Gradual Reforms and the Emergence of Energy Market

Chapter
Part of the Lecture Notes in Energy book series (LNEN, volume 13)

Abstract

This chapter is organized as follows: we first use maps and graphs of spatial energy prices and then statistically test, using both conventional unit root tests and panel unit root tests, issued related to convergence in energy prices over time and space. Then, we compare our results and finding with other’s and finally provide some conclusions.

Keywords

Unit Root Test Electricity Price Price Series Panel Unit Root Test Price Trend 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

This chapter is organized as follows: we first use maps and graphs of spatial energy prices and then statistically test, using both conventional unit root tests and panel unit root tests, issued related to convergence in energy prices over time and space. Then, we compare our results and finding with other’s and finally provide some conclusions.

9.1 Spatial Price Trends

Given the likely importance of transport costs when attempting to interpret the evidence on market integration, it is useful to identify which of China’s provinces are energy producers and which are energy consumers. Table  3.9 reports energy outputs and deficits (consumption) for coal, electricity, gasoline and diesel. It can be seen that the major coal producers, ordered by volume, are Taiyuan (581 mmt), Hohhot (298 mmt), Zhengzhou (195 mmt), Xi’an (183 mmt), Jinan (141 mmt), Guiyang (118 mmt), Harbin (103 mmt); The major coal importers, ordered by volume, are Nanjing (153 mmt), Jinan (149 mmt), Shijiazhuang (130 mmt), Hangzhou (113 mmt), and Guangzhou (111 mmt), and major coal exporters, ordered by volume, are Taiyuan (298 mmt), Xi’an (143 mmt), Hohhot (136 mmt), Lanzhou (30 mmt) and Guiyang (19 mmt).

Figure 3.4 of  Chap. 3 shows the distribution of major cities across China. Juxtaposing Fig. 3.4 against Table 3.9 in  Chap. 3, we see a general picture of how energy is transported in China. Based on the assumption that energy prices are lower in the producing areas and higher in the consuming areas due to transportation costs, to understand whether there is an energy market in China, we need to observe the spatial price series ordered from producing to consuming areas. Likewise, the consuming areas can show price differentials caused by transportation costs if ordered by their distances to the producing areas. To simplify our analysis, we choose major energy producers and major energy consumers to compare the price trends and spatial patterns.

9.1.1 Coal

Consider, firstly, the spatial observations and comparisons of coal prices. Recall that coal is mainly produced in the north and the main transport routes are from west to east and from north to south. Note the spatial coal price trends of Figs. 9.1, 9.2 and 9.3, of which six are spatial coal price trends ordered from producing areas to consuming areas and one is from three southeast consuming areas ordered by their distances to major producing areas. There are several points that can be made based on these data on spatial coal price trends. Firstly, all spatial coal price trends demonstrate some evidence of the Law of One Price (LOP), which suggests that price differentials are mainly caused by transportation costs. Secondly, the spatial price trends do not illustrate consistent evidence of the LOP over the whole sample period as price differentials are not consistent over time for example, the price differentials were small from 1999 to 2001 for some changes (e.g., Panel A and Panel B of Fig. 9.1, Panel B of Fig. 9.2, Fig. 9.3). Finally, some price differentials display the LOP relation only late in the sample and in an irregular fashion for example, the spatial coal price trend from Xi’an via Wuhan to Changsha does not show a LOP relation during 1997–1999 and 2003–2004 (Panel A of Fig. 9.2). This finding probably indicates that the coal price deregulations may be heterogeneous across the country. It is noted that coal prices in consuming areas are divergent from in producing areas after 2002 (Panel A and Panel C of Figs. 9.1, 9.2), see the next section for an explanation.
Fig. 9.1

The trends of coal spot prices from major producing to consuming areas. Panel A: From Harbin to Shanghai and Guangzhou. Panel B: From Taiyuan to Wuhan and Guangzhou. Panel C: From Xi’an to Jinan and Shanghai. Note: Unit is ¥/t

Fig. 9.2

The trends of coal spot prices from major producing areas to consuming areas. Panel A: From Xi’an to Wuhan and Changsha. Panel B: From Datong to Jinan. Panel C: From Hohhot to Wuhan and Guangzhou. Note: Unit is ¥/t

Fig. 9.3

The trends of coal spot prices along southeast coastal major consuming areas. Note: Hangzhou, Fuzhou and Guangzhou ordered by distance to major producing areas (¥/t)

9.1.2 Gasoline

Figures 9.4 and 9.5 present the spatial price trends for gasoline from producing to consuming areas (Fig. 9.4), along consuming areas (Panel A and Panel B of Fig. 9.5) and along ports areas (Panel C of Fig. 9.5). It is notable that the spatial gasoline price trends are very consistent for not only ‘from gasoline producing areas to consuming areas’, but for ‘all consuming areas’ as well. Secondly, although spatial gasoline prices are very similar, the averaged statistics still show some spatial price differentials caused by transportation distance for example, Panel A of Fig. 9.4 shows the price differentials from producing areas, Shenyang where the gasoline average price is ¥4,182 per ton from 2000 to 2005, to consuming areas by distance, Hohhot where the gasoline average price is ¥4,244 per ton and Wuhan where the gasoline average price is ¥4,364 per ton during the same period. Likewise, Panel B and Panel C of Fig. 9.4 also clearly shows the transportation costs from producing areas, Xi’an and Lanzhou where average gasoline prices are ¥4,169 and ¥4,174 per ton from 2000 to 2005, respectively, to consuming areas, via Chengdu where average gasoline price is ¥4,258 per ton to Xining where average gasoline price is ¥4,285 during the same period. Thirdly, gasoline spot prices are similar in the consuming areas, while price differentials can be still clearly be seen due to their different distances to the producing areas for example, along the east coastal areas, Nanjing is closer to gasoline producing areas (e.g., Shenyang, Xi’an and Lanzhou) so that its gasoline spot price is lowest among the east coastal areas of Nanjing, Shanghai, Hangzhou and Fuzhou (Panel B of Fig. 9.5, also refer to Fig. 3.4  Chap. 3). Note that Panel A of Fig. 9.5 shows Guangdong as having the lowest gasoline spot price even though it is the most southern, which may reflect oil smuggling in southern China and if so, Guangzhou might be taken as a gasoline producing area. Fourthly, gasoline spot prices are homogeneous around port areas (e.g., Shanghai, Tianjin, Qingdao and Dalian, Panel C of Fig. 9.5), which might reflect similar oil import prices. Finally, it seems that there are two different periods of price differentials, separated around 1999–2000. This may indicate that government intervention played a more important role in energy price formation during the first sub-period, while market forces were more powerful during the second sub-period.
Fig. 9.4

The trends of gasoline spot prices from major producing to consuming areas. Panel A: From Shenyang to Hohhot and Wuhan. Panel B: From Xi’an to Chengdu and Nanning. Panel C: From Lanzhou to Chengdu and Nanning. Note: Unit is ¥/t

Fig. 9.5

The trends of gasoline spot prices around major consuming and port areas. Panel A: Along southern Guangzhou, Nanning, Fuzhou and Kunming. Panel B: Along east coastal Nanjing, Shanghai, Hangzhou and Fuzhou. Panel C: Around major port, Shanghai, Tianjin, Qingdao and Dalian. Note: Unit is ¥/t

9.1.3 Diesel

Figures 9.6 and 9.7 present the spatial price trends from producing areas to consuming areas (Fig. 9.6), along the east coastal consuming areas (Panel A of Fig. 9.7), along southern consuming areas (Panel B of Fig. 9.7), and along northern consuming areas (Panel C of Fig. 9.7). There are several points that can be made. Firstly we observe that the spatial diesel spot price trends are consistent for not only the producing areas to consuming areas, but for all consuming areas, although there is still some variation in time and space. Secondly, diesel spot prices appear homogeneous, while the data support the hypothesis that spatial price differentials are due to transport costs for example, Panel A of Fig. 9.6 shows that between 2000 and 2005 the average diesel price was ¥3,659 per ton from producing areas (Shenyang) to ¥3,725 per ton to consuming areas (Hohhot). Likewise, Panel B of Fig. 9.6 shows the 2000–2005 average prices from producing areas (Urumqi, ¥3,469 per ton) to consuming areas via Chengdu (¥3,606 per ton) to Guiyang (¥3,646 per ton). Panel C of Fig. 9.6 presents the 2000–2005 average spot prices from producing area (Xi’an, ¥3,580 per ton) to consuming areas via Chengdu (¥3,606 per ton) to Kunming (¥3,651 per ton). Thirdly, diesel spot prices appear homogeneous in the consuming areas and it seems there is an oil importing centre in the south for example, along the east coastal areas and the south areas, both Panel A and Panel B of Fig. 9.7 shows that Guangdong acts like a producing area. Similarly, we suspect that there might be large quantity of oil smuggling in southern China. If so, Guangzhou might be taken as a producing area. Fourthly, it seems that there are two different periods of price differential trends. The first period up to 2000 shows spot prices as flat with little variation. However, the second period from 2000 shows much more variability.
Fig. 9.6

The trends of diesel spot prices from major producing to consuming areas. Panel A: From Shenyang to Hohhot. Panel B: From Urumqi to Chengdu and Guiyang. Panel C: From Xi’an to Chengdu and Kunming. Note: Unit is ¥/t

Fig. 9.7

The trends of diesel spot prices along major consuming areas. Panel A: Along east coastal Guangzhou, Fuzhou and Shanghai. Panel B: Along southern Guangzhou, Nanning, Fuzhou and Guiyang. Panel C: Along northern Taiyuan, Hohhot, Shijiazhuang and Jinan. Note: Unit is ¥/t

