Human Capital Investment and Work Incentives

• Published date: 01 September 2016
• DOI: http://doi.org/10.1111/jems.12155
• Author: Matthias Kräkel
• Date: 01 September 2016

Human Capital Investment and Work Incentives

MATTHIAS KR¨

AKEL

Department of Economics

University of Bonn Adenauerallee

24-42 53113 Bonn Germany

m.kraekel@uni-bonn.de

Traditionalhuman capital theory based on the work by Gary Becker shows that firms do not invest

in general human capital but offer firm-specific training that is only useful for the training firm.

I extend the traditional approach by adding two natural assumptions—workers cannot be forced

to acquire new knowledge, and they exert unobservable effort to produce valuable output for their

employer. I show under which conditions firms do not offer firm-specific training but invest in

general human capital, which increases the workers’ outside option. This investment behavior is

well in line with the documented prevalence of industry-specific and occupation-specific human

capital over firm-specific human capital.

1. Introduction

In his seminal article, Becker (1962) shows that a firm does not invest in the knowledge

of its workers if this knowledge is also useful for other firms (general human capital).

However, in case of firm-specific human capital, which can only be used in the training

firm, investment costs and returns are shared between the worker and the firm (see also

Hashimoto, 1981), resulting in a sharing of the respective quasi-rent (Hutchens 1989,

p. 52).

In this paper, I extend traditional human capital theory by including two assump-

tions that do not seem to be unrealistic—workers cannot be forced to acquire human

capital, and they choose unobservable effort to produce valuable output for their em-

ployers. In the basic model, I show that in this extended setting, firms do not invest in

purely firm-specific human capital but are often willing to invest in general training.1

Traditional human capital theory assumes that a worker has no choice whether to

acquire knowledge or not. When a firm decides to invest in human capital, a worker

is considered more like a robot to be programmed rather than a human being who is

free to learn or not. Often, however, such programming is not possible. For example,

teachers typically do their best to transfer knowledge to their students, but the latter

do not always appreciate the teachers’ efforts. In the following, I consider a situation in

which firm and worker need to cooperate for human capital investment to be successful.

As a second extension, I add a moral hazard problem to the traditional approach. The

worker has to choose between working hard or not working hard, which is not directly

observable by the firm. As the worker is protected by limited liability,creating incentives

is costly for the firm.

I like to thank two anonymous referees, an anonymous co-editor, Pia Pinger, and Anja Sch¨

ottner for helpful

comments and suggestions.

1. Empirical studies also show the prevalence of general over firm-specific human capital; see Sch¨

onberg

(2007).

C2015 Wiley Periodicals, Inc.

Journal of Economics & Management Strategy, Volume25, Number 3, Autumn 2016, 627–651

628 Journal of Economics & Management Strategy

In the extended setting, I show that the firm does not invest in purely firm-specific

training but may prefer investment in general human capital. The intuition is the follow-

ing. Additional human capital helps a worker to carry out his task. As a consequence,

the firm only needs lower powered incentives to motivate the worker in case of moral

hazard, resulting in a lower rent for the worker. Firm-specific human capital solely

leads to this rent reduction without affecting the worker’s outside option. Hence, the

worker will always disagree to make use of the offered training. The firm anticipates

this behavior and decides against training to save cost. If, however, human capital is

sufficiently general, it can lead to a significant increase in the worker’s outside option

with the result that the worker is now willing to learn.2In turn, to make the firm willing

to offer training, the outside option should not be allowed to become prohibitively large

because the worker has to be compensated for it when staying with the training firm.

To check the robustness of the main finding, I discuss four variants of the basic

model. First, I consider the possibility that the firm may commit to a long-term wage

profile offered to the worker. If such commitment is possible, investment in purely firm-

specific human capital will be successful. However, long-term commitment can also

lead to considerable disadvantages for a firm in an uncertain environment. Second, I

deviate from the basic model by assuming that the firm does not have all the bargaining

power. Instead, wages are determined via the Nash bargaining solution. In this setting,

purely firm-specific training will be successful if the efficiency gains from training are

sufficiently large and the worker’s bargaining power is not too high. Third, I modify

the assumption of the basic model that human capital lowers a worker’s effort cost. In

that case, additional human capital makes completion of the task easier for the worker,

enabling him to save effort costs (e.g., the worker needs less time to complete the task).

As an alternative, I consider a scenario in which human capital increases the worker’s

success probability. For example, we can think of knowledge that makes the machine

or technique used by the worker more effective but leaves his effort costs unchanged

(e.g., his working time). I show that a result similar to that of the basic model will hold

if human capital and effort are complements. If, however, human capital and effort

are substitutes, the worker’s rent will always increase with training, making any kind

of human capital investment successful. Finally, I let the firm’s training decision be

described by a continuous variable. As in the basic model with discrete training, purely

firm-specific human capital investment will not be successful.

The major finding of the paper fits well with the empirical human capital liter-

ature. Several empirical studies document that often human capital is neither purely

firm specific nor purely general.3Instead, it is productive in several but not all firms

and can often be classified as either occupation specific (i.e., knowledge is valuable

in a certain type of job that can be found in different firms) or industry specific (i.e.,

knowledge is only valuable in firms that belong to a certain industry). Neal (1995) uses

data from the Displaced Worker Surveys and shows that the wage losses of displaced

workers switching industries are substantially larger than the wage losses of workers

staying in their predisplacement industries. Hence, workers’ human capital is for the

most part industry specific. Neal (1995, p. 653) concludes, “Workers apparently receive

compensation for some skills that are neither completely general nor firm-specific but

rather specific to their industry or line of work.” The parameter condition for the main

2. For example, general human capital increases the worker’s income by the worker becoming self-

employed.

3. See, for example, Poletaev and Robinson (2008), Zangelidis (2008), and Suleman and Lagoa (2013).