Public Private Partnerships for Infrastructure Development: Risk, Reward and Rents

The Three Rs of Public Private Partnerships for Infrastructure Development: Risk, Reward and Rents

Introduction

‘The links between infrastructure and development are well established, including the impact of infrastructure on poverty alleviation, equality, growth and specific development outcomes such as job creation, market access, health and education’(WBG 2015a, p3)

The link between infrastructure development and economic growth is almost universally accepted (first Aschauer 1989 and followed by Watson 1998, Calderón and Servén 2004, Bhattacharya et al. 2012, Garsous 2012, amongst others). The WBG estimate the annual infrastructure financing gap to low and middle-income countries to be approximately $1trn (WBG 2015b, p4). Within that group, the worst affected region is SSA whose annual gap is estimated at $100bn (BCG & AFC 2017, p8). Forecasts estimate that closing the infrastructure gap would lead to a cumulative increase in growth for the region of 2.6% of GDP per year[1] (BCG & AFC 2017, p8). Unsurprisingly therefore, infrastructure development generally, and specifically in SSA, is high on the development. For example, the 2017 edition of the WBG’s high profile ‘Africa’s Pulse’ was focussed on infrastructure in SSA (WBG 2017a).

Traditionally infrastructure has been seen as the archetypal public good, in some cases a merit good. Historically that has meant public provision, and therefore public financing. This line started to blur in the early 1980s as public infrastructure in many countries was privatised. In the post 2008 financial crisis era public budgets have become stretched in both advanced and developing countries, and bank lending has become more constrained (Inderst and Stewart 2014, p2). Public financing of the infrastructure gap appears unrealistic. The WBG’s plan to close the gap is clear; PPP. With an estimated $8.7bn of private investment in SSA infrastructure in 2013 versus $0.1bn in 1989 the policy has increased resources available, however comparing this to the funding secured by London’s Crossrail project in 2009, $23bn (Economist 2015, p5), it is clear there is significant work to do in terms of improving SSA’s access to infrastructure finance.

PPP refers to a range of arrangements between public provision and full privatisation, however there is no universally accepted definition. For the purpose of considering a broad body of literature and opinion from different regions and organisations this document uses a broad ‘catch all’ definition of PPP, in line with Kwak et al. (2009, p52) ‘a cooperative arrangement between the public and private sector that involves the sharing of resources, risks, responsibilities and rewards with others for the achievement of joint objectives’[2].

PPPs as a concept are not new. Some point to concessions for road construction in the Roman empire as the first example of private provision of public services (UNDESA 2016, p2). In terms of the development discourse the concept of PPPs is a relatively recent innovation. Privatisation of government services became in vogue in advanced countries in the early 1980s, with PPP in the modern sense following in the early 1990s. This had a knock-on effect on the development discourse with the WBG endorsing the policy in its 1994 WDR (WBG 1994).

As noted above, in PPP transactions the public and private sector share ‘resources, risks, responsibilities and rewards (Kwak, Chih and Ibbs 2009, p52)’, which can be simplified for the purpose of analysis to risk and reward (resources and responsibilities are operational concerns, risk and reward determine investment decisions). We define risk in a technical corporate finance sense as ‘variance in actual returns around expected returns’ (Damoradan 2009). Institutional investors require compensation (reward) proportional to the risk they accept in a transaction. To the extent that they receive reward in excess of accepted risk, we define that as a rent.

This paper does not step into the moral and ideological quagmire of whether the private sector should have a role in infrastructure provision. To the extent that one accepts that there is a private role in infrastructure provision one must also accept that private enterprise is profit driven, if the sector is not profitable then there will be no private investment.

We can be relatively certain that individual sunk investments in infrastructure do not earn lower rewards than the risks they accept, if that were the case they would not have invested. However, can we be sure that investors do not earn rents?

