Introduction

For many years, forecasts of the economy of Hong Kong have been made utilizing a macroeconometric model using annual economic data. The annual model forecasts medium-term trends in the economy and makes three- to five-year forecasts for such economic aggregates as the gross domestic product (GDP) and the general price level (Lin and Chou 1985). However, it says very little about the short-term fluctuations in these aggregates.

Thus, to fill this lacuna, a model designed more specifically to explain the short-run changes in output and prices is presented here. The quarterly econometric model of the Hong Kong economy focusses on the determination of GDP and its components as well as on the general price level. It consists of 31 equations, of which 17 are stochastic and the rest are identities. The equations depict behaviors in six sectors of the economy: private consumption, private investment, foreign trade, manufacturing, price level, and money.

The model contains 71 variables, of which 31 are endogenous and 40 are exogenous, and is estimated by using 48 quarterly observations from 1974 Ql to 1985 Q4. The endogenous variables with equation references, and the exogenous variables are listed in appendix A.

Raw data of the variables come from two sources. The unpublished quarterly GDP series were obtained directly from the Census and Statistics Department (CSD) of the Hong Kong government. And with the exception of value added in the manufacturing sector (XM), the employment, price, and money variables were taken from various issues of the Hong Kong Monthly Digest of Statistics. Data for XM are compiled by interpolating related series.

In the following section, the economic-theoretical content of the model is presented. Then three historical simulations and ex-post forecasts are performed to help evaluate the model. Finally, the paper ends with a multiplier analysis.

The Structure of the Model

There are 31 equations in the model, of which 17 are stochastic and 14 are identities. All identities of the model define one variable in terms of the others. A more detailed explanation of these identities is given in appendix A. This section provides a theoretical discussion of the 17 stochastic equations. One of the most difficult problems that was encountered in the process of constructing the model was the specification of appropriate lag structures.