Since the 19th century, economists have been speculating about the existence of a “Giffen good”—a good with an upward-sloping demand curve, meaning that demand for the good increases as its price increases (all else equal). At first blush, such a good seems impossible.

A normal demand curve slopes downward for two reasons: First and most obvious, as the price of the good increases, consumers will substitute other products in its place—what is known as the “substitution effect.” That is, if the price of a hamburger goes up, consumers will shift to eating more pizza and chicken sandwiches.

A second reason is that if the price of a commonly purchased good increases, consumers are left with less real income. Thus, consumers simply won’t be able to buy the same amount of goods and services—including the costlier good—that they did before. This latter phenomenon is called the “income effect,” because consumers’ purchasing behavior changes depending on their real income.

Yet, some goods clearly do have a negative income elasticity, meaning that when a person’s income increases, he buys less of the good. In essence, in these cases the income effect works against the substitution effect: suddenly wealthier consumers purchase less Ramen noodles and fast food even though they can now afford to buy more of those goods. Instead, those consumers buy more steak and more meals at upscale restaurants. Because of this characteristic, economists refer to goods like Ramen and fast food as “inferior goods,” and steak and upscale restaurants as “superior goods.” Notice that if a person’s real income were to fall, Ramen and fast food consumption should increase.

Given the existence of inferior goods, it’s possible to imagine a Giffen good. A good could be so inferior that, if the consumer’s income were to fall at a time when the inferior good’s price is rising, the income effect on the inferior good could fully overcome its substitution effect. Voilà: consumption of the inferior good increases despite its rising price, and the good is a Giffen good.

Despite the theory, it’s difficult to identify a real-world case of a Giffen good. The most plausible candidate in history seems to be potatoes in Ireland during the 1845–1849 Great Famine, when consumption seemed to not lose pace despite soaring prices. But even in that extreme case, scholars are uncertain if potatoes actually passed the positive-slope threshold of a Giffen good.

However, some new evidence suggests that cigarettes may approach this dubious distinction for an important group of society: low-income pregnant women. It’s a possibility that suggests the government should hesitate before it considers future “sin” taxes.

Substituting away from health / The data come from a new paper on cigarette taxes published in the journal Pediatrics. The paper’s top-line finding is the estimate that a recent $1 per pack federal cigarette tax increase reduced smoking in expectant mothers, which in turn reduced infant mortality by 1 per 1,000 live births. Despite that effect, the authors did not recommend further tax increases, worrying that some expectant mothers may simply be unable to stop smoking. If their taxes were increased further, these women may be forced to reduce their consumption of healthy goods and services to be able to afford (and perhaps even augment) their smoking habit. As a result, further reductions in infant mortality from higher taxes may not be forthcoming—in fact, some of the gains from the last tax increase could potentially be undone.

The authors’ show of discretion when it comes to their study’s implications is rare and welcome. I attribute their wisdom to their being doctors and not policymakers, and thus they are familiar with unwelcome side effects. When it comes to tax policy as well as regulatory policy, the notion of diminishing returns often gets ignored by government.

For instance, a decade ago, government considered lowering the maximum allowable concentration of arsenic in drinking water from 50 parts per billion to 30. (See “Special Report: The Arsenic Controversy,” Fall 2001.) Evidence that the proposed standard would have yielded significant health benefits was thin and cost estimates for compliance with the proposed standard were enormous. That led to the worry that the proposed standard would, on net, harm public health because money devoted to arsenic reduction would be reallocated from other uses—including other initiatives to improve public health and safety.

Likewise, higher cigarette taxes can come with a whole host of unwelcome outcomes. Besides a potentially deleterious effect on pregnant smokers and their children, we might also expect a greater demand for black-market untaxed cigarettes—something that inevitably occurs with each cigarette tax increase—or smokers substituting toward other substances that can be just as harmful to one’s health as cigarettes, if not more so. For instance, while I welcome the slowly spreading legalization of marijuana, there’s no real health difference between smoking a tobacco cigarette and a marijuana cigarette. The tax on both should be, apropos Pigou, high enough to correct any external costs of consumption that are borne by society, but no higher.

Giving up cigarettes is easier said than done, of course. Nicotine is an addictive substance and its effects on the nervous system differ greatly from person to person. Some people find kicking the habit to be extremely difficult, no matter how strong their resolve to avoid tobacco use, while others find it easier to quit. What’s more, some low-income people may not have knowledge or ready access to tobacco substitutes or may not appreciate the potential costs of continuing to smoke during a pregnancy.

In other words, the decision to smoke during a pregnancy can be complicated, and a prohibitive tax alone is not going to eliminate it. In a similar vein, a recent National Bureau of Economic Research paper found a virtual nonresponse to higher cigarette taxes in smoking among teenagers following the most recent tax increases at the state level. (See “In Review: Working Papers,” Winter 2015–2016.)

Policymakers tend to reflexively treat a tax on goods with high social costs as an easy and uncontroversial method to improve societal outcomes, but the world is more complicated than that. At some point the gains from additional cigarette tax increases diminish and the tax becomes little more than a way for government to extract money from the low-income Americans who now constitute the majority of smokers in the United States. Cigarette taxes may be approaching that stage.

Readings

  • “Cigarette Tax and Infant Mortality,” by Stephen W. Patrick, Kenneth E. Warner, Elisabeth Pordes, and Matthew M. Davis. Pediatrics, Vol. 137, No. 1 (January 2016).
  • “Cigarette Taxes and Youth Smoking: Updated Estimates using YRBS Data,” by Benjamin Hanlin, Joseph Sabla, and Joseph Rees. National Bureau of Economic Research Working Paper #21331, June 2015.