The WBG argue that efficiently run bidding processes and competition prevent the existence of rents. Theoretically competition for investment should compete returns down to the lowest level that investors are willing to accept, which we theorise is where rewards are equal to risk accepted. This is currently impossible to prove. The returns and (crucially) expected returns of investors in infrastructure are covered by ‘commercial confidentiality’, with few minor exceptions[3] PPP investors are not required to disclose expected or realised returns on their investments. There are ways of estimating average returns, for example as of August 10^th 2017 BlackRock’s exchange traded emerging market infrastructure fund was trading 19.55% up year to date (Blackrock 2017), but observing returns on these funds can only tell us broad averages across a wide sector, it tells us nothing about individual projects. Rents in PPP are in effect transfers from the taxpayer to investors. In many developed countries with privatised utilities the profits of the utility providers are under constant scrutiny and intense regulation, the UK announced another independent energy price review in 2017 (FT 2017). The same level of scrutiny (typically) does not exist in developing countries. Not only is there a moral issue surrounding international investors earning rents at the expense of citizens in developing countries, but also a question of opportunity cost, the resources extracted as rents could have been invested in alternate productive capacities.

This paper draws on a range of WBG publications on PPP to demonstrate how the willingness to conduct deep and meaningful analysis into core structural elements of PPP, such as risk and return characteristics, as opposed to individual implementation issues, has disappeared over time. It then argues there is reason to believe rents may exist in the PPP sector.

In the absence of risk and return data to conclusively prove the rents hypothesis, the paper builds a case by considering tangential but relevant evidence.  It considers evidence from work surrounding the determinants of FDI and the experience of the UK PPP sector to argue that there is sufficient reason to believe rents may exist to justify further transparency and analysis. Where it is relevant to consider specific regions, we focus on SSA as this is where in the infrastructure gap is the greatest. However, this piece is also to some extent a critique of WBG PPP policy, and the WBG PPP policy is targets developing countries generally.

Methodology

Section 1 outlines the evolution of thinking surrounding PPPs from the key development organisations, largely drawing on literature produced by the WBG[4]. It proposes that since PPPs entered the development discourse in the early 1990s we have observed two generations of thinking.

In the first instance the thinking was ideologically driven, fitting in with the NPM agenda and the global drive for small government. Often ideological policy making precludes thorough analysis; for PPPs this was not the case. Private provision of infrastructure was hypothesized to be superior to public provision primarily in terms of efficiency. Such a claim produces opportunity for analysis. During this period WBG PPP publications were mainly concerned with identifying the scenarios in which PPP could be most effective.

The second generation of PPP thinking can be traced to the post 2008 financial crisis period and became ‘supercharged’ in the SDG era. The new agenda is no longer ideological. The argument has become less sophisticated, it states that there is not enough public funding available to pay for the world’s infrastructure needs, private finance is the only viable option. Throughout this piece we will refer to this as the ‘only option’ argument. During this period (which is still ongoing) the WBG are mainly concerned with how to make PPPs work best, regardless of the scenario.

The purpose of this paper is not to question the ‘only option’ assertion in its own right, it presents no alternative infrastructure financing mechanism. Instead it outlines how the argument has led to a backward step in terms of the quality and rigour of analysis into the core elements of PPP, in its place focussing on the conditions to make them attractive to investors. We question whether this drive to improve conditions for investors may have swung the balance of risk and reward too far in their favour, at the expense of the state and tax payers.

Specifically, we highlight the lack of analysis surrounding the appropriate level of returns for private sector investors in the development PPP literature, despite its prominence in PPP discussions in advanced economies. Ideally this piece would have been a rigorous empirical investigation proving or disproving the hypothesis that significant rents exist within the PPP sector, however the opaque nature of PPP deals and the infrastructure funds who invest in them prevents such a study from being a reality. This study finds no smoking gun.

Having set out how PPP thinking has become unsophisticated we adopt a two-pronged argument, highlighting evidence from other areas of study to provide logical reason to expect rents in the development PPP sector. We propose that these arguments offer sufficient reason to justify altering transparency rules and conducting closer analysis of PPP risk and return characteristics.

Section 2 presents evidence of the unrealistic valuations of risk from institutional investors in Africa. It argues that if investors overestimate risk, they will require returns to compensate them for the perceived risk level, not the actual risk level, and therefore earn de facto rents.

Section 3 considers evidence of rents in UK PPPs, evidence which has led to significant change in PPP policy in the UK. If rents exist, or have existed in the UK PPP sector, we argue that is sufficient reason to believe that the same is a possibility in the development PPP sector.

