Estimating the benefits and costs of forming business partnerships

• Date: 01 June 2020
• Author: Jungho Lee
• Published date: 01 June 2020
• DOI: http://doi.org/10.1111/1756-2171.12324

RAND Journal of Economics

Vol.51, No. 2, Summer 2020

pp. 531–562

Estimating the benefits and costs of forming

business partnerships

Jungho Lee∗

I estimate a matching model of business-partnership formation to quantify the relative impor-

tance of productivity gains, financing gains, and the coordination failure of effort provision

(moral hazard) among partners. Productivity gains account for 61% of the gain from the ob-

served partnerships. For partnersin the first quartile of the wealth distribution, however, financ-

ing accounts for 93% of the gain. The cost of moral hazardcorresponds to 42% of the entire gain

from partnerships. A loan policy specifically targeting partnerships is less effective in improving

welfare than a conventional loan policy that provides loans to individual entrepreneurs.

1. Introduction

Finding a partner is one of the most important decisions for potential start-up owners. Suc-

cessful business owners often argue that finding the right partner is the key to a firm’s success

(e.g., Cohen and Eisner 2010; Kawasaki 2004). Indeed, many successful companies started as

partnerships. Examples include Hewlett-Packard, Procter & Gamble, and Ben & Jerry’s. Yet, de-

spite successful examples, only 18% of nonfamily-owned businesses began as partnerships.1

Who forms business partnerships, and why? Answering these questions can be valuable not

only for potential start-up owners, but also for policymakers who try to boost entrepreneurship

via partnerships.

Theoretical studies have identified gains and losses from business partnerships. For exam-

ple, working with a partner may increase a firm’s productivity through a knowledge transfer

between partners. Financially constrained entrepreneurs can increase their financing capacity by

finding a wealthy partner.The gains, however, can be offset by coordination failure such as moral

hazard in teams—the problem of inducing optimal effort by partners when the private return to

effort is smaller than the marginal product of effort. The overall impact on entrepreneurs of the

∗School of Economics, Singapore Management University; jungholee@smu.edu.sg.

I thank the editor, Chad Syverson, and two anonymous referees for their helpful comments. This article is based on the

first chapter of my PhD dissertation at WashingtonUniversity in St. Louis, and was circulated under the title “Why Form

Business Partnerships.” I am greatly indebted to Barton H. Hamilton for his guidance and advice. I am also very grateful

to Tat Chan, George-Levi Gayle, and Carl Sanders for their help and feedback on this project. I thank Mariagiovanna

Baccara, Kyoung Jin Choi, Ignacio Esponda, Nicolas L. Jacquet, Yumi Koh, Takashi Kunimoto, Gea M. Lee, Patrick

Legros, Sunha Myong, JungJae Park, Thomas Sargent, Bernardo Silveira, AloysiusSiow, Jingyi Xue, and Yichong Zhang

for helpful discussions. All errors are my own.

1The Survey of Income and Program Participation (SIPP). See Section 2 in this article.

© 2020, The RAND Corporation. 531

532 / THE RAND JOURNAL OF ECONOMICS

option to form a partnership will depend on the magnitude of productivity gains, financing gains,

and the cost of moral hazard.

In this article, I quantify the relative importance of productivity gains, financing gains,

and the cost of moral hazard in business-partnership formation by developing and estimating

a model of partnership formation. The model is an extension of Evans and Jovanovic (1989), an

occupational-choice model between a worker and an entrepreneur with a borrowing constraint.

I modify their model to incorporate an option to form a partnership.2By forming a partnership,

financially constrained agents can increase their borrowing capacity. Forming a partnership may

also increase productivity. A partnership can be formed if mutual gains in productivity exist or

if one partner gains in productivity and the other partner gains in financing despite the cost of

moral hazard. Who is matched with whom is determined in a stable matching at which no partner

would prefer to be a single owner, and one cannot find two partners who would prefer to form a

partnership with each other than to remain with their current partners. The model is estimated by

the method of simulated moments using the SIPP, a nationally representative household-based

survey of the US population.

The estimated model implies the increase in productivity accounts for the major gains from

partnerships among those who choose to be partners: About 61% of the aggregate gains for all

partners are attributed to the productivity gains. However, the gains from financing are the ma-

jor gains for low-wealth partners. For the partners in the first quartile of the wealth distribution,

financing accounts for about 93% of the entire gain from partnerships. The cost of moral haz-

ard is high among the observed partnerships. This cost corresponds to 42% of the entire gain

from partnerships.

I use the estimated model to evaluate a policy aimed at boosting entrepreneurship via part-

nerships. To better understand the effectiveness of such a policy, I first discuss the welfare im-

plication of financial friction when the option to form a partnership is available. In my model,

financial friction generates inefficiency through two channels. The first one is from financially

constrained single owners who could not invest optimally. In addition to this conventional chan-

nel, my model predicts the sorting pattern among start-up owners changes due to financial fric-

tion. The welfare loss associated with the former channel is 74.18%, and that with the latter

channel is 25.82% of the aggregate welfare losses due to financial friction. The major part of

welfare loss associated with changes in the sorting pattern comes from previously single owners

who become partners once financial friction is imposed. The welfare loss by this transition group

accounts for 24.14% of the aggregate welfare losses due to financial friction.

