Real exchange rate and productivity in China
Sylviane Guillaumont Jeanneney* and Ping Hua*
CERDI-IDREC, CNRS-Université d’Auvergne
65, boulevard François Mitterrand,
63000 Clermont-Ferrand, France.
Tel: 33 4 73 43 12 17
Fax: 33 4 73 43 12 28
Email: @, @.
Summary: This article investigates the impact that the appreciation of real exchange rate in China since 1994 has exerted on total factor productivity growth and its two components. We explain respectively the arguments that may explain a positive and a negative impact of real exchange rate on efficiency change and on technical progress. Then for twenty-nine Chinese provinces we calculate DEA Malmquist indices of productivity growth and its two components for the period from 1993 to 2001. Finally we present a panel estimation of productivity growth and we show that the appreciation of the real exchange rate had an unfavourable effect on technical progress but a favourable effect on efficiency growth, and these two effects offset each other partially to give a lesser negative effect on productivity growth.
.
Key words: China, DEA Malmquist index, productivity, real exchange rate.
JEL O4, F31,F43
Résumé : Cet article étudie l’impact que l’appréciation du taux de change réel en Chine entre 1993 et 2001 a exercé sur la croissance de la productivité globale des facteurs de production.. On expose les arguments susceptibles d’expliquer un impact positif ou négatif sur l’efficience technique et sur le progrès technique Puis, grâce à un indice de Malmquist calculé avec la méthode DEA, on décompose la croissance de la productivité entre ses deux composantes pour l’ensemble des provinces chinoises. Enfin on présente une estimation en panel de la croissance de la productivité des facteurs et l’on montre que l’appréciation du taux de change réel a exercé une action défavorable sur le progrès technique en partie compensée par une action favorable sur la croissance de l’efficience.
Mots clefs : Chine, indice de DEA Malmquist, productivité, taux de change réel
JEL O4, F31,F43
1 Introduction
Industrialized countries have experienced large fluctuations of real exchange rates during the last thirty years so that many authors have taken an interest in the link between the level of real exchange rates and productivity growth in these economies. So Paul Krugman (1989) suggested that the large appreciation of the dollar between 1979 and 1985 might have induced the acceleration of industrial productivity growth in the United-States because the rise of the dollar would have pushed firms to increase their productivity. In the same way Porter (1990), in his well-known book about competitiveness and growth, upheld that an overvalued exchange rate might contribute to increase productivity. Inversely it is likely that the large real depreciation of the Canadian dollar during the 1990’s explains the widening productivity gap between the United-States and Canada (Courchene and Harris, 1999, Grubel, 1999).
However for developing countries one does not generally make the hypothesis that a real appreciation of one currency has a positive impact on productivity. Most authors think that an overvaluation of currency, by reducing the competitiveness of tradable goods sector, negatively affect productivity growth . China is a very appropriate field to investigate the effects of an appreciation of the real exchange rate on productivity growth. Indeed since the beginning of the last decade China has really become a market economy (Guillaumont and Hua, 2002). In 1992, after several years of break in the reforms, the process of liberalization and trade openness is launched again, following the famous trip of Deng Xiaoping in the South of China. At this time there were still two exchange rates of dollar in terms of Renminbi for commercial operations, an official rate and a higher swap rate which was determined on foreign exchange markets but was strictly controlled by the central rulers. The firms should sell 20% of their exports receipts at the official rate and might use the remaining 80% to finance their own imports or sell them on the foreign exchange markets at the swap rate. The planned imports were supported by priority foreign allowances at the official rate while the other imports were financed at the swap rate. However, since the beginning of 1994, Only swap rate exists. While previously the two exchange rates were periodically devalued, the rate of exchange of the Renminbi in terms of dollar, henceforth single, has remained constant. Therefore the real effective exchange rate of China, which was strongly depreciated in 1992 and 1993, has known some appreciation until 1998 and then a stabilization (see Figure 1) . The real appreciation of the Chinese currency did not prevent the economy from growing at a rate which has been permanently higher than 7% per annum.
Moreover, the variation of the real exchange rate differs from one province to one another because the provinces have very different rates of inflation as well as different foreign trading partners (Guillaumont Jeanneney and Hua, 2002). From 1993 to 2001, the annual average appreciation of the real effective exchange rates of Chinese provinces has ranged from in Hainan to in Beijing municipality (see graph 2)
NB .A rise means an appreciation of the Renminbi.
It would not be convenient to measure the impact of the real exchange rate on factor productivity without taking in account the inverse causality stated by Balassa (1964) and Samuelson (1964). These authors have shown that real exchange rate tends to appreciate in countries where productivity growth (which occurs chiefly in the tradable goods sector) is faster than in the rest of the world. Indeed, “under the assumption that prices equal marginal costs, inter-country wages differences in the sector of tradable goods will correspond to productivity differentials, while the internal mobility of labor will tend to equalize the wages of comparable labor within each economy” (Balassa, 1964, ). So the growth of productivity tends to raise the price of non-tradable goods domestically determined. Regarding Chinese provinces as different economies, we have shown elsewhere (Guillaumont Jeanneney and Hua, 2002) that, in the long run, the differentials between the real exchange rates of Chinese provinces may be explained by the Balassa-Samuelson effect. This effect is captured by the gap between the per capita product of each province and respectively that of its foreign trading partners and that of China as a whole. In the following econometric estimation of the impact of real exchange rate on productivity growth, we shall use these gaps as instruments for the real exchange rate.
This paper is organized as follows. In part 2 we present theoretical assumptions, which may justify a relationship between the level of real exchange rate and the growth of factor productivity and we set out their empirical counterparts. This analysis suggests that real exchange rate does not affect productivity through the same channels whether productivity growth is due to efficiency change or technical progress. It is why in the third part we decompose the productivity growth into its two components thanks to DEA Malmquist indices (Data Envelopment Analysis). Then in the fourth part we present our empirical results, based on an estimation using fixed effects on a panel data of 29 provinces over the period 1993-2001. We estimate successively efficiency change, technical progress and total factor productivity change as a function of real exchange rate, intermediary variables representing the main channels through which real exchange rate affects productivity growth and some other control variables. The results corroborate our assumption that the real appreciation has exerted a positive effect on efficiency growth, but a negative one on technological change.
2. How does the appreciation of real exchange rate affect the productivity growth: a theoretical analysis
The literature presents opposite arguments about the effects of real exchange rate on total factor productivity, whether productivity growth is due to efficiency improvement or technical progress. We successively recall the arguments which explain positive or negative impacts of the appreciation of the real exchange rate, namely of the fall of the relative price of tradable goods, on productivity
. Why real appreciation should be good for productivity growth?
We may stress several reasons which support the idea that real appreciation is good for productivity growth.
On the one hand, real appreciation decreases the relative cost of imported capital goods and then induces a rise of the capital-labour ratio. It is likely that this rise supports technical progress but simultaneously induces a lesser efficiency due to the drawbacks in the management of more capitalistic and sophisticated technologies. On the other hand a real appreciation means an increase in the real labour remuneration which may induce an improvement of workers productivity particularly in a country where the wages of unskilled workers are still very low. This assumption was presented as soon as the 1957 by H. Leibenstein who stressed that in developing countries a too weak remuneration of labour might spoil workers’ health and work capacity and showed that the motivation of workers acts on production factor efficiency, what he called the “X-efficiency”. However skilled workers are also concerned by the increase of remunerations induced by a real appreciation of exchange rate. We may suppose that this last one slows down the emigration of this kind of workers (Harris, 2001). In fact China endures a significant brain drain and in the 1990’s we observe some Chinese workers coming back with the improvement of skilled work remuneration .
Finally it is likely that the real appreciation has exerted a positive effect on the productivity of industrial enterprises in the extent that it has exacerbated the foreign competition. In the case of appreciation, firms may be compelled to close their less efficient factories. A phenomenon of “creative destruction” benefits to the most efficient enterprises. It is also possible that real appreciation pushes firms to improve their technical efficiency in a context of monopoly or collusive oligopoly (Krugman, 1989). The argument is the following. Managers benefit from only a part of the profit induced by a better management or a stronger exertion since a part of the profit goes to the owners of the enterprise. In the case of monopoly managers do not choose the exertion that maximises the profit. As Marshall told, the better profit of a monopoly is a quiet live. Then in a situation of oligopoly (due to the new foreign competitors and, in the case of China, due to competitors localised in the other provinces) the managers will choose a higher level of effort, non only because in the short run this behaviour may increase the profit, but also because the decrease of costs dissuades competitors from producing and thus avoids a fall in the price. Due to this strategic yield, there exists an additional benefit induced by the effort which may push manager’s effort nearer to its optimum.
. Why real appreciation should be deleterious to productivity growth?
The most current argument in favour of a harmful effect of real exchange appreciation on productivity growth is based on the fact that real appreciation tends to slow down exports growth. Actually since the beginning of China’s transition towards a market economy exports in current dollars have increased rapidly. However annual export growth has been slightly reduced during the 1990’s. It has passed from % on average during the period 1985-1993, when the exchange rate depreciated, to % in 1994-2002, when on the contrary the exchange rate appreciated. One expects that an increasing trade openness exerts several kinds of favourable effects on productivity.