Section 4 briefly considers up to date empirical work to argue that alongside changes in transparency rules, a change of MDB project prioritisation towards projects in countries with low levels of PPP experience may also reduce the possibility of rents.

Finally, section 5 concludes.

Section 1- WBG PPP thinking: then and now

The WBG, or more specifically the IBRD, has been concerned with infrastructure development since its inception, the renewed focus on infrastructure and the $1trn infrastructure gap has something of a ‘back to the future’ about it. In 2014 approximately 40% of the WBG’s total capital employed was on infrastructure projects (Kim 2014), infrastructure is a significant percentage of the WPG’s current activity. Their strategy is clear, in the words of the WBG president ‘to fill this [infrastructure] gap, we need to tap into the trillions of dollars held by institutional investors’ (Kim 2014), in other words – PPPs.

PPP for infrastructure is not a new phenomenon, UNDESA (2016, p2) traces them back to the Roman empire. In the early 90s however the concept of PPP for infrastructure came to the fore in both developed and developing countries. UNDESA (2016, p2) and others link PPP’s re-emergence in developed countries to the prevailing neoliberal economics and new ideas of NPM. They outline the big picture justification, ‘PPPs were often invoked as alternatives to bureaucratic public services and inefficient state enterprises…to enhance the efficiency of the public administration and public service provision’. The adoption of PPP as a policy by the WBG and the development orthodoxy can be linked the Washington Consensus (Akinkugbe 2013, p9-10), the set of development policies born out of neo-liberalism which dominated development thinking in the 1980s and 1990s. The key arguments in favour of PPPs are their potential efficiency savings and the opportunity for capital constrained governments to outsource financing. We argue that the perceived first order benefit has changed over time.

Despite the central position they hold in development policy, PPPs are not universally adored in the literature. Separate from ideological challenges concerned with issues such as the conversion of public services into an ‘asset class’ (Whitfield 2010) there are fundamental structural challenges to the concept of PPPs. Political economy arguments claim that PPPs will never deliver on goals due to conflicting interests and incentives of public and private partners (Vining and Boardman 2014, p17 and Liu, et al. 2016, p1102). They employ a principle agent model to argue that the moment a PPP has been agreed the private sector has an incentive to minimise spending on, for example, maintenance to maximise profit, which they are often able to do due to an asymmetric information relationship. They operate the infrastructure asset, and therefore possess more knowledge. UNDESA (2016, p22) confirm this point in the context of development PPPs by noting how regularly PPPs prove more expensive than alternative procurement options, fail to deliver promised efficiency and cannot demonstrate VFM. Private provision of infrastructure can also lead to inequality of access (Walsh 1995). In the case of utilities, sales can rarely cover costs, meaning projects have to rely on government subsidies to remain profitable, which can prove problematic if the government falls into financial difficulty (Klein 2015, p1).

Aside from these fundamental structural issues, the literature has identified a broad range of impediments to efficient PPP implementation in developing countries. These include social, political and legal risks, unfavourable economic and commercial conditions, inefficient public procurement frameworks, lack of technical engineering expertise and public sector related problems such as inexperienced negotiators (Zhang 2005, p73 and Yong 2010, p69). Transaction costs can also be prohibatively expensive (Välilä 2005) rendering efficiency savings irrelavent. A further argument proposes that there is some mimimum level of development required before PPPs become appropriate tools (Tan 2011).

Despite these significant challenges and criticisms PPP has remained at the centre of infrastructure policy in advanced and developing countries for more than 20 years. This section seeks to demonstrate the shift in the WBG’s justification for PPPs, and the way in which the depth of analysis surrounding development PPPs has changed in response to this shift. We compare two periods, the first is the period loosely beginning in 1994 and ending around 2007. The second period begins in the aftermath to the global financial crisis, and is ‘supercharged’ in 2015 by the SDGs.