9.1.4 Electricity

Finally, we look at changes in electricity spot prices in China. Figures 9.8 and 9.9 display the spatial price trends from producing areas to consuming areas (Fig. 9.8), along east coastal consuming areas (Panel A of Fig. 9.9), and along northern major consuming areas (Panel B of Fig. 9.9). The first three figures allow us to observe whether the LOP exists for spatial electricity spot prices in China, while the second two figures allow us to observe the possibility of convergence of electricity spot prices given similar transmission distance. Several points can be made for the spatial electricity spot prices. Firstly, it can be seen from Fig. 9.8 that electricity spot prices are lower in producing areas, rising with increasing transmission distance for example, Panel A of Fig. 9.8 shows that electricity spot prices are lower in two producing areas (Kunming and Guiyang) than in a major consuming area (Guangdong). It also shows that spot prices tend to become more convergent in two producing areas when they are very close in location using hydropower. The same can be observed for coal-burning power plants, Panel B and Panel C of Fig. 9.8 (for hydroelectricity mainly located along the Yangzi River). Next, the electricity spot prices in major consuming areas tend to converge given similar distances to producers, for example, Panel A of Fig. 9.9 shows that electricity spot price differentials along the east coastal major consuming cities converge in the later periods, particularly after 2002. A clearer convergent scenario of electricity spot prices can be seen from Panel B of Fig. 9.9 based on northern major consuming cities (Shijiazhuang, Jinan, Nanjing and Zhengzhou). It is apparent that more stable (flat) price trends can be seen in the early periods while more volatile price patterns emerge in the later years. Again, this finding suggests that market forces have been playing a more important role in electricity price formation in China. Finally, it should be noted, however, that electricity spot prices appear to involve more government intervention than other energy types. This suggests that the emergence of an electricity market in China may be a more recent and embryonic process.
Fig. 9.8

The trends of electricity spot prices from major producing to consuming areas. Panel A: From Guiyang and Kunming to Guangzhou. Panel B: From Taiyuan to Beijing. Panel C: From Wuhan to Shanghai and Guangzhou. Note: Unit is ¥/1,000 kWh

Fig. 9.9

The trends of electricity spot prices along major consuming areas. Panel A: Along east coastal Shanghai, Hangzhou, Fuzhou and Guangzhou. Panel B: Along northern Shijiazhuang, Jinan, Nanjing and Zhengzhou. Note: Unit is ¥/1,000 kWh

In summary, although we are unable to draw hard conclusions based on our descriptive statistics, the patterns of energy price movements would seem to indicate the potential convergence of spatial energy prices and provide support for the emergence of an energy market. The alternative would be to suggest that a ‘central planner’ is generating an observationally equivalent set of integrated, spatially determined, energy prices whose variation is determined solely by the distance of consumption from production.

9.2 Unit Root Tests

The previous section used ‘ocular tests’ for convergence based upon time series graphs and patterns. In this section we use parametric time-series based tests for convergence. We first briefly, introduce the methods used and then present the empirical results. Given the apparent changes in price regimes identified above, we will consider several sub-periods.

A powerful approach used to investigate price convergence applies unit root tests to examine whether price differentials are stationary (see for example, Bernard and Durlauf 1996; Greasley and Oxley 1997). Rejection of the unit root hypothesis implies that the time series of relative prices are stationary, such that relative prices will converge in the long run. Otherwise, if the tests fail to reject the null hypothesis, the relative price series will follow a random walk (Fan and Wei 2006).

9.2.1 The ADF Unit Root Tests

The methodology used here commences with the ADF unit root tests on the raw price data. The individual ADF unit root test results are displayed in Table 9.1 for the price level series and Table 9.2 for the first differenced price series for the raw data series during the whole sample period (1995–2005). The ADF unit root tests show that each of the 35 city raw price data series exhibits unit roots for all four energy sources. All the tests suggest that the first differences of the series are stationary and therefore that all series are integrated of order 1 or I (1).
Table 9.1

ADF unit root tests for raw prices (level, 1995–2005)

City market

Coal

Electricity

Gasoline

Diesel

t-stat.

p-values

t-stat.

p-values

t-stat.

p-values

t-stat.

p-values

Beijing

−1.45

0.845

−2.06

0.564

−1.45

0.844

−2.82

0.190

Tianjin

−2.26

0.453

−3.08

0.111

−2.13

0.526

−2.97

0.141

Shijiazhuang

−2.29

0.436

−1.87

0.666

−1.67

0.761

−3.00

0.133

Taiyuan

−1.49

0.829

−1.65

0.771

−1.35

0.874

−2.06

0.565

Hohhot

−1.04

0.936

−2.54

0.309

−1.37

0.869

−2.69

0.239

Shenyang

−2.52

0.320

−1.96

0.620

−2.28

0.441

−2.79

0.200

Changchun

−1.82

0.693

−1.29

0.889

−2.52

0.317

−3.09

0.111

Harbin

−1.59

0.792

−1.38

0.866

−2.02

0.591

−3.47

0.044

Shanghai

−0.97

0.946

−1.52

0.820

−2.28

0.443

−2.28

0.444

Nanjing

−1.76

0.720

−1.73

0.735

−1.89

0.659

−2.66

0.253

Hangzhou

−1.01

0.940

−2.99

0.134

−1.99

0.603

−2.49

0.331

Hefei

−1.21

0.906

−1.97

0.615

−1.32

0.882

−2.48

0.337

Fuzhou

−0.77

0.966

−1.20

0.908

−1.50

0.827

−3.01

0.132

Nanchang

−1.86

0.672

−2.83

0.187

−1.97

0.616

−2.37

0.393

Jinan

−1.52

0.822

−1.71

0.746

−1.69

0.754

−2.18

0.499

Zhengzhou

−1.44

0.899

−3.25

0.077

−1.86

0.672

−2.31

0.424

Wuhan

−0.79

0.964

−2.69

0.240

−1.52

0.822

−2.56

0.298

Changsha

−0.51

0.983

−1.91

0.649

−1.84

0.685

−2.93

0.155

Guangzhou

−1.34

0.876

−1.13

0.922

−2.05

0.569

−2.83

0.188

Nanning

−1.13

0.922

−2.39

0.384

−1.93

0.638

−2.62

0.270

Haikou

−1.73

0.736

−1.44

0.849

−1.62

0.784

−2.44

0.359

Chongqing

−2.13

0.531

−2.53

0.315

−2.09

0.545

−2.12

0.530

Chengdu

−1.47

0.838

−2.88

0.170

−1.75

0.728

−2.52

0.321

Guiyang

−1.48

0.836

−1.86

0.673

−2.56

0.301

−2.25

0.458

Kunming

−1.49

0.829

−2.75

0.216

−2.19

0.493

−2.68

0.246

Lhasa

−2.23

0.469

−2.84

0.186

−2.97

0.143

Xi’an

−2.54

0.308

−3.19

0.087

−1.36

0.870

−3.34

0.061

Lanzhou

−2.81

0.196

−3.09

0.110

−1.84

0.683

−3.12

0.104

Xining

−0.43

0.986

−2.27

0.449

−1.55

0.811

−2.47

0.343

Yinchuan

−0.84

0.960

−1.51

0.824

−1.78

0.714

−3.06

0.118

Urumqi

−1.37

0.867

−2.53

0.315

−1.48

0.834

−3.25

0.077

Qingdao

−2.39

0.382

−2.08

0.555

−2.52

0.320

Dalian

−1.73

0.738

−1.81

0.697

−2.21

0.485

−2.58

0.289

Xiamen

−1.98

0.612

−1.65

0.770

−2.17

0.507

−3.01

0.129

Ningbo

−1.70

0.749

−1.73

0.738

−1.99

0.602

−2.17

0.503

Note: Null hypothesis is that each series contains a unit root; ADF is the Augmented Dickey-Fuller test; MacKinnon (1996) one-sided p-values; lag length chosen via Hannan-Quinn information criteria

Table 9.2

ADF unit root tests for first differences (1995–2005)

City market

Coal

Electricity

Gasoline

Diesel

t-stat.

p-values

t-stat.

p-values

t-stat.

p-values

t-stat.

p-values

Beijing

−4.93

0.000

−29.65

0.000

−28.40

0.000

−6.99

0.000

Tianjin

−19.77

0.000

−32.86

0.000

−25.44

0.000

−26.61

0.000

Shijiazhuang

−14.06

0.000

−9.83

0.000

−25.81

0.000

−7.35

0.000

Taiyuan

−22.06

0.000

−33.22

0.000

−27.10

0.000

−27.33

0.000

Hohhot

−9.20

0.000

−30.45

0.000

−24.22

0.000

−7.73

0.000

Shenyang

−11.37

0.000

−11.13

0.000

−24.91

0.000

−7.25

0.000

Changchun

−4.52

0.002

−33.52

0.000

−26.60

0.000

−26.06

0.000

Harbin

−21.36

0.000

−27.74

0.000

−25.58

0.000

−23.36

0.000

Shanghai

−21.06

0.000

−31.29

0.000

−25.64

0.000

−25.30

0.000

Nanjing

−20.89

0.000

−29.63

0.000

−28.30

0.000

−30.88

0.000

Hangzhou

−5.70

0.000

−30.50

0.000

−26.21

0.000

−28.14

0.000

Hefei

−18.47

0.000

−33.14

0.000

−26.31

0.000

−28.35

0.000

Fuzhou

−10.03

0.000

−32.61

0.000

−28.73

0.000

−24.90

0.000

Nanchang

−23.39

0.000

−11.51

0.000

−27.44

0.000

−28.51

0.000

Jinan

−5.53

0.000

−27.18

0.000

−25.82

0.000

−23.26

0.000

Zhengzhou

−19.86

0.000

−6.98

0.000

−26.99

0.000

−28.32

0.000

Wuhan

−8.10

0.000

−28.84

0.000

−27.21

0.000

−26.96

0.000

Changsha

−12.11

0.000

−23.62

0.000

−24.48

0.000

−27.74

0.000

Guangzhou

−19.44

0.000

−29.34

0.000

−26.46

0.000

−27.79

0.000

Nanning

−20.47

0.000

−26.90

0.000

−30.71

0.000

−9.01

0.000

Haikou

−10.93

0.000

−32.77

0.000

−27.64

0.000

−25.48

0.000

Chongqing

−7.88

0.000

−10.91

0.000

−30.11

0.000

−25.47

0.000

Chengdu

−6.93

0.000

−32.20

0.000

−5.54

0.000

−6.74

0.000

Guiyang

−6.73

0.000

−28.91

0.000

−28.20

0.000

−26.16

0.000

Kunming

−14.13

0.000

−26.04

0.000

−28.50

0.000

−26.50

0.000

Lhasa

−11.28

0.000

−33.41

0.000

−31.73

0.000

Xi’an

−26.18

0.000

−34.10

0.000

−28.69

0.000

−9.29

0.000

Lanzhou

−6.66

0.000

−30.67

0.000

−26.42

0.000

−6.27

0.000

Xining

−19.20

0.000

−35.04

0.000

−27.34

0.000

−26.52

0.000

Yinchuan

−8.13

0.000

−21.56

0.000

−25.29

0.000

−26.12

0.000

Urumqi

−27.73

0.000

−31.08

0.000

−28.49

0.000

−28.34

0.000

Qingdao

−19.80

0.000

−20.01

0.000

−19.52

0.000

Dalian

−7.78

0.000

−30.61

0.000

−25.20

0.000

−7.12

0.000

Xiamen

−10.09

0.000

−28.34

0.000

−27.68

0.000

−9.10

0.000

Ningbo

−21.20

0.000

−31.29

0.000

−25.83

0.000

−26.63

0.000

Note: Null hypothesis is that each series contains a unit root; ADF is the Augmented Dickey-Fuller test; MacKinnon (1996) one-sided p-values; lag length chosen via Hannan-Quinn information criteria

