Consider the proposal to regulate our nation's large food retailers. For these purposes, large food retailers include huge enterprises with national multiproduct chain stores, like Walmart, Target, Walgreens, and Costco; small food stores attached to national chain gasoline stations and regional or national chain grocery stores, like Safeway, Wegmans, Trader Joe's, and Whole Foods; plus chain restaurants including both fast food sellers, like McDonalds, Taco Bell, and Wendy's, and longer-stay restaurants, like Applebee's, Outback Steakhouse, and Olive Garden. Although these enterprises are heterogeneous in some respects, they are all large and together account for a massive share of the food we buy.
Retailer Performance Goals: Fewer Calories Sold and Reduced Sales of Added Sugar, Sodium, and Saturated Fat
This proposal would oblige our nation's large retailers of the sort described above to reduce the amount of calories, added sugar, sodium, and saturated fat that pass through their cash registers on an annual basis. The overall industry reduction proposed is an annual calorie decrease of 2 percent and a 5 percent annual decrease in the particular troublesome ingredients (added sugar, sodium, and saturated fat) for each of five years. By the end of the period, the regulated firms as a group should be selling one-quarter less added sugar, sodium, and saturated fat and 10 percent fewer calories than at the start of the program. These reductions, if achieved and maintained (or even expanded) in future years, would transform the American diet by reducing the overall number of calories, added sugars, sodium, and saturated fat consumed by Americans as a whole. To be sure, what is sold in stores or restaurants does not precisely equate with what is consumed. And so it is imaginable that lowered sales of certain ingredients could also result in less waste so as to yield no change in consumption of the regulated components. But this is quite unlikely (and if it were to happen to any significant extent, that change could be offset by requiring larger reductions by retailers).
It is vital to note that this plan does not rely on changing individual consumer behavior one person at a time. The plan can be successful at improving the American diet even if many individuals make no dietary changes under the plan. Note well that rather than targeting the obesity rate, the plan aims to reduce the overall number of calories, sugar, sodium, and saturated fat consumed by Americans in the aggregate.
Some of the Fine Print: Administrative Details
There are a number of important administrative details to consider that will be explored here. First, the emphasis proposed is on added sugar, and so, natural sugar in fruit, for example, should be exempted from the sugar-reduction portion of the regulatory regime. However, added sugar would have to be broadly defined to include, for example, high fructose corn syrup. Also, the required fat reduction should not apply to healthier fats such as those found in nuts. These are some of the many crucial details that would have to be worked out in the design of any actual plan.
The goal here is not to defend precisely which ingredients or nutrients would be the object of the regulation, although added sugar, sodium, and saturated fat are widely pointed to as the basic categories of our dietary excess (World Health Organization 2011). Were scientific experts to conclude, for example, that the plan should start with added sugar alone, its potential application to sodium and saturated fat could be put off (Lustig 2009). Nor is the goal here to defend a specific annual calorie reduction amount, although the increase in our average calorie intake has been at least 10 percent in the years coinciding with the obesity explosion (Duffey and Popkin 2011), and so that explains the initial recommendation of a 2 percent reduction each year for five years. It should suffice for now to work with plausible targets that are roughly what one would expect public health policymakers to adopt.
The legislature could adopt this proposal in broad terms and then assign the job of filling in the plan's details to an appropriate administrative agency. Assume for now that this would be done by the U.S. Congress as a national plan (although there is no reason why this regulatory reform needs to start in the United States). Although the current federal responsibility for regulating food and beverages is strewn across a large number of bodies, if we are to stick with existing agencies, experience suggests that it probably makes most sense to assign the responsibility to the Food and Drug Administration (FDA) (rather than, for example, the USDA or the Federal Trade Commission).
