The New York City Ban on Trans Fats

New York City's Board of Health has decided to ban trans fats in food sold in restaurants (also in food sold by catering and meal services), the ban to become fully effective in mid-2008. The ban raises a fundamental issue of economic policy.

Trans fats are largely synthetic fats widely used in fried foods and baked goods. There is substantial medical evidence that they are significant contributors to heart disease (perhaps increasing the incidence of heart disease by as much as 6 percent) because they both raise the cholesterol that is bad for you (LDL) and lower the cholesterol that helps to protect your arteries against the effects of the bad cholesterol (HDL). About half of New York City's 20,000 restaurants use trans fats in their cooking; and roughly a third of the caloric intake of New Yorkers comes from restaurant meals.

A strict Chicago School economic analysis of the ban would deem it inefficient. The restaurant industry in New York is highly competitive, and so if consumers are willing to pay a higher price for meals that do not contain trans fats, the industry will oblige them; to force them to shell out more money, rather than leaving it to their decision, is thus paternalistic, indeed gratuitous. Restaurants catering to health-conscious eaters will advertise that they do not use any trans fats in their meal preparations, or will state on the menu the amount of trans fats in each item. Other restaurants will cater to diners who prefer a cheaper meal to a heathier one. The ban thus forces people who want to eat in restaurants to pay higher prices even if they would prefer to pay less and take the risk of an increased likelihood of heart disease. Some of these would be people who eat in restaurants rarely, and avoid trans fats when they cook at home, so that the health risk to them of a restaurant meal containing trans fats is small. Others would be people who disbelieve the medical opinion--and such opinion often is wrong--or think that trans fats improve the taste of food or that the ban is the result of political pressure from producers of substitutes for trans fats, such as corn oil, or from the restaurants that have voluntarily abandoned the use of trans fats and don't want to be put at a competitive disadvantage by restaurants that have lower costs because they do use trans fats. Moreover, the enforcement of the ban will increase the costs of New York City government, resulting in higher taxes on an already heavily taxed population. Since half the restaurants in New York City continue to use trans fats, this shows that a majority of consumers would not support the ban.

What is missing in this analysis is a cost that, ironically, a great Chicago economist, George Stigler, did more than any other economist to make a part of mainstream economic analysis: the cost of information. It might seem, however, that the cost of informing consumers about trans fats would be trivial--a restaurant would tell its customers whether or not it used trans fats, if that is what they're interested in, and if it lied it would invite class action suits for fraud. But there is a crucial difference between the cost of disseminating information and the cost of absorbing it. If gasoline stations in the same neighborhood charge slightly different prices for the same grade of gasoline, the reason may be that the price difference is smaller than the time (and gasoline!) cost to the consumer of driving to the different stations to see which has the lowest price. But if the consumer did bother to conduct that search, he would have no difficulty in understanding the information that he obtained. It is different with trans fats. Many people have never heard of them; many who have don't know that they are (very probably) harmful to health; and, above all, almost no one outside the medical and nutrition communities knows how harmful trans fats are, and in what quantity. That is, they do not know what a dangerous level of trans fats is, what their own consumption of trans fats is relative to that level, and how much their restaurant-going increases the total amount of trans fats that they consume. They have, in short, no idea of the benefit of avoiding trans fats in restaurants. And except for a few hypochondriacs and people who already have heart disease, no one wants his restaurant experience poisoned by having to read a menu that lists beside each item the number of grams of trans fats it contains and indicates (perhaps with a skull and crossbones) the danger created by consuming the item. Actually the danger would be impossible to explain to diners, because it would depend on the diner's average daily consumption of trans fats, which neither the diner nor the restaurant knows.

In such a situation, even those of us who distrust government regulation of the economy should be open to the possibility that the ban on trans fats would produce a net improvement in the welfare of New Yorkers by satisfying a preference that most of them would have if the cost of absorbing information about the good in question were not prohibitive.

A very crude cost-benefit analysis suggests that this possibility is real. Proponents of the ban estimate that it will reduce the annual number of heart attack deaths in New York City by 500. That can be taken as an upper-bound estimate. It seems high to me, as the total annual number of deaths from heart disease in New York City is only 25,000, and it seems unlikely that removing trans fats from restaurant meals alone would cause a 2 percent drop in the heart disease death rate. If that 500 figure holds up, then if one uses the consensus economic estimate of the value of an American life (an estimate based on behavior toward risk, behavior that reveals the cost that the average American is willing to pay to reduce the risk of death), which is $7 million, a saving of 500 lives confers a benefit of $3.5 billion. (This figure is too high, but I will adjust it later.) On the cost side, although the restaurant industry is up in arms about the ban, and although the ban's proponents cannot be correct that the industry would incur no cost at all to substitute other fats for trans fats--for if there were no cost, the substitution would have been made years ago, when trans fats began to be implicated in heart disease--I have not seen evidence that the cost would be great. Remember that half the restaurants in New York City have already phased out trans fats, without anyone noticing a big jump in restaurant prices. And the manufacturing cost of the substitutes for trans fats does not appear to be higher--the only advantage of trans fats is that they increase the shelf life of foods somewhat. This is important to restaurants, by enabling them to economize on spoilage costs, but surely not critical.