However, the ADF test statistics are biased towards the non-rejection of a unit root when there are structural breaks in the data (Nelson and Plosser 1982; Perron 1989; Enders 1995). Thus, testing for the presence of the structural break is important, although the approach taken here is to consider sub-periods where breaks are imposed based upon our prior expectations informed by known changes in regimes where, energy economic reforms have been carried out since the early 1990s. Therefore, energy reform would likely produce some structural changes of price series and at worse, spurious breaks would lead only to a reduction in the power of the test (Nelson and Plosser 1982; Perron 1989; Enders 1995). We therefore split the whole sample into two sub-periods (1995–1999 and 2000–2005) based on the previous analysis of price trend figures and our understanding of China’s energy price reforms. The ADF unit root tests are displayed in Table 9.3 for the period of 1995–1999 and in Table 9.4 for the period of 2000–2005, which suggests that the first differences of the series are stationary and therefore that all series are integrated of order 1 or I (1). This means that, even if there are structural breaks in the series, their time series properties are not affected.
Table 9.3

ADF unit root tests for raw prices (p-values, 1995–1999)

City market

Level

First difference

Coal

Electricity

Gasoline

Diesel

Coal

Electricity

Gasoline

Diesel

Beijing

0.858

0.591

0.053

0.591

0.000

0.000

0.000

0.000

Tianjin

0.844

0.275

0.611

0.275

0.000

0.000

0.000

0.000

Shijiazhuang

0.932

0.564

0.408

0.564

0.000

0.000

0.000

0.000

Taiyuan

0.855

0.630

0.516

0.630

0.000

0.000

0.000

0.000

Hohhot

0.540

0.453

0.362

0.453

0.000

0.000

0.000

0.000

Shenyang

0.239

0.265

0.044

0.265

0.000

0.000

0.000

0.000

Changchun

0.913

0.650

0.664

0.650

0.000

0.000

0.000

0.000

Harbin

0.897

0.764

0.046

0.764

0.000

0.000

0.000

0.000

Shanghai

0.691

0.475

0.008

0.475

0.000

0.000

0.000

0.000

Nanjing

0.803

0.582

0.854

0.582

0.000

0.041

0.000

0.041

Hangzhou

0.619

0.608

0.940

0.608

0.000

0.001

0.258

0.001

Hefei

0.407

0.403

0.319

0.403

0.000

0.000

0.000

0.000

Fuzhou

0.894

0.710

0.162

0.710

0.000

0.000

0.000

0.000

Nanchang

0.879

0.289

0.117

0.289

0.000

0.000

0.000

0.000

Jinan

0.011

0.624

0.756

0.624

0.000

0.000

0.000

0.000

Zhengzhou

0.646

0.253

0.000

0.253

0.000

0.002

0.000

0.002

Wuhan

0.984

0.901

0.894

0.901

0.000

0.000

0.729

0.000

Changsha

0.652

0.083

0.156

0.083

0.000

0.000

0.000

0.000

Guangzhou

0.990

0.605

0.592

0.605

0.000

0.000

0.000

0.000

Nanning

0.635

0.801

0.928

0.801

0.000

0.000

0.000

0.000

Haikou

0.903

0.480

0.088

0.480

0.000

0.000

0.000

0.000

Chongqing

0.721

0.652

0.326

0.652

0.000

0.000

0.000

0.000

Chengdu

0.727

0.783

0.665

0.783

0.000

0.000

0.171

0.000

Guiyang

0.626

0.895

0.466

0.895

0.000

0.000

0.000

0.000

Kunming

0.902

0.494

0.034

0.494

0.000

0.000

0.000

0.000

Lhasa

0.652

0.395

0.652

0.000

0.000

0.000

Xi’an

0.354

0.397

0.038

0.397

0.000

0.000

0.000

0.000

Lanzhou

0.124

0.001

0.235

0.001

0.002

0.000

0.000

0.000

Xining

0.000

0.840

0.529

0.840

0.000

0.000

0.000

0.000

Yinchuan

0.442

0.193

0.454

0.193

0.000

0.000

0.000

0.000

Urumqi

0.913

0.870

0.645

0.870

0.000

0.104

0.000

0.104

Qingdao

0.652

0.840

0.652

0.000

0.000

0.000

Dalian

0.238

0.932

0.651

0.932

0.000

0.000

0.000

0.000

Xiamen

0.397

0.617

0.073

0.617

0.000

0.000

0.000

0.000

Ningbo

0.847

0.719

0.868

0.719

0.000

0.000

0.000

0.000

Note: Null hypothesis is that each series contains a unit root; ADF is the Augmented Dickey-Fuller test; lag length chosen via Hannan-Quinn information criteria

Table 9.4

ADF unit root tests for raw prices (p-values, 2000–2005)

City market

Level

First difference

Coal

Electricity

Gasoline

Diesel

Coal

Electricity

Gasoline

Diesel

Beijing

0.762

0.000

0.881

0.416

0.141

0.000

0.000

0.001

Tianjin

0.192

0.252

0.711

0.269

0.000

0.000

0.000

0.000

Shijiazhuang

0.271

0.554

0.659

0.287

0.000

0.000

0.000

0.000

Taiyuan

0.572

0.450

0.908

0.631

0.000

0.000

0.000

0.000

Hohhot

0.891

0.320

0.789

0.595

0.000

0.000

0.000

0.000

Shenyang

0.820

0.578

0.686

0.383

0.153

0.000

0.000

0.000

Changchun

0.846

0.099

0.702

0.431

0.004

0.000

0.000

0.000

Harbin

0.522

0.164

0.718

0.258

0.000

0.000

0.000

0.000

Shanghai

0.608

0.125

0.663

0.541

0.035

0.000

0.000

0.000

Nanjing

0.053

0.242

0.683

0.502

0.000

0.000

0.000

0.000

Hangzhou

0.971

0.369

0.614

0.212

0.001

0.000

0.000

0.000

Hefei

0.612

0.166

0.764

0.586

0.098

0.000

0.000

0.000

Fuzhou

0.936

0.225

0.698

0.707

0.112

0.000

0.000

0.000

Nanchang

0.593

0.003

0.694

0.388

0.000

0.000

0.000

0.000

Jinan

0.360

0.810

0.847

0.301

0.008

0.000

0.000

0.000

Zhengzhou

0.764

0.476

0.602

0.594

0.000

0.000

0.000

0.000

Wuhan

0.484

0.003

0.852

0.333

0.000

0.000

0.013

0.000

Changsha

0.913

0.065

0.582

0.350

0.000

0.000

0.000

0.000

Guangzhou

0.658

0.719

0.841

0.420

0.000

0.000

0.000

0.000

Nanning

0.884

0.239

0.586

0.590

0.000

0.000

0.000

0.000

Haikou

0.052

0.667

0.302

0.000

0.000

0.000

Chongqing

0.783

0.319

0.711

0.685

0.000

0.000

0.000

0.000

Chengdu

0.722

0.607

0.741

0.560

0.000

0.000

0.012

0.002

Guiyang

0.588

0.460

0.678

0.604

0.000

0.000

0.000

0.000

Kunming

0.552

0.233

0.516

0.366

0.000

0.000

0.000

0.000

Lhasa

0.767

0.172

0.647

0.377

0.236

0.000

0.000

0.000

Xi’an

0.484

0.381

0.668

0.175

0.000

0.000

0.000

0.000

Lanzhou

0.481

0.187

0.699

0.334

0.000

0.000

0.000

0.000

Xining

0.670

0.378

0.780

0.676

0.000

0.000

0.000

0.000

Yinchuan

0.604

0.751

0.678

0.152

0.000

0.000

0.000

0.000

Urumqi

0.620

0.328

0.729

0.215

0.033

0.000

0.000

0.003

Qingdao

0.767

0.459

0.331

0.522

0.236

0.000

0.000

0.000

Dalian

0.507

0.386

0.751

0.356

0.000

0.000

0.000

0.000

Xiamen

0.430

0.609

0.744

0.451

0.000

0.000

0.000

0.000

Ningbo

0.425

0.604

0.548

0.592

0.155

0.000

0.000

0.000

Note: Null hypothesis is that each series contains a unit root; ADF is the Augmented Dickey-Fuller test; lag length chosen via Hannan-Quinn information criteria

Table 9.5 presents unit root tests of price convergence for the whole sample period and Table 9.6 for the two sub-periods (1995–1999 and 2000–2005) for 35 provincial capital city markets using the group averages of all 35 markets as a benchmark. It can be seen from Table 9.5 that of the 35 relative price series there are five instances of convergence for coal, two for electricity, 13 for gasoline and 14 suggest convergence for diesel. The general results suggest that there are fewer instances of convergence for coal and electricity than for gasoline and diesel. From these simple tests we cannot conclude that China’s energy markets are integrated. To test the proposition further consider the results in Table 9.6, which show the number of rejections of the null for each of the four energy types and each of the 35 markets during the two sub-periods. Based on these results it appears that there are no apparent differences in the number of rejects across the two sub-periods except for coal. This suggests that not all market prices are convergent to the national average prices and that national aggregated markets might not exist, but instead regional markets exist. However, there is some evidence that market forces are gradually having some effect in determining energy price formation as there are more instances supporting convergence in the second sub-period than the first sub-period, except in the case of coal. The most obvious example of this can be found for the case of diesel where there are only 15 pairs of convergent relative price series in the first sub-period (1995–1999), but 20 pairs of convergent relative price series in the second sub-period (2000–2005).
Table 9.5

