• Search

The Middle Eastern and African Arbitration Review 2018

Business Valuation in Countries Without Developed Financial Markets

Introduction

Many transnational disputes have taken place in countries without developed financial markets. If a tribunal finds in favour of the claimant on liability, it often confronts difficult questions about the appropriate quantum to award. In the absence of agreed facts about the economic situation of a country, there is considerable latitude for expert opinion to differ as to the appropriate damages to award.

Valuation of going-concern businesses is usually conducted either by reference to their market value – the price that somebody would pay for them – or by reference to their economic value, the sum of their expected future cash flows, adjusted for risk and expressed in today’s terms.

In principle, these two approaches to value should reach similar conclusions. In other words, the prices revealed in financial markets for financial assets should be closely related to investors’ expectations about the amount and risk of the expected future cash flows from owning the asset at any given time.

In countries with developed financial markets, the market value of financial assets is readily available from stock markets, bond markets and derivatives markets. Information about the prices of listed businesses, or about the prices at which (listed or unlisted) businesses have changed hands, may also be used to infer the value of similar untraded businesses.

In a liquid and well-regulated financial market, price and value should normally be closely related to one another. Because of what they reveal about actual and expected risks and returns, financial markets are also a vital source of information about investors’ past experience and future expectations at any given date. In contrast, investors in countries without developed financial markets and the associated financial statistics and transaction data face significant barriers to assessing the value of businesses.1

Countries that lack developed financial markets place foreign investors at a particular disadvantage because they are less able to quantify the risks to which their investment might be exposed. Locals’ greater familiarity with a country and its business environment means that they are better able to assess the risks and opportunities associated with operating in their country than foreign investors who lack access both to public data (in the absence of formal financial markets) and the informal information sources that may guide locals’ investment decisions.

This article seeks, first, to consider how developed financial markets contribute to assessments of value. Second, we look at some of the techniques that practitioners use to draw conclusions about countries without developed financial markets by inference from those countries that do have them. Third, we consider the assumptions implied by using data derived in one country to draw conclusions about the value of assets in another country. Finally, we consider the steps that practitioners might take to adopt a robust approach to assessing business value in countries without developed financial markets.

The role of financial markets in assessments of value

Financial markets play a direct role in assessments of market value and an indirect role in assessments of economic value.

At the most obvious, financial market prices are an assessment of the value of financial assets traded on exchanges. Shares, bonds, derivatives, exchange-traded funds, currencies, commodities and other assets are freely traded in public markets and their value is normally assumed to be equal to the price that investors are willing to pay for them.

Market prices may have wider effects. Even though only a small fraction of the world’s commodities change hands on public exchanges, the prices observed on those exchanges usually determine the contractual prices at which trading occurs between counterparties off the market. Those prices can then ripple out further in the economy: market-determined gas prices, for example, may drive the prices of chemicals, such as ammonia and certain plastics, that are derived from natural gas.

At one remove, one can draw inferences about the value of financial assets that are not traded on public exchanges by reference to the market prices of similar assets that are so traded. For example, if the shares in a publicly listed Texan oil company trade at 10 times next year’s expected earnings before interest, taxes, depreciation and amortisation (EBITDA), it may be reasonable to infer (perhaps with certain adjustments) that an unlisted Texan oil company might also be valued at 10 times next year’s expected EBITDA.

In assessments of economic value, a valuer must first create a set of expected cash flows and then discount it at a rate that reflects the cash flows’ relevant risks. The cost of equity of a company is not directly observable and is often derived by reference to the Capital Asset Pricing Model (CAPM), which expresses it as a function of:

  • the risk-free rate, calculated as the yield to maturity of high credit quality government bonds;
  • the market risk premium, the return in excess of the risk-free rate expected by investors in non-risk-free assets, such as shares or corporate bonds; and
  • beta, a statistical measure of the relationship between the returns on the asset being studied and changes in the returns of the market as a whole.