Export growth induces a shift of production factors into export sector which is generally considered as more efficient than the other sectors (Feder,1983, Guillaumont, 1994). This argument seems to be relevant for China where light industry of consumer goods constitutes the bulk of the export sector (55% of exportations in 2001 ). This productive sector, very labour intensive, corresponds to the comparative advantage of China (Yue and Hua, 2002) and technical efficiency is probably higher in this sector than in heavy industry or in agriculture as well as in service sector. This first argument is based on dualist view of the economy according to which the labour marginal productivity is unequal in the different sectors. This assumption seems to be relevant, as Chinese workers cannot freely choose their work place. The relative advantage, in terms of efficiency, of the manufacturing sector in China may have been progressively increased by learning-by-doing effect and by scale economies due to market expansion. This advantage is probably still present several years after the beginning of the transition of China towards a market economy and consequently in the 1990’s. Indeed the export sector provides external economies to the whole economy: improvement of management skill and labour formation.
On the other hand, the impact of the export sector on technological change is uncertain. The industry of consumer goods or of small equipments, on which the export expansion is based in China, is less capable to generate technical progress than the heavy industry, all the more since a great part of this industry is only an activity of imported components assembling. Consequently, it is possible that an increasing ratio of exports to GDP would be associated to a less technical progress.
The favourable effect of openness also passes through foreign direct investments (Dayal-Gulati and Husain, 2002). In China, as in other developing countries, foreign investments are concentrated in the sector of tradable goods, chiefly in industry. They have been stimulated by the real depreciation of the currency until 1994. Actually, as soon as 1993, China became the country which received the most foreign direct investments among developing countries (Démurger, 2002). Then, it is likely that the real appreciation would have slowed down direct investments as that was observed in particular after the Asian crisis in 1997. One supposes that foreign firms bring in technological improvements and their know-how . This positive action occurs through the creation of subsidiary companies more productive than domestic firms and through the diffusion towards the last ones of technical innovations and better management. This imitation effect occurs in competing domestic firms, but still more in firms which are suppliers or buyers of foreign enterprises (Sun, 1998).
However the negative potential effect of the appreciation of the real exchange rate on the productivity is not exclusively linked to lesser growth of exports or of foreign direct investments. Indeed real appreciation is a hindrance to import competing products as well as exports. It reduces the profits and the capacity of self-financing, and therefore the investment of the tradable goods sector (industry) for the benefit of services and protected sectors. If the industrial sector is most innovative, real appreciation may slow down innovation well beyond the only export industry or foreign enterprises.
In short, real exchange rate has many effects on total factor productivity which are different whether productivity growth results from technical progress or from efficiency improvement. A great part of these effects occurs through commercial and financial openness which are hindered by a real appreciation. We expect a positive impact of trade openness (measured by the export ratio) principally on efficiency while foreign direct investments should be good for technical progress as well as for efficiency. On the other hand, real appreciation, by rising capital labour intensity, should be favourable to technological innovations, but unfavourable to efficiency. Moreover it is likely that the export ratio, the rate of foreign direct investments and the capital-labour ratio do not capture all the effects of an appreciation of the real effective exchange rate. This last one, corresponding to a relative fall of tradable good prices, deters the investment in this most innovative part of the economy and therefore is unfavourable to technical progress. In return real appreciation incites to improve efficiency with a higher labour remuneration and an intensification of competition.
The effects of real appreciation that we have identified until now are long run phenomena (Harris, 2001). In the short run real appreciation, by decreasing tradable goods demand, may reduce factor utilisation and lead to a lesser transitory efficiency.
Econometric model
As the waited effects of real exchange rate on technical efficiency and progress are different, our econometric model, which uses twenty-nine provinces as a sample for the period from 1993 to 2001, distinguishes between technical efficiency and progress. Consequently, three functions relative to technical efficiency change (), technical progress (), and total factor productivity change () will be successively estimated for the different Chinese provinces.
Among explanatory variables, next to real effective exchange rate (ER), we introduce several control variables, considered as independent of real exchange rate, relative to the importance of public enterprises (ENP) and to education level (EDU). Indeed we may suppose that public enterprises, for which the essential of bank funding is reserved in China, make important investments (with high contents of technological innovations) more easily than other enterprises, but in return their technical efficiency is constrained by an excess of workers difficult to be made redundant. Also the presence of well-educated people is favorable to a good management and thus to an improvement of technical efficiency. Consequently, we introduced simultaneously three variables of human capital, concerning the proportion of the population having attained at most primary education level (EDUP), secondary education level (EDUS) or university education level (EDUU). Next to these variables of structural nature, we introduced real GDP per capita lagged one period (YRP-1) to test an eventual convergence phenomenon, conformably to traditional growth theory. In order to control for transitory effect of business cycle on the utilization rate of production factors, we introduced the gap of the ratio of changes in inventories to GDP to its trend (INV), in the equation of technical efficiency change and thus in the equation of total factor productivity change. In fact, this ratio has met a decreasing trend during the estimated period which probably reflects a better efficiency in input utilization and in the management of finished goods, as this is normal in a transition economy. So we consider that a positive gap relative to trend reflects a flat overall economic situation and inversely.
All the theoretical hypotheses relative to the effects of real exchange rate presented before and the waited signs for the different variables which result from these hypotheses, are resumed in table 1 where, as in our following estimation, an increase of real exchange rate corresponds to a real appreciation. To test these hypotheses we carry out our estimation in three stages. At first we introduce the intermediate variables, identified as transmission channels of real exchange rate to productivity change, namely export rate (X), foreign direct investments ratio (FDI) and capital intensity (KL). In this model the coefficient associated with real exchange rate (ER) will capture direct effects on productivity change of the relative price change associated to a variation of real exchange rate, those which pass neither through export rate, nor through foreign direct investments, nor through capital intensity. We wait for a negative coefficient for technical progress change and a positive one for efficiency change.
Three functions can be therefore written as follow:
For technical efficiency change (equation 1):
The waited signs are such as:
a1>0, a2<0, a3<0, a4<0, a5>0, a6> or <0, a7<0, a8>0, a9>0, a10<0
For technical progress (equation 2):
The waited signs are such as:
b1<0, b2>0, b3<0, b4<0, b5>0, b6> or <0, b7>or <0, b8>0, b9 >0
For total factor productivity change (equation 3):
The waited signs are such as:
c1< or >0, c2< or>0, c3<0, c4<0, c5>0, c6>or <0, c7<0, c8<or>0, c9 >0, c10<or>0
Then, in order to measure the total effect of real exchange rate on efficiency change, technical progress and total factor productivity change, we regress the variables representing transmission channels of real exchange rate to productivity, respectively export rate, foreign direct investments and capital intensity, on real exchange rate. In equations 1, 2 and 3 relative to total factor productivity and its two components, we then replace these variables by estimated residuals of the three previous functions. In fact these residuals represent the parts of trade openness, of foreign direct investments or of capital intensity that are not explained by real exchange rate. The only consequence of this substitution of the residuals to the intermediate variables themselves is the modification of coefficient associated with real exchange rate that from now on captures the total effects of real exchange rate on total factor productivity growth and on its two components.
Table 1
Waited impacts of real exchange rate (an increase corresponds to an appreciation) on efficiency change, technical progress and productivity change
Impact of exchange rate on intermediate variables
Impact of intermediate variables on productivity growth
Impact of exchange rate on productivity growth
Indirect impact of exchange rate
Openness to outside
Technical efficiency
Technical efficiency
Technical progress
Technical progress
Total productivity
Total productivity
Foreign direct investments
Technical efficiency
Technical efficiency
Technical progress
Technical progress
Total productivity
Total productivity
Capital intensity
Technical efficiency
Technical efficiency
Technical progress
Technical progress
Total productivity
Total productivity
Direct impact of exchange rate
Technical efficiency
Technical progress
Total productivity
Total impact of exchange rate
Technical efficiency
Technical progress
Total productivity
3. Measurement of total factor productivity and its two components
Several methods can be used to calculate productivity. We can use partial productivity, calculated simply as production divided by one production factor, but this measurement is biased because of possible substitution between production factors. In order to avoid this bias, we can calculate total factor productivity (TFP), measured as ratio between production (added value) and weighting sum of production factors. The traditional method of TFP measurement consists to estimate one Cobb-Douglas production function and to consider the part of product non explained by production factors or the residual of the function as TFP measurement. But the residual represents really technological level only with parfait technical efficiency hypothesis. This last hypothesis is contestable in particular for transition countries as China. Malmquist index we will use here allows ignoring this hypothesis and decomposing total factor productivity into technical efficiency and technical progress.