ADF unit root tests for relative prices (p-values, 1995–2005)

City market

Coal

Electricity

Gasoline

Diesel

Beijing

0.334

0.598

0.130

0.085

Tianjin

0.711

0.694

0.402

0.221

Shijiazhuang

0.011

0.047

0.040

0.096

Taiyuan

0.662

0.863

0.186

0.001

Hohhot

0.085

0.197

0.008

0.102

Shenyang

0.323

0.576

0.212

0.002

Changchun

0.728

0.915

0.146

0.000

Harbin

0.401

0.893

0.099

0.004

Shanghai

0.429

0.866

0.000

0.004

Nanjing

0.750

0.511

0.044

0.236

Hangzhou

0.040

0.187

0.466

0.180

Hefei

0.056

0.758

0.557

0.866

Fuzhou

0.124

0.880

0.005

0.056

Nanchang

0.800

0.369

0.000

0.065

Jinan

0.378

0.642

0.148

0.156

Zhengzhou

0.088

0.025

0.035

0.055

Wuhan

0.328

0.271

0.207

0.004

Changsha

0.635

0.077

0.018

0.055

Guangzhou

0.630

0.721

0.218

0.335

Nanning

0.000

0.534

0.553

0.340

Haikou

0.761

0.633

0.346

0.570

Chongqing

0.190

0.356

0.398

0.005

Chengdu

0.383

0.076

0.123

0.356

Guiyang

0.966

0.799

0.033

0.267

Kunming

0.489

0.439

0.000

0.002

Lhasa

0.981

0.532

0.678

0.345

Xi’an

0.090

0.418

0.061

0.017

Lanzhou

0.253

0.341

0.070

0.050

Xining

0.060

0.727

0.366

0.247

Yinchuan

0.680

0.883

0.000

0.088

Urumqi

0.749

0.462

0.378

0.512

Qingdao

0.981

0.417

0.035

0.002

Dalian

0.355

0.721

0.091

0.140

Xiamen

0.037

0.557

0.001

0.012

Ningbo

0.110

0.523

0.219

0.174

Number of rejecting null

5

2

13

14

Note: Null hypothesis is that each relative price series contains a unit root. ADF is the Augmented Dickey-Fuller test; MacKinnon (1996) one-sided p-values; group is used a benchmark; lag length chosen via Hannan-Quinn information criteria

Table 9.6

ADF unit root tests for relative price (p-value)

City market

Coal

Electricity

Gasoline

Diesel

2000–2005

1995–1999

2000–2005

1995–1999

2000–2005

1995–1999

2000–2005

1995–1999

Beijing

0.446

0.334

0.591

0.598

0.592

0.130

0.127

0.085

Tianjin

0.618

0.711

0.282

0.694

0.409

0.402

0.021

0.221

Shijiazhuang

0.415

0.011

0.271

0.047

0.001

0.040

0.002

0.096

Taiyuan

0.385

0.662

0.443

0.863

0.596

0.186

0.034

0.001

Hohhot

0.215

0.085

0.271

0.197

0.440

0.008

0.099

0.102

Shenyang

0.082

0.323

0.589

0.576

0.208

0.212

0.011

0.002

Changchun

0.450

0.728

0.230

0.915

0.292

0.146

0.015

0.000

Harbin

0.731

0.401

0.155

0.893

0.269

0.099

0.005

0.004

Shanghai

0.165

0.429

0.533

0.866

0.215

0.000

0.320

0.004

Nanjing

0.872

0.750

0.412

0.511

0.433

0.044

0.192

0.236

Hangzhou

0.844

0.040

0.702

0.187

0.002

0.466

0.004

0.180

Hefei

0.679

0.056

0.214

0.758

0.116

0.557

0.644

0.866

Fuzhou

0.860

0.124

0.530

0.880

0.001

0.005

0.000

0.056

Nanchang

0.122

0.800

0.198

0.369

0.045

0.000

0.300

0.065

Jinan

0.383

0.378

0.967

0.642

0.003

0.148

0.382

0.156

Zhengzhou

0.335

0.088

0.363

0.025

0.242

0.035

0.012

0.055

Wuhan

0.116

0.328

0.033

0.271

0.719

0.207

0.000

0.004

Changsha

0.685

0.635

0.047

0.077

0.001

0.018

0.194

0.055

Guangzhou

0.945

0.630

0.836

0.721

0.068

0.218

0.002

0.335

Nanning

0.117

0.000

0.400

0.534

0.000

0.553

0.005

0.340

Haikou

0.675

0.761

0.184

0.633

0.158

0.346

0.252

0.570

Chongqing

0.448

0.190

0.322

0.356

0.266

0.398

0.058

0.005

Chengdu

0.478

0.383

0.513

0.076

0.327

0.123

0.250

0.356

Guiyang

0.623

0.966

0.552

0.799

0.088

0.033

0.252

0.267

Kunming

0.674

0.489

0.095

0.439

0.002

0.000

0.001

0.002

Lhasa

0.675

0.981

0.035

0.532

0.465

0.678

0.043

0.345

Xi’an

0.000

0.090

0.447

0.418

0.472

0.061

0.033

0.017

Lanzhou

0.593

0.253

0.144

0.341

0.170

0.070

0.019

0.050

Xining

0.591

0.060

0.486

0.727

0.000

0.366

0.414

0.247

Yinchuan

0.636

0.680

0.681

0.883

0.005

0.000

0.003

0.088

Urumqi

0.553

0.749

0.278

0.462

0.043

0.378

0.540

0.512

Qingdao

0.675

0.981

0.414

0.417

0.070

0.035

0.064

0.002

Dalian

0.027

0.355

0.388

0.721

0.003

0.091

0.004

0.140

Xiamen

0.611

0.037

0.686

0.557

0.000

0.001

0.015

0.012

Ningbo

0.130

0.110

0.437

0.523

0.000

0.219

0.455

0.174

Number of rejecting null

2a

6

3

2

14

13

20

15

Note: Null hypothesis is that each relative price series contains a unit root; ADF is the Augmented Dickey-Fuller test; MacKinnon (1996) one-sided p-values; lag length chosen via Hannan-Quinn information criteria

aSee text for explanation why the number is smaller in the second sub-period than in the first sub-period

To analyze more deeply the role of the market mechanism in determining the formation of energy prices, we undertake ‘city-by-city’ based unit root tests for all four fuel types. The results are presented in Table 9.7, which displays the number of the city-by-city based relative price series that reject the null for each of the 35 city markets and each of the three periods (1995–2005, 1995–1999 and 2000–2005). We firstly observe that the number of pairs of relative price series rejecting the null from the first sub-period to the second sub-period increases (except for coal). For example, in the case of gasoline, the number of pairs of relative price series rejecting the null was 242 in the first subsample period, but this rises to 478 in the second subsample period Similarly, for diesel, there were 164 in the first subsample, this rising to 484 in the second period. As a result, a total of 630 pairs of city-by-city based relative price series the percentage rejecting the null hypothesis increased from 19% in the first subsample period (1995–1999) to 38% in the second subsample period (2000–2005) for gasoline and from 13% in the first subsample period (1995–1999) to 39% in the second subsample period (2000–2005) for diesel. For electricity, the numbers of pairs of relative price series rejecting the null are 42 in the first subsample period (1995–1999), but 70 in the second subsample period (2000–2005).
Table 9.7

Numbers rejecting the null based on city-by-city ADF unit root tests for relative prices

Benchmark city

Coal

Electricity

Gasoline

Diesel

2000–2005

1995–1999

2000–2005

1995–1999

2000–2005

1995–1999

2000–2005

1995–1999

Beijing

1

7

7

1

0

9

8

4

Tianjin

6

3

3

2

15

0

17

4

Shijiazhuang

0

5

2

1

23

7

19

10

Taiyuan

3

0

0

1

1

9

16

7

Hohhot

3

6

1

0

9

7

7

2

Shenyang

1

8

0

0

12

16

17

4

Changchun

1

6

3

0

20

0

19

17

Harbin

1

10

2

2

13

8

14

3

Shanghai

2

7

3

0

10

12

10

2

Nanjing

0

6

2

1

6

7

8

0

Hangzhou

1

3

4

2

23

2

18

4

Hefei

2

9

1

1

10

5

6

2

Fuzhou

0

5

1

1

20

13

20

0

Nanchang

3

10

4

2

16

7

5

2

Jinan

0

3

0

0

15

0

3

3

Zhengzhou

1

4

1

2

10

11

27

4

Wuhan

1

8

6

1

0

8

23

6

Changsha

0

1

3

1

26

11

16

4

Guangzhou

0

11

0

0

17

15

13

7

Nanning

3

1

0

3

17

2

19

7

Haikou

0

1

7

1

21

8

12

5

Chongqing

1

1

2

0

18

2

5

7

Chengdu

2

3

1

2

5

1

7

2

Guiyang

2

0

1

1

3

4

14

7

Kunming

0

4

7

2

27

16

20

4

Lhasa

2

1

2

0

0

1

15

2

Xi’an

12

5

1

0

13

14

10

5

Lanzhou

1

2

2

12

18

12

18

6

Xining

1

3

1

1

13

4

16

1

Yinchuan

0

6

0

1

25

10

24

12

Urumqi

0

0

1

0

19

10

10

7

Qingdao

2

1

0

0

4

0

10

3

Dalian

5

2

0

1

20

1

19

2

Xiamen

4

2

0

0

12

8

17

6

Ningbo

3

4

2

0

17

2

2

3

Sum

64

148

70

42

478

242

484

164

Percentage of rejecting null

5 a

12

6

3

38

19

38

13

Note: Null hypothesis is that each relative price series contains a unit root; ADF is the Augmented Dickey-Fuller test; MacKinnon (1996) one-sided p-values; lag length chosen via Hannan-Quinn information criteria

aSee text for explanation why the number is smaller in the second sub-period than in the first sub-period

In the case of coal there are more pairs of convergent relative price series in the first sub-period than the second sub-period, 148 and 64, respectively. However, several points should be noted. Firstly, the price of coal varies significantly spatially due to the extremely unbalanced distribution of reserves and the high long distance transport costs, which may comprise a large percentage of final user prices. Secondly, final user prices of coal may be significantly distorted if the transportation sector (especially rail) is itself not market-oriented. Thirdly, transport costs apparently increased substantially in the second subsample period, especially after 2002, from the major producing areas to the major consuming areas (refer to Figs. 9.1, 9.2). Finally and also potentially most importantly, major coal consumers (e.g., power plants) used to pay a lower price to coal producers before 2002 due to coal price regulation. However, with coal market deregulation, they had to offer higher prices to coal producers after 2002. Given these points, it is dangerous to conclude that the market mechanism deteriorated in the second sub-period.

