China's Foreign Trade Behavior in the 1980's
An Empirical Analysis
Author: Adi Brender1
  • 1 0000000404811396https://isni.org/isni/0000000404811396International Monetary Fund

This paper studies the behavior of China’s foreign trade in the 1980s. Chinese customs data are used to construct, for the first time, quarterly unit value and volume series that are then used to estimate foreign trade income and price elasticities. The estimated export supply price elasticities are negative and became even more negative after the reforms in 1985. The import price elasticity is very close to minus one. These results suggest that administrative controls still have a dominant effect on China’s trade and may also suggest that reforms in the management system of Chinese enterprises were not sufficient to create profit maximizing behavior by managers.

Abstract

This paper studies the behavior of China’s foreign trade in the 1980s. Chinese customs data are used to construct, for the first time, quarterly unit value and volume series that are then used to estimate foreign trade income and price elasticities. The estimated export supply price elasticities are negative and became even more negative after the reforms in 1985. The import price elasticity is very close to minus one. These results suggest that administrative controls still have a dominant effect on China’s trade and may also suggest that reforms in the management system of Chinese enterprises were not sufficient to create profit maximizing behavior by managers.

I. Introduction 1/

Since the adoption of outward-oriented reform policies, beginning in 1978, China’s foreign trade has expanded at a dramatic rate. Between 1980 and 1990, the annual value of trade tripled (from $39 billion to $115 billion), and China’s share of world exports nearly doubled (from 1.0 percent to 1.9 percent) (Tables 1 and 2).

Table 1:

China: Trade Growth, 1978-90

article image
Sources: IMF, International Financial Statistics, June 1990; and IMF, International Financial Statistics Yearbook, 1989.
Table 2:

Growth and Share in World Trade, 1980-90

(In billions of U.S. dollars)

article image
Sources: For China: General Administration of Customs, various issues; and for World: IMF, International Financial Statistics Yearbook, 1989 and IMF, International Financial Statistics, May 1991.

This paper uses Chinese customs data to analyze the behavior of China’s trade during the 1980s. The disaggregated data are used to construct, for the first time, quarterly unit value and volume series that are then used to estimate foreign trade price and income elasticities. Close relations are found between the sequence of economic reforms and trade developments. Administrative controls and changes in the management system of Chinese enterprises and foreign trade corporations (FTCs) are also shown to have an important effect on the behavior of the Chinese entities involved in foreign trade. The results suggest that the decentralization of decision making in the absence of strong links between managerial performance and rewards may lead managers to reduce the volume of exports as prices increase. The results also suggest that, despite the far-reaching reforms, China in 1990 was still far from being a free market economy. One of the reasons for this is that much of the delegation of decision making was not to enterprises but to local governments.

The high savings rate in China 2/ allowed for large investments without a large current account deficit. Still, the need for imported technology required the development of exportable products to finance technology imports. Primary products led export growth early in the decade (in volume terms) while industrial products exports grew faster since 1985. Some diversification of industrial exports towards home electronics and other mechanical and electronic products began to take place in the late 1980s after a stage of import substitution earlier in the decade. Direct foreign investment was another important source of foreign exchange earnings, while the deteriorating terms of trade, and especially the decline in oil prices, levied a heavy burden on the development process.

Chapter II of this paper lays the background for the trade analysis with an overview of the reforms, particularly in the external sector. Chapter III presents an analysis of unit values and volume developments and the relations between domestic production and trade. Chapter IV derives and analyzes income and price elasticities and the effects of other factors on China’s trade. Chapter V offers some concluding remarks. Finally, Annex I describes the data base, its coverage, and its limitations, and explains the methods used to calculate the indices presented in this paper.

II. Background: Economic Reforms

1. Decentralization of decision making

Beginning in 1978, the Chinese economy entered a phase of reforms that featured the relaxation of direct planning controls, decentralization of decision making in state-owned enterprises, and the emergence of a non-state sector. 1/ The reforms had important implications for the external sector, and a short description of them is therefore in order prior to the discussion of China’s foreign trade. 2/

Reforms began in the rural areas in 1979. The reforms allowed more independence to households in the agricultural sector regarding production decisions and gradually permitted the long-term private leasing of land. 3/ Households were still required to enter into contracts with state or local government agencies, but any surplus belonged to the household and could be sold directly in the market. At the same time, the operations of nonagricultural enterprises in the rural sector were also liberalized. 4/ A large number of both collectively and privately owned enterprises were established and, by 1987, these accounted for 25 percent of industrial output and 17 percent of industrial employment.