. Production frontier notion and distinction between efficiency and technical progress
Farrell (1957), followed by Aigner, Lovell and Schmidt (1977) and Meeusen and Van des Broeck (1977), defined technical efficiency as the ratio of observed production to maximum or potential production feasible in relation to available technologies, with the same quantities of production factors. This maximum production is also called a production frontier. Thus technical efficiency may be introduced into a model in panel from a classical production function as follows:
, with t = 1,…,T and i = 1 …N
where represents observed production level, production level on the frontier, technical efficiency level, technological level and Xit production factors, for province i during period t.
This production function can thus be transformed into growth rate such as:
where , , and represent respectively growth rates of production, production factors, technical progress and efficiency change; and represent respectively production elasticities relative to production factors and to technological level.
Malmquist Indices of productivity
Malmquist index is a total factor productivity index, calculated relative to previous year. It is a geometric average of efficiency index and technical progress index. Efficiency index is the ratio between observed production to potential one, taken account of available technologies. Technical progress index is potential production, which may be measured on the base of the production factors in current or in previous year. Malmquist index of technical progress is then calculated as geometric average of the two indices.
Malmquist index is illustrated in Figure 3 which represents observed and frontier productions according to a combination of production factors (Nishimizu et Page, 1982; Kalirajan et al ., 1996 ; Wu, 2002). The points A and B represent observed production levels in two periods or successive years t et t+1, . Yt and Yt+1, as well as the points C and D represent respectively frontier or potential production levels, . Y*t and Y*t+1, with different quantities of production factors and different technologies in each of the two periods. The change of the distance from observed production to frontier one between period t and period t+1 represents technical efficiency change. Frontier production movement represents technical progress from one period to another one, which may be measured at the level of production factors either in period t (Xt), or in period t+1 (Xt+1).
Figure 3: Decomposition of output growth rate
Y Production frontier t+1
Y*t+1 D
Yt+1 B
E Production frontier t
F
Y*t C
Yt A
0
Xt Xt+1 X
In practice, Malmquist index calculation uses distance functions. It consists to calculate the ratio of realized production to frontier production, by combining successively the levels of production factors and available technologies. This leads to measure four distance functions. Two first functions are calculated by considering the technologies of period t and successively the amounts of production factors in t and t+1, . and . On figure 3 these two elements correspond respectively to ordinate ratios of points A and C, . A/C and of points B and F, . B/F. We obtain the first index of productivity such as:
.
According to the same principle, two other functions are calculated by considering the technologies of period t+1 with the amounts of production factors in t, then in t+1, such as: , represented on figure by A/E and represented by B/D. We obtain as before productivity index such as:
In order to avoid choosing an arbitrary benchmark or technology reference, we follow Färe et al (1994) to calculate Malmquist index of total factor productivity as a geometric mean of the two preceding indices, such as:
The precedent equation may be written as following:
The first term represents technical efficiency change between two periods, . convergence of provinces towards frontier production. On figure 3, it concerns the rapport of ordinate ratio between points B and D relative to that between points A and C, . . The second represents technical progress or production frontier movement, . on the figure . Malmquist index may be inferior, equal or superior to one, corresponding respectively to deterioration, stagnation or improvement of total factor productivity.
. DEA method (Data Envelopment Analysis)
The calculation of Malmquist index implicates to measure production frontier or efficiency frontier. To calculate this frontier, the most used non-parametric method is DEA method (Data Envelopment Analysis). It consists to use linear programming methods to construct a non-parametric piecewise surface (or frontier) over the data, so that to be able to calculate efficiencies relative to this surface, with hypotheses relative to convex and monotony of all production possibilities. Consequently, with DEA method, we can build an empirical production frontier by piecewise surfaces that are constituted by the most efficient provinces and measure the distance of each province to this frontier as efficiency (Battest et al., 1997). In other words production frontier or the best practice (Fare et allii, 1994) is common for all provinces. The last ones have different indices of technical progress because they do not use the same production factors and thus do not have same innovation level.
The advantage of the DEA non-parameter method is that it does not impose same production function on all Chinese provinces, as should be parametric method. This is why we choose to use it. Its inconvenient is however not to take into account measurement errors and random shocks as in return parametric method allows .
. Productivity measurement of Chinese provinces
Malmquist index of total factor productivity and its two components, technical progress and efficiency, are calculated from 1993 to 2001 for twenty-nine Chinese provinces . DEAP software version is applied (Coelli, 1998).
The data on GDP and employment are issued from annual editions of China Statistical Yearbook. Real GDP is nominal GDP divided by its deflator. On the other hand, we do not have data on capital stock for each province. We have calculated it from gross fixed capital formation (GFCF), which is obtained from Comprehensive statistical data and materials on 50 years of New China for 1952-1998, and completed by China statistical yearbook. .
We have calculated capital stock in two stages. At first we have evaluated initial capital stock for the estimation period, . 1992, by inventory permanent method, supposing an annual depreciation rate of 5 %, such as: where KR and IR represent capital stock and investment in constant prices. This formulation compels to know GFCF during the preceding twenty years and amounts to consider that in 1972 capital stock is equal to investment. The capital stock in 1992 (KR92) is equal to the sum of all past twenty years’ investments in constant prices, net of depreciations. The calculation formula is such as:
where KR72=IR72
Once estimated initial capital stock in 1992 and as capital depreciation data for each province are available from 1993, capital stock estimation for the period from 1993 to 2001 is calculated such as: , where DR represents real depreciations which are nominal depreciations deflated by price indices of investment in fixed assets.
On average, total factor productivity of Chinese provinces increased at an annual average rate of % from 1993 to 2001. It has improved during the whole estimated period, but with a decreasing trend (figure 4). The most important growth rate of productivity is observed in 1993 when it attained 5 %. Productivity growth rate decreased until 1 % in 1998, and then stayed at this level until 2001. Productivity improvement is due to technical progress, which has met an annual average growth rate of % from 1993 to 2001. Inversely, technical efficiency has deteriorated at a rate of % per year on average during the same period.
Total factor productivity does not increase at the same pace in Chinese provinces. In eastern, central and western regions , growth rate is respectively of 3 %, 2 % and 1 % per year on average from 1993 to 2001 (see table A1 in annex). But in central and western regions annual growth rate becomes null since 1998. Eastern region has the fastest average annual growth rate of technical progress (4 %), . the double of central and Western regions’ one. Concerning technical efficiency, Eastern region is not better than Center region. The annual average growth rate of technical efficiency is null in Eastern as in Center, while Western region has deteriorated at the pace of 1 % per year on average during studied period.
Note: A value greater than one indicates productivity improvement; and inversely a value less than one
Means productivity deterioration.
During the estimation period from 1993 to 2001, annual average growth rate of total factor productivity is between % for Guangxi and % for Shanghai, that of technical progress between 0 % for Gansu and % for Shanghai, and that of technical efficiency between % for Shanxi and % for Anhui (see table A2 in annex).
4. Econometric estimation of total factor productivity and of its two components
Econometric estimations of the impact of real effective exchange rate on productivity change and on its two components are based on provincial annual data for the period from 1993 to 2001. They are panel estimations and all variables are expressed in logarithms. The data concerning Chinese provinces are issued from China Statistical Yearbook, apart from contrary indication. Several arguments justify the choice of the estimation period. Firstly, it is since 1992-1993 that China is really working as a market economy, while before domestic prices were quite disconnected from world prices (Guillaumont Jeanneney and Hua 2002). Secondly, real exchange rate is stable or appreciates since 1993 (cf. figure 1 in introduction). Thirdly, this period choice allows using relative homogenous data, particularly concerning exports and price indices of investments in fixed assets.
. Variables definition and calculation
. Dependant variable: Malmquist index
Malmquist indices relative to TFP change and to its two components, defined in previous section, are calculated by DEAP software. As their values are around one, we multiplied them by 100 and then expressed in logarithms to obtain an approximation of productivity change.
Real effective exchange rate
As in 1993 China still had two exchange rates, an official rate and a swap rate. Renminbi exchange rate against dollars is calculated for this year as weighted average of these two exchange rates, taking foreign exchange retention rate as weighting. The real effective exchange rate indices of Chinese provinces are calculated, with base 1995 =100, as ratios between the consumer price index of each province and the average of consumer price indices of its fifteen most important trading partners (defined according to geographical origins of imports in 1998 ), converted into Renminbi. Thus, an increase of real effective exchange rate corresponds to an appreciation of Renminbi. The weighted nominal exchange rates calculated for 1993 are not the same for all provinces because the swap exchange rate was different for each province (Khor, 1993). Despite that for the rest of the estimation period Chinese provinces have a same nominal exchange rate, from now on single, their real effective exchange rates are different because their trading partners and their inflation rates are different (see figure 2 in introduction).