Primarily we argue that while early PPP advice was based largely on ideology, the promise of efficiency gains led to inevitable consideration of the gravity of these gains. The policy may have been ideologically driven, but it opened itself to analysis. Both development agencies ostensibly in favour of PPPs and critics from a variety of fields wrote papers either outright damning the policy, or highlighting unsuccessful projects and drawing conclusions as to which types of infrastructure may be most suitable for PPP. In contrast, the latter period is coloured by the ‘no option’ argument which has resulted in skin-deep analysis surrounding how to make the best of PPP, as opposed to any effort to consider its wider viability. Although he does not identify a change in rhetoric over time, Whitfield (2010, p7) sums up this argument by describing the WBG’s publications on PPP, they ‘focus on individual projects, the PPP process, deal flow, procurement, legislation, and project finance. This effectively limits, either by design or default, debate of broader public policy issues and longer-term implications’.

Period 1: 1994-2007

The 1994 WDR was specifically focussed on infrastructure and included sections on PPI. It was not the first time the WBG had written about PPP, Kessides (1993) (writing for the WBG) foreshadows many of the arguments in the WDR, there were already developing countries using PPPs prior to 1994 (e.g. Mexico Toll road). However, PPP being placed in the most high-profile WBG publication is significant and is an appropriate sign post for the idea’s entry into main stream development discourse.

The WDR summarises its key arguments into six points (p2). The first point notes, ‘Infrastructure can deliver major benefits in economic growth, poverty alleviation, and environmental sustainability – but only when it provides services that respond to the effective demand and does so efficiently.’ The implication being that the private sector is the group able to respond effectively to demand. Of the six key takeaways listed, five discuss the efficiency benefits of private sector provision (the final one notes the government’s role in actively pursuing PPP). The primary argument for the adoption of PPPs in 1994 was efficiency gains.

The second key issue, constraints to government funding, is mentioned, but only once (p93) and not in the key takeaways. The idea of substituting constrained government finance for private funds was a factor, but in 1994 it was a secondary concern.

In 1994 the WBG were also conscious of the challenges associated with ensuring returns to investment remain proportional, albeit they choose a ‘performance’ as opposed to risk measure; ‘investors’ returns should be linked to project performance, and any government guarantees needed should be carefully scrutinized’. It is important that they consider the incentivisation of governments. Governments can lower the observed costs of a PPP by accepting more of the risks via government guarantees. Government guarantees are off balance sheet contingent obligations, governments seeking to report an observationally good deal to their electorate may choose to accept more risk than is prudent or efficient. While this study is primarily interested in examples of private sector rents, it is important to note that in 1994 the WBG recognised how both party’s incentives can negatively affect the outcome of a PPP.

Many publications have followed, many of them regionally focussed designed to learn from success and failure of PPP. PPPs are hugely complex, there was a desire to learn what did and didn’t work. For example, a thorough assessment of a number of programs in Bangladesh led the WBG to conclude ‘the introduction of public private partnership … is more likely to be successful in neighbourhoods which are high in social capital’ (Pargal, Huq and Gilligan 2000, p24). Key here is there is no proposal of how to make neighbourhoods with low social capital more accepting of PPPs. The aim was to identify the areas best able to take advantage of PPP efficiency savings, not work out how to shoehorn PPPs into any situation.

In 2001 the IFC looked deeper into the promise of greater efficiency in the private sector (Everhart and Sumlinski 2001). They note a vast literature identifying both crowd in and crowd out effects of public investment in both developing and OECD countries[5]. Their analysis seeks to deal with the idea that arguments concerning the inefficiency of public investment could be overstated if there are additional catalytic crowd in effect which should be considered in parallel.

The papers discussed are only a window into the literature (a search for the unique phrase ‘public-private partnership’ in the WBG archives returned in excess of 500 articles for the period 1994 – 2007), however they highlight the view at the time. There was a willingness to see PPP as a tool, not the tool and a willingness to question the claim of relative efficiency.

In 2004 the WBG (specifically the PPIAF) published an article in response to an argument which was gaining attention in the press that infrastructure investors in Latin America were making excess profits from their investments (Sirtaine, et al. 2004), the very claim we aim to consider in this paper.

The PPIAF are given returns data by a range of investors. They note that there are three theoretical opportunities for investors to extract rents, but find no evidence of inappropriate returns. They identify that ‘a rate of return in excess of the cost of capital inappropriately penalizes consumers, while a rate of return beneath the cost of capital inappropriately discourages further investment’ (p.37).

