Changes in China's Influence on Korean Exports by Production Phase

I estimate the impact of China’s GDP on Korean exports to the world and Korean exports to China based on the VAR model. It is found that the responses of Korean exports to shocks in China’s GDP are very different by export type (exports of raw materials, capital goods, and consumption goods). Reflecting the trend that the proportion of capital goods in Korean exports is increasing, Korean exports of capital goods respond most sensitively to these shocks among the three export types. Another finding is that Korean exports of raw materials do not respond very sensitively relative to the other export types. These findings are almost same as those for Korean exports to China in terms of response to GDP shocks. In addition, Korean exports to the world and Korean exports to China are more responsive to China’s GDP growth shocks over time.

China's economy has been growing rapidly at an annual average growth rate of over 8% for the last 20 years. As a result, China's economic changes are affecting other economies, especially other Asian economies such as those of Korea and Southeast Asia, which are closely linked with China through trade.
Currently, changes in China's macroeconomic variables, such as economic growth rate, exchange rates, and prices, are very important factors for these economies to monitor and respond to accordingly.
Economists and economic agents such as firms and governments have great interest in the effect of China's economy on their economies. Many economists have studied the effect of the Chinese economy on other countries' exports or economic growth rates. Abeysinghe and Forbes (2001), Abeysinghe and Lu (2003), and Girardin (2005) show how China's economic growth has affected other economies' growth. According to Abeysinghe and Forbes (2001), one unit of positive shock in China made cumulative (through the fourth quarter) impacts of 0.09, 0.52, 0.38, and 1.08 percentage points for the Japanese, Malaysian, Thai, and Korean growth rates, respectively. Yang and Vines (2000), Eichengreen et al. (2004), and Greenaway et al. (2008) investigated the impact of Chinese exports on other economies' exports. Eichengreen et al. (2004)'s analysis showed that in third-market countries, such as the United States and Japan, China's exports crowded out other Asian countries' exports. On the other hand, for several Asian countries, such as Korea, Japan, and Singapore, which export capital and intermediate goods to China, exports to China offset their losses in the third markets. However, Greenaway et al. (2008) argue that although the increase in China's imports from Asian countries partly offsets their losses in the third markets, the loss is larger than the offsetting gain. In conclusion, China's exports negatively affect other Asian countries' exports.  Changes in Chinese economic growth affect the Korean economy through many paths, such as trade, prices, and exchange rates. Trade is one of the major paths. As mentioned in the introduction, there are two channels through which China's growth affects Korea's exports. Figure 1 shows the trend in the growth rate of Chinese real GDP and Korea's exports to China. In general, it looks like a stable relationship exists between the two variables in the long run.

Structural Changes in Korean Exports and China's Role
There were major changes in the structure of Korean exports over the last 20 years. Until the early 1990s, Korea's major trading partners were the developed countries, such as the United States, the European Union, and Japan. As a result, Korea's major exporting goods were consumption goods (see Figure 3).

Changes in Korea' s Dependence on China for Exports
China's weight in Korean exports of raw materials, capital goods, and consumption goods increased from 8.1%, 4.0%, and 3.2%, respectively, in the fourth quarter of 1994 to 27.6%, 20.2%, and 16.4%, respectively, in the first quarter of 2009 (see Figure 7). It is also found that the weight of capital and consumption goods is increasing while the weight of raw materials is decreasing in terms of Korea's exports to China. It is noted that capital goods have been exported more than raw materials since 2003 (see Figure 8).   III. Empirical Analysis