Each of these three elements of the CAPM – the risk-free rate, market risk premium and beta – are derived from observation of the prices of government bonds and non-risk-free assets in public markets.2 Together, these three elements are used to calculate the cost of equity according to the formula:

      Ke = Rf + (Rm - Rf) × β

The more assets are traded on public exchanges in published transactions, the greater the information available to a financial practitioner to draw inferences about the value of other assets in the economy. In countries with less developed financial markets, the data may be less reliable because of liquidity issues or sector biases that render calculations of asset betas and market risk premiums unreliable.3 In countries with no financial markets at all, the absence of data silences the CAPM altogether.

How do practitioners extrapolate from developed country financial market data to countries without developed financial markets?

An absence of reliable data does not mean that tribunals (or quantum experts) can evade their responsibility to assess questions of value in countries without financial markets.

In the author’s experience, there are three ways in which the value of an asset might be adjusted to reflect its location in a country without financial market data.

In the first approach, a valuation is conducted as if the asset were located in a developed country. The valuation is then adjusted by a straightforward discount factor (say 30 per cent or 50 per cent) to reflect the fact that it is in a country with higher perceived risks. The discount factor might be based on a ‘rule of thumb’, say that mining investors into Country X normally apply a discount of 30 per cent per tonne of concentrated ore to their projects compared to those in Country Y. Alternatively, the discount might be based on prices in observable markets. Let’s say the price of real estate in the capital of Country X trades at a discount of 30 per cent to that in the capital of Country Y. On that basis, one might infer that prices of other assets in Country X could also be discounted by a similar factor.

The defects of such an approach are obvious. Any empirical underpinnings for the discount factor are either unscientific (because they are based on a small sample of investors) or inappropriate (because they are based on different assets to the one being studied) or both. In other words, the discount is, at best, an educated guess.

The second approach is to add a country risk premium to the discount rate, a rate that is otherwise assessed using developed country financial market data. Given the absence of financial markets in the target country, that premium must be assessed by reference to the country risk premium of a different country on the assumption that the systematic risks in Country X are somewhat similar to those in Country Y.

In countries without developed financial markets, the only financial asset with an observable price may be the foreign debt of the government. If the government’s debts are all in local currency, then they are not directly comparable to bonds issued in a common foreign currency, such as US dollars, because they incorporate currency risk as well as the default risk of the sovereign.4 In those circumstances, all that is available is the government’s rating from one of the main credit rating agencies.

For example, let us say that Country X has a BB- credit rating but no sovereign debt issuance in US dollars. Let us further assume that Country Y also has a BB- credit rating but has outstanding US-dollar-denominated debt with a yield premium of 2 per cent over US Treasuries. On that basis, it might be inferred that the appropriate sovereign risk premium for Country X is also 2 per cent. In those circumstances, the sovereign risk premium of the government – the yield premium of government debt over ‘riskless’ assets such as US Treasuries – is sometimes used as a proxy for the risks faced by investors.

There are, however, several issues, logical or technical, with the use of a sovereign risk premium as the proxy for the country risk premium.

  • It is not necessarily safe to assume that the premium for risk borne by investors in government debt is the same, or closely related to, the one paid by an investor in an oilfield or a drinks factory. For example, one of the risks faced by investors is that of expropriation by the state. Such a risk cannot apply to state-issued debt instruments. Alternatively, default by the government will not necessarily imply difficulties in repayment by private enterprises.
  • The addition of a perpetual risk premium to a discount rate does not closely mirror the actual pattern of certain risks, such as political risk, which rise and fall with the passage of time. Even countries with the same credit rating at a given point in time may be on paths to improving or weakening creditworthiness: applying the same premium to both because they have the same credit rating at the given moment may be inappropriate.
  • There remains dispute among academics and practitioners as to whether country risk premiums are appropriate at all.5
  • Even if one accepts that country risk premiums are appropriate, however, for countries with developed financial markets there often exist several potential approaches that may generate quite dissimilar results. It is not clear that these results can easily be applied to countries without developed financial markets or where certain data is not available.