The first opportunity to earn rents is at the initial negotiation stage, the second group of opportunities are the pre-agreed periodical pricing changes. Contracts will typically outline a range of scenarios, and potential price rises for each scenario, as a result there are ‘incentives for concessionaires to dress up their accounts in order to present the lowest profitability or returns possible’ (Sirtaine, et al. 2004, p10). The third opportunity for excess returns is at renegotiation. The report notes that there may be an incentive for investors in a competitive bidding process to be overly aggressive in their initial bid, knowing that when their investment proves unprofitable the government will be willing to renegotiate in their favour.

Sirtaine et al. propose that these latter two opportunities are the responsibility of the regulator to manage. This places significant emphasis on the skill and experience of the regulator. They argue that the regulators in Latin America were sufficiently skilled to prevent rents. The discussion of the UK in section 3 will question whether this can be assumed to be the case generally in developing countries.

The paper suffers from an inevitable challenge that ‘Estimated returns are highly sensitive to estimates of capital costs and investments’ (Fay and Morrison 2007, p37). If one chose a different collection of assumptions then the results would be entirely different, however this does not invalidate the exercise. Conducting the analysis both allows the debate to be based on evidence and shows a willingness to hold the private sector to account. The importance of Sirtaine et al’s work (which was commissioned by the WBG) is that it engages with the concept that there is a necessity to control and regulate the private sector to ensure appropriate levels of returns. While they find no evidence of rents, they accept the incentivisation structure which could produce them, and discusses how good quality regulation and regulators are necessary to keep the incentives of the private sector in check. A report of this type published annually and distributed broadly could act as a check on investor rent seeking. At the time of writing it is 13 years since Sirtaine et al.’s paper was published, given that it recognised opportunity for rents, it is hard to understand why no follow up has been produced.

Space constraints preclude the discussion of every document produced by the WBG concerning PPPs in this period. The four discussed outline the constructive approach to PPPs from the WBG at the time. They were viewed as a useful tool to achieve the infrastructure development which could be more efficient than publicly funded infrastructure. There was however a willingness to subject the fundamentals of the policy to analysis, and a healthy level of scepticism surrounding private sector investors.

Period 2: Post financial crisis – present day

In 2017 analysis is less nuanced. The passage of reasoning is displayed in figure 1. The logic is considered self-evident and irrefutable. As a result the literature has largely shifted away from an analysis of the appropriateness of PPP, and toward methods of encouraging and incentivising private funds into infrastructure. This transition was a two-step process, beginning in the aftermath of the financial crisis, and becoming set in stone around the time of the adoption of the SDGs.

In 2011 the PPIAF published an extensive PPP ‘how to’ guide (PPIAF 2011) which demonstrates a subtle change in attitude from the pre-crisis years, as such it is an appropriate place to begin the analysis of ‘period 2’.

The 2011 document is extremely comprehensive. In the preamble to a section outlining the main advantages of PPPs, the paper states, ‘the estimated demand for investment in public services shows that government and even donor resources cannot fill the investment gap alone, and so harnessing private capital can help to speed up the delivery of public infrastructure’. While the analysis is likely accurate, the way in which the point is deployed, separate from the list of pros and cons of the PPP structures, leads one to believe that it is objective truth, it is neither a pro nor a con, it simply is. This is a precursor to the clearly stated ‘no choice’ argument which we will consider shortly.  It subtly changes the subject of PPP dialogue away from when and where is PPP an appropriate financing structure to – how can we make PPP an appropriate financing structure.

It is significant that PPIAF are willing to offer advice on a vast range of topics surrounding the successful implementation of a PPP program from how to set up a PPP agency within the government, to how to communicate government objectives and priorities with investors, but the only advice on VFM is about how to run a competitive tender process, the assumption being that competition will inevitably lead to a good deal for the public sector. This is surprising given the organisation’s acceptance 7 years previously that an overly competitive bidding process could lead firms to undercut the levels at which they can meet their own costs, in the expectation that they will be able to renegotiate a better deal a few years into the contract (Sirtaine, et al. 2004, pxi).