Variables representing transmission channels of real effective exchange rate impact on productivity
As we explained in section 2, the impact of real effective exchange rate on productivity passes partially through exports, foreign direct investments and capital intensity. The first independent variable is thus export ratio of each province to its GDP. The data on provincial exports are available only since 1992, year when China began to use International harmonized Commodity Description and coding system (HS) allowing a better classification of data. These data are established by General Administration of Customs of the People’s Republic of China, which classifies foreign trade by province (conformably to international practice) according to production origins (for exports) and to final destination of products (for imports) (S. Guillaumont and Hua, 2001). They are quietly different from those established for whole transition period by Ministry of Foreign trade and Economic Cooperation (cf. China Regional Economy: A Profile of 17 Years of Reform and Opening Up et Almanac of China’s Foreign Economic Relations and Trade). The differences between these two series seem to come principally from the fact that imports and exports realized by the societies under direct control of central government are not taken into account in Ministry’s data . In fact, these societies import large quantities of commodities (grains, fertilizer etc.), and then send them in domestic market. These imports are without doubt considered as domestic goods from the point view of provinces (Naughton, 1999) .
For each province foreign direct investments ratio (FDI) is direct investments relative to gross fixed capital formation. Capital intensity is the ratio of capital in constant prices to total employment.
. Other control variables
Education variables, representing human capital for each province, are calculated as population proportion having attained at best primary, secondary or university education levels. The data for the years 1990, 1996, 1999 and 2000 are available in China Statistical Yearbook. For the other years, we have used calculation method presented in Démurger (1988). This method consists to calculate the number of diploma holders by adding to accumulated diploma holders newly holders of diploma and by removing the number of death of correspondent year (supposing that mortality ratio is same for the three categories of degrees) and then to divide total number of diploma holders by population. The employment rate of public enterprises is measured by workers number in public enterprises relative to total employment of each province. With regards to change in inventories, we use the gap to the trend of the ratio of changes in inventories relative to GDP. The GDP per capita of each province is calculated from GDP expressed in 1995 constant yuans, converted into dollars by Renminbi exchange rate vis-à-vis dollar in 1995 (. according to World Bank method in World Development indicators), and divided by population.
. Econometric tests
The stationnarity test of Im-Pesaran-Shin allows us to reject unit root hypothesis for all variables in our estimation (see table 2). The results of Breusch and Pagan LM test and Hausman specific test indicate that we cannot reject the hypothesis of one model with fixed effect (see table 2).
Main potential econometric problem concerns the endogeneity of independent variables. We recalled in introduction Balassa-Samuelson effect that supposes an inverse relation to that tested here between productivity change and real exchange rate. The endogeneity problem is also possible for the other independent variables which are macro-economic. We have shown (Guillaumont Jeanneney and Hua, 2002) that by applying Balassa-Samuelson effect, real effective exchange rate of provinces is explained by the ratio of GDP of each province to that of its foreign trading partners on the one hand and to GDP of China as a whole on the other hand. These variables are used here as instruments of exchange rate. For export rate, ratio of foreign direct investments, capital intensity and ratio of changes in inventories, instruments are constituted by variables lagged one year and by dummy variable equal to one for coastal provinces. The results of DWH test do not allow us to reject endogeneities of these variables (see table 3). The results of Pagan/hall heteroskedasticiy test, which is the most pertinent in estimation with instrumental variables, allow us to prefer Generalized Moments Model with instrumental variables to model with fixed effects (Baum, Schaffer and Stillman, 2003). Finally, the pertinence and the validity of instruments are tested using Sargan over-identification test. The results do not allow us to reject the hypothesis that the instruments are in dependant of error terms.
Table 2. Stationnarity test of Im-Pesaran-Shina
Technical efficiency change
***
Technical progress
***
Total factor productivity change
***
Real effective exchange rate
***
Ratio of public employment
**
Primary education
***
Secondary education
***
University education
***
Real GDP per capita lagged one year
**
Gap to trend of changes in inventories /GDP
***
Export ratio
***
Capital intensity
***
FDI/FBCF
***
a. Panel t-statistics
. Results of econometric estimations
Econometric results are reported in tables 3 to 5. Table 3 presents regression results of basic model that includes all intermediate variables corresponding to transmission channels of real exchange rate to productivity. Thus the coefficients relative to real exchange rate represent its direct impact on productivity growth, which does not pass through intermediate variables. Table 4 presents the regressions of intermediate variables on real exchange rate. The results show that actually they are transmission channels of real exchange rate to productivity growth. The residuals of these regressions are substituted to intermediate variables for productivity growth estimations presented in table 5. The results of these estimations with residuals show the total impact of real exchange rate on productivity growth and on its two components.
From the results reported in these tables, most coefficients are significant with waited signs.
We can see from tables 3 and 5 that greater population proportion of one province attaining university education level is, faster efficiency change and technical progress are, and thus total factor productivity growth rate, while the impact of other two education levels is negative or non significant. As waited, the more the public enterprises are important in one province, faster technical progress is and weaker efficiency change is.
Table 3. Estimation of productivity growth and its components: basic model
Technical efficiency change
Technical
progress
Total factor productivity
Real effective exchange rate
***
()
***
()
()
Ratio of public employment
***
()
***
()
**
()
Primary education
***
()
()
()
Secondary education
*
()
()
()
University education
**
()
***
()
**
()
Real GDP per capita lagged one period
()
()
()
Gap to trend of changes in inventories/GDP
*
()
*
()
Export ratio
**
()
***
()
()
Capital intensity
***
()
***
()
()
FDI/FBCF
()
**
()
*
()
Number of observations
228
2280
228
Adjusted R²
Breusch and Pagan LM test
Hausman specific test
Pagan / Hall heteroskedasticity testb
DWH test of endogeneity b
Sargan over-identification test b
b. P value.
Table 4: Estimation of transmission channels of real exchange rate to productivity
Capital intensity
Export ratio
Ratio of foreign direct investments
Gap to trend of changes in inventories /GDP
Real effective exchange rate
***
()
**
()
***
()
**
()
Constant
***
()
***
()
***
()
**
()
Number of observations
261
261
261
261
Adjusted R²
Table 5. Estimation of total impact (direct and indirect) of real exchange rate on productivity change and its components
Technical efficiency change
Technical progress
Total factor productivity change
Real effective exchange rate
***
()
***
()
***
()
Ratio of public employment
***
()
***
()
**
()
Primary education
***
()
()
()
Secondary education
*
()
()
()
University education
**
()
***
()
**
()
Real GDP per capita lagged one period
()
()
()
Residual of gap to trend of changes in inventories /GDP
*
()
*
()
Residual of export ratio
**
()
***
()
()
Residual of capital intensity
***
()
***
()
()
Residual of FDI/FBCF
()
**
()
*
()
Number of observations
228
230
228
Adjusted R²
Breusch and Pagan LM test
Hausman specific test
Pagan / Hall heteroskedasticity testb
DWH test of endogeneity b
Sargan over-identification test b
b. P value.
The results relative to intermediate variables or transmission channels of real exchange rate to productivity are also instructive. Export ratio exercises a positive effect on efficiency, but negative on technical progress. Foreign direct investments favour technical progress as well as capital intensity does, while capital intensity is a factor of less efficiency. Finally, the diminution of total demand related to business cycle (measured by changes in inventories) lessens technical efficiency by reducing utilization ratio of production capacities.
Table 3 allows measuring residual impact of real exchange rate on productivity, which means the impacts that does not pass through intermediate variables; while table 5 allows estimating total effects. We observe that, according to table 3, real effective exchange rate exerts a positive impact on technical efficiency and thus a real appreciation does increase efficiency, probably due to higher competition pressure. On contrary real exchange rate exerts a negative impact on technical progress, and thus real appreciation hinders technical progress likely by decreasing industrial sector progression. This leads that the direct impact of real exchange rate on total factor productivity is not significantly different from zero.
Total impact of real exchange rate on productivity change depends evidently on the impact of real exchange rate on intermediate variables identified as channels of transmission. The table 4 indicates that real appreciation as waited exerts a negative effect on exports and foreign direct investments by decreasing competitiveness of province and a positive effect on capital intensity due to relative price decrease of imported equipment goods (cf. table 1). Moreover real appreciation exerts an unfavorable short run effect on economic activity which is captured by the positive sign of the variable concerning changes in inventories.
By comparing tables 3 and 5, we observe that if total impact of real exchange rate remains significant, positive for efficiency and negative for technical progress, this total impact is significantly less important than the sole direct impact, since the coefficients pass respectively from + and for efficiency and technical progress (in table 3) to + and (in table 5). It results that total impact of real exchange rate on total factor productivity is significantly negative (). In other words, real appreciation of exchange rate in the nineties have contributed to improve technical efficiency, to decrease technical progress, and finally to decease total factor productivity.
5. Conclusion
Three main lessons emerge from our analyses.
The development of university education is actually one priority objective of Chinese government. This strategy is completely justified by the positive impact of this education level on productivity growth, in particular on efficiency.
On the other hand, Japan and the United–States make strong pressures on China government in order that it re-values the Renminbi. As regard the possible impact of a new appreciation of Renminbi on the productivity in China, the implications of our analysis are ambiguous. In the past it seems that the real appreciation has contributed to slow down productivity growth by curbing the path of technical progress. But this negative impact on technical progress has been dampened because the real appreciation has increased the capital- labour ratio and reduced the growth of exports. Moreover it has been partially compensated by a positive impact of the appreciation on efficiency. However, in the future, it is likely that export growth lead to more innovations, as China will increase higher technology exports (by instance electronic products).