There are significant variations in the number of pairs of convergent relative price series across cities and fuels in the whole sample period. For example, there are 6 city markets (Taiyuan, Changchun, Harbin, Chongqing, Qingdao and Xiamen) that are integrated with more than 20 other city markets for diesel; and Chongqing and Qingdao are integrated with almost every other city (30 and 29, respectively). For gasoline, there are 8 city markets (Shanghai, Fuzhou, Nanchang, Guiyang, Kunming, Qingdao and Xiamen) that are integrated with more than 20 other city markets. But fewer city markets are convergent with many other city markets for electricity and coal for example, 4 cities are convergent with only 5–6 other city markets for electricity. Nanning and Shijiazhuang are integrated with 17 and 11 other city markets, respectively, and 4 city markets (Hefei, Guangzhou, Xi’an and Xiamen) are integrated with 7–9 other city markets for coal. Based on these observations, it might be concluded that there are more regional energy markets in China.

Compared with the ADF tests based on a group as the benchmark (Table 9.6), there are many more pairs of relative price series that reject the null hypothesis (i.e. suggest convergence) based on the city-by-city ADF unit root tests (Table 9.7) for example, the numbers of pairs of convergent electricity relative price series are 70 for 2000–2005 and 42 for 1995–1999 based on the city-by-city ADF unit root tests (Table 9.7), but correspondingly, only two and three based on group as a benchmark (Table 9.6). Similarly, the numbers of pairs of convergent gasoline relative price series are 478 for 2000–2005 and 242 for 1995–1999 based on the city-by-city ADF unit root tests (Table 9.7), but only 14 and 13 based on group as benchmark (Table 9.6), correspondingly.

The numbers of pairs of convergent diesel relative price series are 484 for 2000–2005 and 164 for 1995–1999 based on the city-by-city ADF unit root tests (Table 9.7), but only 20 and 15 based on group as benchmark (Table 9.6), correspondingly. The same can be observed for coal.

9.2.2 Panel Unit Root Tests

9.2.2.1 National Panel Unit Root Tests

Using the same procedure, we next conduct panel unit root tests to ascertain whether there is a national integrated energy market in China. Similarly, we conduct panel unit root tests on the raw data series to ascertain their order of integration. These results are presented as Table 9.8, which display tests for the raw data in both levels and the first difference for three sample periods. The results show some variation, but in general the evidence is in favour of an order of integration of 1 or I (1).
Table 9.8

Panel unit root tests of raw data for all 35 city markets

Tests

Coal

Electricity

Gasoline

Diesel

 

Stat.

Prob.

Stat.

Prob.

Stat.

Prob.

Stat.

Prob.

1995–1999

Level

LLC

7.62

0.995

1.48

0.931

4.38

1.000

4.81

0.998

Breitung

−2.10

0.078

−3.69

0.000

3.73

1.000

1.23

0.891

IPS

0.85

0.803

−1.51

0.085

−3.54

0.000

−0.99

0.160

First difference

LLC

−100.69

0.000

−128.57

0.000

−122.34

0.000

−120.97

0.000

Breitung

−47.04

0.000

−76.67

0.000

−22.64

0.000

−27.22

0.000

IPS

−77.04

0.000

−96.73

0.000

−97.01

0.000

−99.99

0.000

2000–2005

Level

LLC

−0.64

0.260

4.83

1.000

4.67

0.998

2.60

0.995

Breitung

0.14

0.557

−3.12

0.001

−1.95

0.026

−4.59

0.000

IPS

1.10

0.865

−3.48

0.000

2.13

0.983

−1.51

0.066

First difference

LLC

−83.52

0.000

−189.13

0.000

−169.78

0.000

−104.97

0.000

Breitung

−12.19

0.000

−43.87

0.000

−44.11

0.000

−30.64

0.000

IPS

−52.22

0.000

−143.93

0.000

−136.05

0.000

−71.64

0.000

Note: Null hypothesis is common unit root for LLC and Breitung, and individual unit root for IPS. Exogenous variables include Individual effect and individual linear trend

Panel unit root tests for the relative price series for 35 city markets based on the group average as benchmark are presented as Table 9.9, which displays three periods and five types of tests for all four types of energy. Several points can be made based on the panel unit root test results from Table 9.9. Firstly, it should be noted that most of the LLC panel unit root tests fail to reject the null hypothesis, which might be due to the assumption of a common unit root process. Secondly, most of the remaining panel unit root tests reject the null hypothesis of a unit root, which suggests that energy prices are convergent as a whole. Thirdly, there are more tests that reject the null in the second subsample period than in the first, suggesting that it is more likely that relative energy prices are convergent in the second subsample period than in the first. Finally, fewer panel unit root tests reject the null for coal and electricity than for gasoline and diesel, suggesting that gasoline and diesel are more likely to be market-oriented than coal and electricity.
Table 9.9

Panel unit root tests of relative price data for all 35 city markets

Tests

Coal

Electricity

Gasoline

Diesel

 

Stat.

Prob.

Stat.

Prob.

Stat.

Prob.

Stat.

Prob.

1995–2005

LLC

4.97

0.99

0.62

0.73

−3.45

0.00

−4.74

0.00

Breitung

−1.79

0.04

−4.73

0.00

−6.72

0.00

−9.86

0.00

IPS

−2.26

0.01

−0.29

0.39

−8.69

0.00

−8.83

0.00

Fisher ADF

100.61

0.01

61.42

0.76

226.67

0.00

221.77

0.00

Fisher PP

302.93

0.00

951.95

0.00

1,728.9

0.00

1,403.0

0.00

2000–2005

LLC

−1.41

0.08

5.22

0.99

−2.65

0.00

−2.89

0.00

Breitung

−2.57

0.01

−3.93

0.00

−5.05

0.00

−4.72

0.00

IPS

−2.01

0.02

−2.24

0.01

−11.40

0.00

−10.05

0.00

Fisher ADF

103.76

0.01

82.31

0.15

343.16

0.00

249.23

0.00

Fisher PP

160.99

0.00

2,123.1

0.00

1,950.4

0.00

1,292.7

0.00

1995–1999

LLC

9.05

0.99

2.97

0.99

3.94

0.99

4.32

0.99

Breitung

−2.90

0.00

−3.66

0.00

−0.12

0.45

−0.83

0.20

IPS

−0.38

0.35

−0.13

0.45

−4.26

0.00

−2.48

0.01

Fisher ADF

78.21

0.23

64.03

0.68

148.33

0.00

106.62

0.00

Fisher PP

306.25

0.00

407.42

0.00

792.52

0.00

602.41

0.00

Note: Null hypothesis is common unit root for LLC and Breitung tests, and individual unit root for IPS. Fisher ADF and Fisher PP tests. Exogenous variables include Individual effect and individual linear trend

9.2.2.2 Regional Panel Unit Root Tests

Here we conduct regional panel unit root tests only with exogenous variables of individual effects and a linear trend in the test equation because most of price series seem to contain a trend, especially in the second sub-period, either deterministic or stochastic (Hamilton 1994). Firstly, we conduct various regional panel unit root tests for the raw data, which are presented in Appendix Tables  A.8,  A.9,  A.10,  A.11,  A.12,  A.13 and  A.14 for regions 1–7 respectively. Then we conduct various regional panel unit root tests for relative price series using regional average price as a benchmark. These regional panel unit root test results for relative prices are present in Appendix Tables  A.15,  A.16,  A.17,  A.18,  A.19,  A.20 and  A.21. Similarly, most of the LLC panel unit root tests fail to reject the null hypothesis, which might be due to the assumption of a common unit root process, but note the bold numbers.

To better consider the regional panel unit root test results, we just abstract the IPS test for an example, which is provided as Table 9.10. Clearly, some general results can be now seen for the tests for all markets, but variations in test results obviously do exist across regions.
Table 9.10

Panel IPS unit root tests of relative prices by region

Region

Coal

Electricity

Gasoline

Diesel

 

Stat.

Prob.

Stat.

Prob.

Stat.

Prob.

Stat.

Prob.

1995–2005

Region 1

0.09

0.54

−0.59

0.28

−1.46

0.07

−2.10

0.02

Region 2

−2.18

0.01

0.57

0.71

−3.00

0.00

−2.87

0.00

Region 3

1.03

0.85

0.14

0.56

−1.47

0.07

−2.79

0.00

Region 4

−0.15

0.44

0.28

0.61

−2.37

0.01

−4.09

0.00

Region 5

−0.26

0.40

0.26

0.60

−2.33

0.01

−2.65

0.00

Region 6

−1.33

0.09

−0.38

0.35

−3.06

0.00

−3.82

0.00

Region 7

−2.65

0.00

−2.06

0.02

−6.46

0.00

−2.75

0.00

2000–2005

Region 1

−0.10

0.46

−2.49

0.01

−1.94

0.03

−3.01

0.00

Region 2

−2.61

0.00

−0.28

0.39

−0.62

0.27

−2.20

0.01

Region 3

0.24

0.59

−0.96

0.17

−2.42

0.01

−2.02

0.02

Region 4

0.72

0.77

−1.06

0.15

−3.85

0.00

−4.11

0.00

Region 5

−0.09

0.47

−0.89

0.19

−7.12

0.00

−5.88

0.00

Region 6

−1.23

0.11

−2.18

0.01

−6.42

0.00

−3.99

0.00

Region 7

−1.32

0.09

−0.53

0.30

−3.04

0.00

−3.38

0.00

1995–1999

Region 1

−0.10

0.46

0.18

0.57

0.35

0.64

−0.54

0.29

Region 2

−0.99

0.16

−0.25

0.40

−1.89

0.03

−1.14

0.13

Region 3

2.09

0.98

0.76

0.78

−0.19

0.42

0.23

0.59

Region 4

−3.47

0.00

−0.32

0.37

−1.45

0.07

−3.64

0.00

Region 5

1.29

0.90

−0.13

0.45

−0.34

0.37

−0.92

0.18

Region 6

0.19

0.58

−1.08

0.14

−1.77

0.04

−0.63

0.27

Region 7

−1.24

0.11

−1.03

0.15

−3.30

0.00

−2.63

0.00

Note: Null hypothesis is individual unit root. Group is used as a benchmark. Individual effect and linear trend are included. Regional classification is referred to Table 1.2 of  Chap. 1