In 2015 the MDBs and the IMF met to discuss a new way of managing development finance in a world of constrained resources and ambitious SDGs. They declared that it was a ‘paradigm shift’ in development finance (The Development Committee 2015). This is the point the ‘supercharged’ ‘only option’ argument began. The Development Committee’s publication is not exclusively about PPP, or infrastructure, it is about how development finance should be approached in the face of a new set of goals, however those goals do include infrastructure development, and as such the publication does discuss infrastructure. The institutional view on the private sector, and how the government should interact with it is set out very clearly.

“The drivers of private finance are distinctly different from the motivations of domestic public finance. Private sector firms seek investment opportunities based on risk-return considerations. To be effective, public sector measures to encourage private investment need either to decrease perceived risk or increase anticipated returns. Governments play a critical role in providing a conducive investment climate.”(p12)

The Development Committee observes insufficient levels of private investment and concludes that this is a result of rewards available being lower than risks present to the private sector. Governments’ (and the IFIs’) role is to de-risk the private sector so as to balance out this equation. A scenario where reward may exceed risk is not considered.

The assumption that competition will compete away rents is open to debate given low levels of competition often observed during tendering processes (Pulido and Hirschhorn 2015). More striking is how the opportunities identified in 2004 (Sirtaine, et al.) for firms to earn rents (via pricing changes and renegotiations) have been forgotten in 2011.

Analysis surrounding PPPs is now almost entirely focussed on how to increase private investment. How can governments and development agencies further de-risk investments? How can more palatable regulatory regimes be built? How can governments make credible long-term policy promises which are believed by the market? The WBG now spends limited time writing about how to maximise the benefits of privatisation, instead column inches are occupied with the task of increasing investment at all costs.

To demonstrate the weakness of analysis, a study of 18 post crisis WBG case studies of ‘successful’ PPPs found that only 1 utilised a non-PPP counterfactual for comparison, all the others declared success on the basis of before/after analysis (Bouman, et al. 2013, p28). The focus on case studies and ‘lessons’ learned in the literature, especially that being produced by the WBG, has been weakened by the ‘only option’ argument which assumes there is no plausible counterfactual. Efforts to learn from success and failure are common to both periods, but the framing of discussion differs. In the pre-crisis period case studies allowed us to understand where PPPs are most relevant, whereas in the post crisis period case studies are focussed on how to make PPPs work in a variety of scenarios.

In a 2003 paper focussed on PPP theory the WBG explained very clearly that the best results were achieved when the private sector was responsible for commercial risk (Harris 2003, p25). The ability to allow private firms to manage commercial risks was one of the key sources of efficiency gains. In 2017 the WBG accept that it may sometimes be ‘difficult’ for the private sector to accept commercial risk and that governments may have to accept all or part of the risk (WBG 2017b, p16). Marques, Cruz and Cunha (2011) note how governements are increasingly willing to accept commercial risk in ensure deals are closed, allowing investors to receive high returns, despite only accepting minimal risk. Governments are incentivised to accept these risks for the same reasons as they are incentivised to engage in PPP to begin with, the risks are off balance sheet contingent obligations and do not harm their surface level observed financial position.

One of the WBG’s recent trade mark programs is their MCPP facility. The Financial Times summarises the program as ‘World Bank arm aims to tempt institutions into emerging market projects with promise to limit losses’ (Sender 2016). The concept is that institutional investors invest in a fund, managed by the IFC, and the WBG accept the first 10% or any losses experienced. Essentially it is a tool to de-risk investment in infrastructure in order to encourage further investment. The strategy in itself is difficult to criticise, but it is representative of the fundamental 2017 conviction that the private sector must be incentivised to invest at any cost, and that to achieve this, the risk return equation must be tilted in their favour.

Governments are incentivised to be seen to be building infrastructure, it is good way of gaining attention from voters (Hodge and Greve 2010, p347). It is important to recognise the WBG has similar incentives. The WBG is judged on how many roads have been built in a year, or how much investment has been secured. Reports do not judge the WBG on the basis of VFM or investor rents, it stands to reason therefore that the WBG should be more concerned with attracting investment to projects than ensuring that investment is secured at a fair price.