Actually the negative impact on productivity growth has been small. Knowing that the annual average growth of the real effective exchange rate has ranged from % to % in the different provinces, the productivity growth has been reduced from to points of percentage for an average annual growth of productivity equal to %. However the striking fact is that according to the Malmquist index the efficiency has fallen in China from 1993 to 2001 at a rate of % a year. So the true stake of the transition of China towards a market economy is the efficiency improvement. We have shown that the real appreciation is a factor of better efficiency.
Finally, (and we join the theme of the previous seminar) the significant relationships between real exchange rate on the one hand, exports, foreign direct investments, capital-labour ratio, business cycle and productivity growth on the other hand, clearly show that during last decade China has become a market economy where economic agents adjust their behaviour to price signals.
References
Aigner, ., Lovell . and Schmidt . (1977), “Formulation and estimation of stochastic frontier models,” Journal of Econometrics, 6 (2), 21-37.
Baum ., Schaffer . and Stillman S. (2003), “Instrumental Variables and GMM Estimation and Testing,” Working Paper No. 545, February, Department of Economics, Boston College.
Balassa B. (1964), “The Purchasing Power Parity Doctrine: a Reappraisal,” Journal of Political Economy, vol. 72, p. 584-596.
Balassa B. (1994), “La théorie de la parité du pouvoir d’achat, un réexamen,” Revue d’Economie du Développement, vol. 2, n° 1, mars, -34.
Caves, . Christensen . and Diewert . (1982), “Multilateral comparisons of output, input and productivity using superlative index numbers,” Economic Journal, 92, p. 73-86.
Chen Y. and S. Démurger (2002), “La croissance de la productivité dans l’industrie manufacturière chinoise, le rôle de l’investissement direct étranger,” Economie internationale, n° 92, 4 trimestre, p. 131-164.
Cortèse L. and P. Hua (forthcoming), “The Effect of the Real Exchange Rate on Technological Progress
Courchene T and R. Harris (1999), “Canada and North American Monetary Union” Canadian Business Economics, , no 4, p. 5-14
Dayal-Gulati A. and . Husain (2002), “Centripetal Forces in China’s Economic Takeoff” IMF Staff Papers, ,no 3, -393.
Dées S. (2002), “Compétitivité-prix et hétérogénéité des échanges extérieurs chinois,” Economie internationale, n° 92, 4 trimestre, p 41-66.
Démurger S. (1999), “Infrastructures, éducation et croissance régionale en Chine”, Revue d'Economie du Développement, n° spécial "Economie chinoise" : croissance et disparités, n° 1-2, p. 71-93.
Färe R., Grssskopf S., Norris M., and Z. Zhang (1994), “Productivity Growth, Technical progress, and Efficiency Change in Industrialized Countries”, American Economic Review, Vol. 84, no. 1, -83.
Farrell, M. J. (1957), “The measurement of productive efficiency,” Journal of the Royal Statistical Society, series A, general 120, p. 253-82.
Feder G. (1983), "On Exports and Economic Growth," Journal of Development Economics, Vol. 12, -2, p. 59-73.
Grubel .(1999), The Case for the Amero: The Merit of Creating a North American Monetary Union, Fraser Institute, Vancouver, Canada
Guillaumont P. (1994), « Politique d’ouverture et croissance économique: les effets de la croissance et de l’instabilité des exportations» Revue d’économie du développement, no 1, -114.
Guillaumont S. and P. Hua (2002), "The Balassa–Samuelson effect and inflation in the Chinese provinces", China Economic Review, 108, 1–27
Harris . (2001), “Is there a Case for Exchange Rate Induced Productivity Changes” mimeo Department of Economics, Simon Fraser University, Canadian Institute for Advanced Research
Kalirajan ., Obwona . and S. Zhao (1996), “A decomposition of total factor productivity growth: the case of Chinese agricultural growth before and after reforms,” American Journal of Agricultural Economics, 78, 331-38.
Krugman P. (1989), « Surévaluation et accélération des productivités : un modèle spéculatif » in Laussel D. and C. Montet, Commerce international et concurrence parfaite, Paris, Economica, -135.
Leibenstein H. (1957), Economic Backwardness and Economic Growth, New-York, Wiley
Leibenstein H. (1966), “Allocative Efficiency versus X-Efficiency” American Economic Review, June -415.,
Lemoine F. (2002), “La Chine dans l’économie mondiale : présentation,” Economie internationale, n° 92, 4 trimestre, p. 5-10.
Lin . and . Liu (2002), “The strategy of development of Chinese economy and the regional income disparity,” CCER, working paper, (in Chinese), no. 2002015.
Lu D. and Y. Qiao (1999) “Hong Kong’s Exchange Rate Regime: Lessons from Singapore,” China Economic Review, vol. 10, p. 122-140.
Mzzusen W. and J. Van des Broeck (1977), “Efficiency estimation from Cobb-Douglas production functions with composed error,” International Economic Review, no 2, p. 435-44.
Naughton, B. (1999), “How much can regional integration do to unify China’s Markets?”, Paper for the Conference on Policy Reform in China, Center for Research on Economic Development and Policy Research, Working Paper, Stanford University, 18-20, November.
Nishimizu M. and . Page (1982), “Total factor productivity growth, technological progress and technical efficiency change: dimensions of productivity change in Yugoslavia, 1965-78,” Economic Journal, vol. 92, 920-36.
Porter . (1990) The Competitive advantage of Nations, Cambridge, Mass, Harvard
University Press
Samuelson P. (1964) “Theoretical Notes on Trade Problems,” Review of Economics and Statistics, vol. 46, March, 145-154.
Sun H. (1998) Foreign investment and economic development in China 1979-1996, Ashgate.
Wu Y. [1999], “Productivity and Efficiency in China's Regional Economics”, in Tsu-Tan Fu et al., Economic Efficiency and Productivity Growth in the Asian-Pacific Region, Edward Elgar.
Yue . and Hua P. (2002), “Does comparative advantage explains export patterns in China?” China Economic Review, vol. 13, 276-296.
Table A1. Evolution of total factor productivity and its two components in three Chinese regions
Eastern
Center
Western
Technical
Efficiency
Technical progress
Total
Productivity
Technical
Efficiency
Technical progress
Total
Productivity
Technical
Efficiency
Technical progress
Total
Productivity
1993
0,99
1,07
1,06
0,98
1,07
1,04
0,97
1,07
1,04
1994
0,98
1,06
1,04
1,00
1,04
1,04
0,98
1,05
1,03
1995
1,00
1,03
1,03
1,02
1,00
1,02
1,01
1,01
1,02
1996
0,99
1,03
1,02
1,02
1,02
1,04
1,01
1,02
1,03
1997
0,99
1,03
1,03
1,01
1,01
1,02
1,00
1,01
1,01
1998
1,00
1,02
1,02
1,01
0,99
1,00
1,01
1,00
1,00
1999
1,00
1,05
1,04
0,99
1,01
1,00
0,98
1,01
0,99
2000
1,00
1,03
1,02
1,00
1,00
1,00
0,99
1,01
1,00
2001
1,00
1,02
1,02
1,00
1,01
1,01
0,99
1,01
1,00
Average
1,00
1,04
1,03
1,00
1,02
1,02
0,99
1,02
1,01
Table A2. Geometric average of total factor productivity of Chinese provinces and its two component from 1993 to 2001
Technical efficiency
Technical progress
Total factor productivity
BEIJING
0,986
1,032
1,017
TIANJIN
1,009
1,039
1,048
HEBEI
0,983
1,015
0,998
SHANXI
0,975
1,014
0,988
INNER MONGOLIA
0,991
1,014
1,005
LIAONING
1,009
1,049
1,059
JILIN
1,012
1,017
1,029
HEILONGJIANG
1,002
1,037
1,038
SHANGHAI
1
1,081
1,081
JIANGSU
1,004
1,042
1,046
ZHEJIANG
0,984
1,038
1,022
ANHUI
1,028
1,015
1,044
FUJIAN
1,005
1,024
1,03
JIANGXI
1,001
1,004
1,005
SHANDONG
1,01
1,017
1,027
HENAN
1,002
1,015
1,017
HUBEI
0,998
1,015
1,014
HUNAN
1,013
1,015
1,028
GUANGDONG
1
1,041
1,041
GUANGXI
0,975
1,007
0,982
SICHUAN
0,995
1,016
1,011
GUIZHOU
0,99
1,011
1,001
YUNNAN
0,984
1,015
0,998
SHAANXI
1,002
1,015
1,017
GANSU
1,017
1
1,017
QINGHAI
0,988
1,018
1,006
NINGXIA
0,997
1,034
1,03
XINJIANG
0,976
1,046
1,021
HAINAN
0,979
1,049
1,027
Average
0,997
1,025
1,022
*The authors would thank Laurent Cortèse for his technical help about DEA Malmquist Index calculation and Martine Bouchut for real effective exchange rate calculation.
An exception is Lu and Qiao (1999) who assume a positive relationship between the appreciation of the real exchange rate and productivity growth in Singapore.
