International project finance: review and implications for international finance and international business

Throughout history, mankind has strived to create monumental landmarks in infrastructure and engineering, consistently exceeding the frontiers of what has been technically, organizationally and financially possible. The construction of the Suez Canal, for example, involved complexities that eventually led to cost-overruns of 1900% (Flyvbjerg et al. 2009); and yet the project was eventually realized, in part, because of innovative project finance (PF) structures in place (Beidleman et al. 1990). Modern history provides abundant examples for equally impressive landmarks such as the Burj Khalifa skyscraper, the EURO-Tunnel, the Hong Kong-Zhuhai Bridge, or the Ichthys oil field, offshore of Australia in which a consortium of oil MNCs financed the exploitation of one of the world’s largest underwater natural gas resources, using almost 34 billion USD in capital. In 2015 alone, a total of 275 billion USD loans were invested in PFs (Project Finance International 2016). Many of these projects were implemented in high-risk environments, such as the Chad-Cameroon oil pipeline (Esty 2004a), or the Petrozuata oil field in Venezuela (Esty 1999b, 2004a).

As a natural consequence of the technical, managerial and political risks involved in many of these megaprojects, both management and financing require equally innovative strategies. While PF practice has evolved considerably in recent decades, Management and Finance scholarship have not kept up with this development. Despite globally growing importance, PF has been somewhat under-represented in research (Esty 2003, p. 29; Sawant 2010, p. 1037). This is even more surprising since several reputable authors including Harvard Finance scholar Benjamin Esty and Oxford’s Management professor Bent Flyvbjerg have highlighted the unique, empirical benefits of PF for academic research (Esty 2004b; Flyvbjerg et al. 2009; Kardes et al. 2013; Subramanian and Tung 2016).

The separation of a large and complex infrastructure investment into a financially and organizationally independent project company allows researchers to observe managerial strategies and their outcomes more clearly, and without the empirical distortions emanating from multiple investments within a corporate balance sheet (Esty 2004b; Gatti et al. 2013; Subramanian and Tung 2016). At the same time, the enormous complexity and scale of PF investments make them an ideal research site for managerial decisions at the frontiers of what is economically possible (Flyvbjerg 2014a; Flyvbjerg et al. 2003, 2009; Kardes et al. 2013).

In this structured review, I seek to analyze PF from an interdisciplinary perspective, but with a focus on International business (IB) research. As such, the paper fills important gaps in the literature. Previous reviews on PF have mostly been textbook publications on either financial (Esty 2003, 2004a; Gatti 2013; Yescombe 2014) or managerial practice (Flyvbjerg et al. 2003; Flyvbjerg 2014a, b, c, d), without cross-fertilization or discussion of reciprocal implications. In particular, there is no literature review to date that focuses explicitly on the implications of PF on IB scholarship. Second, all of these publications were conceptual in nature, focusing on practical characteristic of PF. To date, there is no systematic literature review that discusses empirical findings in a structured manner. Finally, most of these publications were directed at practitioners and students. As such, they did not elaborate on theoretical implications that emanate from PF for scientific research (two finance related exceptions being Esty (1999a, b, 2004a, b)).

The aim of this literature review is to integrate finance, management and IB research on PF, to identify reciprocal theoretical implications and derive an interdisciplinary research agenda. Along with PF practice, the fields have developed considerably over the years, with changing emphasis, theories and applications. In addition to this within-field heterogeneity, the questions posed by researchers across disciplines vary considerably. A quantitative literature review would not adequately capture such complexities. Therefore, I rely strongly on a qualitative literature review to develop the theoretical contributions and the broader research agenda. However, I also present aggregate and quantitative data on PF research to give a broad overview of the field.

Following a literature-based definition and discussion of PF as a distinctive form of investment, I systematically review the literature on PF in Management, IB and Finance disciplines. Overall, these findings provide considerable support for the benefits of PF in high risk environments. I then discuss the implications of PF research for the most relevant theories in these disciplines. Therein, I make two important contributions: First, by applying theory to the phenomenon, I deductively identify important questions for PF practice and ultimately for PF research. Second, and in a more inductive approach, I analyze the structures used in PF to derive important theoretical implications for the disciplines under review. In closing, I identify the most promising avenues for further scientific inquiry and make a case for more interdisciplinary research on PF.