Section 2 – Africa’s image problem

‘More than any other continent, Africa has an image problem’(UNCTAD 1999, p182).

‘The HLP believes that the actual risk of infrastructure projects in LICs and MICs is frequently lower than the perceived risk amongst foreign providers of capital’(HLP 2011, piii)

There is a view, supported by research from UNCTAD (1999) and a study for the G20 (HLP 2011), that investors are put off from investing in the African continent because they see it as a place of civil unrest, poverty and corruption. Investors are put off by perceived risk. A logical extension to this argument is that when investors do elect to invest, they expect returns to compensate them from this perceived level of risk. Whether gaining return for an imagined but believed source of risk is a ‘rent’ is a semantic point, however for the purpose of this piece we can treat it as such, as it is a transfer from public to private sector in excess of the actual risk accepted.

In 2013 McKinsey interviewed a number of key players in emerging market infrastructure investment. The then CEO of the Canada Pension Plan Investment board (Mark Wiseman) explained the ways in which his fund make investment decisions (McKinsey&Company 2013). His fund is exactly the kind of ‘long’ fund which the WBG target as potential development infrastructure investors.  Pension funds have long term liabilities, and therefore make long term investments (to match asset and liability duration). As a result, when investing in infrastructure they need confidence that ‘the regulatory framework … have to actually transcend any given government’. The challenge of assessing not only the likelihood of expropriation in the current political regimes, but also future regimes, is the reason Wiseman limits his investment in emerging market infrastructure. However, what if Wiseman and his fund are not well equipped to estimate this risk? They may neglect to invest in projects which they would have invested in if they could better estimate risk. When they do invest they may require compensation for a higher level of risk than really exists. This section will seek to identify evidence that investors overestimate risks to their investments in SSA.

Risk and return in Africa

The paradox of high returns and low investment in Africa has puzzled observers ever since the policies of the ‘Washington Consensus’ contributed to the opening up of the continent to international financial markets. Warnholz (2008) shows empirically that there is no statistically significant relationship between prior realised profits and subsequent investment decisions for investors in Africa which appears to make limited logical sense.

Risk is not easy to measure. In advanced countries with sophisticated financial markets financial institutions have become relatively skilled (although not infallible e.g. the global financial crisis) at estimating and hedging certain risks such as exchange rate risk, however estimating political risk in any given developing country is more challenging. Many of the financial obligations connected to infrastructure have a duration of 25 years or more, investors must consider the risk not only associated to a political regime, but also those that follow. The task is all the more difficult due to lack of resources. Investors in advanced countries benefit from a vast array of financial journalism and data which is not matched in individual developing countries.

Investors may to some extent assign a blanket view of risk in Africa, and more specifically SSA (Tawia and Walker 2009), failing to consider the significant differences between economies. Further there is a view that Africa and SSA are somehow structurally different to the rest of the world economy, a concept which is often referred to in the literature as the ‘Africa Dummy’. While this is not a subject which has previously been discussed in any great depth in the PPP literature, there are several quantitative and qualitative studies related to the more general field of determinants of FDI which address the issue.

If FDI investors overestimate the level of risk in SSA generally, then the same is likely true for investors in infrastructure. If this is the case then the WBG policy of decreasing risk may be effective in increasing investment, however it also facilitate firms earning de facto rents. Schleifer (2000) even suggests that this gap in risk perception offers arbitrage opportunities that certain investors actively seek.

Quantitative evidence

Haque et al. (2000) use regression analysis to assess the theory that investors systematically overestimate risks to investment in Africa. They analyse the sovereign credit ratings of the universe of developing countries. Sovereign credit ratings and investor risk perception are not necessarily identical; however investors are unlikely to deviate dramatically from the agencies when making their own political and country risk assessments.

In a series of regressions including economic variables and omitting political variables there is evidence of a significant Africa Dummy, that is African countries are rated as riskier than countries in other continents with identical economic indicators.

In an alternative specification, without region dummies, but with political variables (coups, government crises and revolutions amongst others) the authors found that only about half of the proposed measures were statistically significant, and that they added very little to the R² of the regression, they argue that political variables are second order concerns when it comes to ratings calculation.

