Product safety, contracts, and liability

• Author: Xinyu Hua,Kathryn E. Spier
• Date: 01 March 2020
• DOI: http://doi.org/10.1111/1756-2171.12311
• Published date: 01 March 2020

RAND Journal of Economics

Vol.51, No. 1, Spring 2020

pp. 233–259

Product safety, contracts, and liability

Xinyu Hua∗

and

Kathryn E. Spier∗∗

A firm sells a dangerous product to heterogeneous consumers. Higher consumer types suffer

accidents more often but may enjoy higher grossbenefits. The firm invests resources to reduce the

frequency of accidents. When the consumer’s net benefit function (gross benefits minus expected

harms) is decreasing in consumer type, the firm contractually accepts liabilityfor accident losses

and invests efficiently. When the consumer’s net benefit function is increasing in consumer type,

the firm contractuallydisclaims liability and underinvests. Legal interventions, including products

liability and limits on contractual waivers and disclaimers, are necessary to raise the level of

product safety.

1. Introduction

Should the manufacturer of a dangerous product be held liable when consumers, while

using the product, suffer injury or harm? Although one’sintuition might suggest that firms should

be held responsible—especially when the product malfunctioned in some way—the case for

products liability is not as strong as it first appears. After all, sophisticated, forward-looking

consumers would be willing to pay a premium for safer products that malfunction less often. In

a well-functioning market, firms can turn a handsome profit by giving consumers the types of

goods and services that they desire and by designing private contracts to assure performance and

allocate risk. Thus, on reflection, the logical case for products liability must hinge on the failure

of private markets and private contracts.

In this article, the divergence between the privateand social incentives to design safer prod-

ucts stems from the exercise of market power. A firm with market power will make cost-justified

∗Hong Kong University of Science and Technology; xyhua@ust.hk.

∗∗Har vardLaw School and NBER.

Wewould like to thank Ying Chen, Eric Claeys, James Dana, Oliver Hart, Louis Kaplow,Susan Norton, Marco Ottaviani,

Ron Siegel, Holger Spamann, the Editor, and anonymous referees, participants in the 7th Annual Law and Economics

Theory Conference at the Rice University (2017), the Law and Economics Conference at the Hong Kong Universityof

Science and Technology(2017), the 28th Annual Meeting of the American Law and Economics Association (2018), the

REStud TourReunion Conference at the University of Copenhagen (2018), the Workshop on Industrial Organization and

Competition Policy at UIBE (2018), and seminar audiences at George Mason University (2018), Harvard Law School

(2018), Harvard Economics Department (2018), National University of Singapore (2018), VanderbiltUniversity (2019),

UniBG (2019), and University of Mannheim (2019).

C2020, The RAND Corporation. 233

234 / THE RAND JOURNAL OF ECONOMICS

improvements in product safety to suit the needs and preferences of the marginal consumer who

is just indifferent between purchasing the good and not. Problems may arise when the marginal

consumer’s needs and preferences are not representative of the average buyer of the product

(Spence, 1975). This is true in many economic settings involving dangerous products, and poses

a particular problem for the firm and for society as a whole when consumers who place a higher

gross value on the product also suffer accidents more frequently. We argue that the problem

of inadequate product safety is pernicious and cannot be overcome by private contracts in a

free market.

The theoretical insights of this article are empirically relevant. Many dangerous products are

supplied by imperfectly competitive markets. According to the Consumer Product Safety Com-

mission (CPSC), all-terrain vehicles (ATVs) have been associated with approximately 100,000

emergency-department treated injuries and more than 600 fatalities each year.1The ATV industry

is dominated by several large competitors, with Polaris and Honda jointly accounting for more

than 60% of sales.2According to the Centers for Disease Control and Prevention (CDC), nearly

half a million people die in the United States each year from smoking-related diseases, and 16

million more suffer from smoking-related illnesses.3With 41% of the market, Philip Morris’

Marlboro brand has more sales than the next eight cigarette brands combined.4Many medical

device makers and pharmaceutical companies have considerable market power as well. Before

Merck pulled Vioxx from the market in 2004, the blockbuster arthritis drug enjoyed a large

market share.

We begin with a simple benchmark model, where a monopolist sells a potentially danger-

ous product to a population of heterogeneous, risk-neutral consumers. Product safety is fully

observable to consumers at the time of sale. By investing more resources, the firm can reduce

the likelihood of accidents. Taking the level of sales as fixed, the socially optimal investment

in product safety would minimize the aggregate production costs plus the aggregate harms to

the consumers. A consumer’s type affects both the consumer’s propensity to suffer harm and the

gross benefit of consumption: higher consumer types suffer accidents more frequently, but may

enjoy larger gross benefits. Note that a consumer’s net benefit from consuming the product—the

gross benefit of consumption minus the expected harms—may be either increasing or decreasing

in the consumer’s type.5The firm cannot obser ve consumer types and therefore cannot price

discriminate directly.

If the firm bears no financial responsibility for accidents, then the private and social incentives

to invest in product safety diverge. When the consumer’s net benefit function is decreasing in

the consumer’s type, then (absent liability) the firm overinvests in product safety. The reason for

this result is that the consumer type who is just indifferent between purchasing the good and

not purchasing it, the “marginal consumer,” is someone who places a relatively high value on

product safety.The fir m chooses a high levelof product safety to match the needs of this marginal

consumer. Conversely, when the consumer’s net benefit is an increasing function of the consumer’s

type, then the marginal consumer places a relatively low value on product safety. In this case,

the firm chooses a low level of product safety to suit the needs of the low-type consumer. Thus,

absent products liability or financial responsibility for consumer harms, product safety may be

either too high or too low.

1Many of the victims are children. See 2016 Annual Report of ATV-Related Deaths and Injuries available at

www.cpsc.gov.These statistics do not include dune buggies or golf carts.

2In 2017, worldwide market shares were Polaris (36%), Honda (28%), Yamaha (13%), and Can-Am (13%). See

www.statista.com/statistics/438085/global-all-terrain-vehicle-market-share/.

3See www.cdc.gov/tobacco/about/osh/index.htm.

4See www.statista.com/statistics/603940/market-share-leading-cigarette-brands-us/.

5Many products fit this description. Consumers of durable goods such as cars, table saws, and pressure cookers

often vary in their frequency of product use. Intensive users of Facebook or Amazon enjoy more benefits but are more

likely to suffer harm from data leakage.

C

The RAND Corporation 2020.