SOURCES OF ECONOMIC GROWTH AND CHANGES IN ENERGY CONSUMPTION: EMPIRICAL EVIDENCE FOR TAIWAN (2004-2016)

Since the 1990s, Taiwan has promoted the upgrading of industrial structure, and the financial crisis has brought about major economic shocks. After Taiwan became a WTO member in 2002, the total trade volume increased significantly under the trade liberalization economic system. The trade sector has become the main source of Taiwan's economic growth. As energy consumption expands with economic growth, the instability of international energy prices has become an important factor hindering Taiwan's future economic development. Taiwan’s energy demand is over 99% dependent on imports, making Taiwan’s economic growth and energy consumption really closely related. This study has found that the source of growth has transformed from domestic final demand effect and exports effect to technical coefficients effect and self-sufficiency Coefficients effect. The financial crisis has increased the “self-sufficiency” coefficient effect and the technical coefficient effect, and all other industries except the “agricultural sector” have shown significant growth. On the other hand, the source of energy consumption has changed after the financial crisis. This also confirms that Taiwan's economic growth is closely related to energy consumption, indicating that energy shortage has indeed had an important impact on Taiwan's economic development.


INTRODUCTION
Taiwan has experienced rapid economic growth since the 1960s. Although it suffered from the impact of two world oil crises, it still maintained a high level of economic growth. In 1985, the Gross domestic product (GDP) per capita reached $3,314, and in 1995 it increased to $12,906.
In the process of economic development that Taiwan has experienced many challenges came from international economic environment. Although Taiwan has overcome the "middle income trap," yet the transformation of the industrial structure remained to be a major challenge. After entering the 21 st century, Taiwan's economic growth has gradually slowed down. In 2005, the GDP per capita was US$19,278. After 10 years, it only increased to US$22,288 in 2015. Although Taiwan's GDP per capita increased to US$24,318 in 2017, it faces rising costs due to high international energy prices and uncertainty about electricity supply.
The financial crisis in year 2007 has caused severe damage to the Taiwan's economy. Such event has indirectly indicated that the industrial restructuring has not been successfully accomplished (Hong and Li, 2015). This was because Taiwan's economic growth has been dependent on export sector for a long period of time, which is highly vulnerable to international financial crisis. The primary objective of this paper is to investigate that whether the industrial structure growth pattern and source has altered 10 years after the mentioned financial crisis. This Journal is licensed under a Creative Commons Attribution 4.0 International License In year 2016 the proportion of Taiwan's agriculture, industry and services industries was 1.82%, 35.06% and 63.13%, respectively. Taiwan's industrial sector has reached the peak of 46.06% in year 1986 and then remained declining every year. This trend is considered to be a phenomenon of "post-industrial society." This study analyzes the relationship between the source of economic growth and energy consumption in Taiwan after becoming a member of the WTO. Although the expansion of free trade has brought economic growth, it has also contributed to the fragility of Taiwan's economic structure. In addition to causing economic shocks, has the financial crisis of 10 years brought about a turning point in the upgrading of industrial structure? This study will compare the changes of the factors in the "Review" economic output before and after the financial crisis and summarize the development characteristics of the industry. It also analyzes whether the factors of energy consumption during this period have also changed, which will help to put forward more specific views and suggestions on Taiwan's economic development in the future.

LITERATURE REVIEW
There were many studies focused on economic growth, among which Kuznets (1966) regards the degree of industrialization as an indicator of economic growth. Some studies have pointed out that economic growth requires structural changes through interindustry correlations (Chenery, 1979;Matthews et al., 1982;Abramowitz,1983). In addition to the above studies, the early literatures were most representative of Aka-Matsu (1962) and Vernon (1966). On the other hand, some studies argue that the redistribution of factors of production is an important factor in economic growth (Duarte and Restuccia, 2010). In addition, the main literatures that analyze economic growth from sources of output were Torii and Fukasaku (1979), Fujita and William (1997), and etc. Fujita and William (1997) estimated the impact of Indonesian manufacturing exports on employment. Their results showed that the employment position generated in the early stage is larger in manufacturing sector.
As a result of globalization that the pattern of economic growth has also changed. Gordon (2015) pointed out that the long-term stagnation of the economy may be caused by the lack of innovative technologies. Innovative technologies were deemed to be one of the important source of economic growth. Coulibaly et al. (2018) studied BRICs countries and analyzed the impact of innovationdriven policies on economic growth. Their results after the model test shows that the variables were all positive and statistically significant.
On the other hand, the analysis between economic growth and energy consumption has also accumulated numerous results (Ozturk and Acaravci, 2010;Chandran et al., 2010;Hossain, 2011;Hossein et al., 2012;Li and Zheng, 2012;Omri and Kahouli, 2014;Bildirici, 2016;Wang et al., 2016;Khan et al., 2017;Nuryartono and Rifai, 2017;Tan and Tan, 2018).  analyze the relationship between energy consumption (EC) and economic growth GDP in low-income countries and middle-income countries. The study pointed out that low-income countries have long-run Granger causality relationships from GDP to EC. In the middle-income countries, there is a bidirectional causality relationship between EC and GDP. Khan et al. (2017) Analyze energy consumption from financial development, and the results show a mixed phenomenon between different countries and regions. There are also some studies that indicate that there is no obvious causal relationship between economic growth and energy consumption, such as Dlamini et.al. (2015;2016), Bah and Azam (2017) and other studies.