By 1984, the focus of reforms shifted to the state enterprise sector. 5/ Managers received greater responsibility for operations of their enterprise, and profits above an agreed level remained in the enterprise. However, a number of factors--adjustment taxes on profitable enterprises, renegotiation of contracts under the management responsibility system, and limitations on bonus payments--potentially diminish the connection between managerial performance and rewards.

The decade of reforms did increase the autonomy of enterprises, but the Chinese enterprise is still far from operating only according to market signals. State ownership and government (especially local) administrative controls, which vary in intensity, still play a major role in economic decisions, especially in the industrial sector. Managers still have to negotiate with the authorities about enterprise operations, and government officials have a final say in many decisions. The state also controls the supply of certain crucial inputs, a large number of input and output prices, and exerts substantial influence on the allocation of credit. As a result, managerial decisions are not necessarily aimed at growth or profit maximization. 1/

2. Reforms in the external sector

The reforms in China’s external sector had a significant impact on the structure and size of foreign trade. This section describes the developments in three major fields: (a) access to foreign markets; (b) access to foreign exchange; and (c) exchange rate determination. These developments, in turn, affected the behavior of the trading entities and the evolution of China’s trade itself.

a. Access to foreign markets 2/

Prior to 1978, China’s trade was conducted by 12 foreign trade corporations (FTCs), all under the Ministry of Foreign Economic Relations and Trade (MOFERT). 3/ These corporations traded the quantities directed by the central plan and all their losses or profits were covered or absorbed by the state. Enterprises, retailers, and local governments did not have access to foreign markets. As part of the production plan, enterprises were assigned target quantities that they had to supply to the FTCs for export.

After 1978, the number of entities involved in foreign trade increased substantially and their discretion in choosing traded goods increased as well. The central government introduced export and import licensing and duties as part of steps to reduce the role of mandatory planning in economic management. The plan still assigned mandatory quantitative targets for trade in some commodities, but for others only value targets were set to local governments and FTCs, and they had flexibility in determining how to achieve the targets. Over the years, the number of products subject to quantitative or value targets set by the central government declined and a system based on value quotas for total exports by the local governments was introduced. However, local governments still tend to assign value targets to their enterprises in order to achieve their aggregate target.

After 1978, ministries other than MOFERT were also allowed to establish FTCs. The existing FTCs were also permitted to delegate authority to their branch offices. In 1984, the monopolistic positions of most FTCs were abolished and the right to engage directly in trade was granted to some enterprises. In addition, a system allowing FTCs to operate as agents of enterprises was introduced. In this agency system the FTCs charge a fee for their services but they do not assume responsibility for profits or losses. In early 1988 the system was extended, and at the same time, more domestic enterprises were given the authority to export directly and import inputs.

In early 1988, local branches of most of the FTCs became independent profit units under the management contract system and were made accountable to the local governments. This, together with the establishment of many new FTCs, increased the number of FTCs to over 5,000. However, during 1989-90, as regulations on their activities were tightened, the number of FTCs was reduced to about 4,000 (World Bank, 1990).

The increased access to foreign markets has the potential to increase the efficiency of the trade sector and improve the flow of information from foreign markets to Chinese enterprises. However, since imports are still, in general, subject to quantitative restrictions, the direct import of inputs by enterprises may be motivated in part as a means of bypassing restrictions. 1/ In the presence of these and other distortions in the domestic and external sectors, it is unclear whether partial trade liberalization increases the efficiency of the economy. 2/

b. Access to foreign exchange

Until 1984, all foreign exchange earnings in China were transferred to the central government. The government then allocated the foreign exchange according to the trade plan. In early 1984, local governments were granted the right to a share (usually 12.5 percent) of the foreign exchange earned in their region, and by early 1985 enterprises were allowed to retain a similar proportion of their foreign exchange earnings. These shares were subject to negotiations and changes throughout the years and varied between provinces and enterprises.

To encourage exports, enterprises that exported more than their target received a larger share (up to 100 percent) of their incremental foreign exchange earnings. In 1988, the retention rate was increased, especially for high domestic value-added products, and reached 60 percent for garments and 100 percent for home electronics. Also in 1988, access to foreign exchange adjustment centers (FEACs) 3/ was permitted for all entities with retention quotas. This allows them to realize larger profits from exports since the exchange rate at the FEACs is more depreciated than the official rate. It also allows enterprises with low priority at the allocation process of foreign exchange to attain the currency they need to import inputs.

