earticle

영어

This paper is aimed at developing models to forecast the demands of the Korean apparel markets. Specifically, the development is focused on multivariate models using time series data for the dependent and the independent variables. The data are collected twice a year, once for the first half and the other for the second half of each year from 1998 to 2008. The independent variables are selected from a set of 43 variables which are thought to affect the demands of the apparel markets. The 43 variables are chosen by referring to the previous studies. Demands are forecasted for 7 markets which include 6 segmented markets and the total market which is the sum of all the segmented markets. The segmented markets are the men's formal wear market, the women's formal wear market, the casual wear market, the sportswear market, the underwear market and the children's wear market. Actually the total of 21 forecasting models corresponding to 3 different time bands, i.e. the first half, the second half, and the whole year, for each of 7 markets have been developed. The time lags and the durations of the influences of the independent variables on the dependent variable are also considered in the models, which is the most difficult and time-taking part of this research. Basically the multiple regression method is applied while using time series data. Therefore, the models of this paper are called the multivariate time series regression models. This study of multivariate forecasting models is unique in respect of the methodology applied to forecast the apparel market demands. And this study is one of a few that deal with the whole apparel market of a nation. The models developed in this study are for the markets of the Korean apparel industry. This is the major difference of this research from the most of the previous ones which only deal with the demands at the level of an individual firm or at the level of a specific store. The results obtained in this study are found out to be more accurate than those of the previous study by Lee et al.(2011) where the univariate method was applied. And the results of this study is expected to contribute to enhancing the efficiency of the apparel companies and invigorate the related future studies in the following aspects:First, the demands of the apparel markets have high variability due to several reasons which make it difficult to forecast. In this respect, the proposed models in this paper can be utilized with minor modifications to forecast the demands of other consumer goods which also have high variability in demands. Second, multivariate time series forecasting models introduced in this study can make up the major drawbacks of the univariate time series models and improve the forecasting power. Consequently the models of this paper can be used by apparel companies to respond timely to rapidly changing market demands. Third, the temporal correlation analysis between apparel market demands and various economic variables showed that there are some variances in the time-lags and the durations of the influences of the economic variables on apparel market demands. In other words, it was found that economic conditions may influence the apparel market either immediately or after some time interval, and the duration of the influence may either be short-term or long-term. Such findings of this paper provide apparel companies with some insights to set up appropriate long-term plans to cope with rapidly changing economic environments.

한국어

의류시장의 수요는 다양한 요인에 의해 영향을 받아 그 변동성이 매우 크기 때문에 예측에 어려움이 있다. 본 논문은 이러한 특성이 있는 국내 의류시장의 수요예측모형을 개발하는 것을 목적으로 한다. 본 연구에서는 독 립변수들과 종속변수 각각에 대해 1998년부터 2008년까지 연 2회(상반기와 하반기) 측정한 22개의 시계열 자 료를 이용한 다변량 시계열 회귀 수요예측모형을 개발하였다. 독립변수는 선행연구에서 의류시장 수요에 영 향을 미친다고 파악된 43개의 후보변수군에서 선정하였다. 수요예측 대상은 6개의 세분 의류시장과 이들을 합한 전체 의류시장으로 총 7개의 시장이다. 게다가 계절성이 강한 특성이 있는 의류시장 수요임을 감안하여 7개 시장 각각에 대해 상반기, 하반기, 상하반기의 3개 시장으 로 분류하여 총 21개 시장에 대한 수요예측모형을 개발하였다. 연구과정에서 독립변수의 종속변수에 대한 영향이 독립변수에 따라 짧거나 혹은 긴 시차를 두고 나타난다는 점과 영향을 주는 기간도 단기인 경우와 장기인 경우가 있다는 점을 발견하였다. 본 논문의 수요예측모형은 이러한 측면까지 고려하여 만들어졌다. 수요예측 결과는 MAPE 기준으로 정확도가 높았으며, 과거의 단일변 량 시계열 수요예측모형의 예측치와 비교해도 우수한 결과가 나왔다.