America's twenty largest retailers alone account for more than 60 percent of the food we purchase in traditional food stores, and Walmart sales alone approach nearly 20 percent of the food sold in stores (USDA 2013). More than 10 percent of the calories Americans consume are sold by fast food restaurants (Fryar and Ervin 2013); and Subway, McDonalds, and Yum! Brands (Taco Bell, KFC, Pizza Hut, and more) together account for a substantial share of all the fast food sales and have more than 50,000 outlets in the United States (Cosper 2013). At the outset, therefore, it might make good administrative sense to focus the plan on a limited number of regulatory targets, say a few hundred, which an agency like the FDA should be able to manage. If one were worried about leakage or escape hatches via sales by smaller retailers, those could be largely picked up by imposing the performance-based targets as well on a modest number of large wholesalers who are typically the major suppliers to these smaller firms. So, too, if need be, independent vending machine operations could be added to the regime as well. Those potential wrinkles in the plan will be put aside for now.
It is crucial to understand that individual regulated retailers would not all be treated identically. Each would be given a somewhat different reduction target based on the nature of what it sells at present. So, for example, if Whole Foods, as compared with industry averages, already sells less in the way of high calorie, saturated fat, sodium, and added sugar containing items, its reduction targets would be less than those imposed, say, on Safeway, were Safeway's typical basket of food sales today above average in calories, fat, sodium, and sugar. The same point applies to Subway and KFC, if their typical sales today were quite different in terms of the measures covered by the plan. In other words, over five years some firms would have to reduce the calories they sell by only 7 percent while others would have to reduce theirs by as much as 14 percent. The same goes for the specific required reductions in added sugar, saturated fat, and sodium.
Other adjustments would also be made in the fine tuning of the plan. Firms should not be penalized for increasing their food sales overall; nor should they benefit simply by reducing their market share and simply selling less of the same basket of items as in the past. Although the administering agency could use several strategies in dealing with this issue, one fairly easy to understand option would be to take a representative sample of $1 million worth of a firm's food sales today and then keep sampling an inflation-adjusted basket of $1 million of the firm's product sales as the plan is put into action (with care given that the sampling time frame does not allow firms to game the regime by significantly altering what they sell only at measurement times). The individual firm's annual reduction requirements, then, would apply to its representative basket, and aggregating those baskets across the entire nation would equal overall countrywide target reductions of the sort described above.
Because of the sophistication of today's bar code technology, it ought to be fairly easy for the FDA to require all items passing through a large firm's cash registers to carry in their bar codes the relevant information about the amount of calories, added sugar, sodium, and saturated fat in the products sold. Every processed food item already has a Universal Product Code, and because of data connected to each product code, firms already embed lots of information in bar codes that is captured through their cash registers. By getting their suppliers to give them data in the proper format that is needed for the proposed regulation to work, that information could be routinely added to the information gathered in the cash register records. Indeed, much of the information that the plan would require is already being generated through government-required label information on processed foods. Once the agency determined how to assure accuracy and honesty in cash register records, it would then be in a position to monitor a firm's sales against its plan targets (Kagan 1989). Given this use of high-tech oversight, generating and auditing the measurements required by this proposal should not be expensive for either the FDA or the regulated firms.
How the Regulated Firms Might Achieve Their Targets
Regulated firms would have many choices in how to meet their goals, and different enterprises might well adopt quite different strategies in seeking to meet their targets. With respect to the large supermarket chains, Walmart, for example, could in turn pressure its packaged goods suppliers to change the ingredients in their products or to change the size (or portion size) of what they package. Walmart could alter the range of products it stocks on its shelves, adding new healthier items and discontinuing less healthy ones. It could change the proportions of the products it sells, doing so by changing where in the store various goods are placed, altering how different foods are priced, shifting the ways in which different foods are advertised, and so on. Walmart might adopt disclosure regimes for its customers that turn out to make a difference. For example, Walmart might try warning against buying too many of certain items that are high in added sugar, saturated fat, and/or sodium by putting red traffic light icons next to those items on the shelves. Moreover, Walmart would have the flexibility both to come up with other innovative ways to meet their targets and to blend their strategies in ways that achieve their goals while still maximizing profits.