Several regional specific conclusions can be drawn based on the regional panel unit root test results. Firstly, more tests are in favour of price convergence for gasoline and diesel than for coal and electricity, which suggests that the former is likely more market integrated than the later. Secondly, more tests are in favour of price convergence during the second sub-period than during the first sub-period, suggesting that energy market integration was a gradual process in China, which coincides with the institutional evolution and gradual price reforms of China’s energy industry. Thirdly, it seems that diesel markets are more integrated than gasoline markets in some regions, for example, more tests are in favour of price convergence for diesel markets than for gasoline markets in region 2. There may be two reasons for this. The first is that urban areas typically consume more gasoline while rural areas use more diesel. The second is that urban residents might receive more subsidies than rural residents. Fourthly, fewer tests are in favour of price convergence in the case of coal in some regions during the second sub-period than during the first sub-period. Please see the previous section for further explanation of this point. Finally, certain types of energy market seem more integrated in some regions than in other regions. This is likely caused by the variations in the regional energy distribution and economic growth, for example, the coal market is more integrated while the electricity market is the least integrated in Region 2 (Beijing, Tianjin and Shanghai) because coal is not a major determinant of regional economic growth while electricity is.

9.3 Inter-fuel Price Trends and Cointegration Tests

9.3.1 Inter-fuel Price Trends

In this section, we use our price data to sketch a descriptive picture of the emergence of China’s energy market cointegration. To do so, we plot the price data and examine how various fuel prices move together in the same geographical regions or markets. Sub-periods will be distinguished and determined by the historical events of important energy reforms as presented in the previous section. The analysis takes place at national, regional and city levels and for two sub-periods of transition (1997–1998) and the new regime (after 1999). We will also pay particular attention to the price co-movement of coal-electricity and gasoline-diesel because these two pairs of fuel price series are most likely to be cointegrated.

9.3.1.1 National Inter-fuel Price Trends

Panel A of Fig.  6.4 shows the price co-movement of coal and electricity for three sub-periods. It can be seen that the electricity price suddenly jumped to a higher level from 1997 to 1998, from ¥360 per thousand kWh in 1997 to ¥475 per thousand kWh in 1998, while the coal price remained almost unchanged for the whole transition period. Government control played an important role in price formation during this transition period. By contrast, the third sub-period shows a more market-oriented period where electricity prices have increased, particularly after 2004 and coal prices also began to increase, especially after 2004. It seems that an apparent convergence of coal and electricity prices occurred after 2002, which needs to be investigated further.

Turning to gasoline and diesel; during the transition period (01/1997–12/1998), the prices of gasoline and diesel were low and their trends were flat and ‘parallel’ (Panel B of Fig.  6.4). Recall that during this period 1994–19981 retail price levels were much lower than international market prices. Petroleum prices have been set according to the international market since 1999, as can be seen with prices increasing, especially since 2002.

9.3.1.2 Regional Inter-fuel Price Trends

As discussed previously, each region might have its own energy regulation policy due to unbalanced economic growth and unbalanced energy reserves across regions. Under these conditions, even if fuel prices are cointegrated at the national levels, it doesn’t necessarily mean that they are cointegrated in each region. To observe price trends of pairs of fuels for each region and consider whether there are any differences in these price trends of pairs of fuels during the whole study period (01/1995–12/2005), we present Figs. 9.10, 9.11, 9.12, 9.13, 9.14, 9.15 and 9.16. According to these figures, all regional price trends of coal-electricity are generally similar to that at the national level. Secondly, some variations are still evident in price trends across pairs of fuels. The most evident example is that the price trends of gasoline and diesel are more likely cointegrated than those of coal and electricity for each region. Thirdly, the price trends of coal and electricity are more likely inclined to be cointegrated as energy policy reforms progressed. The most likely cointegration for price trends of coal and electricity can be seen after 2003 where coal prices show a strong rising trend, approaching the electricity price trend. Fourthly, the price trends of gasoline and diesel are more similar to those at the national level than the price trends of pairs of coal and electricity. This is because the prices of petroleum products are more likely homogeneous across the country than coal and electricity prices are. Fifthly, the price trends of gasoline and diesel look homogeneously cointegrated during the whole study period for all regions. However, similar to those at the national level, the price trends of gasoline and diesel appear to diverge since 2002 for all regions.
Fig. 9.10

Price trends of pairs of fuels for Region 1 (Region 1: Shijiazhuang, Taiyuan, Hefei, Jinan and Zhengzhou). Panel A: Coal and electricity. Panel B: Gasoline and diesel

Fig. 9.11

Price trends of pairs of fuels for Region 2 (Region 2: Beijing, Tianjin and Shanghai). Panel A: Coal and electricity. Panel B: Gasoline and diesel

Fig. 9.12

Price trends of pairs of fuels for Region 3 (Region 3: Shenyang, Changchun and Harbin). Panel A: Coal and electricity. Panel B: Gasoline and diesel

Fig. 9.13

Price trends of pairs of fuels for Region 4 (Region 4: Nanjing, Hangzhou, Nanchang and Wuhan). Panel A: Coal and electricity. Panel B: Gasoline and diesel

Fig. 9.14

Price trends of pairs of fuels for Region 5 (Region 5: Fuzhou, Changsha, Guangzhou, Nanning and Haikou). Panel A: Coal and electricity. Panel B: Gasoline and diesel

Fig. 9.15

Price trends of pairs of fuels for Region 6 (Region 6: Chongqing, Chengdu, Xi’an, Lanzhou, Guiyang and Kunming). Panel A: Coal and electricity. Panel B: Gasoline and diesel

Fig. 9.16

Price trends of pairs of fuels for Region 7 (Region 7: Hohhot, Xining, Yinchuan and Urumqi). Panel A: Coal and electricity. Panel B: Gasoline and diesel

Having compared the variations over region, we then have the following observations for specific regions:

Firstly, the prices trends of coal and electricity demonstrate the least likely emergence of an energy market in Region 2 (Fig. 9.11) and in Region 3 (Fig. 9.12) for example, during the transition period, electricity prices remain flat, jump to a new level and remain flat until 2000, and then jump and remain flat again roughly until 2003 in Region 2 (compared to Beijing, Shanghai and Tianjin). Correspondingly, however, during the transition period, coal prices remained flat, decline to a new level in 1998 and decline three times until the mid of 2000, and then slowly rise until early 2003 when they accelerate until mid of 2004 in Region 2 (Fig. 9.11). This probably reflected political concerns that stability and the tensions created by ‘reforms’ in this region (Beijing, Shanghai and Tianjin).

Secondly, having compared the trends of electricity and coal prices, we found that electricity prices appeared to be adjusted more frequently and probably responded more sensitively to demand than coal prices in Region 5 (Fig. 9.14, including Fuzhou, Changsha, Guangzhou, Nanning and Haikou). This also can be seen in Region 1 (Fig. 9.10, which includes Shijiazhuang, Taiyuan, Hefei, Jinan and Zhengzhou) and probably in Region 6 (Fig. 9.15, which includes Chongqing, Chengdu, Xi’an, Lanzhou, Guiyang and Kunming).

Thirdly, although the price trends of gasoline and diesel have shown a consistent co-movement for most of regions since the transition period (01/1997, see Panel B of Figs. 9.10, 9.11, 9.12, 9.13, 9.14, 9.15, 9.16), Region 2 (including Beijing, Shanghai and Tianjin) seemingly demonstrates the apparent regulated price trends of gasoline and diesel during the late 1998 to the late 1999 (Panel B of Fig. 9.11).

9.3.1.3 City Inter-fuel Price Trends

Whether inter-fuel price trends at the city level are similar to those at the national and regional levels, needs to be further examined since aggregation might obscure the real relations between pairs of fuel prices at the city market level. To save space, only selected major cities are discussed here, specifically: Harbin, Beijing, Shijiazhuang, Taiyuan, Jinan, Zhengzhou, Wuhan, Nanjing, Shanghai, Hangzhou, Guangzhou, Xi’an, Chengdu, and Urumqi. These cities are evenly located cross the country, and are either important energy production bases or important economic growth zones or both. The GDP in the 14 provinces and the sample cities accounts for almost 70% of the national GDP in 2006. Therefore, these 14 provincial capital city markets should be fairly representative for national energy reforms.

Price trends for pairs of fuel sources are presented in order from the northeast (Harbin) to South (Guangzhou) and to West (Xinjiang) and Southwest (Chengdu) in Appendix Figs.  A.1,  A.2,  A.3,  A.4,  A.5,  A.6,  A.7,  A.8,  A.9,  A.10,  A.11,  A.12,  A.13 and  A.14. The price trends of inter-fuels at the city market level are fundamentally similar to those at both national and regional aggregate levels. The price trends fluctuate more and are also ‘flatter’ price trends can be found at the single city market level for pairs of coal and electricity. In contrast, the price trends fluctuate less and price trends are steeper at the single city market level for pairs of gasoline and diesel. This suggests that markets are more likely to be cointegrated for pairs of gasoline and diesel than for pairs of coal and electricity.

The potential emergence of energy price cointegration appeared 1 year later at the city market level. This may be explained by data aggregation. As observed at both the national and regional aggregate levels, the new regime of energy economic development began in early 1999. However, the city level price data show that the new regime of energy economy emerged in the mid or late 1999 for example, in Beijing (Panel B of Appendix Fig.  A.2), Jinan (Panel B of Appendix Fig.  A.5), Zhengzhou (Panel B of Appendix Fig.  A.6), Wuhan (Panel B of Appendix Fig.  A.7), and Nanjing (Panel B of Appendix Fig.  A.8), gasoline and diesel prices started to display rising and changing trends in late 1999.

The emergence of a potentially cointegrated energy market for coal and electricity seems later than for petroleum products. It seems that coal and electricity prices show a potential cointegrated relationship for most of the cities after 2002. The prices of coal and electricity have also changed more frequently since then. Therefore, we might tentatively propose that the real emergence of coal and electricity cointegration was after 2002.

