Specialized Careers

• Date: 01 September 2014
• DOI: http://doi.org/10.1111/jems.12065
• Published date: 01 September 2014

Specialized Careers

HESKI BAR-ISAAC

Rotman School of Management

University of Toronto

Tor on to, O N M5 S 3E 6

heski.bar-isaac@rotman.utoronto.ca

JOHANNES H¨

ORNER

Department of Economics

Yale University

New Haven, CT 06517

johannes.horner@yale.edu

An agent has different abilities in two types of tasks. These tasks are revealed over time through

his performance. The agent initially decides whether to engage in only one task (specialize) or

to take on any task that arises (be a generalist). This decision trades off the cost of being idle

against staying available for relatively lucrative tasks. We compare specializing with acting as

a generalist in an infinite-horizon model and provide complete characterizations of efforts. We

show how specializing acts as a means of committing to exert more effort. In a two-period version

of the model, this implies that positive costs for switching strategies, through license fees, for

example, may be socially desirable.

1. Introduction

The idea that there can be private and social benefits from specialization is well-known

and much-celebrated. Indeed, in March 2007, the Bank of England began to issue a

banknote featuring Adam Smith and a pin factory.In the famous example of Smith (1776)

and in much subsequent literature, gains from specialization arise for technological

reasons. The literature has tended to abstract from effort incentives and to focus instead

on the market demand and supply of specialized skills (often simply contrasting between

skilled and unskilled labor) to determine the extent of specialization in equilibrium.1

In this paper, we take a more partial approach and highlight a complementary

mechanism. We consider a single individual’s decision to specialize, and we highlight

that specialization (which is relatively easy to monitor and enforce) can provide incen-

tives for effort that might otherwise be absent. Specializing entails spending relatively

more time engaged in a given task, and so the incentives to improve reputation in that

task are greater than they would be for a generalist. The cost of specialization is that the

agent’s lower flexibility might lead to taking fewer assignments.2

1. For a recent overview,see Acemoglu and Autor (2010). In addition to considering the extent or scale of

the market, the literature has also addressed the costs of coordination as limiting the extent of specialization

(Becker and Murphy,1992) and extensively explored complementarities with information and communication

technologies.

2. In the model, an agent who is not engaged in a specialized task is idle; in application this “idle” time

may reflect less lucrative and less specialized work that is always freely available: for example, in a consulting

firm this may be data-gathering, fact-checking, etc., in addition to time without engagement.

C2014 Wiley Periodicals, Inc.

Journal of Economics & Management Strategy, Volume23, Number 3, Fall 2014, 601–627

602 Journal of Economics & Management Strategy

Of course, it may, in fact, turn out that the agent has little ability in that task;

moreover, incentives diminish over time. Therefore, specialization may require com-

mitment. This is in contrast to purely technological reasons and other rationalizations

for specialization, such as comparative advantage or exploiting returns for individual

specific human–capital investments (as in Rosen, 1978, 1983). Enforcing specialization,

which is easier (but less effective) than enforcing effort directly, can be done formally

or informally. Professional organizations often directly constrain the type of work in

which a professional can be engaged. For example, the Law Society of England and

Wales (2004) imposes considerable restrictions on the nonlegal activities that a solicitor

can undertake, especially from the same physical premises or under the same business

name.3Informal and cultural constraints can also play a role in limiting the scope and

variety of an agent’s work. Here, we simply take as given that such formal and informal

institutions provide a commitment device that can generate greater overall efficiency

but may introduce some inefficiency ex post.4Alternatively, our results can be viewed as

illustrating consequences of such rules, even if they arise for other reasons.

Although our primary analysis considers the decision to work as a specialist or

generalist as a one-off once-and-for-all decision, we consider an extension in a two-

period version of the model where an agent can start as a specialist and then at some

cost move on to become a generalist. This cost could be thought of either as an explicit

license fee or certification cost. Such fees and costly licensing requirements are prevalent

in the services industries in particular (Kleiner, 2006). Our analysis then allows us to

characterize an optimal switching cost (license fee) albeit under the assumption that it

plays no other role but to affect agent incentives and allocation.

Specialization in professional services has typically been considered as a strategic

choice at the firm rather than the individual level (Chatain and Zemsky, 2007; Garicano

and Hubbard, 2009), and incentives for effort have not been addressed. However, for the

medical profession specifically, there is a longstanding literature that seeks to address

the extent to which physicians’ choices of specialization respond to lifetime earnings

in these specialities (Sloan, 1970; Nicholson, 2002; Gagn´

eandL

´

eger, 2005). In part, this

interest reflects a policy concern over the aggregate provision of care and shortages in

certain practice areas. Courty and Marschke (2008) present an interesting analysis based

on modeling different specialities as differing in their exposure to moral hazard; the

analysis relates the failure of the market to efficiently allocate health care services with

the recent growth in medical specialization and subspecialization. Our analysis, rather

than being a full market analysis, is more modest in scope, focusing on an individual’s

decision to specialize or not. In this sense, it is perhaps closer to the literature on career

concerns, which typically does not address a full market equilibrium—indeed, it is

typically predicated on the assumption that the agent has market power and earns all

the surplus that he generates.

Formally, we build on the career-concern model introduced in Fama (1980) and

Holmstr¨

om (1999). In this seminal model, an agent’s current performance influences

3. It is hard to imagine that clients would not know of such activities, particularly at the same premises.

Thus, in ascribing such rules to conflicts of interest, one would have to assume some paternalism on the part

of the Law Society.This is a plausible, though not entirely satisfactory explanation—in particular, with respect

to commercial work, where one might suppose that clients are fairly sophisticated and where there may be

benefits to relaxing these restrictions and enjoying economies of scope.

4. In the model, we largely take the somewhat extreme view that agents either fully specialize or fully

generalize and cannot switch strategies. In practice, things may not be so extreme. In a two-period analog

of the model, considered in Section 5, we are able to characterize optimal penalties or switching costs from

moving from specialization to acting as a generalist, and we highlight that similar intuitions apply.