The effects of ambiguity on entrepreneurship

• Author: Pablo A. Gutiérrez Cubillos,Claudio A. Bonilla
• DOI: http://doi.org/10.1111/jems.12402
• Date: 01 February 2021
• Published date: 01 February 2021

J Econ Manage Strat. 2021;30:63–80. wileyonlinelibrary.com/journal/jems © 2020 Wiley Periodicals LLC

|

63

Received: 27 March 2019

|

Revised: 23 August 2020

|

Accepted: 31 August 2020

DOI: 10.1111/jems.12402

ORIGINAL ARTICLE

The effects of ambiguity on entrepreneurship

Claudio A. Bonilla

1

|Pablo A. Gutiérrez Cubillos

2

1

School of Economics and Business,

University of Chile, Santiago, Chile

2

University of British Columbia,

Vancouver, Canada

Correspondence

Claudio A. Bonilla, University of Chile,

257 Diagonal Paraguay Avenue, Santiago

8330015, Chile.

Email: cbonilla@fen.uchile.cl

Abstract

We incorporate ambiguity (Knightian uncertainty) into a classic model of en-

trepreneurship to analyze, among other things, its effects on the optimal level of

business startups, the relation between total assets and the size of the en-

trepreneurial investment, the effects of increasing ambiguity on developing new

ventures, and the decision to self‐select into entrepreneurship for an indifferent

decision maker. We first show that, under the monotone‐likelihood ratio

property, the introduction of ambiguity negatively affects the optimal en-

trepreneurial investment, something thatisconsistentwithmostexperimental

evidence about entrepreneurial choice under ambiguity. Then, we show that the

classical explanations for the positive correlation between total assets and busi-

ness startups based on decreasing absolute risk aversion preferences and prudent

behavior can be challenged when ambiguity is incorporated into the analysis, and

we provide the conditions that guarantee that the traditional comparative static

result under risk is replicated under ambiguity. We also show that increases in

ambiguity aversion reduce entrepreneurial activities. Finally, we discuss our

results under alternative ways of modeling ambiguity.

1|INTRODUCTION

Entrepreneurship is essential to economic development and growth. Entrepreneurship is the engine by which economic

agents look for opportunities to innovate and develop new products and services that intensify competition and increase

productivity. Therefore, understanding the factors (like risk and uncertainty) that affect entrepreneurship is important

not only to economists but also to policymakers and government officials.

Risk (measurable uncertainty) and uncertainty (unmeasurable uncertainty) have been analyzed in the economics of

entrepreneurship since Knight (1921) conceptualized the idea that, in principle, less‐risk‐averse agents will tend to become

entrepreneurs because they can bear the risk of a venture more efficiently than more‐risk‐averse individuals. More‐risk‐

averse agents will then tend to become employees with a sure wage to avoid exposure to entrepreneurial risk. Some authors

have challenged this view about the relationship between entrepreneurship and risk aversion (Kan & Tsai, 2006;

Vereshchagina & Hopenhayn, 2009). However, most literature accepts the negative relation between risk aversion and

entrepreneurship (or self‐employment) as a common feature of the economics‐of‐entrepreneurship field (Ahn, 2009;

Ekelund, Johansson, M, & Lichtermann, 2005;Hsieh&Parker,2017; Kihlstrom & Laffont, 1979;Parker,1997,2006).

An interesting previous empirical work that documented the positive relation between business startups (or

entrepreneurship) and total assets, instead of risk characteristics, is the work of Evans and Jovanovic (1989). These authors

[Correction added on 21 September 2020, after first online publication: the name of the co‐author Pablo G. Cubillos has been corrected to Pablo A.