Finally, a strange phenomenon can be found between the price trends of coal and electricity for most provincial city markets during the transition and new regime, which varies across city market. Namely, electricity prices increase while coal price decline for example, during the early 1998 to the late 2001, electricity prices jumped twice, while coal prices correspondingly dropped twice in Harbin (Panel A of Appendix Fig.  A.1). During the whole of 2000, electricity prices jumped dramatically, while coal prices declined slightly in Beijing (Panel A of Appendix Fig.  A.2). The reasons for this are unclear, but large surpluses of coal may be one of the most important factors that significantly depressed the coal price during this period (Wang 2007).

In summary, we have the following primary assumptions for the emergence of energy price cointegration in China to be statistically tested in the next section. Firstly, the emergence of price cointegraton across homogeneous pairs of fuels is apparent though the intensities of price cointegration vary across homogeneous pairs of fuels. Secondly, the descriptive pictures of fuel price trends indicate that the same co-movement of prices occurs in the case of different petroleum products in the whole study period (01/01/1995–31/12/2005). When reconciling energy price reforms and price co-movement trends, however, we may conclude that it is most likely since 2000 that the emergence of price cointegration across petroleum products has occurred because only since then the prices of petroleum products began to increase and change strongly for most of city markets according to our observations above and price reform time table. Actually, the emergence of energy price cointegration appears to be 1 year later at city level (which is 2000) than at the regional or national aggregate level (which is 1999). This may be the result of price aggregation. Thirdly, the descriptive pictures of fuel price trends also demonstrate that the same co-movement of prices has most likely occurred in the case of pairs of coal and electricity since 2003 when the prices of coal and electricity began to climb and change significantly for most city markets. Similarly, the potential emergence of price cointegration is several years later at city level (which is 2003) than observed at regional or national aggregate level (which is 1999). As can be seen, the emergence of price cointegration is 3 years earlier for gasoline and diesel than for coal and electricity, where price reforms for gasoline and diesel markets are earlier and complete compared to those for the coal and electricity markets. In the next section, we consider formal statistical tests of the existence of price cointegration to consider, among other issues the last two assumptions above.

9.3.2 Panel Cointegration Tests

Given the potential for different pricing periods and reform effects, in the subsequent analysis we test for the existence of inter-fuel price cointegration first for the whole period and then for sub-periods whose dates are informed by institutional and historical changes. The tests also consider national and regional level markets.

We utilise all seven panel cointegration statistics discussed in  Chap. 4 for specific inter-fuel price series and specific time periods of interest. The panel cointegration statistics are presented as Appendix Table  A.22 for the national panel and Appendix Tables  A.23,  A.24,  A.25,  A.26,  A.27,  A.28 and  A.29 for the panel of Regions 1–7, respectively.

If all tests reject or all tests do not reject, the conclusion is clear, however, as is common when using such a battery of tests, the results are potentially ambiguous and care must be exercised in choosing which results to emphasize and why. As discussed in Pedroni (2004), in terms of monthly data, with fewer than 20 years of data it may be possible to distinguish even the most extreme cases from the null of no cointegration when the data are pooled across members of panels with these dimensions. This condition has been met in our case since we have 36 observations each year or three observations each month. Furthermore, if the panel is fairly large so that size distortion is less of an issue, the panel \( \upsilon \)-statistic tends to have the best power relative to the other statistics. In very small panels, however, if the group-rho statistic rejects the null of no cointegration, we can be relatively confident of the conclusion as it is slightly undersized and empirically the most conservative of the tests. The other statistics tend to lie somewhere in between these two extremes and have minor comparative advantages over different ranges of the sample size. The panel-\( \upsilon \) statistic is the strongest panel cointegration test and therefore the next discussions will be focused on this test.

9.3.2.1 National Panel Cointegration Tests

Table 9.11 presents the national panel cointegration tests for the inter-fuels of all four fuels, electricity and coal, and diesel and gasoline during one whole period (1997–2005) and two sub-periods (1997–1999 and 2000–2005). For the full sample period the national panel cointegration tests suggest that all four price series move together in the long-run given the assumption of no deterministic trend (Table 9.11). A priori, however, we would find this result unlikely since we know that for some years and some fuels energy prices were independently controlled and their time series paths appear to vary. If we consider the three sub-period tests, most of the panel \( \upsilon \)-statistic tests do not reject the null of no cointegration. The lack of cointegration for all fuels at the national level is also as we might expect as coal and electricity appear to move, over time, differently to gasoline and diesel prices. Such an expectation is supported in the results. However, it is clear that that all four fuel prices are more cointegrated in the second sub-period than in the first sub-period since one of panel \( \upsilon \)-statistic tests rejects the null hypothesis in the second sub-period.
Table 9.11

Panel cointegration tests for all 35 markets (p values)

Test statistics

All four fuels

Electricity and coal

Diesel and gasoline

 

1997–2005

1997–1999

2000–2005

1997–2005

1997–1999

2000–2005

1997–2005

1997–1999

2000–2005

No deterministic trend

Panel \( \nu \)-statistic

0.399

0.305

0.386

0.229

0.345

0.075

0.000

0.000

0.000

Panel \( \rho \)-statistic

0.000

0.000

0.000

0.000

0.000

0.000

0.000

0.000

0.000

Panel \( t \)-statistica

0.000

0.000

0.000

0.000

0.002

0.000

0.000

0.000

0.000

Panel \( t \)-statisticb

0.394

0.204

0.005

0.001

0.260

0.011

0.000

0.000

0.000

Group \( \rho \)-statistic

0.000

0.001

0.000

0.000

0.000

0.000

0.000

0.000

0.000

Group \( t \)-statistica

0.000

0.001

0.000

0.000

0.001

0.000

0.000

0.000

0.000

Group \( t \)-statisticb

0.344

0.258

0.051

0.000

0.240

0.000

0.000

0.000

0.000

Deterministic intercept and trend

Panel \( \nu \)-statistic

0.004

0.374

0.057

0.014

0.393

0.100

0.000

0.000

0.000

Panel \( \rho \)-statistic

0.000

0.000

0.127

0.000

0.000

0.279

0.000

0.000

0.000

Panel \( t \)-statistica

0.000

0.000

0.036

0.000

0.000

0.182

0.000

0.000

0.000

Panel \( t \)-statisticb

0.047

0.000

0.139

0.090

0.000

0.385

0.000

0.000

0.000

Group \( \rho \)-statistic

0.000

0.000

0.011

0.000

0.000

0.079

0.000

0.000

0.000

Group \( t \)-statistica

0.000

0.000

0.008

0.000

0.000

0.227

0.000

0.000

0.000

Group \( t \)-statisticb

0.008

0.000

0.398

0.011

0.000

0.240

0.000

0.000

0.000

Note: Statistics are asymptotically distributed as normal. The statistic ratio test is right-sided, while the others are left-sided. Null hypothesis is no cointegration among the fuel prices and no exogenous variables are included in test equation. Pedroni panel cointegration test is based Engle-Granger. Pedroni (1999) shows that the panel-ADF and group-ADF statistics have better small sample properties than the other statistics, and hence they are more reliable

aNon-parametric

bParametric

Secondly, for coal and electricity, the national panel cointegration tests provide some weak evidence of cointegration for the full sample period. However, these weak results are not supported when we consider the two sub-periods where the results suggest that the coal and electricity price series did not move together in a long-run during 1997–1999, while the coal and electricity price series may have moved together during 2000–2005. These results are consistent with our previous observations.

Thirdly, the national panel cointegration tests show a different scenario for gasoline and diesel price series. The national panel cointegration tests suggest that gasoline and diesel price series have moved together in a long-run during both for the full sample period and the two sub-periods.

9.3.2.2 Regional Panel Cointegration Tests

Table 9.12 provides regional panel cointegration tests of inter-fuel price series. Based upon similar analysis, we can make the follow conclusions for the regional-based panel cointegration tests:
Table 9.12

Panel \( \nu \)-statistic cointegration tests by region

Region

Electricity, coal, diesel and gasoline

Electricity and coal

Diesel and gasoline

 

1997–1999

2000–2005

1997–1999

2000–2005

1997–1999

2000–2005

 

Stat.

Prob.

Stat.

Prob.

Stat.

Prob.

Stat.

Prob.

Stat.

Prob.

Stat.

Prob.

No deterministic trend

Region 1

0.481

0.355

0.458

0.359

0.863

0.275

2.028

0.051

3.344

0.002

12.178

0.000

Region 2

2.360

0.025

1.315

0.168

1.959

0.059

0.348

0.376

3.959

0.000

11.675

0.000

Region 3

−0.457

0.359

0.600

0.333

−0.811

0.287

0.945

0.255

−0.202

0.391

6.815

0.000

Region 4

0.799

0.290

1.990

0.055

2.537

0.016

1.209

0.192

3.985

0.000

9.478

0.000

Region 5

−0.221

0.389

5.762

0.000

2.316

0.027

2.444

0.020

4.862

0.000

14.336

0.000

Region 6

1.638

0.104

9.793

0.000

5.247

0.000

2.545

0.016

1.496

0.130

13.313

0.000

Region 7

2.698

0.011

2.889

0.006

4.374

0.000

0.927

0.260

1.298

0.172

8.249

0.000

Deterministic intercept and trend

Region 1

0.527

0.347

1.191

0.196

1.008

0.240

2.262

0.031

1.611

0.109

7.511

0.000

Region 2

1.141

0.208

1.027

0.236

0.003

0.399

−0.490

0.354

1.764

0.084

7.702

0.000

Region 3

−1.518

0.126

0.297

0.382

−2.201

0.035

0.679

0.317

−0.797

0.290

4.370

0.000

Region 4

0.587

0.336

1.592

0.112

1.782

0.082

0.026

0.399

1.739

0.088

5.773

0.000

Region 5

−1.235

0.186

5.241

0.000

0.293

0.382

0.346

0.376

2.290

0.029

9.333

0.000

Region 6

1.038

0.233

9.169

0.000

3.747

0.000

1.855

0.071

0.207

0.391

8.693

0.000

Region 7

1.487

0.132

2.103

0.044

1.933

0.062

−0.319

0.379

0.276

0.384

5.711

0.000

Note: Statistics are asymptotically distributed as normal. The statistic ratio test is right-sided, while the others are left-sided. Null hypothesis is no cointegration among the fuel prices and no exogenous variables are included in test equation. Regional classification is referred to Table 1.2 of  Chap. 1

Firstly, some regional panel tests reject the null of no cointegration for coal, electricity, gasoline and diesel prices during the second sub-period (2000–2005), suggesting that inter-fuel prices are cointegrated in some regions even during the transitional energy economy. Given the assumption of no deterministic trend, there are four of seven panel-\( \upsilon \) statistic tests that reject the null of no integration in the second sub-period.

