Private investment with social benefits under uncertainty: The dark side of public financing

• DOI: http://doi.org/10.1111/jpet.12358
• Date: 01 June 2020
• Author: Anne Stenger,Giuseppe Attanasi,Marco Buso,Kene Boun My
• Published date: 01 June 2020

Received: 1 October 2017

|

Revised: 24 December 2018

|

Accepted: 27 December 2018

DOI: 10.1111/jpet.12358

ORIGINAL ARTICLE

Private investment with social benefits under

uncertainty: The dark side of public financing

Giuseppe Attanasi

1

|

Kene Boun My

2

|

Marco Buso

3

|

Anne Stenger

4

1

Université Côte d’Azur, CNRS, GREDEG,

Nice, Sophia‐Antipolis, France

2

CNRS, BETA, Université de Strasbourg,

Strasbourg, France

3

Interuniversity Centre for Public Economics

(CRIEP), Interdepartmental Centre G. Levi

Cases for Energy, Economics and Technology,

University of Padova, Padua, Italy

4

BETA, INRA, Université de Strasbourg,

Strasbourg, France

Correspondence

Giuseppe Attanasi, GREDEG, CNRS,

Université Côte d’Azur, CNRS, GREDEG

250, rue Albert Einstein, 06560 Nice

Sophia‐Antipolis, France.

Email: giuseppe.attanasi@unice.fr

Funding information

Agence Nationale de la Recherche, Grant/

Award Numbers: ANR "GrICRiS", ANR

“ForWind”012‐AGRO 0007

Abstract

We develop a game‐theoretic model of private–public con-

tribution to a long‐term project with sequential actions and

moral hazard. A private agent is one who is in charge of both

the financial contribution and the management effort, these

two actions entailing private costs and uncertain ex‐post private

and social benefits. A public agent is one who decides the

amount of public funding to this quasi‐public good, knowing

that the size and the probability of attaining a surplus ex post

depend on the private agent’s effort. We consider four public‐

funding scenarios: benefit‐sharing versus cost‐sharing crossed

with ex‐ante versus ex‐interim government intervention. We

test our theoretical predictions by means of an experiment that

confirms the main result of the model: Cost‐sharing public

intervention is more effective than benefit‐sharing in boosting

private financial contribution to the project. Furthermore, when

public intervention comes after private contribution (ex‐interim

government intervention), both public‐funding scenarios have a

negative impact on the private management effort. In our

model, the latter result is explained by the private agent’shigh

degree of risk aversion. These results have policy implications

for strategic investments with long‐term social consequences.

In deciding the optimal timing and method of the contribution,

governments should also consider the indirect effects on

agents’long‐term management efforts.

J Public Econ Theory. 2020;22:769–820. wileyonlinelibrary.com/journal/jpet © 2019 Wiley Periodicals, Inc.

|

769

Editor’s note: This paper constitutes part of the two special issues of JPET “Commemorating works of James Andreoni, Theodore Bergstrom, Larry

Blume, and Hal Varian”

1

|

INTRODUCTION

In this paper, we provide a theoretical framework and a related laboratory experiment meant to investigate the

effect of public contribution on the private provision of a quasi‐public good

1

that is characterized by uncertain

benefits and moral hazard. In general, markets for these goods are incomplete, so public intervention is considered to

be necessary to restore efficiency.

The specific quasi‐public goods we are interested in are long‐term projects that yield private and public

benefits—for example, innovation activities that are pursued by private firms. Most of the existing research focuses

on assessing the short‐term impacts of public subsidies to these long‐term projects. However, due to the lack of

reliable data, empirical studies can hardly measure public programs’ability to affect long‐term outcomes and create

value for society. In addition, the behavioral effects of public interventions are difficult to assess from an empirical

point of view because of the difficulty in isolating and measuring private investors’behavioral responses to such

interventions.

We claim that the controlled environment of a theory‐driven laboratory experiment can help shed some light on

both issues. In particular, our main aim is to identify in an experiment the most efficient form and timing of

government’s contribution to privately funded and managed long‐term projects, which are characterized by moral

hazard and uncertain private and social benefits.

Developed countries allocate large portions of their budgets to large‐scale, long‐term programs that (a) facilitate

the formation of long‐term tangible assets (e.g., adaptation to climate change, and ecoinnovation technologies) and

intangible assets (e.g., education and research and development), and that (b) boost innovation and competitiveness

(European Commission, 2013; Olsen & Fearnley, 2014).

Many of these investments have wide public benefits because they generate significant returns for the

society as a whole by supporting essential services and improving living standards (see, e.g., EU program

“Horizon 2020”). However, such investments do not always attract private investors mainly for two reasons.

First, their private outcomes from these projects are characterized by a high level of uncertainty (e.g.,

Fankhauser, Smith, & Tol, 1999). Second, these projects typically imply social benefits that private agents do

not take into account.

2

In this view of privately provided public goods (Andreoni, 1988a), public funding can correct market failures,

giving private agents incentives for considering positive externalities for the society (Gonzalez, Jaumandreu,

& Pazo, 2005). However, public funding is primarily short‐term and focuses on the early stages of investment

(Borgström, Zachrisson, & Eckerberg, 2016) because the high level of uncertainty that characterizes

innovation activities makes both formal and informal contracts technically unfeasible (Tirole, 1999).

Therefore, the presence of incomplete contracting limits the possibility of regulation and control on ex‐post

outcomes.