The New York City restaurant industry has annual sales of $9.5 billion. I do not know what percentage of those sales is accounted for by the restaurants that have already phased out trans fats, so let me assume, conservatively, that the restaurants that have not done so account for $6 billion of the $9.5 billion. Suppose the ban would increase their costs by 1 percent--which seems too high, however, since the major costs of a restaurant are wages, which would be unaffected, and the cost of food, which would be affected only slightly (the shorter the shelf life, the more food must be bought relative to the amount that can be sold). Apparently the substitutes for trans fats do not affect the taste of food.

One percent of $6 billion is $60 million. My $3.5 billion benefit figure is obviously much greater than my $60 million cost figure, and probably it is too great. Many of the 500 deaths may be of people who have advanced heart disease and thus a truncated life expectancy and impaired value of life, quite apart from trans fats. Most of the deaths are of elderly people (only about 12 percent of deaths from heart disease in New York City are of people below the age of 65), whose value of life may be below average, though most elderly people cling pretty tenaciously to life, consistent with studies that find that elderly people are on average actually happier than young people. I suspect too that the figure of 500 deaths due to trans fats in restaurant food is too high. But suppose I slash it to 100, and assume that the average value of life in this group is only $1 million; this still yields a benefit figure, $100 million, that comfortably exceeds the cost figure, comfortably enough to cover the cost of enforcing the ban. Moreover, the benefit figure excludes the benefit to people who have heart disease but do not die of it (or have not yet died of it). Heart disease causes suffering even when it does not kill the sufferer.

I have also excluded from the benefit figure any external benefit, that is, a benefit to people who do not have heart disease (or perhaps never eat in restaurants), but subsidize the medical expenses of those who do, through Medicare, Medicaid, and risk pooling by private insurance companies. I exclude it because I'm not sure it's a net external benefit. Even a total elimination of heart disease might not significantly reduce aggregate expenditures on health care, because it would result in an increase in illness and death caused by other diseases, such as cancer. (Diseases in effect compete with each other; if a person is saved from one disease, this increases the "market" for another disease.) It would also increase the average age of the population, which might result in greater transfer payments and hence heavier taxes.

My cost-benefit analysis is, necessarily, highly tentative. However, it inclines me to a sympathetic view of the trans-fats ban. I anticipate strong opposition from libertarians.

Professor Becker is traveling, and as a result will not be able to post his comment on the trans-fats issue until mid-week.

My piece provoked as I expected a number of interesting comments. Some take a strong libertarian line, which can be traced back to John Stuart Mill: government has no right to regulate private activity that does not have adverse effects on nonconsenting third parties. But I consider myself a Millian, and if I am correct that it is very difficult for people to absorb information about the dangers of trans fats, and if I am further correct that the cost of ill health from heart disease will be shifted in part through programs like Medicaid and Medicare to the as it were nonconsenting taxpayer, then the ban is Millian. I appreciate the concern that adoption of a sound regulation of restaurant meals will encourage further, less justifiable, interferences with consumer autonomy. But one can never expect government to get things just right, given the play of politics. There are always going to be silly regulations, but that is not a compelling argument for having no regulations at all. The commenters who denounce the "nanny state" do not indicate what if any regulations they approve of. Do they think there should be no inspections of restaurants by health inspectors? No regulation at all of food or drug safety by the Food and Drug Administration?

Some commenters think that people should be encouraged to study the dangers of trans fats and make their own judgments about what to eat. But people have limited time to do research on such matters. It makes sense to delegate the research to a central authority, so that instead of 300 million people trying to learn about trans fats and every other lurking menace, a handful of experts conducts the research and when it is reasonably obvious how we would react if we were informed of its results, implement the proper response. Surely our capacity to absorb information is quite limited and we must rely on the research of others for most of what we know and the knowledge of others for our protection.

Some of the comments reflect a (natural) misunderstanding of the concept of "value of life," pointing out correctly that people do not sell their lives for the calculated value. All value of life means is this: if the average person would demand $7,000 to assume a.001 (one in a thousand) risk of immediate death, there would be a net increase in social welfare if the risk could be averted at a cost of less than $7,000. Suppose 10,000 people are exposed to the risk. Then the total cost that the risk imposes is $70 million ($7,000 x 10,000), and net social welfare will be increased by a measure that eliminates the risk at a cost of less than $70 million. Another way to put this, with identical implications, is that a 1 in 1,000 risk of death will result in 10 deaths in a population of 10,000, and the $70 million loss figure amounts to valuing each life at $7 million.

Some comments take issue with the details of the cost-benefit analysis. That is fair, but notice how I scaled down the benefit figure radically in order to allow a generous margin of error, and how I excluded a major benefit--eliminating the suffering, as distinct from death from, heart disease. I doubt that any plausible adjustments could reverse my conclusion that the benefits of the ban exceed the costs.

I do think it worth emphasizing that trans fats seem exceptionally dangerous--almost in the category of poisons. The article in the New England Journal of Medicine that Professor Becker cites estimates a 6 to 19 percent reduction of what the authors call heart disease "events" from eliiminating trans fats, and the research backing it seems solid. Incidentally, I did not see in the article anything about trans fats doing more for the taste of foods than substitute oils--in fact the article discusses a Danish study that finds no quality difference.

I agree with Becker that many young people who are clogging their arteries by eating restaurant meals rich in trans fats will be saved by better cholesterol drugs that we can expect in the future. However, those drugs will doubtless be paid for in large measure by taxpayers through the Medicare and Medicaid programs. This means that the cost of trans fats will be shifted, in part at least, from those who consume them to those who do not--a classic externality, which justifies public intervention (depending on its cost and efficacy) even to Millian liberals such as myself.