Empirical Results

Figure 3.5: IRFs of export and production of automobile to oil price shocks

(The dotted line in the following figures is the 95% confidence interval)

a aThe period of export is 2000M7-2009M12;

a aThe period of production is 1993M1-2009M12.

Step 1 shows two facts. First, consistent with typical reactions to oil price shocks, the Japan’s real GDP decreased before 1985 in Figure 3.3. The two oil price crises in the 1970s brought significant adverse effects to the Japanese economy. In contrast, Japan’s real GDP was almost unaffected over time after 1985. Second, similar to the results of Blanchard and Gali (2007), response scales in the Japanese economy became smaller after 1985.

The reason behind the insignificant change in Japan’s real GDP relative to oil price shocks is considered in Step 2. Henceforth, each term in the expenditure side of the national income is analyzed to explain the characteristics of the Japanese economy

Figure 3.4 shows the results on private consumption, total investment, government ex-penditure, and exports and imports in Step 2. Total investment includes private residential and non-residential investments, private and public inventories, and public investment in Japan’s case.

As seen in the first row on the left-hand side of Figure 3.4, prior to 1985, consumers simultaneously lowered their consumption levels because of anticipated decreases in fu-ture income. In the second row on the left-hand side of Figure 3.4, firm investments only show a small negative effect by oil price shocks. Government expenditure is not affected by oil price shocks in the first four quarters shown in the third row on the left-hand side of Figure 3.4. Consumer demand decreased and the Japanese yen depreciated during the period; hence, imports decreased accordingly, as evident in the fifth row on the left-hand side of Figure 3.4. These responses can be explained by empirical literature, such as Barsky and Kilian (2004) and Rotemberg and Woodford (1996).

The unique increase in Japan’s export sector relative to oil price shocks can be seen in the fourth row on the left-hand side of Figure 3.4. Since the 1970s, Japanese automo-biles have entered the international market and have become an important component of Japan’s export business. The series of energy and environmental laws passed in the US since 1980 has objectively promoted the importation of Japanese energy-efficient auto-mobiles in the country, as discussed in Hamilton (2009b).

However, our real interest is on why Japan’s real GDP was not affected after 1985.

Private consumption of Japan was not significantly affected by the oil price shocks, as shown in the first row on the right-hand side of Figure 3.4, because there was no antic-ipated change in the future income. However, surprisingly, the total investment shows a

constant increase in the second row on the right-hand side of Figure 3.4. This characteris-tic of Japan’s economy is the most important after 1985. Government spending decreased to small levels, as shown in the third row on the right-hand side of Figure 3.4. Exports after 1985 particularly became larger than those were before 1985, as shown in the fourth row on the right-hand side of Figure 3.4. Recently, imports have experienced a slight increase, as shown in the fifth row on the right-hand side of Figure 3.4.

Why did firms choose to boost their investments during the oil price shocks after 1985? In the case of the US, the total investment increased, including the infrastructure and equipment investment of firms. This finding corresponds to the response of American firms, as the investment in the oil-mining sector also increased (Edelstein and Kilian, 2007). However, despite the fact that it has no oil-mining sector, why did Japans total investment also increase?

Figure 3.6: The growth rate of total investment and export before 1986

Figure 3.7: The growth rate of total investment and export after 1986

Note that Japanese exports increased in both sub-periods. Is there a correlation be-tween total investment and exports? Figures 3.6 and 3.7 plot the relationship of total investment and exports. Prior to 1986, an obvious relationship was not evident, as shown in Table 3.6, but after 1986, the Granger causality relationship already began, as shown in Table 3.7. When exports increase, firms will spend more money in investment. Corre-spondingly, when firm investments increase, Japanese export will also increase.

Table 3.6: Granger causality test total investment and exports from 1973 to 1985

Null Hypothesis F-statistic Prob

∆lnexportdose not Granger cause∆lninv 0.85877 0.5183

∆lninvdose not Granger cause∆lnexport 1.27703 0.2958

Table 3.7: Granger causality test total investment and exports from 1986 to 2010

Null Hypothesis F-statistic Prob

∆lnexportdose not Granger cause∆lninv 4.00254 0.0051

∆lninvdose not Granger cause∆lnexport 3.75796 0.0074

Next, we should determine which export sector increases during oil price shocks. Au-tomobile exports have the largest share in Japan’s total exports (18.4% in 1988 and 17.1%

in 2007 based on trade statistics of Japan); hence, the automobile sector can be considered an important factor in Step 3. Figure 3.5 shows the trend of Japanese automobile exports and production for different periods, respectively, both of which exhibit an increase during oil price shocks. This result is in accordance with that of Hamilton (2009b).

Note that quarterly data are the shortest frequency data to describe macroeconomic aggregates in primary statistics. They are appropriate frequencies for portraying impulse-response functions in Figures 3.3 and 3.4. As regards the automobile sector of Japan, the Japan Automobile Manufacturers’ Association has counted automobile production since 1990s with monthly frequency. This sector is easily affected by market condition, such as price of gasoline, so monthly data are the best index to be used. Hamilton (2009b)

also uses monthly data to investigate this problem. The use of monthly frequency data to obtain the impulse-response functions is appropriate, as shown in Figure 3.5.

Table 3.8 shows the average weights of the expenditure side of the national income for the two periods in the Japanese economy. The weights of private consumption and total investment decreased, whereas the weights of expenditure of government, export, and import increased in the second sub-period. Moreover, the positive responses of total investment and export offset the negative response of government expenditure, which only created very minimal effects on Japan’s real GDP (Step 4).

Table 3.8: The each term weights in expenditure side of national income period Pri-consumption Total-inv Gov-expenditure Export Import

1973-1985 0.5968 0.2918 0.1031 0.0923 0.0830

1986-2010 0.5632 0.2591 0.1613 0.1041 0.0877

The reasons behind the weaker responses in Japan’s real GDP can be explained us-ing three factors, namely, the positive response of investment after 1985, the increase in Japanese (automobile) exports, and the low government expenditure in the expenditure side of the national income even if it has a negative response.