From a practitioner’s perspective, the most important definition of PF comes from the Basel Committee on Banking Supervision¹:

Project finance may take the form of financing of the construction of a new capital installation, [...] The lender is usually paid solely or almost exclusively out of the money generated by the contracts for the facility’s output [...]. The borrower is usually an SPE (Special Purpose Entity) that is not permitted to perform any function other than developing, owning, and operating the installation. The consequence is that repayment depends primarily on the project’s cash flow and on the collateral value of the project’s assets. (Basel Committee on Banking Supervision 2004, p. 61)

The primary addressees of BASEL II are banks who provide debt to a project. The accord outlines two constituting characteristics of PF: First and unlike traditional corporate finance, a PF loan is awarded to a separate project company that is financially and organizationally separated from the sponsoring firms. Second, in true PF, banks relinquish the right to recover the loans from the equity sponsors in the event of default. Loan securitization is based solely on the projected cash-flows of the project and not on the creditworthiness of the investing sponsors. As a result, banks take on extensive risks related to the non-performance of the project.

Focusing strongly on cash-flow related lending and non-recourse characteristics, the Basel II definition fails to fully incorporate the equity providers’ view. In this regard, academic scholarship has been somewhat more inclusive, highlighting the pivotal importance of risk sharing strategies as a third component of PF (Brealey et al. 1996, p. 25; Nevitt and Fabozzi 2000, p. 1). Most PF investments involve large scale investments in location-specific assets and in highly uncertain contexts (Esty and Sesia 2011; Hainz and Kleimeier 2006). For corporate finance, this makes projects prohibitively risky. As a result, they seek to identify the most important sources of risk a-priori and devise risk management strategies (e.g. contracts, equity participation) to address them. Such strategies transfer certain sources of risk to parties that are more capable of managing them (Beidleman et al. 1990; Brealey et al. 1996; Farrell 2003; Miller and Lessard 2001). In combination with separation of the project and non-recourse, cash flow-related lending, risk sharing allows sponsors to make investments that otherwise would be economically unfeasible. By outsourcing the project to a legal entity without recourse, sponsors protect their main operations from costs of financial distress (Subramanian and Tung 2016). In addition, the high level of leverage increases return on equity, and makes the PF investment a lucrative business for sponsors (Brealey et al. 1996).

A necessary precondition for PF is that sponsors and lenders reach a-priori agreements on the distribution of cash-flows that is mutually beneficial (Esty 1999b). Such an agreement is difficult to reach, especially when projects involve a large number of participants with different interests. As a result, and as a fourth characteristic, PFs involve very high transaction costs related to a-priori legal and planning costs. Larger projects commonly involve a multitude of parties on the equity side such as facility developers, operators, suppliers and contractors (Brealey et al. 1996; Gatti 2013; Yescombe 2014). These sponsors are complemented on the debt side by relatively powerful lending consortia which often include commercial banks, institutional investors, export credit agencies and multilateral development banks. They contribute almost 80% of capital and seek high levels of control over project management. While equity providers are motivated by the potential upsides of projects, banks capitalize on high fees and relatively fixed interest rates. This asymmetry in incentives creates a potential source of conflict and considerable agency costs, especially in the event of financial distress.

In large projects with high socio-economic importance, additional tiers of agency conflicts arise (Flyvbjerg et al. 2009). For example, PFs often incorporate buyers and local stakeholders in their capital, or contractual structure (e.g. offtake agreements, environmental compensation, and social programs). Most prominently, the local government often takes on a crucial role as capital provider, owner and/or guarantor (in these cases PF is often referred to as private-public partnerships or PPP). In international PF, where foreign firms seek to implement a large-scale project, foreign exploitation collides with local ownership of resources and territorial sovereignty. This results in a fifth characteristic of international PF: high risk of obsolescing bargaining (Vernon 1971). Once the facilities are installed, the local stakeholders can abuse the vulnerability of investors to renegotiate contracts and extract cash-flows from the investment. Large projects, in particular, are exposed to considerable pressure from the media, NGOs, local communities and ultimately taxpayers (Henisz 2002, 2014; Marquis and Raynard 2015). Given these risks, the question remains as to why PF has succeeded in becoming one of the most important financing strategies for high-risk investment.

