Meeting the challenge of impact investing: how can contracting practices secure social impact without sacrificing performance?

Abstract

Impact investors seek financial return and positive economic, social or environmental impact. While certain segments of the impact investing industry see strong returns and high positive impact as conflicting objectives, others argue that strong performance across both objectives can be simultaneously achieved. Despite the importance of this question to the industry, no widely accepted framework exists that enables impact investors to determine whether and how they will be able to invest without a trade-off. This article draws on contract theory and the analysis of incentives in multitask principal–agent relationships in an attempt to fill the void. For impact investors, this article pioneers a theoretical framework for the discussion of strategy, offering a first set of hypotheses that can be elaborated through further theoretical work and tested against actual experience. Within contract theory, this article fits most closely alongside the applied multitask literature concerned with an agent assigned to a financial and an environmental task (Sinclair-Desgagné, B., and H. L. Gabel. 1997. “Environmental Auditing in Management Systems and Public Policy.” Journal of Environmental Economics and Management 33 (3): 331–346; Lothe, S., I. Myrtveit, and T. Trapani. 1999. “Compensation Systems for Improving Environmental Performance.” Business Strategy and the Environment 8 (6): 313–321; Lothe, S., and I. Myrtveit. 2003. “Compensation Systems for Green Strategy Implementation: Parametric and Non-Parametric Approaches.” Business Strategy and the Environment 12 (3): 191–203), while the theory referenced within bridges models from Holmstrom and Milgrom (1991. “Multitask Principal-Agent Analyses: Incentive Contracts, Asset Ownership, and Job Design.” Journal of Law, Economics, & Organization 7: 24–52), Feltham and Xie (1994. “Performance Measure Congruity and Diversity in Multitask Principal/Agent Relations.” The Accounting Review 69 (3): 429–453) and Budde (2007. “Performance Measure Congruity and the Balanced Scorecard.” Journal of Accounting Research 45 (3): 515–539).

Keywords:

• impact investment
• incentives
• social finance
• contracts
• principal–agent

Acknowledgements

I would like to express my gratitude to the supportive team of Gudrun Timm, Marcus Roth, Antje Tetzlaff, Josef Boven, Justus Vitinius, Dr Marc Sasse, Dr Tobias Birkendorf and additional representatives from the Deutsche Investitions- und Entwicklungsgesellschaft (DEG) who provided valuable context for this work over the summer of 2011. In addition, I would like to extend my gratitude to the following impact investors who shared their experience through additional background interviews: Roland Michelitsch and Pablo Verra from the International Finance Corporation (IFC), a representative from the European Bank for Reconstruction and Development (EBRD), Andrew Reicher from the Private Infrastructure Development Group (PIDG), Ed Siegel from Big Issue Invest, Harold Rosen from Grassroots Business Fund, Jonathan Jenkins from UnLtd, and Mark Cheng from Chelwood Capital. Finally, I am indebted to Dr Greg Fischer at the London School of Economics for his insight and support.

Notes

See Mitchell, Kingston, and Goodall (2008) and Joy, de Las Casas, and Rickey (2011) on the case of the UK; Freireich and Fulton (2009) and O'Donohoe et al. (2010) on the supply of capital to impact investing.

See Bolton and Dewatripont (2005) and Tirole (2006).

This constraint prevents the entrepreneur from receiving less than zero and the investor from losing more than she puts in. See Innes (1990).

Bond and Gomes (2009) appear to be the pioneers in the literature in this regard, using a limited liability constraint in a multitask context.

It is worth noting that other bodies of literature have analysed the issue of multi-dimensional performance in an investment context, but are equally imperfect for our purpose. Myriad papers have analysed the empirical relationship between corporate financial and social performance as part of the literature related to socially responsible investing, but statistical analysis has had trouble producing conclusive results (e.g. Garcia-Castro, Ariño, and Canela 2010). The academic literature on the motivation for various forms of ethical, socially responsible, or impact investing does provide an important context for this paper by explaining why impact investors may populate the market (Lewis and Cullis 1990; Anand and Cowton 1993; Lewis 2001; Renneboog, Ter Horst, and Zheng 2008), but does not address our core question as to the drivers of the investment's eventual performance.

Like Feltham and Xie (1994), I use a linear contract but explicitly do not assume its optimality.

The second-order conditions are satisfied by the convexity and concavity assumptions of the cost function and the benefit function, respectively.

See Holmstrom and Milgrom (1991) for the full derivation of the original multitask model.

Holmstrom and Milgrom (1991, 32) establish the condition, assuming a linear contract, such that a partial incentives solution is optimal and generates non-zero impact: Examining the role of the cross-partial effort cost c_xs illuminates the effect of effort–cost complementarities on the optimal financial incentive y_x . The greater the cross-partial, the greater the optimal financial incentive can be.

To avoid confusion with the core model I re-define ϵ_x as a normally distributed random variable with variance σ_x ².

It is necessary to assume that the entrepreneur has a reason to switch technologies. Noting the ‘very important task’ assumption in the core model, which requires the project to produce at least some non-financial output in order to be seen as valuable by the investor, we suggest that the entrepreneur's need for financing will drive this switch.

Budde (2007) assumes a risk-neutral agent for the same purpose. Although congruity is defined in Budde (2007) by the surplus the performance measures produce rather than directly by the measures’ properties, this is not critical for the continuity of intuition across models.

One might ask why the principal does not scrap the explicit contract and only implement a bonus scheme. As Budde (2007, 529) suggests, if some (imperfectly congruent) performance measure is available as a fallback option, it is unlikely that the discounted value of bonus to the principal, in the form of slightly improved congruity, will be greater than the bonus payable to the agent for all efforts exerted. Paying the bonus will no longer make financial sense for the principal. By refusing to pay, she will be forced to rely only on the formal contract and be worse off than she would have been under the hybrid contract.

Additional information

Notes on contributors

Madeleine Evans

The paper was written while the author was in the Master of Public Administration programme at the LSE.