Fast food chain restaurants would have a variety of their own, sometimes different, strategies to call upon in order to achieve their reduction goals (and to be clear the reduction targets would be applied at the chain level—e.g. Burger King—regardless of whether the actual retail outlets were owned by the national brand or by local franchise holders). For example, McDonalds might reduce the portion size of fries and sugar-sweetened beverages they sell. They might shift more customers away from Coke to Diet Coke or water, introduce potato items that are lower in calories, fat, and salt than the fries they now sell, and/or get more customers to buy fruit instead of fries. McDonalds could also alter how it promotes its fat-filled burgers, by changing the price or location on the menu board and in its media ads.
Consequences for Firms That Do not Meet (or Exceed) Their Targets
Firms failing to meet their regulatory targets would be required to pay substantial fees (or fines). In setting these fees, either Congress, or the agency administering the plan, would consider the overall social costs of the sale of excess unhealthy foods (which are very substantial). Put simply, the goal of this strategy is not actually to collect money from firms failing to meet their targets but rather to provide the regulated firms with strong financial incentives to achieve the plan's objectives. An alternative and even stronger sanction scheme would be simply to forbid firms from selling more than their target, a strategy that could be realized by requiring firms to hold permits to sell the regulated foods and then giving the firms permits only up to their target quantities. But, then, what would the penalties be for unpermitted sales? In the end, this might come down to the same thing as the fees/fines proposed above.
Moreover, beyond the direct financial costs, noncompliant firms could earn bad reputations in the marketplace as more Americans seek healthy places to shop. Indeed, the FDA, if put in charge, should be directed to give public praise to firms that meet their targets, while perhaps at the same time to single out for shame those firms that fail to meet targets (Skeel 2001; Graham 2000).
For firms that achieve reductions beyond their targets, the plan could adopt a number of responses. A firm might be allowed to apply the excess to its next year target, or maybe a bonus payment would come from government for accomplishing even more than was required. Firms that exceed their targets perhaps could sell that excess to firms that are otherwise having difficulty meeting their targets. This latter option would make this performance-based plan very much like the “cap and trade” strategy that has been used in the environmental context to address acid rain as well as climate change from global warming. Yet, experience in that realm suggests that policymakers should at least worry about possible undesirable gaming of the program by some retailers if the selling of excess reductions were permitted (Wara 2007).
Possible Pitfalls of the Proposal
Certainly, like any policy proposal, potential pitfalls and unintended consequences, as well as creative compliance that amounts to substantial noncompliance, must be considered (Baldwin, Cave, and Lodge 2012; Farber 1999). Experience with other performance-based regulatory schemes like the federal education NCLB (2001) program has made clear that it is crucial to set the targets wisely and be able to measure compliance reliably. What we want our schools to do is to better educate our children, and NCLB was enacted, among other purposes, to hold schools accountable for that. In the details, what Congress mandated is that students achieve certain success rates on standardized tests. This, however, has resulted in “teaching to the test” that might not coincide with actual learning and arguably detracts from authentic forms of teaching and learning that provide students with a broader education (Volante and Cherubini 2007). Moreover, the NCLB testing mechanism has not proved secure as cheating has stigmatized the regulatory scheme (National Public Radio 2011). Furthermore, imposing penalties on public agencies (school districts) that fail to achieve their targets has proven difficult, especially when it becomes widely believed that, for all too many schools, the targets set are not plausibly achievable (Nation 2011).
However, there is good reason to expect that these problems would not plague the application of performance-based regulation as envisioned here. While it is true that good health is the real social objective, there is a widespread consensus that the reduced consumption of calories in general and certain sorts of calories, in particular, is itself an appropriate social goal and one that will reliably translate into public health gains. The targets proposed here are fairly clear, easy to understand, and would appear to be easily measurable. And since relatively few firms dominate the market, it should be relatively easy to monitor those firms to determine whether or not they are in compliance without much controversy. And, private business, which now profits from the social costs its products generate, can be sold to Congress as a morally acceptable and economically practical target of the penalties proposed for noncompliance. Nonetheless, if it turns out that calories per se do not really matter or that sodium and/or saturated fat consumption really is not a problem—as some argue—then obviously using performance-based regulation to change the American diet would not be the same as making it a healthier diet (Taubes 2011, 2007). But this concern applies as well to most of the alternative regulatory strategies now on the table or recently adopted.