The origins of international PF can be traced back as far as the Middle Ages, when the British Crown financed silver mines through non-recourse loans from Italian merchant banks (Esty et al. 2014; Kensinger and Martin 1988). However, wide-spread practice of PF did not start until the early twentieth century oil field exploration in the United States (Esty et al. 2014). Modern PF, as a financial risk management tool developed in the 1980s, when it became the primary means of financing large energy projects in the US. Between 1991 and 2012, PF raised over 2.5 trillion USD to fund more than 6000 international projects (Subramanian and Tung 2016).

The financial crisis in 2008 has led to a considerable contraction in PF lending and a redistribution of projects towards the Asian continent (Esty et al. 2014). However, with the exception of the project bonds market, PF lending has proven remarkably resilient to financial market turmoil. By 2013 PF loans recovered from 249.3 billion USD in 2009, to a pre-crisis level of 415 billion USD (Esty et al. 2014).² This swift recovery highlights the durability of PF under high risk environments.

Table 1 illustrates the distribution of investments across regions. It shows the comparatively high shares of investments in high risk-regions in Africa and Asia and supports the common claim that PF is particularly useful in in countries with weak institutions. For example, in 2014, more than a quarter of investments in the Americas were directed towards high risk countries in South America (Project Finance International 2016). The most comprehensive sources of PF information are the databases maintained by Dealogic and Thomson Reuters (Byoun et al. 2013; Corielli et al. 2010; Lerner et al. 2008). Within their sample of investments, the average Economist Intelligence Unit political risk score of host-countries (33) was comparable to South Africa.

Table 2 summarizes the sector distribution of PF investments. The largest sector for PF is power generation (38%), followed by transportation (22%) and oil and gas (20%). These sectors share a number of characteristics: First, they require sizeable investments that justify the considerable legal and administrative costs that are involved in PF financing. Second, they involve location-specific assets which, once installed cannot easily be recovered from a potentially risky sovereign territory. Third, their outputs are reasonably stable, allowing for a-priori valuation and cash-flow based lending.

Scholarship predicts a growing importance of PF. On the one hand, developed governments face budgetary constraints and have become increasingly dependent on private sector funding. On the other hand, developing countries are eager to close the infrastructure gap (Esty et al. 2014). At the same time, the re-emergence of expropriation concerns provides a strong incentive for private companies’ incentives to use PF, as a means of addressing political risk.

In addition to economic scale of PF, there are strong socio-political and arguments that highlight the importance of PF. Despite justified criticism of ex-post development effects and the distribution of wealth (Arndt et al. 2012; Lenz et al. 2017), PF has allowed developing economies to construct infrastructure projects, which would not have been possible through public funding (Brealey et al. 1996; Kleimeier and Versteeg 2010; La Cour and Müller 2014). The critical views on PF also underline the pressing need for further research on PF. Therein, researchers can benefit from the unique empirical setting.

The dearth of empirical research on PF is surprising, since there are compelling arguments that PF provides a uniquely suitable “research site” for both financial and managerial research (Esty 2004b, p. 214).

First, the organizational and financial isolation of investments, with limited life-time provide a “clear window” through which to study managerial decisions and their outcomes (Esty 2004b, p. 214). I refer to this as the clarity benefit of PF. In the study of firm behavior, multiple projects are “commingled” within a balance-sheet, and these projects are often exposed to very different risks (Subramanian and Tung 2016, p. 154). The portfolio of projects changes over time and it becomes more difficult for researchers to trace back variations in performance to past decisions. Because of the multiplicity of projects, the risk of omitted variables and spurious findings is higher in the analysis of firms compared to PFs.

