, after all, they're willing to pay more to avoid risk.

Many people would also resist this conclusion; that is, they would think it appropriate for EPA to treat the rich and the poor alike in developing rules. They would likely explain their conclusion by reference to principles of fairness and equality. They would also probably assume that their conclusion implied that we should spend as much to protect the poor as to protect the rich.

Not Mannix. Mannix says that the reason we shouldn't take willingness-to-pay to its logical conclusion is that we'll end up spending too much to protect the poor. This would be bad… why? Because it would be paternalistic, even presumptuous; who are we to tell the poor that they should have the same quality of air that we, the rich, can afford? Especially when they're spending "other people's money."

But remember: the "other people" who would, through regulation, be required to spend "their money" to, say, clean up the air, are the people who are killing the people whose lives are under discussion.

When did telling someone to stop killing someone else become a matter of "spending the [killer's] money"? This is the kind of moral emptiness we get when we start off by talking about not killing people as "benefits of mortality reduction."

As analysts we may feel we can improve the analysis by making adjustments for age or quality, or to incorporate the latest willingness-to-pay data. But as a government official, I am reluctant to go very far down that road. In part, that is because I question whether government has any legitimate business making such adjustments, and in part it is because, if the government did get into that business, the adjustments would likely be made according to the rules of politics, not necessarily those of economic analysis. So my final argument is this: perhaps a flat VSLY is desirable for the same reason people argue that a flat tax would be desirable; it minimizes the opportunities for special pleadings and preferential treatment. Mannix's peroration requires some untangling. Although Mannix questions "adjustments" to economic analysis based on age, he himself has spent most of his speech defending such an adjustment: the use of life-years saved as the sole criterion for judging regulation. In addition, although Mannix clearly disdains making such judgments based on "the rules of politics," the criterion of life-years saved is hardly apolitical. It is certainly not a "scientific" choice.

One can only hope Mannix's favored methodology will suffer the same fate as the flat tax he compares it to.

back to Blog