The retention system was modified in early 1991 and a uniform rate was set throughout the country. The retention rate is 60 percent to the FTC and 10 percent to the supplier for general commodities, and 90 percent and 10 percent, respectively, for mechanical and electronic products. The central government and local governments receive 20 percent and 10 percent, respectively, of foreign exchange earnings on general commodities. The central government is eligible to purchase additional foreign exchange from the FTCs and enterprises at the FEACs exchange rate.

c. Exchange rate determination and developments

The existence in China of many import and export duties and quantitative controls, as well as dual exchange rates through most of the 1980s, made it difficult to determine the exchange rate effectively faced by entities engaged in foreign trade. Indeed, in the late 1980s the variation of retention rates between industries and regions led to an infinite number of effective exchange rates 1/ between the administered and the FEACs’ exchange rate. The narrowing of the difference between the administered and FEAC rates since the December 1989 and November 1990 devaluations (27 and 11 percent respectively), and the more uniform retention rates established in 1991 reduced the problem, but trade duties and quantitative controls still play an important role in determining the relevant price to the trading entities.

From 1981 until 1984, China had dual exchange rates. The official rate depreciated gradually while the secondary rate was fixed at a more depreciated rate. The secondary rate, termed the internal settlement rate, was used for settlement of payments between the FTCs and the enterprises.

The internal settlement rate was abolished in 1985 but in 1986, with the establishment of the FEACs, dual exchange rates prevailed again in China. Between July 1986 and December 1989 the administered (official) exchange rate was fixed at 3.72 yuan per U.S. dollar, leading to a real appreciation of the yuan in the face of rising inflation. 2/ At the same time, the exchange rate in the FEACs depreciated, reflecting the excess demand for foreign currency at the official rate. For exporting enterprises, the ability to sell retention quotas in the FEACs allowed them to maintain the profitability of exports. The devaluations in December 1989 and November 1990 returned the real exchange rate to the level prevailing in 1986. Beginning in April 1991, the administered rate was adjusted more frequently through small periodic adjustments.

d. Summary

Despite the substantial reforms since 1978 the Chinese economy is still characterized by much government intervention, both in the external and the domestic sectors of the economy. Decentralization of decision making stopped, in many cases, at the local government level and was not passed on to the enterprises. The enterprises remained under state ownership and this raises doubts whether managers seek maximization of growth and profits, or want to maximize slack to ensure that targets can easily be met. The results presented in Section IV.4 provide support for the latter hypothesis.

In the external sector minimum targets still prevail for some enterprises and this, in combination with the ownership structure, may lead to suboptimal behavior. The effects of the contract system are explored in Chapter IV. However, the reforms in the external sector were substantial and the increased liberalization facilitated the fast growth of China’s foreign trade during the 1980s. Although several partial reversals of the process took place in periods of balance of payments difficulty, China’s trade in 1991 is much more liberalized than in 1978.

III. Developments in China’s Trade From 1980 to 1991

1. Some general observations

China’s international trade grew rapidly during the 1980s. The dollar value of total trade increased between 1980 and 1990 at an annual rate of 12 percent while the value of world trade increased by 6 percent annually. This led to an increase in China’s share of world exports from less than 1 percent in 1980 to 1.9 percent in 1990. The growth rate of exports was similar to those of the fast-growing economies of Hong Kong, Korea, Singapore, and Taiwan Province of China. While the absolute figures for China’s share in world trade are still small, the dynamics are an indication of its growing importance in world markets.

Chapter II of this paper described the reforms that took place in China’s foreign trade during the period. Policy changes and liberalization of trade were not a continuous process but occurred in discrete steps. The effect of these steps on trade is evident from the import data in Table 2. Following reform measures in 1984 and 1988, one observes increases of 64 percent and 61 percent, respectively, in the value of imports. 1/ In the periods between the adoption of new policy measures, imports grew at a very slow rate. 1/ The fast growth of imports in reaction to policy measures may be an indication of constrained demand. Whenever the restrictions were eased, the demand for imports surfaced immediately and filled the new “quota”. Until a new reform is introduced, all the allowed value is imported and then we see a new “jump” when new reforms are introduced. This possibility is supported by the estimated price elasticities of imports in Chapter IV.

The effects of policy measures on China’s trade were also demonstrated during 1988-91. After the imposition of restraints on imports, together with the austerity plan in 1988 and the devaluation in December 1989, imports declined by 15 percent in the first three quarters of 1990 compared to the same period a year before. Exports continued to grow at an annual rate of 14 percent. These trends continued early in 1991.