PF provides an attractive controlled environment free from various influences that are present in corporate finance. As a stand-alone entity, PF’s structural details [and the performance outcomes] are easily observable to outsiders, whereas structural decisions of corporations can be obscured by other corporate activities (Byoun and Xu 2014, p. 124; Esty 2004b, p. 217).

The same isolation benefit applies to the local context. While firms commonly operate in a multitude of institutional contexts, the institutional environment of a PF can be isolated much more effectively. This facilitates measurement and improves causal inference. As a result, PF lends a uniquely clean setting for the analysis of institutional effects on the governance and performance of foreign investments. In addition, the special purpose nature of PF means that the project company and its’ governance structure is initiated as a clean sheet and for a clearly defined special purpose. Hence, it is influenced to a much lesser extent by pre-existing structures and corporate imprinting (Marquis 2003; Marquis and Tilcsik 2013). In PF, partners join to create a new venture under a tailored governance structure. They are assembled as a function of the specific project rather than past interaction. As a result of superior clarity, PFs lend themselves in a unique way to quantitative empirical studies. Support for the clarity benefit comes from empirical results in quantitative finance. Studies have shown that PF loans are priced more effectively than corporate loans (Blanc-Brude and Strange 2007; Kleimeier and Megginson 2000). The separation of the project from the companies’ reduces information asymmetries between the lender and the investment and allows for more efficient credit appraisal. The same information benefit manifests when researchers use PF, rather than companies, as research sites. In sum, PF provides a superior research setting that is free from portfolio effects, institutional overlap and historic precedents and clearly defined in terms of project context.

In addition to empirical clarity, the application of PF at the frontiers of managerial, financial and technological possibilities provides opportunities for researchers to ask questions which do not apply to day-to-day corporate decisions; or which are simply not observable in less extreme contexts. I refer to this as the complexity benefit of PF. Megaprojects introduce managerial challenges, which are potentially valuable for both Finance and Management research (Esty 2004b; Flyvbjerg 2014a; Kardes et al. 2013). For instance, whereas traditional research on inter-firm cooperation focuses almost exclusively on dyadic forms of cooperation such as R&D alliances (Ahuja et al. 2009; Eisenhardt and Schoonhoven 1996; Gulati 1995; Lavie and Rosenkopf 2006). PFs allow researchers to analyze cooperative investments with a much larger number of participants. For example, there are compelling theoretical arguments that the social dynamics of such multi-partner alliances may be very different from fairly simple dyadic settings (Chung and Beamish 2012; Heidl et al. 2014; Mohr et al. 2016). Flyvbjerg et al. (2009) show that the multi-partner nature of large projects leads to collective delusion and deception, as multiple participants seek self-interest in situations of diluted accountability. This results in what he refers to as, the “iron law of megaprojects: Over budget, over and over again” (Flyvbjerg 2011, 2014b, p. 11)

Unlike the clarity benefit, the complexity benefit materializes most often in qualitative research, which is more capable of capturing such questions. For example, case studies of mega-projects have a long and fruitful tradition in PF research (Ahola and Davies 2012; Chen 2009; Doh and Ramamurti 2003; Esty 1999b, 2001, 2004a; Gkeredakis 2014; van Marrewijk 2007). As Oxford University scholar Bent Flyvbjerg states in his appraisal of megaprojects:

Never has systematic and valid knowledge about megaprojects therefore been more important to inform policy, practice, and public debate in this highly costly area of business and government. (Flyvbjerg 2014b, p. 7)

In sum, PF provides both clarity and complexity benefits for empirical research and promises new applications for Finance and Management theories, as well as managerially relevant insights. Despite the economic relevance and the empirical benefits of PF, research on PF has been scarce and fragmented. In the following, I review PF research in Finance, Management and IB.

I carried out a structured literature review of published journal contributions, using Thompson Reuters Web of Science. The search was limited to papers which either have the terms “project finance” and/or “megaproject” in their title, or as the author-provided topic.³ The search yielded 192 papers that met the criteria. Furthermore, contributions which were not categorized as “Management”, “Business Finance”, “Economics” and “Business” were excluded. I also excluded publications which were not ranked by the Association of Business Schools Ranking (ABA Version 2015), reducing the sample to 78 published articles. I then analyzed the references cited in the papers for articles with more than five occurrences and included these publications if they were sufficiently addressed PF-related topics. This additional step generated a further 21 papers from sources outside Thompson Reuters Web of Science, in particular from the Journal of Applied Corporate Finance. In a first step, I quantitatively analyze the sampled publications in Tables 1, 2, 3, 4, 5 and 6.