Finally, the non-linear effects of a country risk premium on cash flows are difficult to reconcile with the potential effects of a given risk. For example, let us consider a series of US$1,000 payments made at the end of each year. Let us say that the risk of any given payment of US$1,000 in Country Y can be observed from financial market data to be 10 per cent. Now let us add a 5 per cent premium to that figure to reflect the incremental risks of investing in Country Z, resulting in a 15 per cent discount rate. As can be seen in the table below, the present value of the first year’s cash flow is reduced by 4 per cent as a result of the premium, but that rises to 20 per cent by the end of fifth year, 36 per cent by the end of the 10th year, 49 per cent in the 15th year and 59 per cent in the 20th year. Valuers might pause to consider whether this is an appropriate pattern to assume for the specific investment they are considering, let alone for whole industry sectors or the entire economy of Country Z.

Sometimes it may be appropriate to apply inferences derived from an investment for which there is data – say the shares in a publicly listed firm – to an investment for which there is no data – say the shares in a closely held company in the same industry. The greater the differences between the target investment and the investment used as a comparable, the less appropriate such inferences become. When the target investment is in a different country, market-based assessments of value are even more difficult to adjust with any reliability than the cost of capital. The fact that a company trades at 10 times EBITDA in a developed European country provides no useful information about the appropriate multiple for a similar company in central Asia or sub-Saharan Africa. Nor is there any generally accepted way to adjust a multiple to reflect different risks or growth assumptions from the ones embodied in the analysis used to calculate the multiple.

The third approach is to calculate the discount rate by reference to a developed country’s financial markets and to incorporate the additional risks of operating in a specific market directly into the cash flows. In other words, it is to leave the discount rate alone but to model different scenarios that incorporate specifically identified risks and to ascribe potential weights to them. In principle, if the risks are correctly modelled, the present value of expected cash flows should be the same as if the correct premium were applied to the cost of capital (if such a premium could be discerned).

The relevant risks to which an investment is exposed may differ, as may the extent of the investment’s exposure to those risks. To name a few:

  • certain types of activities are inherently more exposed to changes in government policy because, for example, they are regulated in the public interest (such as utilities that exhibit elements of natural monopoly) or pay royalties that compensate the state for the exploitation of national assets (such as minerals or radiomagnetic spectrum) or contract directly with the government (such as construction firms or hospital operators);
  • an oilfield (say) that sells its output on the world market in US dollars per barrel is less exposed to currency instability than a food processor whose output is sold domestically in local currency;
  • manufacturing investments whose products cross frontiers (whether internal or external) are more exposed to delay and local intervention; and
  • a business that exploits significant intellectual property may be less exposed to expropriation risk than one that can be carried out using publicly available know-how.

Each of the risks above, and many more that might be contemplated, has a distinct effect on cash flows: revenues may decline after the regulator’s next price review, currency instability and exchange controls may affect the cash flows to the parent, corruption acts like a tax, expropriation will cause future cash flows to fall to zero and so on.

In principle, it is possible to model all of these potential effects, either in isolation or in combination. One approach is to use ‘deterministic’ scenario analysis to model the effects of different combinations and to calculate an ‘expected value’ for the investment by multiplying the value under each scenario with its assessed probability of occurring. Another approach is to use ‘stochastic’ modelling techniques, such as Monte Carlo simulation, in which each event is assigned a probability of occurring, or a probability distribution, and the cash flows projected (say) 10,000 or 100,000 times to derive a distribution of potential outcomes.

Both approaches suffer from a common defect, which is that it is difficult to assess ex-ante the potential outcomes of the next regulatory review, the extent of possible future currency instability or the probability of expropriation. Nonetheless, both approaches can allow the valuer to assess the valuation effects of using different assumptions about the likelihood, date and consequences of specific risks and to consider the likely potential range of values, excluding the most extreme outcomes.

For countries without developed financial markets, however, these defects may simply have to be managed: it seems preferable to explore the effects of risk on cash flows rather than using arbitrary rules of thumb or data derived from different countries to make adjustments to the discount rate.

Conclusion

As noted above, financial markets translate investors’ perceptions of expected risks and expected returns – the drivers of financial value – into prices. The problem in countries without developed financial markets is that the assessed risks and returns that would be derived from an analysis of this data is either unreliable or non-existent. One solution to that problem is to apply a premium to the discount rate, leaving cash flows as a ‘best estimate’ of what would happen if the investment were operated as expected.