That does not mean, however, that this proposal will come with no difficulties. Therefore, it is worth exploring the potential pitfalls unique to applying performance-based regulation to food and beverages.
First, what if the regulated firms achieve their targets by taking the reductions from food sold to higher-income and higher-educated people, leaving less well off Americans with the same diet they now have? As a practical matter, this is probably an unlikely outcome because many of easiest-to-achieve, healthier-diet improvements might well be achieved from the least educated and lowest earning families. However, to address this class-based concern, one could, for example, apply the targets, not to each firm on an overall basis, but rather by requiring reductions in subgroups of the firm's retail outlets that tend to serve different types of customers. These solutions mirror the strategies that are sometimes suggested or employed to counter the creation of “hot spots” in low-income communities when regulated entities respond to performance-based goals by reducing pollution only in higher income areas (U.S. Environmental Protection Agency 2013). NCLB adopted an analogous approach by requiring test score gains by students from each of a district's major ethnic/racial groups (NCLB 2001).
Second, were the proposal enacted, it would likely stimulate a search, for example, for nonsodium additives that could at least somewhat replace salt in enhancing taste. If such replacements had no negative health consequences, that would be fine, but what if they were just as risky as consuming high levels of sodium? Such a consequence would undermine the overall goal, and the FDA would need to be empowered to take action either to block such substitutes or to treat them as sodium for purposes of measuring whether the plan's targets were being achieved. In a similar vein, new food products created in laboratories might come onto the market to replace sugar and saturated fats, which also might be either ominous or innocuous. If dangerous in new ways, this could mean that the regulatory strategy, even if successful on its face, might be socially counterproductive. Because of these concerns, the FDA would have to monitor regulated entities and be empowered to take action designed to ensure that their solutions were not directly undermining the goals of the plan.
Third, some regulated firms might try to cheat by doctoring their records, tampering with their bar codes, bribing the officials who enforce the scheme and so on. While this sort of criminal evasion is probably not completely preventable, it is probably of relatively lesser concern with respect to large food enterprises that are regularly in the public's eye. Still, perhaps additional deterrence could be achieved by adopting a “whistleblower” feature as part of the regulation, so that insiders who saw cheating by their firms could be encouraged to come forward by the promise of a generous reward.
Fourth, experience from other regulatory endeavors should be drawn on to try to dampen industry efforts to use internal compliance mechanisms and/or litigation strategies to circumvent the plan's social objectives (Edelman et al. 2011). Fortunately, that the performance objectives here are quantifiable should help avoid at least some of the problems we have seen, say, in the field of employment discrimination where, instead of being given outcome goals (i.e., minority hiring quotas), firms are merely required to engage in more amorphous fair hiring practices that depend a great deal on internal firm culture to actually yield substantially greater racial diversity in the workplace. Nonetheless, leaving key details to FDA development, rather than including them in the legislation, can mean not only delays in implementation, but also legal challenges by industry to the details that the FDA adopts.
Fifth, other possible gaming strategies by the regulated parties also need attention. For example, the definition of what items are included and excluded from the scheme could become a point of evasion and litigation (e.g., what counts as “sugar” or a “saturated fat”). Or escape hatches that, in effect, would permit regulated firms to contract out their unhealthy food to others to sell would need to be monitored. These sorts of problems plague any regulatory effort in this realm: for example, when schools are required to serve healthier foods in the cafeteria, we sometimes see junk-food selling food trucks appear at school curbs, or when calorie counts are required on food seller menu boards, we see movie theater snack food bars pressing for exemptions.
