آیا در بحران سال 1998 روسیه می توانست نوسانات نرخ ارز را پیش بینی کند؟

The purpose of this paper is to examine the exchange rate volatility of the ruble, before the August 1998 crisis, using a macro-economic model of the Russian Federation developed by Basdevant (2000). A main focus of this paper is to use this simulation exercise to find out what was the origin of this financial crisis. This paper will show how expectations of the exchange rate played a crucial role. Nevertheless, this analysis does not seek to only explain the Russian crisis by a specific shock on expectations the objective is also to provide an analysis of the timing of the crisis. A further objective is to demonstrate a new methodology that allows an econometric model to be used to forecast the possibility of financial crises. Finally this paper demonstrates that a model can be used to assess the likelihood of an unusual event such as the Russian financial crisis by exploiting the information, which is contained in the model’s own residual process. Using this technique we demonstrate that the variance of exchange rates produced by the model rose dramatically in anticipation of the full height of crises in 1988.

The purpose of this paper is to examine the exchange rate volatility of the ruble, before the August 1998 crisis, using a macro-economic model of the Russian Federation developed by Basdevant (2000). The main issue is to use this simulation exercise to find out what was the origin of this financial crisis. This paper will show how expectations of the exchange rate played a crucial role. Nevertheless this analysis does not seek to only explain the Russian crisis by a specific shock on expectations. The objective is also to provide an analysis of the timing of the crisis rather than the crisis itself. A further objective is to demonstrate a new methodology, which allows an econometric model to be used to forecast the possibility of financial crises. The origin of the Russian crisis lay in structural problems: mainly capital stock obsolescence and shortage (see Basdevant, 2000 or Hall & Basdevant, 1999) as well as institutional ones. In particular the two main aspects are: the co-operation (or lack of) between industry and the banking system, and the inefficiency of markets (there is no real market clearing, and non-cash transactions are widespread). Before presenting the model and simulation results, it is necessary to characterise the crisis with some stylised facts, as the model should obviously reflect these. The crisis began on August 17, 1998, when Prime Minister Kirienko announced that the government would allow the ruble to be devaluated 34% by the end of the year. He also declared a 90-day foreign debt moratorium, and announced a de-facto default on the government’s domestic bond obligations. On August 26th the Russian Central Bank announced that it would not be able to support the ruble any longer. In less than a month the national currency collapsed by 300%, from 6.2 rubles to the dollar to over 20. Inflation shot up 15% in August compared with 0.2% in July, and has continued to climb. As stressed by Krugman (1979) financial crises occur when countries tend to finance government expenditures by printing money, which can lead to a currency crisis, as the fixed exchange rate regime becomes unsustainable. It is quite remarkable that in Russia the financial crisis came after the currency crisis. This lead to a total collapse of the banking system. Kaminsky and Reinhart (1999) link currency crises to banking system crises for a number of industrial and developing countries. They emphasise that the peak of a banking crisis most often comes after a currency crisis. Moreover they insist on the fact that most often banking crises have preceded currency crises. In the case of Russia, the crisis came mainly from the inability of the government to raise enough taxes to reduce the public deficit. Hence it became indebted towards international creditors, and thus rendered the Russian Federation dependent on capital inflows. The ruble was kept at a fixed exchange rate because of a fall in reserves. According to most Russians official the question before the 1998 crisis was not whether a crisis was going to happen — as the ruble was obviously over evaluated — but simply when the crisis was going to happen. Therefore this model will be used to evaluate the timing of the crisis, emphasising that the crisis came from the continuation of a policy of a fixed exchange rate (even with a relative flexibility between a certain band) that was unsustainable. Therefore we implicitly consider that — contrarily to what usually happens in financial crisis — the collapse of the banking system was not the cause of the crash. This approach can be related to other studies as Mishkin (1996) or Stoker (1994) who consider the same causal relation as the one defended in this paper. To analyse the timing of the crisis we propose a new methodology which draws on ideas from the financial econometric literature on ARCH and GARCH modelling and the standard macro-economic literature on stochastic simulations to derive a time varying measure of the probability of a financial crises occurring. Further details are given below but essentially the idea is that the model and its errors are a complete description of the data generation process. Any mis-specification or omitted effect from the model is, by definition captured in the errors of each equation. So unless the crisis happens without any warning either the model, or its error processes should be able to predict that something is going to occur. We therefore propose using the technique of stochastic simulation to calculate the standard error bands of the model variables. But instead of doing this in the conventional way where the shocks are drawn from a constant distribution we will base the distribution of the shocks on the properties of the very recent model residuals, much as is done in a standard ARCH model. We will thus be able to calculate a time varying volatility profile for each variable in the model. If the technique is successful we would expect the volatility (or standard error) of variables such as the exchange rate to increase dramatically just prior to the crises. Thus indicating the rising probability of such a crisis. The rest of the paper is organised as the following: Section 2 describes the modelling of the exchange rate, Section 3 presents stochastic simulations results, and Section 4 concludes. In Section 5 a summarised structure of the model is presented. The completed model is presented and discussed in Basdevant (2000).

In this paper we have demonstrated that a model can be used to assess the likelihood of an unusual event such as the Russian financial crisis by exploiting the information which is contained in the models own residual process. Using this technique we have demonstrated that the variance of exchange rates produced by the model rose dramatically in anticipation of the full crises in 1988. This paper has also shown how expectations formation is crucial in understanding the timing the magnitude of the crisis. Following the crisis the progress made were rather impressive with a strong output recovery and the achievement of a financial stability. A favourable external environment has contributed to the recent success, and Russian producers have benefited from the devaluation to reuse a part of the capital stock. However, many of the structural problems that were present before the 1998 crisis still remain to be tackled. Hence, recent achievements, particularly on growth, are vulnerable to a major deterioration in the environment. The main challenge facing the authorities remains the need to implement a broad-based acceleration of structural reform, to ensure that the recovery and the stabilisation gains will be sustained over the medium term. It is a major challenge for the Putin’s administration to define and conduct a policy addressing long-run issues faced by the Russian economy. This would form the basis to prevent another financial crisis, but should also provide the stability and credibility that is needed to attract private investors who should provide ultimately the only solution to Russian economic issues. This not only means to attract foreign investors and to benefit from technology transfers (see Bayoumi, Coe, & Helpman, 1996; Coe, Helpman, & Hoffmaister, 1995; Temple, 1999), but also simply to attract Russian investors and to stop capital flight.