Smoking

• “The Effects of E‑Cigarette Taxes on E‑Cigarette Prices and Tobacco Product Sales: Evidence from Retail Panel Data,” by Chad D. Cotti, Charles J. Courtemanche, Johanna Catherine Maclean, et al. January 2020. NBER #26724.

• “Intended and Unintended Effects of Banning Menthol Cigarettes,” by Christopher Carpenter and Hai V. Nguyen. February 2020. NBER #26811.

In late October 2019, the U.S. House Ways and Means Committee approved—with bipartisan support—an e‑cigarette tax on nicotine content that would be proportional to the federal tax on cigarettes. The current House bill specifies a tax rate of $0.028 per milligram of nicotine. How much would the tax increase e‑cigarette prices and how would e‑cigarette and traditional cigarette consumption change as a result?

Using retail scanner data from the years 2011–2017 and existing state taxation rates, the authors of the first paper calculate that pass-through of taxes on e‑cigarettes to the retail price is more than the tax (160% of the tax), which the authors attribute to the highly concentrated, less-than-competitive structure of the industry. The authors calculate that the proposed House tax increase would be $2.54 per milliliter of vaping liquid. One pod—good for about 200 puffs—from e‑cigarette maker Juul contains 0.7 ml of nicotine and is currently priced at $4, so the price increase would be at least 44% with a conservative 100% pass-through rate. The authors also find that the demand for e‑cigarettes is elastic, with an estimated price elasticity of –1.5 (the quantity of e‑cigarettes purchased is reduced by 1.5% for every 1% increase in price). Thus, a 44% increase in price would result in a 66% decrease in Juul pod purchases, a result consistent with the goal of those who support the tax.

Decreased e‑cigarette sales are not the only consumer response, however. The authors also determine that traditional cigarette sales increase following an increase in e‑cigarette prices, with a cross-price elasticity of demand of 0.9. The authors calculate that if the House bill were to become law, the purchase of traditional cigarettes would increase by 6.2 extra packs for every one standard e‑cigarette pod no longer purchased, a result probably not consistent with the goals of those who support the tax.

Turning to the second paper, the medical literature finds that menthol reduces the harshness and irritation of smoking and changes the structure of nicotine receptors. The result is that menthol makes traditional smoking easier to start and harder to quit. For that reason, anti-smoking groups have long advocated banning menthol flavor in traditional cigarettes. In late February, the U.S. House of Representatives approved a bill banning the use of menthol in both traditional cigarettes and e‑cigarettes.

What effects would such a ban have? The paper discusses evidence from Canada. Between May 2015 and July 2017, seven Canadian provinces enacted bans on menthol (Nova Scotia and Alberta in 2015, Quebec and New Brunswick in 2016, and Ontario, Prince Edward Island, and Newfoundland and Labrador in 2017) while three provinces (Saskatchewan, British Columbia, and Manitoba) did not.

Provincial menthol bans eliminated legal traditional menthol cigarette sales. But young people substituted non-menthol cigarette smoking. Adults avoided the ban by evading it: they bought menthol cigarettes on Indian reservations unaffected by the ban. The net result was that the provincial menthol bans had no significant effects on population rates of cigarette smoking or quit behaviors for either youths or adults. —Peter Van Doren

Soda Taxes

• “The Effectiveness of Sin Food Taxes: Evidence from Mexico,” by Arturo Aguilar, Emilio Gutierrez, and Enrique Seira. December 2019. SSRN #3510243.

Mexico’s tax on non-diet soda has been hailed for reducing consumption with possible long-term health benefits. In a previous Working Papers column (Winter 2017–2018) I discussed a paper that argued that the reduced-soda-consumption effect was likely overstated because of substitution from expensive soda brands to cheaper store brands.

This paper utilizes Mexican retail scanner data containing weekly purchases of 47,973 barcodes by 8,130 households to examine the effect of the soda tax as well as a companion 8% sales tax on high-caloric-density food (defined as containing more than 275 kilocalories per 100 grams). The taxes are relatively large (about three times that of the average state-level soda tax in the United States) and have a fairly broad base: they apply to 39.4% of food products and 46.3% of beverage products in the data. Taxed products account for about 23% of expenditures, 33% of total calories from packaged goods, and 39% of total food expenditures. But certain staples such as oil, milk, and bread are exempt from the tax.

