انتخاب روش قیمت گذاری و سرمایه گذاری خارجی تحت خطر سیاسی

The paper analyses asset prices in a context of uncertainty over future government policy. As current policy is maintained, perceived risk abates thus leading to a gradual appreciation of asset prices and a gradual decrease in their conditional variance. Option values computed under this process have time series and the term structure of conditional volatility, which, in general, are downward sloping. In price series without a policy reversal, implied volatility from option prices will exceed actual volatility, with this wedge progressively disappearing. This may be viewed as the volatility analogue of the ‘peso premium’ for assets subject to large, infrequent price drops.

In recent years, the emphasis on the financing of development has shifted from debt to equity and from governments to the private sector. A remarkable rise in foreign direct investment has taken place in reforming economies, at first tentatively and then progressively accelerating, until the recent crisis. The time path of foreign direct investment is quite comparable across countries. The initial flows were small, accelerating over time, reflecting high ex post profitability of early experiences, and leading into a dramatic climb at a late stage. Fig. 1 describes the time series of foreign direct investment capital committed to the People’s Republic of China between 1983 and 1998. This gradual buildup, mirrored in the time pattern of returns on the local capital markets, is consistent with various explanations. A simple argument has to do with logistics: it takes time to train and build capacity. However, this simple explanation cannot account for the parallel progressive acceleration in portfolio flows. Investment delays may arise because there is value in waiting to invest when some fundamental uncertainty is resolved only over time (see McDonald and Siegel, 1986). Another explanation is that learning about local productivity takes place sequentially (e.g., Chamley and Gale, 1994 and Thimann and Thum, 1998). The latter two explanations raise the question of what is the source of local uncertainty, and how this ‘learning hypothesis’ may be tested against models without a gradual resolution of uncertainty.1 The fastest rising destination of foreign investment has been economies whose governments had announced market-oriented reforms. In our opinion, this trend reflects diffused expectations that these countries are progressively moving towards a market economy, and a more reliable legal and taxation system. However, since sovereign governments may reverse economic reforms, policy and political risk considerations remain paramount for foreign investors in developing economies, as the recent crisis in Latin America and financial instability in South-East Asia confirmed. We believe that the progressive decline in perceived political risk has played an important role in explaining this pattern, as the evidence in Perotti and van Oijen (1999) suggests.2 In a recent paper studying the impact of capital market liberalizations on various emerging markets, Bekaert and Harvey (1998) find a reduction in the cost of capital after liberalization, while volatility is hardly impacted. This paper tackles explicitly the modelling of political risk and its implications for investment and asset pricing, drawing inspiration from the literature on time-inconsistency of monetary and fiscal policy (e.g., Barro and Gordon, 1983, Barro, 1986 and Fischer, 1986). We model explicitly the dynamics of policy uncertainty (for a static treatment, see Eaton et al. (1986) and Perotti (1995)). Our model shows that a reform policy maintained for some time may in fact conceal an opportunistic policymaker ‘biding its time’ prior to changing the rules of the game.3 As a result, investors will be reassured only gradually about the stability of any market reform. In our model, lack of news is good news, and asset prices, representing equity claims on foreign investment, increase if nothing happens. Thus we obtain a so-called ‘peso premium’: the time series of prices can exhibit excess returns and positive serial correlation in the absence of policy changes. Since it is difficult to draw precise inferences from realized returns, we focus on the empirical implications in terms of expected future volatility as obtained from the implied volatility embedded in option prices on assets exposed to political risk. Section 2 presents a model of capital investment which endogenizes the probabilistic timing of a possible policy shift, and derives the time profile of foreign investment, its valuation and perceived risk. In Section 3 we compute option prices based on a lattice of potential asset prices generated by the theoretical political risk model augmented by a generic market risk factor.4 We then describe the empirical implications of the model in terms of implied volatility. Section 4 concludes.

The paper endogenizes the gradual buildup of policy credibility in an economy moving towards market reforms but facing residual policy risk. We derive the intertemporal price process under such political risk and offer an intuitive interpretation for the gradual expansion of foreign investment and its acceleration over time. It is descriptive of an early stage in foreign direct investment in emerging economies, when reforms are still not fully established and foreign investors act as tentative, cautious pioneers. The standard assumption in asset pricing that new information is released continuously as a flow of a large number of small, uncorrelated events is inappropriate in certain contexts such as in emerging markets, when policy shifts have a very large systematic effect on asset values. Because investors’ expectations affect their investment decisions, the timing of policy decisions needs to be treated as a strategic variable. The resulting information process is asymmetric: absence of news is good news. This can produce positive serial correlation, sustained trends, and excess returns in price series in the absence of policy changes. Our option pricing approach offers several other empirical implications: in particular, the medium term evolution of implied volatility is shown in general to be monotonically decreasing and the wedge between realized and implied volatility falls gradually over time. The implications of introducing risk aversion in our model can be deduced rather intuitively — it would introduce a risk premium in the discounting of payoffs as a function of perceived volatility. In our context, the risk premium would gradually fall as the risk of expropriation falls; this would lead to lower initial prices and a steeper price appreciation over time. Thus our time series results on implied volatility would be strengthened: price volatility would be higher at first (when the conditional price series would have higher capital gains to reflect the required risk premium) and would decline now even faster (since not only the risk of the potential fall, but also the associated risk premium, would gradually fall). An interesting observation is that under political risk the time series of prices may resemble a speculative bubble, although the true cause of the rapid buildup is simply rational updating of policy credibility. Our proposed methodology enables one to discriminate between rational price updating and a deviating bubble: namely, while in the case of rational updating implied volatility should ultimately fall, in the case of a stochastic bubble it should increase over time. In future research we plan to investigate empirically the option pricing model and to refine possible instruments for policy analysis. Testing this model would require a time series of implied volatilities from a sample of option prices across several countries facing political risk. The main implications to be assessed are that: (1) implied volatility would exceed historical volatility; (2) the wedge between the two measures should decline over time; (3) the wedge would be correlated with the degree of political risk as proxied by market surveys or other measures commercially available; (4) the term structure of implied volatility and the realized price volatility path would be downward-sloping; and (5) the speed of the decline may be correlated with measured improvements in perceived political risk. While some tests can be carried out using realized volatility, most tests involve implied volatility data, which is not yet available as such option markets are not yet fully developed. However, the rapid evolution in these markets and the increasing need for active risk management suggest that such tests will become feasible in the near future.