As evident from the distribution of papers in Tables 3, 4, 5 and 6, academic research picked up on PF as an important research context long after its economic boom in the 1980s. One practical reason for this delay was data availability. During a time in which empirical research using larger datasets and more sophisticated computational methods flourished, company level-data was becoming more publicly available. Project data, on the other hand, was collected by a selected number of banking services firms, which were available only at high cost.

The breakdown according to publication sources also illustrates the leadership role of Finance Journals have taken in PF research. Using the R bibliometrix package, a summary of author-supplied keywords was generated and is presented in Table 7. The analysis reveals a broad mix of managerial and financial topics. It lends support to the pivotal roles of risk management, governance, capital structure, and contracts in PF. It also highlights the important IB elements of foreign direct investment and political risk.

In a second and more qualitative step, I analyze the publications in historical order before summarizing their implications for related fields. The earliest publications on PF related to practitioner articles in Harvard Business Review and Sloan Management Review from Wynant (1980), Beidleman et al. (1990) and Wells and Gleason (1995). These publications emphasize the leadership role of practice in PF research. From an academic perspective, the origins of PF research can be traced back to two seminal publications: On the Finance side, Shah and Thakor (1987) derived formal proof that PF is a superior form of investment in high-risk context, because it allows sponsoring firms to create a tailored capital structure and isolate project risk. At around the same time, but on a conceptual level, Williamson (1988) published a seminal article encouraging the integration of corporate governance and financial theory in the Journal of Finance. The article bridged interdisciplinary gaps and explicitly connected PF to managerial question of governance, using transaction costs theory. Williamson complements the restrictive formal model of Shah and Thakor (1987) with managerial arguments such as opportunistic discretion, incomplete contracts and ex-post transaction costs. In addition, Williamson makes a strong and explicit case for the empirical superiority of project investments:

Whereas most prior studies of corporate finance have worked out of a composite-capital setup, I argue that investment attributes of different projects need to be distinguished. (Williamson 1988, p. 576)

Despite this prominent encouragement, PF research found much more traction in the financial research community.

The phenomenon of PF has been identified as an important research context by both Finance and Management. However, the theoretical implications that result from PF have not been explicitly outlined, especially for the field of IB. The two research streams have evolved in isolation without mutual acknowledgement of the valuable insights of the other. In the following section, I analyze the phenomenon of PF from both perspectives and derive important theoretical implications. While the focus of the analysis is on Management and IB theory, I begin the review with interdisciplinary and financial perspectives on PF.

Financial research on PF has focused very strongly on the a-priori benefits of structuring investments as PF. Therein, it has not sufficiently acknowledged the ex-post dynamics in complex, multi-partner investments, especially when a project fails to achieve its pre-defined financial goals and as a-priori structures lose their efficiency. Finance research would benefit greatly by integrating findings from Management research and seeking to more thoroughly address deadlocks, cost-overruns and harmful social-dynamics. Ultimately, this would allow Finance to devise more flexible and responsive structures that could improve the efficiency of PF. Including game-theory and real options perspectives more strongly, would enable Finance researchers to learn more about ex-post dynamics of PF and the pitfalls that have resulted in so many project failures.

In addition, institutional perspectives in Finance research have focused almost exclusively on the legal environment of the host-country of the project. IB and Management research, however, have identified inter-partner differences as important barriers to efficient collaboration. Following the seminal contributions of Esty and Megginson (2003) and Esty and Megginson (2000), which focused more on the number of banks in a banking syndicate, research should incorporate distances between banks and ultimately between banks and sponsors, as potential determinants of PF structures. In addition, IB scholarship has made considerable advances in measuring more normative institutions in host countries, which could complement the dominant, formal institutions studied by Finance researchers. Acknowledging inter-partner differences with regard to informal institutions and culture would allow finance researchers to explain arising conflicts in PFs and would ultimately lead to a more nuanced understanding of multi-partner collaboration in uncertain environments.