The government’s limited ability to affect durable outcomes is confirmed by empirical studies that focus on the

effect of government contribution to innovative activities that are pursued by private firms. These empirical

analyses find no direct effect of public contribution on durable outcomes like workers’and firms’productivity or

the capacity to innovate (Cerqua & Pellegrini, 2014; de Blasio, Fantino, & Pellegrini, 2015; Lerner, 1999). Moreover,

the timing of the intervention (Georghiou, 1998) and private investors’expectations about the level of contribution

1

Quasi‐public goods are goods that benefit both public and private interests. They have the characteristics of both private and public goods, as they may

be semi‐non‐excludable and/or semi‐non‐rival. Examples include roads, social infrastructures and innovations, parks, and public attractions.

2

Consider as an example the issue of adaptation to climate change. Many natural disasters that provoke economic damages are consequences of climate

change (e.g., drought, insect plagues, and windstorms). In the short term, interested private actors (e.g., farmers) should make financial investments to

start adaptation projects (e.g., implementation of new adequate infrastructures). Ecological restoration would be achieved only in the long term, through

efficient management of these projects. However, during the management stage private actors might underinvest in adaptation efforts. In fact, even if

these efforts do benefit first the agents who invest in adaptation, high uncertain effects of climate change in the future and understatement of adaptation

benefits for the society may lead to “maladaptation.”

770

|

ATTANASI ET AL.

(Gonzalez et al., 2005) are found to explain the ex‐post (in)effectiveness of government support. A narrower strand

of literature focuses on the effect of public funds—not just on the total level of private investment/innovation, but

on how investors behave during and/or after the project is implemented (behavioral additionality impact; see

Clarysse, Wright, & Mustar, 2009; Falk, 2007; Luukkonen, 2000). Empirical findings confirm that public incentives

can also have an indirect effect, either positive by facilitating exchanges through informal contracts that allow the

investor to learn the innovation process, or negative by creating a crowding‐out effect on private innovation

investment (Falk, 2007; Luukkonen, 2000).

3

Our theoretical framework models this interaction between public intervention and private agents’behavior in

the long‐term project as a two‐stage game. In the first stage—“investment stage”—a representative private agent (a

consortium of private firms that is in charge of the negotiation process with the State) chooses the initial

investment, that is, the financial contribution to a quasi‐public good (long‐term project) whose public and private

outcomes are uncertain. We assume that these outcomes are perfectly correlated among them,

4

and positively

correlated with the amount of effort the private agent makes during the second stage of the project, namely the

“management stage.”Indeed, the private effort increases the probability that the project will yield high rather than

low private and social outcomes ex‐post. To avoid confounding effects, we assume that this probability is

independent of the level of initial investment (private contribution to the quasi‐public good). However, this level

increases the value of ex‐post outcomes.

5

In line with the above‐discussed short‐term nature of public funding in long‐term projects with social

benefits, we assume that the public agent is active only in the investment stage. During this stage, he or she

sets the amount of public funding to the project. Furthermore, while he or she may observe and always verify

the private agent’s initial investment, the level of effort is neither observable nor verifiable (moral hazard)

because of long‐term and contingent outcomes. As a consequence, we assume that the public agent can

contribute in the short term only by providing monetary incentives that are based on the verifiable level of

investment. Considering this contractual constraint, we model four public‐funding scenarios: benefit ‐sharing

versus cost‐sharing, crossed with ex‐ante versus ex‐interim public intervention. The first variable describes

the type of public funding: The public agent may either grant a transfer that increases the private agent’s

marginal benefit from the investment, or sustain a fraction of the investment cost. The second variable

indicates the timing of this intervention in the investmentstage,thatis,beforeoraftertheprivateagentsets

the initial investment: in the former case—ex‐ante intervention—the level of investment is only verifiable,

while in the latter—ex‐interim intervention—it is also observable.

Our theoretical approach to the private provision of public goods is different from that followed by the

seminal papers of Andreoni (1988a) and Bergstrom, Blume, and Varian (1986). As is well known, they

consider public‐good games with multiple private agents facing a trade‐off between using their wealth to

consume a private good or contributing to a pure public good, with no nature uncertainty about private or

social outcomes. We instead study the private provision of quasi‐public goods whose realizations are

3

In the case of intellectual property, public incentives can, paradoxically, lead to fewer health‐improving products (Heller & Eisenberg, 1998).

4

The assumption of a positive correlation between private and public benefits from the project better represents the kind of long‐term projects analyzed

in this paper, especially when dealing with adaptation to climate change. Consider, as an example, the Noor Solar Power Station (Ouarzazate, Morocco). It

has been built to increase power generation from solar power in Morocco (construction began in May 2013, commercial operations began in June 2017).

Due to a public–private partnership contributing to the project—the Moroccan Agency for Solar Energy and the Spanish consortium TSK‐Acciona‐Sener—

the public benefits—mitigation of greenhouse gas emissions and local environment impacts, green growth of local firms, improvement of the country’s

energy security—are strongly correlated with private ones—revenues from solar energy provision. We introduce the further assumption of perfect

correlation so as to nullify the discrepancy between ex‐ante equality and ex‐post equality of private and public benefits, which might generate undesired

confounds in the behavior of agents with social preferences, for example, motivated by inequity aversion (see Sections 2.4 and 2.5).

5

Andreoni, Kuhn, and Samuelson (2018) analyze a two‐stage game where in the second stage a two‐period prisoner’s dilemma is played, and in the first

stage the players can choose the allocation of the stakes across the two periods of the prisoner’s dilemma. They provide experimental evidence of a higher

joint payoff in the second stage if in the first one the stake allocation is endogenously chosen rather than randomly determined. Although in a completely

different setting, also in our two‐stage game the first‐stage choice—initial investment—positively affects the (private–public) joint payoff in the second

stage.

ATTANASI ET AL.

|

771