Results of the Impulse Response Analysis in the MVTEC Model

In this section we study the impact of energy prices on macroeconomic activities by analyzing impulse response functions. Figures 4.2 to 4.4 present the impulse response functions of each energy price (DWTI, DCOAL, and DNG) from one-standard deviation shocks to industrial production (DIP), stock price (DSP), real interest rates (DR), unemployment rates (DUN), exports (DEX), and imports (DIM) in the two-regime models.

The left of Figure 4.2 presents the impulse responses of macroeconomic variables to an oil price shock in regime one. When an oil price change is below the threshold value c^* (regime one), it can be seen that an oil price shock has a positive impact on industrial production. The response of industrial production to oil price shocks is rising in periods 1 and 3. After 5 to 6 periods, the effect gradually dies out.

Moreover, an oil price shock has a persistently negative impact on stock prices over 11 periods. The oil price shock has an immediate positive response in the interest rate, and then falls. This result can be expected as increases in oil price create inflationary effects in the economy which consequently bring an upward pressure on interest rates.

The results for the unemployment rate are somewhat stronger. Except for the first one minor negative response, the graph shows persistent positive responses of unemployment to a shock in oil price. The maximum effect is reached in the second period when the unemployment rate increases by 2%.

The left of Figure 4.3 presents the impulse responses of macroeconomic variables to an oil price shock in regime two. When an oil price change exceeds the threshold value c^* (regime two), an oil price shock has an immediate positive impact on industrial production, and after a minor negative shock it tends to remain for a significant period of time. The IRF analysis shows that oil price shocks exhibit more volatility in the one-regime model than in the two-regime model.

-2

Response of DIP to DW TI

-2

Response of DIP to DCOAL

-2

Response of DIP to DNG

-2

Response of DSP to DW TI

-2

Response of DSP to DCOAL

-2

Response of DSP to DNG

-8

Response of DR to DW TI

-10

Response of DR to DCOAL

-10

Response of DR to DNG

-8

Response of DUN to DW TI

-8

Response of DUN to DCOAL

-10

Response of DUN to DNG

-2

Response of DEX to DW TI

-3

Response of DEX to DCOAL

-3

Response of DEX to DNG

-3

Response of DIM to DW TI

-3

Response of DIM to DCOAL

-3

Response of DIM to DNG

Figure 4.2 Impulse Responses from One Standard Deviation Shock of Energy Price Change in the Regime One VAR Model (12 Periods Forward)

-12

Response of DIP to DW TI

-4

Response of DIP to DCOAL

-4

Response of DIP to DNG

-12

Response of DSP to DW TI

-3

Response of DSP to DCOAL

-10

Response of DSP to DNG

-80

Response of DR to DW TI

-12

Response of DR to DCOAL

-30

Response of DR to DNG

-30

Response of DUN to DW TI

-12

Response of DUN to DCOAL

-20

Response of DUN to DNG

-12

Response of DEX to DW TI

-4

Response of DEX to DCOAL

-10

Response of DEX to DNG

-10

Response of DIM to DW TI

-6

Response of DIM to DCOAL

-12

Response of DIM to DNG

Figure 4.3 Impulse Responses from One Standard Deviation Shock of Energy Price Change in the Regime Two VAR Model (12 Periods Forward)

The response of stock prices to shocks in oil price is positive up to the first period and eventually declines. An oil price shock has an initial minor negative impact on interest rates and then the increase lasts for approximately 12 periods. The responses of the unemployment rate are only after an initial slight positive impact and then fall.

The middle of Figures 4.3 and 4.4 shows the responses of macroeconomic variables to a one-standard deviation coal price shock. When a coal price change is below the threshold value c^* (regime one), it can be observed that a positive coal price shock increases industrial production in period 3. However, the responses of industrial production increase for one period and then fall. After 3-4 periods, the effect gradually dies out. This result suggests that coal price shocks have a delayed positive impact on industrial production. A coal price shock has a slight negative impact on stock prices that lasts for approximately 4 periods. In particular, the unemployment rate initial rise lasts for approximately 2 periods and then decreases in period 3. The results show that following a 10% increase in coal price, the unemployment rate increases immediately by 3–4%. The maximum effect is reached in the first period, after which the effect gradually dies out. By the same token, a positive coal price shock increases exports and imports for the two periods and affects Taiwan economy positively.

When a coal price change exceeds the threshold value c^* (regime two), a coal price shock has a slight negative impact on industrial production that lasts for approximately 3 periods. In periods 4 and 6, the response of industrial production to shocks in coal price is increasing and then the effect dies out. As expected, a coal price shock has a negative impact on stock prices. The graph presents that the response of stock prices to shocks in coal price is positive up to the first period and it eventually declines. A coal price shock has a negative impact on the interest rate expected for periods 4 and 8.

Similar to stock prices, there is a lag effect on the unemployment rate.

Figures 4.3-4 show the responses of six variables to a natural gas price shock.

When a natural gas price change is below the threshold value c^* (regime one), a natural gas price shock has a significant positive impact on industrial production. The industrial production initial rise lasts for approximately 2 periods and increases in periods 4-5. The figure also shows that following a natural gas price shock, stock prices increase immediately by 0.5%. After that, stock prices decrease persistently for 2-4 periods. The maximum effect is reached in period 4, when stock prices have decreased by 1%. The responses exhibit more volatility and last for a long term. As expected, a natural gas price shock has a positive impact on the interest rate.

When a natural gas price change exceeds the threshold value c^* (regime two), a natural gas price shock has a small negative impact on the interest rate. After 3-4 periods, the effect gradually dies out. As expected, stock prices initially rise, probably affected by the rise in economic activity, and then die out in period 3.

The results of the impulse response of energy price changes are summarized: (1) In the two-regime model, energy price shocks have negative impact on industrial production. When energy price changes are below the threshold value c^*, the impulse responses of oil price shocks exhibit more volatility than coal and natural gas price shocks. In addition, the threshold values reflect periods of industrial production for each energy price: natural gas, oil, and coal. It also finds that an oil price shock has a delayed negative impact on industrial production with one lag in regime two. In particular, the impulse responses of energy price shocks in regime one are more than that of energy price shocks in regime two. (2) As evident from the IRFs, energy price shocks have a negative impact on stock prices in regime two. The threshold values reflect responses of stock prices for each energy price: natural gas, oil, and coal.

Moreover, the longest response period for stock prices is in an oil price shock. When energy price changes exceed the threshold value c^*, energy price shocks have delayed

negative impacts on the stock market. Within the framework of the one-regime model, a significant relationship between energy price change and stock prices is few and far between. The findings speak to the fact that the two-regime model seems to offer more detailed responses. (3) As expected, interest rates initially rise, probably affected by the rise in economic activity, and then fall. However, an oil price shock or a natural gas price shock has an immediate positive response from the interest rate in regime two.

(4) In the two-regime model, energy price shocks have positive response on the unemployment rate. (5) As indicated from the above results, when an energy price change is below the threshold value c^*, it is capable of explaining a significant portion of macroeconomic activities.