Given the considerably shorter tradition of PF research, there are numerous possible research avenues for IB and Management research. In the following, I focus my discussions on the three most promising, from my perspective. First, PF provides a practical example of how firms devise strategies to address clearly identifiable sources of risk in international investments. In its analysis, IB research has been very much focused on the much broader concept of uncertainty as a determinant of strategy. Early economic scholarship, however, has been very clear on the difference between unmeasurable uncertainty and measurable (and manageable) risks (Knight 1921). IB research would benefit greatly from sharpening its perspective to differentiate between very broad uncertainty and specific sources of risk that are both measureable and manageable (Müllner 2016). This would allow IB to depart from very broad uncertainty related mechanisms such as hierarchy, control equity towards more micro, contractual risk management tools, with high practical relevance. Finance research has found early on that “financial arrangements cannot be viewed in isolation from other parts of the nexus of contracts” (Dailami and Hauswald 2007, p. 275; Fama 1990). A similar, more holistic approach, including contractual and socio-economic risk management, should be applied to internationalization strategy, in order to achieve higher validity and practical relevance. Using the example of PF as a practical reference point, IB research could learn how firms orchestrate complementary risk management strategies in a way that allows firms to realize investments without the use of control or hierarchy.

Second and in a similar vein, PF provides evidence that, contrary to most applications of transaction cost theory, firms do not minimize transaction costs. In fact, transaction cost economics has been criticized for low managerial practicability, its static nature and for providing false normative prescriptions (Ghoshal and Moran 1996; Rugman and Verbeke 2003, p. 130). PF illustrates that firms base their strategy on risk-adjusted transaction costs and real options (Brouthers and Hennart 2007, p. 403; D’Aveni and Ilinitch 1992, p. 597). Under certain circumstances, as found in large infrastructure investments, risk reduction can be the key driver of strategy, rather than transaction cost efficiency. Since firms differ in the degree to which they can manage a certain sources of risk, cooperating with firms that dispose of superior risk-management capabilities can be economically feasible, despite potentially higher costs. Further research in IB should put more emphasis on important firm-specific advantages in terms of risk management and finance (Rugman 1980). Acknowledging these firm-specific advantages would help IB researchers to learn why some investments are made despite high transaction costs and why some companies are more capable of taking on risks from uncertain (risky) environments.

Third and finally, PF clearly illustrates the pivotal role of lenders in financing foreign investments in risky contexts. This is in stark contrast to contemporary IB research, which commonly focuses very strongly on equity side strategies (e.g. joint ventures). The role of the liabilities’ side and the actors that it inevitably connects to the investment is often disregarded (Agmon 2006; Krugman 1985). PF is practical proof that lenders can serve as political allies and providers expertise. Future IB and Management research needs to address the important role of the liabilities’ side as a strategic complement to equity-based strategies. Additionally, Management research on international investments would benefit from addressing the pivotal role of debt providers in the dynamics of megaprojects’ failures. In his ground-breaking work on finance and governance, Williamson (1988, p. 580) recognized that “debt is unforgiving if things go poorly. Failure to make scheduled payments thus results in liquidation” and yet, the pivotal role of lenders in the failure of projects has not been sufficiently acknowledged in Management or in IB research. In addressing the role of lenders, research could gain important insights on the boundary conditions of strategy. In the long run, the opportunities and resources that are available companies in the process of internationalization depend not only on their assets, but also on their ability to finance risky investments.

Concluding, it can be summarized that PF offers not only a superior empirical setting for Finance, Management and IB research. It also instigates important theoretical contributions for both disciplines. Interdisciplinary perspectives, in particular, can help to advance our understanding of the financing and management of risky investments. PF involves the most complex and challenging projects of our time, often in high-risk environments. It has repeatedly challenged the frontiers of financing and management practice, and it is time that research makes appropriate use of PF in order to test the boundary conditions of existing theories.