The authors do not then elaborate on their findings. They do not present a specification combining both the Africa Dummy and political variables, they also don’t include any interaction terms between the dummy and the explanatory variables which may have allowed us to see how individual economic and political factors affect risk perception in Africa differently. We can conclude from Haque’s work that investors do view Africa as inherently riskier than other apparently economically similar developing regions, but can make little or no inference surrounding which factors contribute to this perception.

If one accepts the idea that the primary variable preventing increased FDI in the African continent is risk, or perception of risk (Collier and Pattillo 2000, p3) then FDI level and risk perception are not quite two sides of the same coin, but they are heavily linked. While Haque et al. appear to be the only authors to identify the existence of an Africa Dummy in sovereign credit ratings, there is a wider body of work looking at the determinants of FDI, and identifying an Africa Dummy there.

Asiedu has undertaken extensive work in this area, also summarising a number of existing studies[6] (Asiedu 2002). Her 2002 article uses regression analysis to show how changes in key economic and political variables identified in the literature affect FDI in SSA differently than they do in other regions. She shows how changes in return characteristics and physical and regulatory infrastructure make no significant difference to FDI levels in SSA, when they do in other regions. Even for those which do affect FDI in SSA, such as trade openness, the marginal benefit in SSA is lower than other regions. This she proposes is evidence that Africa is perceived as ‘different’ by investors.

Asiedu then considers a number of interaction terms to show how the impact of openness, infrastructure and returns to investment remain positive and significant for a population of emerging market countries which excludes SSA, yet these are far smaller or statistically insignificant in SSA. She proposes that ‘the inability of countries in Sub-Saharan Africa to attract FDI may be partly blamed on the fact that these countries are located in a continent which happens to have a bad reputation’.  She argues that observed trends are due to a lack of knowledge, risk perception is not calculated on the basis of country specific conditions, but a broad view of the region.

Tawia and Walker (2009) attempt an ambitious effort to show this miscalculation of risk empirically in the context of SSA. Their concept is familiar, international investors take a broad-brush approach to country risk and political risk when it comes to SSA, therefore overestimating risk in the relative good performers in the region. They use a variety of corporate finance models (primarily the ‘Fama-French’ model) to show how models which can accurately predict levels of risk, return and volatility in developed market stock exchanges cannot explain the same factors in a selection of African stock exchanges. They propose this is due to an overestimation of risk in the market. In reality, average stock market returns are remarkably flat in the African countries, leading the authors to believe that risk is overestimated by investors.

Summary

Wiseman speaking in 2013 undoubtedly had access to more market information about African markets than investors did in the early 2000s when Haque, Collier, Patillo and Assiedu were conducting their respective research. Sichei and Kinyondo (2012) employ more modern panel data techniques to identify evidence that while the Africa Dummy remains significant, it is falling in value over time, suggesting that policy and or improved information is having a gradual effect. However annual FDI flows into SSA are now approximately equal to 2000 levels, albeit after significant falls from highs of 2009. During the same time period, global FDI has more than doubled. Whatever progress is being made in terms of policy and information flows has still resulted in SSA falling further behind as a destination for FDI.

One cannot and should not uncouple the wider concept of FDI into SSA from PPPs in SSA infrastructure. To reiterate the importance for the infrastructure agenda, there is reason to believe that investors systematically overestimate risks in SSA and Africa more generally. If this is true then it means they require returns in excess of those which will fairly compensate them for the risks they accept. In PPP this means that investors require a greater share of the returns than they merit, at direct cost to the tax payer.

[1] SSA 2016 GDP growth 1.2% (World Bank Data)

[2] For reference, a table of alternate definitions and their sources compiled by Kwak et al (2009) is included in Appendix 1

[3] Listed investors in some jurisdictions may be legally required to publish high level project returns data

[4] In recent years the WBG have also published ‘Policy Research Working Papers’ which for the purpose of this paper we do not consider expressions of WBG opinion or policy plans. They are published by the WBG under the names of the authors and with the explicit caveat that they represent the views of their authors, not the organisation.

[5] Evenhart and Sumlinski’s summary table can be seen in Appendix 2

[6] Summary table Appendix 3