The calculation of the real effective exchange rate is explicated below in the section
The rates of appreciation are calculated on the base of the price of the Renminbi in terms of foreign currencies, namely on the base of the ratio of consumer prices in China expressed in foreign currencies to consumer prices in foreign countries.
Indeed one province trades with foreign countries and with other Chinese provinces, what justifies these two gaps..
An appreciation may increase the return to skilled labour in a Stolper-Samuelson effect if the tradable sector is human-capital intensive relative to the non tradable sector (Harris, 2001, )
China Statistical yearbook
About the numerous studies of these effects, see Chen and Démurger (2002)
The growth rate of GDP non explained by the change of production factors can be related to some phenomena which have any relationship with TFP, such as measurement errors, climate changes, etc.
China is composed of 22 provinces (Hebei, Liaoning, Jiangsu, Zhejiang, Fujian, Shangdong, Guangdong, Hainan, Shanxi, Jilin, Heilongjiang, Henan, Anhui, Hubei, Hunan, Jiangxi, Gansu, Shaanxi, Sichuan, Guizhou, Yunnan and Qinghai), four autonomous municipalities under the direct control of central government (Beijing, Tianjin, Shanghai et Chongqing), and five autonomous regions (Guangxi, Inner Mongolia, Ningxia, Xinjiang and Tibet). In our econometric analysis, the autonomous region of Tibet is absent for lack of data, the statistics of Chongqing, province, created in 1997, have been included in those of Sichuan, this leads to retain 29 provinces in large sense.
Gross fixed capital formation in constant prices is calculated as GFCF in current prices divided respectively by its prices for the period from 1972 to 1991 and prices of investment in fixed assets from 1992 to 2001, which correspond in China two different series. GFCF prices are obtained from Zhongguo Guorei ShengShang Zongzhi Hesuan Lishi Ziliao, 1952-1995. The lacking data for several provinces are replaced by detail prices (see Lin and Liu, 2002). The price indices of investment in fixed assets are originated from China Statistical Yearbook. It should be better to use the same deflator for the whole calculation period, but GFCF prices are available only until 1995, and the price indices of investments in fixed assets are available only since 1992. We have used the same deflator (prices of fixed investments) for the whole estimation period (1993-2001).
Eastern region includes Beijing, Tianjin, Hebei, Liaoning, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong and Hainan. Central region includes Shanxi, Inner Mongolia, Jilin, Heilongjiang, Anhui, Jiangxi, Henan, Hunan, Hubei and Guangxi; Weastern region concerns Sichuan (including Chongqi), Guizhou, Yunnan, Shannxi, Gansu, Qinghai, Ningxia and Xinjiang.
We have to eliminate unfortunately several countries of ex-soviet union for which we do not have the data on exchange rate. The consumer price indices of foreign partners are obtained from IMF, International Financial Statistics. The price indices of each province are originated from China Statistical Yearbook. Swap rates of each province in 1992 and 1993 are obtained in Khor (1993).
We only obtained import origins for different provinces for this year close to China’s Customs General Administration.
The differences of openness rates are also very important, particularly for three autonomous cities, Beijing (81 % according to custom data and 26% according to Ministry data), Tianjin (43 % and 33 %), Shanghai (62 % and 23%), Guangdong (124% and 60%) and Hainan (50% and 62 %).
The choice of custom data is justified particularly since 2000, Ministry refers to the same data.
Source IMF International Financial Statistics
As real effective exchange rate has negative effect on exports and foreign direct investments, and these variables exercise the first one a positive effect on efficiency and the second one technical progress, the introduction of the residuals of functions in table 4 should decrease the coefficient relative to real effective exchange rate (should lead it less positive for efficiency but more negative for technical progress). The real appreciation exercises inversely positive effect on capital intensity. The introduction of residuals increases the coefficients of real effective exchange rate (more positive for efficiency, but less negative for technical progress). These effects offset each other partially to give the results in table 5.
PAGE
PAGE 28
Graph1
Technical efficiency
Technical progress
TFP
Figure 4. Malmquist indices relative to total factor productivity and its two components (geometric average of Chinese provinces)
Feuil1
effch techch tfpch
4 Shanxi 25 Gansu 100 20 Guangxi
20 Guangxi 14 Jiangxi 4 Shanxi
28 Xinjiang 20 Guangxi 3 Hebei
29 Hainan 22 Guizhou 23 Yunnan
3 Hebei 4 Shanxi 22 Guizhou
11 Zhejiang 5 Mongolie intérieure 5 Mongolie intérieure
23 Yunnan 3 Hebei 14 Jiangxi
1 Beijing 12 Anhui 26 Qinghai
26 Qinghai 16 Henan 21 Sichuan
22 Guizhou 99 17 Hubei 17 Hubei
5 Mongolie intérieure 18 Hunan 1 Beijing
21 Sichuan 23 Yunnan 16 Henan
27 Ningxia 24 Shaanxi 24 Shaanxi
17 Hubei 21 Sichuan 25 Gansu
9 Shanghai 100 7 Jilin 28 Xinjiang
19 Guangdong 100 15 Shandong 11 Zhejiang
14 Jiangxi 26 Qinghai 15 Shandong
8 Heilongjiang 13 Fujian 29 Hainan
16 Henan 1 Beijing 18 Hunan
24 Shaanxi 27 Ningxia 7 Jilin
10 Jiangsu 8 Heilongjiang 13 Fujian 103
13 Fujian 11 Zhejiang 27 Ningxia 103
2 Tianjin 2 Tianjin 8 Heilongjiang
6 Liaoning 19 Guangdong 19 Guangdong
15 Shandong 101 10 Jiangsu 12 Anhui
7 Jilin 28 Xinjiang 10 Jiangsu
18 Hunan 6 Liaoning 2 Tianjin
25 Gansu 29 Hainan 6 Liaoning
12 Anhui 9 Shanghai 9 Shanghai
Feuil1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
effch
Graphique 7. Taux de croissance annuel de l'efficience technique
STOCKD1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
techch
Graphique 6. Taux de croissance annuel du progrès technique
grap3regions
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
tfpch
Graphique de croissance moyen annuel de la PTF
Results from DEA P Versio nom de fichier dans DEA est STOCKD1
MALMQUI ST INDEX SUMMARY
year = 2
firm effch techch tfpch
1 BEIJING est
2 TIANJIN est
3 HEBEI est
4 SHANXI centre
5 INNER MONGOLIA centre
6 LIAONING est
7 JILIN centre
8 HEILONGJIANG centre
9 SHANGAI est 1