Secondly, the majority of the regional panel tests do not reject the null hypothesis of no cointegration for coal and electricity prices during both sub-periods, especially in the second sub-period. Although most of the previous results display some cointegration for coal and electricity prices after 2000, the regional panel statistical tests do not confirm these results. However, there seem to be three exceptions; Regions 1, 5 and 6 for the sub-period 2000–2005, for which the strongest panel \( \upsilon \)-statistic seemingly tends to reject the null hypothesis of no cointegration. It may appear strange for coal and electricity prices to move together during the earlier period (1997–1999), but not during the latest period (2000–2005), however, this was a period of state controlled prices where some common movements would be expected. One might therefore expect this regulated link to disappear as a consequence of the gradualist reforms.

Thirdly, all regional panel tests reject the null hypothesis of no cointegration for gasoline and diesel prices during the latest sub-period of 2000–2005, which suggests that gasoline and diesel prices move together in the long-run after 2000 in all regions. However, the regional panel tests for the 1997–1999 sub-period (equivalent to the period of transition) suggest that gasoline and diesel prices move together in a long-run in some regions for example, four regional panel tests reject the null hypothesis of no cointegration for gasoline and diesel prices in Regions 1, 2, 4 and 5 while three regional panel tests do not reject the null hypothesis of no cointegration for gasoline and diesel prices in Regions 3, 6 and 7. There are several points to consider here. Firstly, geographically, gasoline and diesel prices appear to move together even during the transition period for those regions located in the center, east and south, but not for those regions located in the remote areas, such as northeast, west, and southwest. Secondly, regional petroleum product markets are evident in China. Thirdly, gasoline and diesel prices have moved together since 1997 in relatively developed areas, which are circled by Shijiazhuang, Taiyuan, Xi’an, Wuhan, Changsha, and east coastal areas.

At this stage, it is potentially interesting to ask why the price series of gasoline and diesel are more cointegrated than those of coal and electricity, both statistically and economically. There may be many answers to this, but the following may be the most important:
  • Gasoline and diesel are more homogeneous energy products than coal and electricity. In this case, it is expected that the former price series are more likely cointegrated than the latter.

  • The intensity and time of reforms are different over the two groups of energy sources. According our review of the energy policy reform in China, the prices of petroleum products and coal were deregulated earlier than that of electricity.

  • The price reforms were almost simultaneous for gasoline and diesel while they were not synchronous for coal and electricity. Typically, price deregulation was earlier for the coal industry than for the electricity industry. One might expect that the non synchronous price reforms in the coal industry and electricity industry would not likely lead to observed cointegration and probably contributed to the later emergence of cointegration of the price series of coal and electricity in China.

  • Coal and electricity are categorized in the same energy group in this study, but they are not a homogeneous commodity although most electricity is generated from coal.

  • Substitutability is significantly different between gasoline and diesel and coal and electricity though they are both substitutable. Gasoline and diesel may be easily substitutable while coal and electricity may be complements.

  • Differences in price deregulation over energy types are closely related to their effects on the national economic growth and consumer consequences. Typically, changes in electricity price appear more related to the cost of living than input costs. Hence, electricity price deregulation was deferred in China. Correspondingly, price reforms for other commodities closely related to electricity production might be also delayed or overdue. This is particularly true for reform of coal prices where most of it is used to generate electricity.

9.4 Comparisons with Other Studies

Whether China is a market economy has attracted attention from both domestic and international scholars, however, few have focused empirically on this question and even fewer are focused on China’s energy market. As the early economic reforms were initiated in crop production, Huang and Rozelle (2006) have shown empirically the emergence of an agricultural commodity markets in China during the past decade. They have also claimed that the power of markets to continue to integrate perhaps, more than anything, shows the power of China’s gradual method of transition. Park et al. (2002) demonstrate that China’s grain markets have grown dramatically over time. As China rejoined the WTO, however, Poncet (2003, 2005) investigates the determinants of inter-provincial trade barriers and concludes in favor of a disintegrated domestic market in China. It should be noted that her studies are not based on the formal unit root tests of price data. Fan and Wei (2006) conduct a detailed investigation using unit root tests for spot price series for China. For the energy market, they consider only gasoline and diesel price tests. They only provide a general picture of market integration. This current study not only provides more robust tests, but demonstrates that the market mechanism is playing an increasing role in determining energy prices.

Although there are many studies on the emergence of China’s agricultural commodity markets (e.g., Huang and Rozelle 2006), there are few studies on inter-fuel price panel cointegration tests. Therefore, it is difficult for us to compare our finding with others.

9.5 Conclusions and Implications

In this chapter, we have shown, in a number of ways, the steady emergence of energy commodity markets that have occurred in China during the study period. Regardless of whether we use descriptive statistics or more formal techniques, our results are consistent with the emergence of markets for coal, electricity, gasoline and diesel. Moreover, energy markets are robust when viewed across space and time.

Although those who visit China are not surprised, such a picture of integrated energy markets may be surprising when juxtaposed against the policy background. Even during the first subsample period, China took a gradualist approach to reforming its energy markets. Our results show that despite the gradualist policy, the operation of energy markets have steadily strengthened in China.

China’s market reforms have really been based on entry-driven competition. In the case of China entry has come from both the dismantlement of the state-own enterprises and the emergence of more energy companies. While this has produced an increase in integration and fall in transaction costs that has been documented in the chapter, it is also eroded the power of the state to control the energy markets with traditional command methods. Our results suggest that if policymakers actually want to control energy markets in the future, they need to devise new ways to intervene in the energy sector, otherwise the reforms they have introduced have clearly led to a more market oriented energy sector. However, if they want the market to function freely, it appears that the reforms to date have moved somewhat in this direction.

Although we have tested for energy price convergence, our results suggest China’s energy economy is still in a state of transition. However, as the market economy is more efficient in resource distribution, one would expect that China’s high energy intensity will be affected by energy market reforms despite the fact that other factors still play an important role in improving a firm’s performance and reducing energy consumption. This suggests that further energy market reforms can reduce China’s energy intensity and in turn energy imports and the impact of China’s energy import on the world energy supply and prices.

It should be noted that although energy prices are convergent across markets, the market process is apparently different across energy types. The results show that gasoline and diesel are more likely market-oriented than coal and electricity. The price reforms of electricity and coal relevant to electricity generation were late and slow. Therefore, how to speed-up price reforms of electricity and coal relevant electricity generation is a great challenge China has to face.

Panel tests demonstrate the convergence of energy prices, however, univariate ADF unit root tests clearly display there are still many regional markets in China for certain types of fuel. This is likely related to the unbalanced distribution of energy reserves, especially for coal. As a result, transportation plays an important role in final user price formation of coal due to huge long distance transportation cost. Reforms of the transportation sector, therefore, particularly the railway, may become a major determinant in the process of price convergence for coal prices across markets and regions.

Finally, it is surprising that the unit root tests and panel cointegration tests accept the null of no cointegration for all four fuel prices at the national level and do so for electricity and coal prices. However, the tests do suggest some clues to the emergence of some developing areas of inter-fuel prices cointegration in China. It also seems that as energy reforms take place, inter-fuel prices are becoming more cointegrated with oil prices apparently cointegrated even during the transition period in some areas.

Footnotes

  1. 1.

    There are likely two reasons for the low prices of petroleum products. The first is low domestic production cost of petroleum products. The second is low quality of both domestically processed and imported petroleum products.

References

  1. Bernard AB, Durlauf SN (1996) Interpreting tests of the convergence hypothesis. J Econometrics 71:161–173MathSciNetMATHCrossRefGoogle Scholar
  2. Enders W (1995) Applied econometric time series. Wiley, New York, 211, 243–251, 376–377Google Scholar
  3. Fan CS, Wei X (2006) The law of one price: evidence from the transitional economy of China. Rev Econ Stat 88:682–697CrossRefGoogle Scholar
  4. Greasley D, Oxley L (1997) Time-series based tests of the convergence hypothesis: some positive results. Econ Lett 56:146–147CrossRefGoogle Scholar
  5. Hamilton JD (1994) Time series analysis. Princeton University Press, PrincetonMATHGoogle Scholar
  6. Huang J, Rozelle S (2006) The emergence of agricultural commodity markets in China. China Econ Rev 17:266–280CrossRefGoogle Scholar
  7. Nelson C, Plosser C (1982) Trends and random walks in macroeconomic time series: some evidence and implications. J Monetary Econ 10:130–162CrossRefGoogle Scholar
  8. Park A, Jin HH, Rozelle S, Huang JK (2002) Market emergence and transaction: arbitrage, transaction costs and autarky in China’s grain markets. Am J Agr Econ 84:67–82CrossRefGoogle Scholar
  9. Pedroni P (1999) Critical values for cointegration tests in heterogeneous panels with multiple regressors. Oxford B Econ Stat 61:653–670CrossRefGoogle Scholar
  10. Pedroni P (2004) Panel cointegration: asymptotic and finite sample properties of pooled time series tests with an application to the PPP hypothesis. Economet Theor 20:597–625MathSciNetMATHGoogle Scholar
  11. Perron P (1989) The great crash, the oil price shock and the unit root hypothesis. Econometrica 57:1361–1401MATHCrossRefGoogle Scholar
  12. Poncet S (2003) Measuring Chinese domestic and international integration. China Econ Rev 14:1–21CrossRefGoogle Scholar
  13. Poncet S (2005) A fragmented China: measure and determinants of Chinese domestic market disintegration. Rev Int Econ 13:409–430CrossRefGoogle Scholar
  14. Wang B (2007) An imbalanced development of coal and electricity industries in China. Energy Policy 35:4959–4968CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  1. 1.College of Economics and ManagementHenan Agricultural UniversityZhengzhouChina, People’s Republic
  2. 2.Department of Economics & FinanceUniversity of CanterburyChristchurchNew Zealand

Personalised recommendations