2. Unit values

This paper uses, for the first time, unit value indices to analyze China’s foreign trade. The use of unit values allows price and quantity developments in China’s foreign trade to be separately identified. The unit value data thus facilitate a better understanding of the sources of growth in China’s trade, the sources of trade deficits, and the performance of Chinese exports in world markets. Some of these issues were discussed in the past 2/ but the availability of unit value data offers new dimensions to the analysis.

a. Terms of trade developments

China’s terms of trade index (Table 3) shows a deterioration of 30 percent during the 1981-90 period. Most of the decline, 23 percentage points, took place in 1986 as a result of the drop in world oil prices and an increase in import prices. However, excluding petroleum prices, the data show a decline of only 13 percent in the terms of trade during 1981-90. This decrease took place after 1985 and was associated with a sharp increase in import prices. The terms of trade of non-oil exporting developing countries improved by 5 percent during 1981-88. China’s terms of trade, excluding petroleum exports, decreased by 2 percent during the same period owing to the much smaller decrease in China’s import prices relative to these countries. The effect of the decline in oil prices on the import unit value index (UVI) of non-oil exporting developing countries can explain the difference. The increase in China’s import UVI during 1985-90 is also in line with the increase in export prices of the industrial economies, which are the sources of most of China’s imports.

Table 3.

Terms of Trade Developments 1981-901/

(1981=100)

article image
Sources: Calculated from data in General Administration of Customs, various issues; and from IMF, International Financial Statistics.

Yearly indices were calculated as the average of the quarterly indices.

Calculated using the official exchange rate and the retail price index.

Since the indices were calculated using sample weights 1/ the effects of using the weights of each category in total trade should also be mentioned. Adjusting the weight of petroleum products in the export data 2/ turns out to be the most important factor and increases the terms of trade index in 1990 from 70 to 75. However, in the most significantly under-sampled category, “Machinery and Electronics,” it seems, based on the data about industrial countries’ trade, that the sample commodities underestimate the price increases in this category and therefore China’s terms of trade index, as presented in Table 3, is biased upward. 3/

Unit value data, for export and import commodities, appear in Tables 4 and 6, respectively, and are used in the following subsection to focus on the commodity categories that contributed to the changes in the aggregate indices. Particularly, the export analysis investigates, using also the data in Table 5, whether there had been an upgrading of China’s export products and an improvement in the perceived quality of Chinese goods in world markets. This should be reflected in increased prices for China’s industrial exports relative to the prices of other products in the same markets.

Table 4.

China: Export UVI of Selected Commodity Categories 1981-90

(1981 =100)

article image
Source: Calculations based on data in General Administration of Customs, various issues.

1985=100.

Table 5.

China: Performance in the U.S. Textile and Apparel Import Market, 1979-88

article image
Source: Calculated from Peltzman (1990).

China, Hong Kong, Korea, and Taiwan Province of China.

Hong Kong, Korea, and Taiwan Province of China.

Table 6.

China: Dollar-Based Import Unit Value Indices of Selected Commodity Categories, 1981-90

(1981 = 100)

article image
Source: Calculations based on data in General Administration of Customs, various issues.

1985 = 100.

1986 = 100.

b. Export and import unit values by commodity category

Table 4 shows opposing trends in the export prices of Chinese primary and manufactured products. While the prices of primary products declined during 1981-90, the prices of manufactured products increased. This was a general phenomenon in world markets, that limited the ability to increase foreign exchange earnings by expanding primary exports.

The major contributors to the positive effects on China’s export UVI were the textile and apparel industries, especially since 1985. The UVI of these industries increased by 35 percent between 1981 and 1990 and their significance in China’s trade makes it important to find out whether the increase in the unit values of these commodities during 1981-90 was part of a general phenomenon in world markets or specific to Chinese exports.

While a broad analysis of this question is beyond the scope of this paper, some insights can be gathered from data on the U.S. textile and apparel market reported by Peltzman (1990). Since the U.S. market is an important destination for China’s garment exports, an analysis of the developments in this market can help to evaluate the change in China’s unit values.

