Sub-Theme 2 | The National Early Warning System in Korea II
The early warning system for financial markets attempts to detect signs of instability in stock market, bond market, and loan market. A stock market crash or liquidity crunch in bank loans are major concerns of this early warning model. However, predicting the erratic behavior of financial asset prices is very difficult because asset prices can fluctuate by changes in investors’ sentiment not related to any changes in fundamentals.
The early warning system for financial markets, developed by Kang, Kim and Lee (2005), also employs the signal approach. The periods of financial market instability are identified by unusual behaviors in stock price, corporate bond yield, and the rate of dishonored bill. In addition, the trading volumes in stock and bond markets, the volatility of stock price and bond yield, term spread and default spread are also used to determine the periods of financial instability. The identified periods include the period of the collapse of Daewoo in 1999, the liquidity crunch of Hyundai group in 2000~01, the credit card crisis in 2003 and the currency crisis period in 1997. Thus, the crisis periods in the early warning model for financial markets are more broadly identified than the model for currency crisis.
The leading indicators considered in the model are categorized into four groups: monetary and credit variables (such as monetary aggregate and credit growth), stock market variables (such as foreign investors’ net investment and dividend-price ratio), macroeconomic variables (such as foreign exchange reserve, unemployment rate), and foreign variables (such as the US stock return, LIBOR, and oil price).
The early warning system for the real estate market has two different yet closely related models. They are the national model and the regional model which examine housing prices and land prices. The national model deals with aberrant behaviors of the nation-wide real estate market while the regional model focuses on the Seoul metropolitan area. The regional model is important because changes in housing prices in Seoul are ultimately expected to spill-over into other regions. The national model employs signaling approach and the regional model uses a probit model. The leading indicators in the models are financial indicators (such as liquidity in financial institutions and stock price index) and macroeconomic variables (interest rate and business cycle related variables) as well as real estate related variables (such as construction orders). A report of the findings of the early warning models is submitted to a committee to incorporate qualitative judgments. The committee is made up of members from real estate industry, academia, research institutes and the government.
Korea lacks natural resources and thus imports petroleum and other commodities from abroad. Hence, a sudden jump in the prices of such commodities can have huge adverse impacts on the domestic economy. An early warning system in this sector, therefore, is an important part of the national system. Nevertheless, the petroleum and commodity sectors are unique in that the prices are determined in the international markets. So, even though the model sends an early warning signal of an imminent jump in prices, the Korean government cannot prevent or reduce the price changes. Nevertheless, the early warning signals make it possible for the government and the private sector to prepare for emergencies in order to reduce the adverse effects of higher commodity prices.
The early warning system for petroleum uses an artificial neural network model to monitor the global petroleum market. The leading indicators considered in the model are the US petroleum inventory, OPEC’s oil production, business cycle variables, and the net position in futures contracts on crude oil. It is true, however, that forecasting future changes in world oil prices is very difficult partly because non-economic factors such as geopolitical situations, natural disasters, and OPEC’s energy policy are also important determinants.
Unlike other sectors, the results of the early warning system for petroleum are released to the public. This is because international markets are not likely to react to the outcomes of the early warning system. Rather, this information can help firms prepare for potential rises in oil price.
The early warning system for other commodities also monitors global economic variables such as business cycles in major countries, inventory of each commodity, and commodity futures prices. Like petroleum, the model for other commodities may have limited forecasting power as the prices of commodities are also determined by non-economic factors. Thus, qualitative monitoring should be emphasized in this sector.
Two different early warning models are developed to monitor labor markets. The first model is developed to predict a sharp decline in the employment rate in the next three months. The leading indicators in this model are mostly business cycle related variables including production and investment. The second early warning model is designed to monitor labor union-management relations. This model keeps a watch on strikes and other problems between labor union and management in order to reduce the potential cost of strikes.
The National Early Warning System in Korea II
|Subject||Economy < Financial Policy|