شاخص های هشدار دهنده زود هنگام سیکلهای دارایی رونق قیمت / ورشکستگی در بازارهای نوظهور

We apply recently developed early warning indicator systems to a cross-section of emerging markets. We find that, with little or no modification, models designed to predict asset price booms/busts in advanced countries may be useful for emerging markets. The concept of monitoring a set of asset prices, real activity and financial indicators is generally found to be efficacious. We also find that, in addition to this set of variables, early warning indicator systems for emerging countries may be augmented with capital flow indicators.

The recent financial crisis has underscored the growing importance of asset price fluctuations for macroeconomic performance. As is the case for developed countries, a number of emerging market economies (particularly in Central and Eastern Europe1) seem to be quite prone to such shocks, as the rapid rise and subsequent decline of asset prices have presumably contributed significantly to the pre-crisis overheating of these economies as well as to the following contraction. Therefore, a system of early warning indicators that would help with early identification of emerging imbalances in asset markets is a much sought-after tool for policy-makers. Development of such a system via the country-specific approach is often impossible because of data limitations. Therefore the standard approach is to make estimations for a group of countries (that may or may not include the analyzed country) and apply the resulting model to the economy in question.2 A number of recent studies use this method and report on models that can be used to predict asset price booms and busts. These may be valuable to the policy-maker. The caveat here is that most of these models have been fitted to explain asset price fluctuations in industrialized countries. It is not clear how useful these models are for emerging markets, as movements in many of the macroeconomic variables used as early warning indicators are remarkably different in the developed and emerging markets. For example, one may find it difficult to distinguish between excessive credit growth that leads to an asset price bubble and the convergence of an underdeveloped banking sector to a level commensurate with the industrialized countries. For transition economies, it may also be challenging to identify “overheating” based on growth rates of real sector variables that fluctuate dramatically as the economy undergoes substantial transformation. In fact, asset prices as such are known to be volatile in emerging markets and therefore difficult to interpret. For these reasons the early warning indicator approach needs to be thoroughly studied before it finds its use in predicting asset price cycles in emerging markets. Indeed, as emphasized in the work of Reinhart and Rogoff (2009), data coverage is crucial for financial crisis analysis. The main contribution of this paper is the application of asset price boom/bust analysis to the new dataset on emerging markets (most notably former Soviet Union countries). The rest of the paper is structured as follows. Section 2 provides a review of recent contributions in the development of asset price cycle early warning indicator models. Section 3 describes the dataset, comprising a cross-section of emerging market economies, on which we conduct the empirical analysis. Section 4 outlines and implements methods to identify boom and bust events that occurred in emerging markets. Section 5 presents an evaluation of the efficacy of existing models for predicting asset price developments in emerging markets and reports on the models fitted here to predict asset price booms/busts in the purely emerging market dataset. Section 6 concludes.

This paper contributes to the literature by investigating whether early warning indicator models can be used for predicting asset price boom/bust occurrences in a cross-section of 29 emerging markets. We identify booms/busts using different approaches. The results are not fully synchronized but may still be regarded as cohesive. The sample obtained is large enough for interpretable econometric analysis although its informational content is limited since, for the most part, only one (most recent) wave of booms/busts can be analyzed. We employ two modeling approaches (stand-alone indicators and discrete choice models) that were previously applied to a cross-section of developed countries. The results seem promising. In fact even the out-of-sample performance of the unmodified developed countries' models is satisfactory on our emerging markets' dataset. Naturally further enhancement and re-estimation of these models increases their in-sample predictive performance, although these modifications need not be extensive. Our results are generally inconclusive as to which approach to predicting asset price boom/bust is superior. But we argue that the concept that relies on monitoring the combined set of asset prices, real activity and financial indicators is widely applicable to emerging markets and its efficiency is confirmed under the different model setups. According to our estimates credit growth and investment (in either growth rates or ratio to GDP) turned out to be particularly reliable indicators for forecasting asset price cycle. We also find that, in addition to this set of variables, early warning indicator systems for emerging countries may be augmented with capital flow indicators.