Prices of taxed drinks increased by 9.7%, and caloric consumption from taxed drinks decreased by 2.7%. Prices of taxed foods increased 6%, and caloric consumption from taxed foods decreased by 3%. So, the tax did slightly reduce the consumption of high-calorie foods.

But there was substantial substitution from taxed to untaxed food. The result was that total calories in all food purchases (taxed and untaxed) were unchanged. Increased calorie consumption of non-taxed drinks and food offset the decrease in the taxed items. —P.V.D.

Banking

• “Banking without Deposits: Evidence from Shadow Bank Call Reports,” by Erica Jiang, Gregor Matvos, Tomasz Piskorski, and Amit Seru. March 2020. NBER # 26903.

Financial institutions are more highly leveraged than other firms. That is, their capital structure has more debt and less equity than non-financial firms. Why?

Some scholars argue that finance is special, i.e., financial intermediation requires such a capital structure. Others say the high leverage is the result of governmentally provided deposit insurance that socializes the losses from leverage and privatizes the gains. If the former view is correct, then governments reduce efficiency by imposing minimum equity standards on banks, while if the latter is true, then capital requirements are simply an attempt to control the adverse effects of subsidized deposit insurance. (See Working Papers, Winter 2010–2011).

This paper compares the capital structure of traditional regulated banks with unregulated “shadow banks.” Traditional banks accept deposits that are federally insured, issue loans, and are subject to safety and soundness banking regulations and examinations. Shadow banks (Quicken Loans, for example) do not accept deposits; they raise money from investors in the capital markets. And they are not protected by deposit insurance or subject to safety and soundness regulation.

The average equity-to-asset ratio for traditional banks is approximately 11%. Shadow banks’ average equity-to-asset ratio is more than twice as high at 25% and resembles that of pre-deposit-insurance banks in the United States and Germany. But the average for nonfinancial firms is almost 50%.

Short-term debt (i.e., retail depositors) comprises 85% of bank debt funding. Shadow banks’ debt is almost exclusively (98%) short-term (rather than longer-term corporate bonds), but is provided by a few large banks.

These stylized facts suggest that short-term debt is an essential characteristic of financial institutions and not the result of regulation or deposit insurance. Even unregulated financial intermediaries are funded with substantially more debt than non-financial companies.

But other facts suggest the effects of deposit insurance: shadow bank leverage increases substantially with size while bank leverage hardly changes with size. Small banks are highly leveraged relative to small shadow banks and are the main beneficiaries of deposit insurance. (See “Too Small to Succeed,” Winter 2019–2020).

The authors write, “The most parsimonious model that explains why bank capitalization is about half of shadow banks and, at the same time, why capitalization across banks is so homogenous, is the following: banks borrow as much as they can, subject to capital requirements, because they have access to subsidized debt funding.” —P.V.D.

Health Insurance

• “Long-Term Health Insurance: Theory Meets Evidence,” by Juan Pablo Atal, Hanming Fang, Martin Karlsson, and Nicolas R. Ziebarth. March 2020. NBER #26870.

Support for the Affordable Care Act (ACA) relies on the conventional wisdom that private individual health insurance policies don’t and can’t “work” given the skewness and concentration of health care expenditures. The lowest 50% of the population in terms of health care expenditures had average annual spending of $276 per person and accounted for only 2.8% of aggregate health care expenditures in 2016. The top 10% of the population, in contrast, had average annual spending of $33,053 per person and accounted for 66% of aggregate spending.

Many conclude from these facts that in the individual market, the healthy will obtain coverage at low rates while the “sick” will be unable to obtain coverage or only at prohibitively high rates. Thus, they argue, individual insurance coverage must be mandated to ensure “pooling” of healthy and sick so that the price is simply the population average, with limited variability by age (community rating).

Bradley Herring and Mark Pauly demonstrated in a 2006 Journal of Health Economics paper that markets could provide guaranteed renewable individual health insurance contracts because high-cost conditions are rare and do not persist. However, though such contracts are possible, they are hardly ever purchased because of the dominance of tax-subsidized employer-provided coverage in the United States.

The authors of this paper take the discussion further by examining the robust system of private guaranteed renewable contracts that exists in Germany, where 10% of the population has such coverage. Compared to the ACA system in the United States, the German individual market is less regulated. Applicants freely choose their level of coverage in terms of benefits and cost-sharing amounts within some very loose limits. There are thousands of different health plans among the 8.8 million policyholders. And the market has been stable and providing insurance for millions of people for decades.