10 JIANGSU est
11 ZHEJIANG est
12 ANHUI centre
13 FUJIAN est
14 JIANGXI centre
15 SHANDONG est
16 HENAN centre
17 HUBEI centre
18 HUNAN centre
19 GUANGDONG est 1
20 GUANGXI est 1
21 Sichuan+Chongqing ouest
22 GUIZHOU ouest
23 YUNNAN ouest
24 SHAANXI ouest
25 GANSU ouest
26 QINGHAI ouest
27 NINGXIA ouest
28 XINJIANG ouest
29 hainan est est centre Ouest
et pt ptf et pt ptf et pt ptf
mean
year = 3
firm effch techch tfpch
1 BEIJING est
2 TIANJIN est
3 HEBEI est
4 SHANXI centre
5 INNER MONGOLIA centre
6 LIAONING est
7 JILIN centre
8 HEILONGJIANG centre
9 SHANGAI est 1
10 JIANGSU est
11 ZHEJIANG est
12 ANHUI centre
13 FUJIAN est 1
14 JIANGXI centre
15 SHANDONG est
16 HENAN centre
17 HUBEI centre
18 HUNAN centre
19 GUANGDONG est 1
20 GUANGXI est 1
21 Sichuan+Chongqing ouest
22 GUIZHOU ouest
23 YUNNAN ouest
24 SHAANXI ouest
25 GANSU ouest
26 QINGHAI ouest
27 NINGXIA ouest
28 XINJIANG ouest
29 hainan est est centre Ouest
et pt ptf et pt ptf et pt ptf
mean 1
year = 4
firm effch techch tfpch
1 BEIJING est
2 TIANJIN est
3 HEBEI est
4 SHANXI centre
5 INNER MONGOLIA centre
6 LIAONING est
7 JILIN centre
8 HEILONGJIANG centre
9 SHANGAI est 1
10 JIANGSU est
11 ZHEJIANG est
12 ANHUI centre
13 FUJIAN est 1
14 JIANGXI centre
15 SHANDONG est
16 HENAN centre
17 HUBEI centre
18 HUNAN centre
19 GUANGDONG est 1
20 GUANGXI est
21 Sichuan+Chongqing ouest
22 GUIZHOU ouest
23 YUNNAN ouest
24 SHAANXI ouest
25 GANSU ouest
26 QINGHAI ouest
27 NINGXIA ouest
28 XINJIANG ouest
29 hainan est est centre Ouest
et pt ptf et pt ptf et pt ptf
mean
year = 5
firm effch techch tfpch
1 BEIJING est
2 TIANJIN est
3 HEBEI est
4 SHANXI centre
5 INNER MONGOLIA centre
6 LIAONING est
7 JILIN centre
8 HEILONGJIANG centre
9 SHANGAI est 1
10 JIANGSU est
11 ZHEJIANG est
12 ANHUI centre
13 FUJIAN est 1
14 JIANGXI centre 1
15 SHANDONG est
16 HENAN centre
17 HUBEI centre
18 HUNAN centre
19 GUANGDONG est 1
20 GUANGXI est
21 Sichuan+Chongqing ouest
22 GUIZHOU ouest
23 YUNNAN ouest
24 SHAANXI ouest
25 GANSU ouest
26 QINGHAI ouest
27 NINGXIA ouest
28 XINJIANG ouest
29 hainan est est centre Ouest
et pt ptf et pt ptf et pt ptf
mean
year = 6
firm effch techch tfpch
1 BEIJING est
2 TIANJIN est
3 HEBEI est
4 SHANXI centre
5 INNER MONGOLIA centre
6 LIAONING est
7 JILIN centre
8 HEILONGJIANG centre
9 SHANGAI est 1
10 JIANGSU est
11 ZHEJIANG est
12 ANHUI centre
13 FUJIAN est 1
14 JIANGXI centre 1 1 1
15 SHANDONG est
16 HENAN centre
17 HUBEI centre
18 HUNAN centre
19 GUANGDONG est 1
20 GUANGXI est
21 Sichuan+Chongqing ouest
22 GUIZHOU ouest
23 YUNNAN ouest 1
24 SHAANXI ouest
25 GANSU ouest
26 QINGHAI ouest
27 NINGXIA ouest
28 XINJIANG ouest
29 hainan est est centre Ouest
et pt ptf et pt ptf et pt ptf
mean
year = 7
firm effch techch tfpch
1 BEIJING est
2 TIANJIN est
3 HEBEI est
4 SHANXI centre
5 INNER MONGOLIA centre
6 LIAONING est
7 JILIN centre
8 HEILONGJIANG centre
9 SHANGAI est 1
10 JIANGSU est
11 ZHEJIANG est
12 ANHUI centre
13 FUJIAN est 1 1 1
14 JIANGXI centre 1
15 SHANDONG est
16 HENAN centre
17 HUBEI centre
18 HUNAN centre
19 GUANGDONG est 1
20 GUANGXI est
21 Sichuan+Chongqing ouest
22 GUIZHOU ouest
23 YUNNAN ouest
24 SHAANXI ouest
25 GANSU ouest
26 QINGHAI ouest
27 NINGXIA ouest
28 XINJIANG ouest
29 hainan est est centre Ouest
et pt ptf et pt ptf et pt ptf
mean
year = 8
firm effch techch tfpch
1 BEIJING est
2 TIANJIN est
3 HEBEI est
4 SHANXI centre
5 INNER MONGOLIA centre
6 LIAONING est
7 JILIN centre
8 HEILONGJIANG centre
9 SHANGAI est 1
10 JIANGSU est
11 ZHEJIANG est
12 ANHUI centre
13 FUJIAN est 1
14 JIANGXI centre 1
15 SHANDONG est 1
16 HENAN centre
17 HUBEI centre
18 HUNAN centre
19 GUANGDONG est 1
20 GUANGXI est
21 Sichuan+Chongqing ouest
22 GUIZHOU ouest
23 YUNNAN ouest
24 SHAANXI ouest
25 GANSU ouest
26 QINGHAI ouest
27 NINGXIA ouest
28 XINJIANG ouest
29 hainan est est centre Ouest
et pt ptf et pt ptf et pt ptf
mean 1
year = 9
firm effch techch tfpch
1 BEIJING est
2 TIANJIN est
3 HEBEI est
4 SHANXI centre
5 INNER MONGOLIA centre
6 LIAONING est
7 JILIN centre
8 HEILONGJIANG centre
9 SHANGAI est 1
10 JIANGSU est
11 ZHEJIANG est
12 ANHUI centre
13 FUJIAN est 1
14 JIANGXI centre 1
15 SHANDONG est 1
16 HENAN centre
17 HUBEI centre
18 HUNAN centre
19 GUANGDONG est 1
20 GUANGXI est
21 Sichuan+Chongqing ouest
22 GUIZHOU ouest
23 YUNNAN ouest
24 SHAANXI ouest
25 GANSU ouest
26 QINGHAI ouest
27 NINGXIA ouest
28 XINJIANG ouest
29 hainan est est centre Ouest
et pt ptf et pt ptf et pt ptf
mean 1
year = 10
firm effch techch tfpch
1 BEIJING est
2 TIANJIN est
3 HEBEI est
4 SHANXI centre
5 INNER MONGOLIA centre 1
6 LIAONING est
7 JILIN centre
8 HEILONGJIANG centre
9 SHANGAI est 1
10 JIANGSU est
11 ZHEJIANG est
12 ANHUI centre
13 FUJIAN est
14 JIANGXI centre 1
15 SHANDONG est 1
16 HENAN centre
17 HUBEI centre
18 HUNAN centre
19 GUANGDONG est 1
20 GUANGXI est
21 Sichuan+Chongqing ouest
22 GUIZHOU ouest
23 YUNNAN ouest
24 SHAANXI ouest 1
25 GANSU ouest
26 QINGHAI ouest
27 NINGXIA ouest
28 XINJIANG ouest
29 hainan est est centre Ouest
et pt ptf et pt ptf et pt ptf
mean 1
MALMQUI ST INDEX SUMMARY S
year
efficience technique progrès technique PTF
2 1993
3 1994
4 1995
5 1996
6 1997
7 1998
8 1999
9 2000
10 2001
mean
MALMQUI ST INDEX SUMMARY
firm effch techch tfpch
1 BEIJING
2 TIANJIN
3 HEBEI
4 SHANXI
5 INNER MONGOLIA
6 LIAONING
7 JILIN
8 HEILONGJIANG
9 SHANGAI 1
10 JIANGSU
11 ZHEJIANG
12 ANHUI
13 FUJIAN
14 JIANGXI
15 SHANDONG
16 HENAN
17 HUBEI
18 HUNAN
19 GUANGDONG 1
20 GUANGXI
21 Sichuan+Chongqing
22 GUIZHOU
23 YUNNAN
24 SHAANXI
25 GANSU 1
26 QINGHAI
27 NINGXIA
28 XINJIANG
29 hainan
mean
[Note th at all M almquist are geometric means]
Technical efficiency Technical progress TFP
1993
1994
1995
1996
1997
1998
1999
2000
2001
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
efficience technique
progrès technique
PTF
Graphique 4. Evolution de la productivité totale des facteurs et ses deux composantes
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
Technical efficiency
Technical progress
TFP
Figure 4. Malquist indices relative to total factor productivity and its txo compoents (geometric average of Chinese provinces)
est centre ouest
et pt ptf et pt ptf et pt ptf
1993 1993 1993
1994 1994 1994
1995 1995 1995
1996 1996 1996
1997 1997 1997
1998 1998 1998
1999 1999 1999
2000 2000 2000
2001 2001 2001
1
²
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
et
pt
ptf
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