Currently, that approach is the preferred one in international arbitration. Over the last 15 years, investment arbitration tribunals (the only ones from which awards are generally made public) have been presented with a range of conflicting evidence about country risk premiums by investors and states. As far as is known from those cases where claimants have been awarded compensation, it appears that the acceptance of country risk premiums by tribunals is a growing trend. Further, tribunals have more often favoured the higher country risk premiums put forward by states (which tend to reduce the value of any award, all else equal) than the lower ones put forward by investors.6 This illustrates one of the curiosities of investment arbitration: it is one of the few forums in which states may consider that their public’s interest is favored by an emphasis on the elevated risks associated with doing business in their territory.

As noted above, additional risk is often expressed in international arbitration as a country risk premium, an abstract financial concept. It might be illuminating if, in the alternative, states and investors were obliged to explain how and when they expected specific risks affecting the investment to materialise and to calculate the effects of those risks. It might also crystallise the extent to which those risks were general business risks, to which investors are presumed to be exposed, and those that result from specific government actions, whose lawfulness can be explicitly considered by the tribunal.

To conclude, as a matter of financial theory, discount rates should reflect systematic factors, leaving cash flows to reflect the idiosyncratic risks to which a particular project is exposed. That principle applies equally to all investments. A requirement to explain clearly which risks are taken into account in a valuation demands greater rigour than an adjustment to the discount rate even if, in principle, the two ought to arrive at the same result. Although desirable in general, detailed consideration of cash flows and risks is particularly important in countries without financial markets.

This article should not be interpreted as suggesting that systematic risks can be ignored in the discount rate.

Rather, it aims to suggest that where there is no financial market data available to make any quantitative assessment of risks, greater weight should be placed on the adjustment of cash flows. If the alternative is to use a simple premium to the discount rate, a reasoned exploration of relevant idiosyncratic risks, and their probabilities, timing and effects, seems preferable. At a minimum, it would enable tribunals to understand the sensitivity of the investment to different risks and to take their own view as to the weight to give to different scenarios.

Year

Face value of amount

PV at 10% discount rate

PV at 15% discount rate

Difference

1

1,000

909

870

4%

2

1,000

826

756

9%

3

1,000

751

658

12%

4

1,000

683

572

16%

5

1,000

621

497

20%

10

1,000

386

247

36%

15

1,000

239

123

49%

20

1,000

149

61

59%

Notes

1 I draw a distinction between countries with ‘liquid and well-regulated’ financial markets and those with no financial markets. The concept of liquidity is a slippery one and, even if liquidity can be quantified, there is no agreement about how much liquidity makes a market ‘liquid’. Similar considerations apply to the issue of regulation, which may be as much about how rules are applied as what they say. This article is not intended to address those points. Nor is it the intention of the paper to address countries that possess formal financial markets but where the reliability of those markets’ price discovery may be open to question. Rather, it is addressed to places without any substantial, formal financial markets at all.

2 In addition, practitioners use data from financial markets to derive other statistics sometimes used in the cost of capital, such as the small stock risk premium. This article does not consider the appropriateness of using other adjustments to the CAPM.

3 Liquidity constraints occur when there is insufficient trading to observe how prices change in response to new information. Sector biases sometimes arise when economic sectors, such as mining or oil and gas, constitute a large fraction of the listed assets in a country. In that case, the market data used to calculate betas may be heavily weighted to one or two sectors that render it unreliable as a measure of how prices in the subject asset respond to changes in a broad range of financial assets.

4 In other words, although a foreign currency translation is available using the spot market exchange rate and the forward curve, or an exchange rate forecast, the expected exchange rate still contains a risk that the actual outturn will not be the same as forecast. In principle, one might be able to use swaps to neutralise that risk, but swaps carry a transaction cost and may not be available for less widely traded currencies such as those of countries without developed financial markets.

5 For example, Brealey, Myers and Allen’s textbook refers to premiums that are exogenous to the CAPM framework as ‘fudge factors’.

6 Beharry, C. Ed., Contemporary and Emerging Issues on the Law of Damages and Valuation in International Invesatment Arbitration, Chapter 9, table 9.1, p 258.