The authors write: “An unquantified advantage of the German long-term contract is its simple design, combined with low information requirements. Moreover, the market has been stable and providing insurance for millions of people for decades. We believe that our findings, coupled with these facts, strengthen the case of the German design as an appealing policy option.” —P.V.D.

IP Boxes

• “IP Boxes and the Activities of Foreign-Owned U.S. Corporations,” by Tim Dowd, Paul Landefeld, and Anne Moore. Joint Committee on Taxation working paper, 2020.

An “intellectual property box” (or IP box) is a preferential tax rate regime for income accruing to intellectual property such as royalties from patents. The rationale for an IP box is simple: if mobile capital can escape a country’s tax jurisdiction easily while fixed capital cannot, then it may make sense for the country to tax the return to mobile capital less heavily than fixed capital.

IP boxes have been around in some form for at least a decade, and there are over 20 countries with IP boxes (which these days are more often referred to as “innovation boxes”) across the world. Many of the EU countries, as well as China, employ some form of one. They vary widely in terms of qualifying income and rates.

The Joint Committee on Taxation staffers who wrote this paper want to know how foreign IP boxes affect the real economic activity of multinational firms. Knowing that would help U.S. policymakers better understand the efficacy of our own tax policy. Because countries adopted IP boxes in different years, the authors used a difference-in-differences approach to estimate the effect the tax policies had on sales, wages, and investment of multinational firms in the United States.

There are two ways a U.S.-based multinational could react to the creation of an IP box in a different country. First, the corporation could transfer activity currently done in the United States to the IP-box country. For instance, some pharmaceutical companies have indicated that Switzerland’s generous IP box is why they (re)located research and development there. This is more than a simple “P.O. Box” change; it is a boost in investment in the firms’ Switzerland operations at the expense of operations elsewhere.

But that substitution effect may bring with it a scale effect: a firm may react to the reduction in its worldwide tax bill from the IP box by increasing economic activity both in the IP-box country and in the rest of the world. In other words, while an IP box encourages more capital investment in the country that implements one, it may also boost economic activity at company locations around the world. The result is that countries that don’t implement an IP box may lose tax revenue but not much overall economic activity.

The paper notes that while it may seem intuitive that the substitution effect would invariably reign supreme with an IP box, Duke University economist Juan Carlos Suarez Serrato found that the 2006 ending of generous tax breaks for U.S. manufacturers operating in Puerto Rico served to reduce the overall activity of U.S.-based corporations. The IP box may have the mirror-opposite effect.

A recent working paper—“The Effect of Innovation Box Regimes on Income Shifting and Real Activity,” by Shannon Chen, Lisa De Simone, Michelle Hanlon, and Rebecca Lester (November 2019)—finds considerable evidence that IP boxes do what they are intended to do: retain and attract mobile capital. It also notes that it is unclear whether attracting this capital results in more real economic activity, as that would depend on whether firms find it advantageous to co-locate production activities with their intellectual property. In research that I published with Hanlon, we found evidence that for the pharmaceutical sector there does appear to be economic incentives to co-locate production and certain intellectual property.

The authors of the Joint Committee paper obtained data on Foreign-Owned Domestic Corporations (FODCs) by examining Internal Revenue Service statistics of income reported on Form 5472 (filed by businesses that are at least one-quarter foreign-owned) and then linked those data to IRS Form 1120, the U.S. corporation income tax form. FODC assets represent 19% of corporate assets and receipts and 14% of taxes paid in 2012.

The authors find that foreign IP-box implementation does appear to increase overall economic activity, which they presume is due to the FODCs reducing their effective U.S. tax rate via IP boxes. In effect, they manage to increase gross receipts in the United States but not net income because their deductions increase. The authors conclude that the firms leverage up their U.S. operations with debt from the parent corporation, which boosts their (deductible) interest payments, thereby transferring income elsewhere.

In the 2017 tax reform, the United States took steps both to make it more attractive to keep capital in the country—by sharply reducing the corporate income tax—as well as make it less remunerative to shift profits abroad. Whether this combination serves to maintain both the stock of U.S. corporate investment and the corporate tax base is unclear, but the rapid changes in both corporate tax rates and regimes in the last few years indicate how difficult it is to impose taxes on corporation income.

—Ike Brannon