Some regulated retailers might engage in a creative compliance strategy of slowly acquiring the equivalent of smaller health food stores that are not subject to regulation by the scheme so as to make the regulated firm's overall basket of goods sold look healthier, and regulations might be needed to counter this practice.
Even if the regulated firms are good citizens and act in good faith, consumers will still be able to turn to unregulated food retailers with some of their food dollars, and it is imaginable that those would be concentrated on retailers who specialize in the very items subject to the performance targets. Surely, some leakage of this sort is likely to occur. But it probably will not be large, and were that to begin to happen at that point the regulated firms will have an incentive both to dissuade their customers from such unhealthy shopping and to bring more firms into the regulatory scheme.
Probably not all of the compliance shenanigans will be anticipated or even blocked by the plan's statute and regulations so that ultimately the plan's accomplishment is likely to be less than its aim. Yet, performance-based regulation might be able to counter this in part by simply raising the annual target reduction numbers from 5 percent a year to 6 percent (Hawkins 1983).
Political Issues and Prospects
It would be important for government to commit to and enforce the plan's targets rigorously through the first round of the plan's operation, say, five years. Otherwise, firms might invest, not in changing what they sell, but in changing the law (Edelman et al. 1991). That is, in order to succeed, performance-based regulation must be viewed by industry as real and solidly in place; this makes adequate funding for plan oversight essential from the get-go. Recent experience with upgraded food safety legislation makes this clear; it is hard to expect the FDA to assure the public a safer food supply if no money is appropriated for the staff needed to bring this about.
Once it takes hold for a few years, then compliant firms, that have invested resources to meeting early-year targets, should support the plan's continuation. Yet, that does not mean that the initial five-year targets must be the targets forever. The government agency in charge of the plan would engage in program evaluation from the outset with the objective of deciding both whether some plan parameters might wisely be altered going forward and whether even greater reductions should be demanded of retailers in later years. The lesson from experience with performance-based regulation here is that the initial goals must be both substantial enough to make a real difference but at the same time be achievable. The numbers proposed here are meant to capture both those features.
Industry is likely to push for an implementation plan that would defer the major reductions until later years of the scheme or allow firms to petition for exceptions or delayed application of the rules. For example, they might accept a 10 percent reduction in calories over five years but push for only 1 percent reductions in each of the first three years, leaving 7 percent to be achieved in the final two. This is how industry has responded to “cap and trade” plans with respect to greenhouse gas emissions as industry attempts to adjust the annual schedule of cap and trade targets within the range allowed by regulators (Deason and Friedman 2010). This delayed target strategy is to be resisted. Not only does it put off real impact in the shorter run, but it also allows industry to perhaps effortlessly pick off the low hanging fruit at the start and then complain later on that the now larger annual improvements are unattainable. Better to seriously press retailers to carefully plan from the outset on how they are going to meet their longer-run target.
This discussion has so far assumed a national plan adopted by Congress. In gauging how food retailers might respond to such a proposal, it is important to consider “as compared to what.” If performance-based regulation is layered on top of a number of other national initiatives that have similar goals of improving the American diet, the business community might be very tenacious in fighting a performance-based regulation plan. By contrast, if such a plan were adopted in lieu of many of the other regulatory strategies on the table, then industry might find it an appealing alternative as it allows flexibility in meeting the targets (Lobel 2004). To win business support in this way admittedly runs the risk that, were the performance-based regulation to fail, what government primarily would be doing is collecting a lot of fines (although that would turn the plan into something like a tax on junk food). Hence, an exemption from other regulatory requirements might only apply to firms that meet their performance-based goals. This benefit of freedom from other regulation (both existing and proposed) could include the preemption of certain state and local regulation as well (Etienne 2011; Gunningham and Sinclair 1999).
A different approach would be for a few enterprising states, such as New York and California, to experiment with performance-based regulation within their borders by regulating the retailers who do business there. Perhaps Congress could facilitate that by exempting complying retailers in such states from some other federal regulation. This complex matter of food regulation by multiple levels of government is otherwise set aside for now.