Gutiérrez Cubillos]

argue that as the economy develops, the greater quantity of assets relaxes the credit‐rationing constraint for some individuals,

increasing their likelihood of obtaining financing to forge new ventures, which makes them switch from paid employment to

self‐employment (entrepreneurship). Cressy (2000) challenged the intuition of that result by arguing that the positive

correlation between assets and the rate of business startups comes from decreasing absolute risk aversion (DARA) pre-

ferences. In other words, as the economy develops, some individuals accumulate more assets and, consequently, their level of

risk aversion decreases. This leads them to switch from safe employment to risky entrepreneurial activities, thus increasing

the rate of business startups. More recently, Bonilla and Vergara (2013) have shown that the positive correlation between

assets and entrepreneurship is actually due to a more general characteristic of preferences. In particular, they argue that

prudent behavior

1

is the main reason behind the positive correlation between assets and entrepreneurship. Prudent behavior

has been previously studied in savings theory as a mechanism that explains the empirical finding that agents desire to use a

positive fraction of their wealth in more savings when they face an increase in future risk, which is known as the

precautionary savings motive (Baiardi, Magnani, & Menegatti, 2020; Kimball, 1990).

Intuitively, risk emerges when an agent makes a choice that makes him face different states of nature, with a given

probability associated with each state. In contrast, ambiguity is present when there are different probability distribution

functions in the environment, and we make a choice without knowing which distribution is the one we are going to

face. This makes optimal decisions more complex.

Recently, the concept of ambiguity has been incorporated into a wide range of economic modeling since Klibanoff,

Marinacci, and Mukerji (KMM) (2005) developed a smooth way to work with ambiguity. The smoothness idea provided

a simpler manner to deal with ambiguity and expanded its use to a wider range of applications compared to the

traditional approach. The traditional approach involved using the maxmin model of Gilboa and Schmeidler (1989)—

where individuals maximize the minimum expected utility over the distributions—and the

α

‐maxmin expected utility

model of Ghirardato, Maccheroni, and Marinacci (2004)—in which welfare is measured as a weighted average between

the minimum and the maximum expected utility levels, given the set of second‐order distributions.

It is easy to justify the inclusion of ambiguity in entrepreneurial decisions. For example, consider the case of an

agent who decides to start her own business. In this case, she has two levels of uncertainty. The first level is related to

the amount of demand for her product. For instance, assume a situation with two possible outcomes: high and low

demand as the first (basic) level of uncertainty. The second level of uncertainty relates to how likely it is that demand

will be high or low. In other words, the entrepreneur knows neither the final outcome nor the probability distribution

underlying the possible outcomes and, therefore, this entrepreneur is facing ambiguity in her venture. Another example

of ambiguity applied to entrepreneurship is the usual robustness analysis in the evaluation of new projects. In this case,

the net present value of projects can be estimated for bad, average, or good macroeconomic conditions under different

probability distributions representing each scenario. The robustness exercise is a way to deal not only with the risk

embedded in a new venture, but also with the ambiguity associated with unknown probability distributions of results.

In this context, some natural questions arise. For instance, what is the effect of ambiguity on entrepreneurship? Is

that effect different to what we already know from the traditional economic theory of entrepreneurship under risk?

Should we have a separate entrepreneurial theory under ambiguity?

The main contribution of this paper is to extend the theory of entrepreneurship under risk, incorporating the

concept of ambiguity. In doing so, we provide new insights about the effects of ambiguity—sometimes challenging

previous results in the field, but also offering conditions to restore the intuition found in the previous literature. We first

show that under certain conditions (monotone‐likelihood ratio property [MLRP])

1

the inclusion of ambiguity reduces

the optimal level of the entrepreneurial investment, something that negatively affects economic growth and develop-

ment. Second, we show that, even in the case where we assume that agents are prudent and ambiguity averse, an

additional condition is required to guarantee a positive monotonic relationship between total assets and business

startups. This condition tells us that there is an upper bound for the absolute ambiguity aversion coefficient that

guarantees the positive monotonic relation between total assets and entrepreneurship. Third, we show that increases in

ambiguity aversion reduce the pool of entrepreneurs, and we provide the conditions that guarantee that increases in

wealth increase the expected return of the marginal entrepreneur. Fourth, we provide conditions on the priors and

preferences that guarantee that increases in wealth increase entrepreneurial investment. Finally, we discuss our results

under alternative ways of modeling ambiguity.

Our results highlight that the relevance of the study of ambiguity in entrepreneurship is threefold. First, it brings the

theory and practice of entrepreneurship closer, since the cases in which the decision maker knows the “real”prob-

ability distribution of results are rare. Ambiguity about the distribution of results and making decisions without all the

64

|

BONILLA AND CUBILLOS