et
pt
ptf
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
et
pt
ptf
Graph1
tcer9301
Figure 2: Annual average rate of appreciation of the real effective exchange rate from 1993 to 2001
Feuil1
prov nprov an tcer: monter=dépréciation 0 tcer9301 tcer9301 prov tcer9301
BEIJING 1 1993 Hainan 29 Hainan
BEIJING 1 1994 Yunnan 23 Yunnan
BEIJING 1 1995 Guizhou 22 Guizhou
BEIJING 1 1996 Guangdong 19 Guangdong
BEIJING 1 1997 Guangxi 20 Guangxi
BEIJING 1 1998 Henan 16 Henan
BEIJING 1 1999 Fujian 13 Fujian
BEIJING 1 2000 Sichuan 21 Sichuan
BEIJING 1 2001 tcer9301 tcer9301 Hebei 3 Hebei
TIANJIN 2 1993 Zhejiang 11 Zhejiang
TIANJIN 2 1994 Ningxia 27 Ningxia
TIANJIN 2 1995 Tianjin 2 Tianjin
TIANJIN 2 1996 Hubei 17 Hubei
TIANJIN 2 1997 Jilin 7 Jilin
TIANJIN 2 1998 Xinjiang 28 Xinjiang
TIANJIN 2 1999 Heilongjiang 8 Heilongjiang
TIANJIN 2 2000 Jiangsu 10 Jiangsu
TIANJIN 2 2001 tcer9301 tcer9301 Mongolie intérieure 5 Mongolie intérieure
HEBEI 3 1993 Anhui 12 Anhui
HEBEI 3 1994 Gansu 25 Gansu
HEBEI 3 1995 Liaoning 6 Liaoning
HEBEI 3 1996 Qinghai 26 Qinghai
HEBEI 3 1997 Shanxi 4 Shanxi
HEBEI 3 1998 Shaanxi 24 Shaanxi
HEBEI 3 1999 Hunan 18 Hunan
HEBEI 3 2000 Jiangxi 14 Jiangxi
HEBEI 3 2001 tcer9301 tcer9301 Shandong 15 Shandong
SHANXI 4 1993 Shanghai 9 Shanghai
SHANXI 4 1994 Beijing 1 Beijing
SHANXI 4 1995
SHANXI 4 1996
SHANXI 4 1997
SHANXI 4 1998
SHANXI 4 1999
SHANXI 4 2000
SHANXI 4 2001 tcer9301 tcer9301
INNER MONGOLIA 5 1993
INNER MONGOLIA 5 1994
INNER MONGOLIA 5 1995
INNER MONGOLIA 5 1996
INNER MONGOLIA 5 1997
INNER MONGOLIA 5 1998
INNER MONGOLIA 5 1999
INNER MONGOLIA 5 2000
INNER MONGOLIA 5 2001 tcer9301 tcer9301
LIAONING 6 1993
LIAONING 6 1994
LIAONING 6 1995
LIAONING 6 1996
LIAONING 6 1997
LIAONING 6 1998
LIAONING 6 1999
LIAONING 6 2000
LIAONING 6 2001 tcer9301 tcer9301
JILIN 7 1993
JILIN 7 1994
JILIN 7 1995
JILIN 7 1996
JILIN 7 1997
JILIN 7 1998
JILIN 7 1999
JILIN 7 2000
JILIN 7 2001 tcer9301 tcer9301
HEILONGJIANG 8 1993
HEILONGJIANG 8 1994
HEILONGJIANG 8 1995
HEILONGJIANG 8 1996
HEILONGJIANG 8 1997
HEILONGJIANG 8 1998
HEILONGJIANG 8 1999
HEILONGJIANG 8 2000
HEILONGJIANG 8 2001 tcer9301 tcer9301
SHANGAI 9 1993
SHANGAI 9 1994
SHANGAI 9 1995
SHANGAI 9 1996
SHANGAI 9 1997
SHANGAI 9 1998
SHANGAI 9 1999
SHANGAI 9 2000
SHANGAI 9 2001 tcer9301 tcer9301
JIANGSU 10 1993
JIANGSU 10 1994
JIANGSU 10 1995
JIANGSU 10 1996
JIANGSU 10 1997
JIANGSU 10 1998
JIANGSU 10 1999
JIANGSU 10 2000
JIANGSU 10 2001 tcer9301 tcer9301
ZHEJIANG 11 1993
ZHEJIANG 11 1994
ZHEJIANG 11 1995
ZHEJIANG 11 1996
ZHEJIANG 11 1997
ZHEJIANG 11 1998
ZHEJIANG 11 1999
ZHEJIANG 11 2000
ZHEJIANG 11 2001 tcer9301 tcer9301
ANHUI 12 1993
ANHUI 12 1994
ANHUI 12 1995
ANHUI 12 1996
ANHUI 12 1997
ANHUI 12 1998
ANHUI 12 1999
ANHUI 12 2000
ANHUI 12 2001 tcer9301 tcer9301
FUJIAN 13 1993
FUJIAN 13 1994
FUJIAN 13 1995
FUJIAN 13 1996
FUJIAN 13 1997
FUJIAN 13 1998
FUJIAN 13 1999
FUJIAN 13 2000
FUJIAN 13 2001 tcer9301 tcer9301
JIANGXI 14 1993
JIANGXI 14 1994
JIANGXI 14 1995
JIANGXI 14 1996
JIANGXI 14 1997
JIANGXI 14 1998
JIANGXI 14 1999
JIANGXI 14 2000
JIANGXI 14 2001 tcer9301 tcer9301
SHANDONG 15 1993
SHANDONG 15 1994
SHANDONG 15 1995
SHANDONG 15 1996
SHANDONG 15 1997
SHANDONG 15 1998
SHANDONG 15 1999
SHANDONG 15 2000
SHANDONG 15 2001 tcer9301 tcer9301
HENAN 16 1993
HENAN 16 1994
HENAN 16 1995
HENAN 16 1996
HENAN 16 1997
HENAN 16 1998
HENAN 16 1999
HENAN 16 2000
HENAN 16 2001 tcer9301 tcer9301
HUBEI 17 1993
HUBEI 17 1994
HUBEI 17 1995
HUBEI 17 1996
HUBEI 17 1997
HUBEI 17 1998
HUBEI 17 1999
HUBEI 17 2000
HUBEI 17 2001 tcer9301 tcer9301
HUNAN 18 1993
HUNAN 18 1994
HUNAN 18 1995
HUNAN 18 1996
HUNAN 18 1997
HUNAN 18 1998
HUNAN 18 1999
HUNAN 18 2000
HUNAN 18 2001 tcer9301 tcer9301
GUANGDONG 19 1993
GUANGDONG 19 1994
GUANGDONG 19 1995
GUANGDONG 19 1996
GUANGDONG 19 1997
GUANGDONG 19 1998
GUANGDONG 19 1999
GUANGDONG 19 2000
GUANGDONG 19 2001 tcer9301 tcer9301
GUANGXI 20 1993
GUANGXI 20 1994
GUANGXI 20 1995
GUANGXI 20 1996
GUANGXI 20 1997
GUANGXI 20 1998
GUANGXI 20 1999
GUANGXI 20 2000
GUANGXI 20 2001 tcer9301 tcer9301
Sichuan+Chongqing 21 1993
Sichuan+Chongqing 21 1994
Sichuan+Chongqing 21 1995
Sichuan+Chongqing 21 1996
Sichuan+Chongqing 21 1997
Sichuan+Chongqing 21 1998
Sichuan+Chongqing 21 1999
Sichuan+Chongqing 21 2000
Sichuan+Chongqing 21 2001 tcer9301 tcer9301
GUIZHOU 22 1993
GUIZHOU 22 1994
GUIZHOU 22 1995
GUIZHOU 22 1996
GUIZHOU 22 1997
GUIZHOU 22 1998
GUIZHOU 22 1999
GUIZHOU 22 2000
GUIZHOU 22 2001 tcer9301 tcer9301
YUNNAN 23 1993
YUNNAN 23 1994
YUNNAN 23 1995
YUNNAN 23 1996
YUNNAN 23 1997
YUNNAN 23 1998
YUNNAN 23 1999
YUNNAN 23 2000
YUNNAN 23 2001 tcer9301 tcer9301
SHAANXI 24 1993
SHAANXI 24 1994
SHAANXI 24 1995
SHAANXI 24 1996
SHAANXI 24 1997
SHAANXI 24 1998
SHAANXI 24 1999
SHAANXI 24 2000
SHAANXI 24 2001 tcer9301 tcer9301
GANSU 25 1993
GANSU 25 1994
GANSU 25 1995
GANSU 25 1996
GANSU 25 1997
GANSU 25 1998
GANSU 25 1999
GANSU 25 2000
GANSU 25 2001 tcer9301 tcer9301
QINGHAI 26 1993
QINGHAI 26 1994
QINGHAI 26 1995
QINGHAI 26 1996
QINGHAI 26 1997
QINGHAI 26 1998
QINGHAI 26 1999
QINGHAI 26 2000
QINGHAI 26 2001 tcer9301 tcer9301
NINGXIA 27 1993
NINGXIA 27 1994
NINGXIA 27 1995
NINGXIA 27 1996
NINGXIA 27 1997
NINGXIA 27 1998
NINGXIA 27 1999
NINGXIA 27 2000
NINGXIA 27 2001 tcer9301 tcer9301
XINJIANG 28 1993
XINJIANG 28 1994
XINJIANG 28 1995
XINJIANG 28 1996
XINJIANG 28 1997
XINJIANG 28 1998
XINJIANG 28 1999
XINJIANG 28 2000
XINJIANG 28 2001 tcer9301 tcer9301
hainan 29 1993
hainan 29 1994
hainan 29 1995
hainan 29 1996
hainan 29 1997
hainan 29 1998
hainan 29 1999
hainan 29 2000
hainan 29 2001
Feuil1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
tcer9301
Graphique 2: Taux de croissance annuel moyen du taux de change effectif réel de 1993 à 2001
Feuil2
Feuil3
Graph1
100
TCERa
Figure 1
Real Effective Exchange Rate of China (1995=100)
Feuil1
gdp95/pop Taux de croissance annuel du PIB réel par tête (1995=100)
1952 1992
1953 1993
1954 1994
1955 1995
1956 1996
1957 1997
1958 1998
1959 1999
1960 2000
1961 2001
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990 TCERa PIB
1991 1991
1992 1992
1993 1993
1994 1994
1995 1995 100
1996 1996
1997 1997
1998 1998
1999 1999
2000 2000
2001 2001
Feuil1
0
0
0
0
0
0
0
0
0
0
Taux de croissance annuel du PIB réel par tête (1995=100)
Feuil2
0
0
0
0
0
0
0
0
0
0
0
TCERa
Graphique 1
Evolution du taux de change effectif réel de la Chine (1995=100)
Feuil3
