In our work, we study such externalities, policies, and private mechanisms. We set out by estimating the size of a specific type of externality: the spillovers produced by big‐box retail stores. The question of whether and how much such spillovers affect surrounding businesses is a contentious one that is of interest to economists and policymakers alike.
To estimate the size of these externalities, we exploit the nationwide store closings associated with the bankruptcies of national chains of bookstores (Borders), department stores (Mervyn’s), electronics stores (Circuit City and CompUSA), and housewares stores (Linens ’n Things). We contrast the sudden disappearance of these chain stores with the continuing presence of comparable chains in the same sectors (Barnes & Noble, Kohl’s, J. C. Penney, Best Buy, and Bed Bath & Beyond). Using geographically detailed data on establishments, we find robust, positive spillovers at the short distances relevant for municipalities — half of the towns in the country span less than 3 miles at their widest point. After the disappearance of a big‐box store, nearby businesses flounder: there are fewer of them, and they employ fewer workers. On the basis of data on household shopping behavior, we establish that this decline happens in part because when consumers stop visiting the big‐box store, they also visit nearby stores less frequently. This decline is absolute, not just relative to stores that are near big‐box stores that did not close.
We then explore how local governments deal with these externalities. Externalities are a common justification for government intervention. If local governments are aware of the spillovers we observe, we might expect those local governments that face the largest positive externalities to be the ones that address them by implementing the most generous development policies. We test this hypothesis by exploiting the fact that some localities have shapes that allow them to capture more of these local spillovers, under the assumption that local governments care disproportionately about economic activity within their borders. We show that although the boundaries of the resident town do not affect the size of a big‐box store’s spillover at a given radius, the “right” geographical shape of the locality makes it more likely that these positive spillovers are contained within city limits. Localities with more compact shapes are more likely to target retail development with subsidies and tax expenditures, conditional on total area.
Of course, the right to locate near one of these large retailers is generally well‐defined. The Coase theorem would predict that in such a situation, the private sector should provide mechanisms for internalizing these spillovers. In fact, prior research has shown that foot‐traffic‐generating big‐box “anchor” stores are heavily cross‐subsidized through private mechanisms like shopping malls, in which they often pay heavily reduced rental rates or even no rent at all. Other research finds that large retailers disproportionately benefit from quasi‐private organizations like business improvement districts as well. We show that cities that are more likely to offer public retail development subsidies are less likely to have these private and quasi‐private incentive provision systems. The public internalization mechanisms crowd out the private‐sector ones.
These estimates are important for several reasons. In a recent International City/County Management Association survey, roughly 40 percent of local governments in the United States reported focusing on retail incentives to spur economic development. Be it through tax increment financing, income and property tax credits, land use subsidies, sales tax rebates, infrastructure assistance, or any of a myriad of other mechanisms, the goal is the same: to make it more attractive for large retailers to set up shop in town. The justification for local development efforts is the notion that these retailers produce positive externalities for the existing local businesses and promote the arrival of new businesses, increasing municipal tax collection and employment.
The amount of public spending involved in these types of subsidies is economically significant. Local governments in the St. Louis area, for example, provided more than $5.8 billion in local subsidies over the past 20 years, of which some 80 percent went to retail stores. Since 1986, Chicago has “lost” $5.5 billion in revenue from committed tax increment financing districts. The perceived impact of these subsidies has made them one of the most common development strategies used by policymakers at the local level, leading to a developing regulatory push for increased transparency regarding their use and cost.
Despite the ubiquity of local economic development policies, earlier research expresses skepticism about the ability of governments to perceive and target true local externalities. Our work sharpens this argument by demonstrating that even if governments are capable of identifying and targeting true externalities, government intervention may crowd out private mechanisms aimed at the same goal. In this sense, we provide rare evidence on the applicability of the Coase theorem: the situation we study features demonstrable external effects and observable variation in the level of public and private efforts dedicated to addressing those effects, but nonetheless similar economic outcomes.
This research brief is based on Daniel Shoag and Stan Veuger, “Shops and the City: Evidence on Local Externalities and Local Government Policy from Big‐Box Bankruptcies,” Review of Economics and Statistics (forthcoming), http://www.mitpressjournals.org/doi/abs/10.1162/REST_a_00703.