As a relevant policy simulation, I investigate the impact of a loan policy specifically target-

ing partnerships. For comparison, I also present the impact of a loan policy that provides loans to

individual entrepreneurs. Given the same amount of aggregate loans, the individual-based loan

policy is more effective in helping financially constrained single owners increase investment.

Moreover, the individual-based loan policy enables partners who would become single owners

without financial friction to switch to single ownership, whereas the partnership loan policy does

not. Overall, a loan policy specifically targetingpar tnerships is less effectivein improving welfare

of start-up owners than a conventional loan policythat provides loans to individual entrepreneurs.

This article is related to the literature on business partnership. Extensive theoretical studies

have examined reasons behind business-partnership formation. In the presence of complemen-

tarity between partners, a partnership arises as the optimal contract if the knowledge or a tacit

human capital between partners is not observable or not contractible (e.g., Teece 1980; Garicano

and Santos 2004; Morrison and Wilhelm Jr 2004; Bar-Isaac 2007). On the other hand, Legros

and Newman (1996) show that agents’ wealthcan play an important role in partnership for mation

in the presence of financial friction. Regarding the cost of partnerships, moral hazard in teams

2To focus on the choice betweensole ownership and partnership, I remove the option of becoming a worker.

C

The RAND Corporation 2020.

LEE / 533

has been the primary theoretical concern since Holmstrom (1982).3I contribute to this line of

literature by quantifying mechanisms discussed in theoretical studies using a parsimonious equi-

librium model of partnership formation. Gaynor and Gertler (1995) and Lang and Gordon (1995)

also empirically examine business partnership. They show that the extent of moral hazard among

partners in professional industries depends on the degree to which the partners spread risk. This

article complements their studies by examining different motives for partnership formation that

are particularly relevant to start-up firms in general industries.4

This article also contributes to the literature on entrepreneurship. Despite the fact that many

entrepreneurs start a business by forming a partnership, most previous models of entrepreneur-

ship have abstracted such an option. A few exceptions include Basaluzzo (2006) and Ševˇ

cík

(2015), who quantify to what extent the option to find a partner can help financially constrained

agents create a business, and hence increase the aggregate outcome. This article differs from

theirs in two important ways. First, Basaluzzo (2006) does not allow mutual gains in produc-

tivity, and therefore, a partnership is formed only if one partner gains in productivity and the

other partner gains in financing. Similarly, Ševˇ

cík (2015) focuses on the efficient resource allo-

cation between two separate establishments in the presence of financial friction, and therefore,

does not allow a possibility that a partnership is formed due to the increase in productivity. By

allowing gains in mutual productivity as another source of partnership formation, I show that the

potential role of financing is small relative to the potential role of productivity gains. Another

closely related article is Åstebro and Serrano (2015), who show independent inventors having a

partner can significantly increase the probability of commercialization success. Different from

Åstebro and Serrano (2015), I allow heterogeneous productivity gains (or losses) and moral haz-

ard across agents induced by a matching model, and use the estimated model to address welfare

and policy questions.

The matching model in this article is classified as the imperfectly transferable utility (ITU)

model in the matching literature (e.g., Chade, Eeckhout, and Smith 2017). Despite its relevance,

practical applications of the ITU models are rare.5I provide one such example that is particularly

relevant for start-up formation. Moreover, I quantify the welfare consequences of changes in

the sorting pattern among start-up owners due to financial friction, and evaluate the equilibrium

consequences of financial subsidies for start-up owners when the option to form a partnership

exists.6

The article is organized as follows. Section 2 documents some facts about business part-

nership formation. A model of partnership formation is presented in Section 3. Section 4 dis-

cusses the identification and estimation of the model. The results are presented in Section 5.

Section 6 concludes.

2. Facts about business-partnership formation

In this section, I describe the main data set for this study and provide some descriptive

patterns related to business partnerships among start-up owners.

Data construction. I use the SIPP for this study. The SIPP is a nationally representative

household-based survey of the US population designed to collect information for income and

program participation. I chose the SIPP for several reasons. First, the sample size is large and

3The articles that study moral hazard in teams include Radner (1986), Rasmusen (1987), Legros and Matsushima

(1991), Legros and Matthews (1993), Miller (1997), Strausz (1999), Battaglini (2006), and Rahman and Obara (2010).

4Another difference is that I focus on the formation of business partnerships among start-up owners, whereas

Gaynor and Gertler (1995) and Lang and Gordon (1995) examine already established partnerships.

5An exception is Legros and Newman (2007).

6Eeckhout and Munshi (2010) present a model of competitive matching in which potential borrowersand lenders

are sorted into groups to exploit gains from trade in the presence of financial friction. They are interested in an informal

financial institution in India called Chit funds, whereas I focus on the business-partnership formation.

C

The RAND Corporation 2020.