Data
The I-O

Methodology
In the framework of I-O model, the following balance equation can be derived (Fujita and William, 1997) Industry (X i ) production equilibrium equation can be expressed by (1). The physical quantity bought by sector j to sector i when j produces the commodity j is denoted as X ij This condition can be expressed as: If coefficients are defined in physical terms, it is assumed that a x X ij ij j = for all i and j are stable. Where F is the amount of the domestic final demand for industry (n × 1). M represents the diagonal matrix of import coefficient (n × n). E is the amount of the exports for industry (n × 1). I is the identity matrix (n × n).
A is the input coefficient matrix (n × n).
t represents the base period while t+1 represents the current period.
The following formula is used to calculate δX: The quantity model (2) can be written as: Denoting by B t , B t+1 and B* in the following matrix: The model in (3) can be decomposed as follows: Each of the term measures the effects of a particular source of economic output growth:

Energy Consumption Factor Decomposition Model
The energy consumption factor decomposition model will be established by the 3.1 source of economic output growth model combined with energy consumption. This study will use this model to estimate the amount of change in energy consumption.
The amount of energy consumptions change between the two periods can be written as (5) δ . represent energy consumptions in t years and t+1 years. Establish coefficient of energy consumption E con. with (4). Energy consumption factor decomposition model can be written as (6) δ E

EMPIRICAL RESULTS AND DISCUSSION
The research findings of our model could be illustrated as the following two parts. The first part reveals the analysis of the changes in the output growth of the overall industry from year 2004 to 2016. The second part reveals the analysis of the changes in the source pattern of output growth before, during and after the year 2007 financial crisis.

Output Changes in 2004-2016: An Over-View
According to Table 1 Table 2 indicates the growth of output before the financial crisis that "Light Industries" (−193.214 NT$ billion) and "Agriculturerelated Industries" (−8.405 NT$ billion) both showed a negative growth. Especially that "Light Industries" has negative growth in all five sources, which clearly indicates Taiwan's traditional industry lacks sufficient competitiveness when facing such financial crisis.

Patterns of Changes in the Various Sub-Periods
During the same period, "Machinery-related Industries" (2,398.520 NT$ billion) and "Service-related Industries" (2,220.813 NT$ billion) were the major growing departments. The growth source of "Machinery-related Industries" was mostly contributed by E factor (1,270.646 NT$ billion) while "Service-related Industries" was mostly affected by DF effect (1,986.085 NT$ billion).
After the financial crisis that Taiwan intended to reduce its damage impact on the economy by expanding public investment. Such attempt was reflected in the growth of industry during this period shown as Table 3. The major growth included "Infrastructure Industries" (4,164.661 NT$ billion), "Iron, Non-Iron Industries" (1,239.647 NT$ billion) and "Machinery-related Industries" (1,003.592 NT$ billion). The factors provided growth for these related industries were contributed by the increase of domestic final demand and hence the effeteness of economic policies. During this period "Agriculture-related Industries" (−256.753NT$ billion) and "Light Industries" (−106.508 NT$ billion) remained to be negative in growth. Such results were most affected by SS effect and TC effect.
After nearly 10 years from the financial crisis that Taiwan's output growth sources have changed and so was the growth industry as illustrated in Table 4.
Due to the impact of the financial crisis that the DF effect (−133.843 NT$ billion) and the E effect (−3,145.846 NT$ billion) have gradually turned into negative value. This indicates that both domestic and foreign markets are affected by "negative wealth effects" to reduce demand for products. On the other hand, contributed by the improvement of SS effect (3,629.214 NT$ billion) and TC effect (4,109.537 NT$ billion) that all other industries except for "Agriculture-related Industries" (−31.573NT$ billion) have shown significant growth. Another noteworthy aspect is the negative growth of "Infrastructure Industries" (−337.664 NT$ billion). This negative growth may have been affected by the reduction of government public works and corporate investment.

