First, a monthly seasonality in conditional mean equations was found. It means that arrivals of foreign tourists into Korea vary with monthly seasonality. The results are consistent with those of other empirical studies, which reported the existence ...
First, a monthly seasonality in conditional mean equations was found. It means that arrivals of foreign tourists into Korea vary with monthly seasonality. The results are consistent with those of other empirical studies, which reported the existence of monthly seasonal variations (Chan 1993; Hiemstra and Wong 2002; Lim and McAleer 2001).
Second, the estimated GARCH model revealed that the impact of news shock on monthly tourist arrivals into Korea is quite persistent as following a quadratic curve. In addition, since the conditional variance in the GARCH model depends only on the square of the unexpected news shock, the news impact curve of the GARCH model follows a quadratic function centered on . This indicates that as the squared spread gets bigger, tourism demand is likely to become more volatile and more difficult to predict. Also, the results of this study confirmed that the symmetric GARCH model tends to overstate the variance for , whereas it tends to underestimate the variance for . That is, if a negative shock causes more volatility than a positive shock of the same size, the GARCH model tends to underestimate the amount of volatility following bad news and overestimate the amount of volatility following good news. Otherwise, if large shocks result in more volatility than a quadratic function allows, then the standard GARCH model tends to under-predict volatility after a large shock and overestimates volatility after a small shock. Thus, the GARCH model showed its limitations in that it cannot capture the asymmetric effect and the leverage effect.
Third, through investigating the EGARCH model and the TARCH model, an asymmetric effect was found. That is, the volatility of monthly inbound tourist demand in Korea varied according to whether the news shock was good or bad. Likewise, the models showed the presence of leverage effects. This means that the negative news shock produced more volatility than positive news shock in the variance of monthly inbound tourism demand in Korea. Based on the evidence from real events in the past years, for example, during the FIFA 2002 World Cup which attracted global attentions, fewer foreign tourists (230,000) visited Korea than the estimated 0.4 million (Ministry of Culture and Tourism 2003). In contrast, total number of inbound tourists from February through June 2003 affected by the SARS was 2,068,061 (KNTO 2004). The number represented a drastic drop of 25.2 percent when compared to that of the same period of 2002.
These results strongly suggest that the Korean government and key tourism industry players (firms, associations, boards, etc.) should focus particularly on the notable and often extraordinary changes in the tourism business environment following the broadcast of bad news because conditional volatility increases more when a negative shock occurs than when a positive shock occurs. However, negative shocks are not easily predicted as they occur suddenly and erratically. This implies the need to plan precisely and seriously for the inevitable negative events. When negative news shocks are predicted, tourism firms, organizations and associations need to, either singularly or collectively, design appropriate tactical corporate and/or business strategies to minimize the adverse effects which may result from a decrease of tourism demand. Swift measures such as development of contingency programs, downsizing supply or reducing investment in consideration of the persistence and dimensions of the bad news shock may be necessary.