Model Specification and Data
Equation (1) represents a country's export function as derived by O'Neill and Ross (1991) and Sato (1967). It consists of world income (WGDP), export country's relative prices to other countries' prices (P D /P O ; P D indicates domestic price and P O indicates other countries' prices), and production capacity (PRD). Relative prices can be interpreted as exchange rates. In this empirical study, I replace the relative prices with real effective exchange rates (REEX). Equation (2) is the first difference of equation (1). Equation (2) shows that a country's export growth depends on the changes in exchange rates, world GDP, and production capacity(production index is used as a proxy of production capacity in this study).
Equation (2) can be re-written as equations (3) and (4) (Kim 2009) by decomposing world economic growth rate into each country's economic growth rate with its weight in the world GDP.
[ ] where s t j indicates the weight of country j (for example, China in this study) in the world GDP. According to equation (3), the export elasticity of a country (for example, Korea) based on another country's GDP growth (for example, For the purpose of this study, world economic growth rate is separated into the growth rate of a country (China) and the rest of the world without China. In equation (4), WGDP t o and s t o refer to the GDP of the world except for country j and its weight of world GDP, respectively.
Based on equation (4), I set up the VAR model to reflect the endogeneity problem of relative prices and the GDP term in the equation [see equation (5)].
The endogeneity of the GDP term should be considered because in this study, Korean exports and China's growth rate can be affected by common factors such as world GDP, as in Eichengreen (2004) , . 0  (6), where the matrix W represents the ratios mentioned above. Matrix B can be estimated as follows. After estimating matrix A based on Y* [see equation (7)], the parameter matrix B p (=A p W p ) is obtained because the ratio W p is known.
Obtaining the estimated matrix B, I can now calculate the impulse response function following the usual processes.
China's GDP, Korea's real effective exchange rates, and Korean exports.
Second, during this period, China's GDP is not affected by Korea's real effective exchange rates and Korean exports. The first and second assumptions are not so restrictive because Korea has a relatively small economy compared to China 2) . Third, in the contemporaneous period, Korea's real exchange rates are not affected by Korean exports. As is well known, while the trend of exchange rates is definitely affected by the changes in the current account in the long run, it is not so sensitive to the current account in the short run. This is because exchange rates in the short run fluctuate mainly as a result of financial factors such as interest rates and stock prices rather than the real factor current account. 3) Further, the payment for exports is not made at the time of purchase. Consider that the data used in this paper are quarterly data.

Taking these factors into account, this assumption is not so restrictive. Based on the LR and Schwartz Criterion, one lag [VAR(1)] and four lags [VAR(4)]
are chosen for the VAR model, respectively. Considering that the data used in the study are quarterly, I used four lags in the estimation. In the case of real effective exchange rates, a possibility of unit root exists in the level value.
However, in this study, in order to estimate how the Chinese GDP growth rate affects the Korean export growth rate, the first differenced data are used.
Therefore, all variables used in the model are probably stable. In the unit root test for the first differenced data, the null hypothesis of unit root is rejected (see Table 1).
The data for Korean exports are extracted from the Trade Statistics database of the Korean International Trade Association (KITA). This database provides trade data by production process: raw materials, capital goods, and consumption goods. The data for GDP growth rate and industrial production index are extracted from the Global Insight database (historical data). The real effective exchange rates are obtained from the Bank of Settlement homepage. The data 2) Korea's share of Chinese exports is 4-5% over the period considered.
3) In the Korean exchange market, which dealers control, it is known that dealers do not consider macro variables in the short run forecasting of exchange rates. 2) *p<0.10, **p<0.05, ***p<0.01.

Estimation Results
(1) Impact of Chinese GDP growth on Korean exports to the world  (2008).
2) Total08, Raw08, Capital08, and Cons08 represent total exports, exports of raw materials, capital goods, consumption goods in 2008, respectively. On the other hand, it can be found from the results that regardless of the type of export, China's impact on Korean exports is increasing over time, reflecting the change in China's economic share in the world (see Table 2). In particular, for exports of capital goods, the impact of China's GDP growth on Korean exports is increasing rapidly over time (see Capital95 and Capital08 in Table 2). This reflects the fact that the proportion of capital goods in Korea's exports to the world is increasing over time. In this paper, I compare the size of the impact based on the trade weights of 1995 and 2008.
(2) Impact of China's GDP growth on Korea's exports to China The responses of Korea's exports to China to shocks to China's GDP growth are shown in the  An interesting result found in the analysis is that out of the three types of export goods, the export response to a shock to China's GDP growth for raw materials is the smallest in the first period after the shock. This difference from final goods, such as capital and consumption goods, can be understood in that the demand for raw materials depends on the demand for final goods.
I compare the relative importance of explanatory variables in explaining the changes in Korea's exports to China based on variance decomposition of its forecasting error (see Table 4). According to the variance decomposition