Changes in Sources of Energy Consumption
It is found from Table 5 that      The period from 2004 to 2006 was the base period, and the coefficient of energy consumption during 2006-2011 was the largest decline with "Agriculture-related Industries", followed by "Infrastructure Industries". During this period, "Chemicalrelated Industries" and "Iron, Non-Iron" Industries' coefficient of energy consumption did not improve much in terms of energy consumption efficiency. The main reason was that it was difficult for capital-intensive industries to significantly update production equipment in a short period of time.
During the period of 2011-2016 after the outbreak of the financial crisis, the efficiency-energy consumption of "Service-related Industries," "Machinery-related Industries" and "Infrastructure Industries" decreased significantly, and all industries fell to 76.42% before the financial crisis. The outbreak of the financial crisis has encouraged industries to reduce energy consumption and improve energy efficiency.
According to the research period (2004-2016) of this paper, Service-related Industries reduced the coefficient of energy consumption to the maximum (26.45%), followed by Machineryrelated Industries (23.50%). Conversely, the decline in the coefficient of energy consumption of Chemical-related Industries was the least, only 7.27%. In addition, the I-O Table is integrated into seven major departments. The source of energy consumption after entering the financial crisis turned to domestic final demand (DF energy effect=253,539,782 KLOE). As shown in Table 7. Although Taiwan's exports were hit by the financial crisis, energy consumption still ranked second (E energy effect= 237,988,280 KLOE). These two factors are most evident in the performance of "Chemical-related Industries", with an increase in energy consumption of 175,549,576 KLOE and 190,143,366 KLOE,respectively. On the other hand, the most effective factor in the reduction of energy consumption during this period is the production of selfsufficiency and the import of final products. During the period of 2006-2011, the energy consumption of 250, 387, 492 KLOE and 438, 489 KLOE was reduced in SS energy effect and FM energy effect, respectively. It is particularly noted that the energy consumption sources of all industries are in a reduced state in the SS energy effect.     After the 2008 financial crisis, Taiwan has changed its sources of energy consumption, as shown in Table 8. The financial crisis has forced the adjustment of production methods of enterprises, although it has contributed to economic growth, but it has also increased consumption of energy. SS energy effect and TC energy effect are the main source of energy consumptions increase, mainly from "Chemical -related Industries." In terms of sources of energy consumption, in addition to "Agriculture-related Industries," "Light Industries" and "Service-related Industries," DF energy effect and E energy effect in other sectors have shown improvements in energy consumption efficiency.

CONCLUSIONS AND POLICY IMPLICATIONS
The above results reveal that the financial crisis has indeed changed the source of output and growth patterns. This study has found that the source of growth has transformed from DF effect and E effect to TC effect and SS effect. This transformation suggests that the future economic growth in Taiwan ought to rely on technological innovation in production to receive higher growth performance. On the other hand, this study also found that the industrial development patterns that drive economic growth have moved from "Machinery-related Industries" to "Infrastructure Industries" and then shifted to "Service-related Industries." This transformation may suggest that Taiwanese economy has entered the so-called "post-industrial society" stage. For the conclusion of the study, we propose that financial crisis has indeed caused changes in Taiwan's economic structure. The subsequences of such transformation have also changed the pattern of industrial development.
However, the financial crisis has made Taiwan's economic development model need to change. Economic sustainable development needs to improve energy consumption efficiency. The study found that the financial crisis caused changes in Taiwan's production process, which also changed the growth factor. These changes are reflected in changes in energy consumption, which indicates that Taiwan's economic development has a certain degree of relationship with energy consumption. The change in coefficient of energy consumption found that 86.20% of Agriculture-related Industries improved most before 2011, and the energy efficiency improvement of Service-related Industries was the most obvious after 2011 (62.83%).This is the financial crisis that has prompted the service industry to move towards efficiency in response to domestic consumer demand. The source of improving energy consumption efficiency is FM energy effect and SS energy effect before 2011, and then changed to DF energy effect and E energy effect after 2011. Such changes indicate that after the financial crisis, Taiwan must adjust its business model in order to cope with the competition between the domestic market and exports.