Between 1982 and 1988 China enjoyed an increase of 18 percent in the unit value of its products in the U.S. textile and apparel market (Table 5). 1/ However, this increase was due to the general price increase in the market. Unlike the other major exporters in this market, China did not experience an increase in the relative unit value of its products. This suggests that there was no perception of improved relative quality of Chinese products in this market and that China had not diversified its exports to higher-priced commodities. A shift to higher-quality and higher-priced products would be the key to continued export growth in the future because many of the quotas in the market are in volume terms. 2/

The increase in China’s import unit values during 1981-90 (Table 6) was a result of a rise in the prices of industrial products in the “Machinery and electronics” and “Goods classified by material” categories. However, during 1986-90, the rise in import unit values was due to increases in the prices of primary products and “Goods classified by material” while the prices of “machinery and electronic products” decreased. 3/

3. Volume of trade

China’s export volume increased during 1981-90 at a higher rate than import volume (Table 7). 4/ The current account deficits between 1987 and 1989 also reflected the deterioration in the terms of trade after 1983, as China’s export volume grew faster than that of imports between 1983 and 1987. 5/ The comparison with world trade volume indicates that China’s increasing share in world trade resulted from faster quantity increases in its trade relative to world trade.

Table 7.

China: Trade Volume Indices, 1981-90

(1981 = 100)

article image
Sources: Calculated from data in General Administration of Customs, various issues; and IMF, International Financial Statistics.

The developments in trade volume (Tables 8 and 9) reflected the sequence of the reforms in the Chinese economy. When the rural sector was reformed during the early 1980s, its production increased (see Appendix Table 15), and exports of agricultural products rose rapidly while imports of these products declined.

Table 8.

China: Export Volume Indices, 1981-90

(1981=100)

article image
Source: Calculations based on data in General Administration of Customs, various issues.

1985 = 100.

Table 9.

China: Import Volume Indices, 1981-90

(1981=100)

article image
Source: Calculations based on data in General Administration of Customs, various issues.

1986 = 100.

1985 = 100.

The focus of reforms shifted to the urban sector by the mid-1980s. At the same time, the growth rate of exports of industrial products increased and was higher than that of primary products. The increase in industrial production (see World Bank 1990) led to higher incomes in the urban sector and, as a result, domestic demand for food rose. This led to a decrease in cereal exports after 1985 and an increase in imports. The production of other agricultural products continued to grow after 1985, facilitating increases both in exports and domestic consumption.

An important process that can be inferred from the data on quantities is that of import substitution. The data in Appendix Table 15 show a decline in the market share of imported durable consumer goods and steel products. In some cases (e.g. television sets and sound recorders), China became a net exporter of products that were imported on a large scale earlier in the decade.

China’s export growth was based on exporting a small proportion of production in many sectors and the development of a few sectors that export a large proportion of their output. The latter is true especially for the textile and apparel industries and for some consumer durable goods. In addition, the processing industries operate mostly for export.

IV. Price and Income Elasticities

In this section, estimated price elasticities are derived for imports and exports and estimated income elasticities are derived for imports. This may help to assess the responsiveness of China’s trade to the different variables and the extent to which market-oriented behavior has developed in China.

Interpreting the estimated elasticities is not an easy task without a complete structural model. Since the data available for this paper were not sufficient to estimate structural equations, only simplified equations for China’s trade were estimated. 1/ Therefore, a discussion of the interpretation of the coefficients is required before presenting the results.

1. Price elasticities

The estimation of export and import price elasticities using reduced form equations is correct only if we accept the “small country assumption” for China. This implies that the demand price elasticity for exports and the supply price elasticity for imports are infinite, and therefore that all the effects of prices on trade volume that one observes are supply effects in the export equations and demand effects in the import equations. Since China’s share in total world trade is still small and this is also true for most of the individual commodities, 1/ it seems reasonable to adopt the small country assumption as a preliminary working hypothesis. Attempts to test this hypothesis are made in section 4.b., but it will be useful to test this assumption in the future along the lines of Bond (1987) and Goldstein and Khan (1978, 1984) using a more complete model.

The interpretation of the import elasticities requires caution. It is not clear to what extent the estimation reflects the demand for imports, and to what extent it reflects the willingness of the Government to allow imports. Some attempts are made to explore the authorities’ reaction to foreign currency constraints but only a complete structural model can separate the government and private responses.

Saracoglu and Zaidi (1986) estimated a system for developing countries 2/ that includes the government “supply” of imports by calculating the probability of restrictions. In China’s case, we know that restrictions exist; hence, a similar method should be used to estimate the probability of liberalization in each period. 3/ Since we have information on the dates of the reforms, slope dummy variables 4/ are used to test if there are any changes in the price elasticities after the reforms. Some other variables that may affect government policy are also included in the regressions, but a more complete model should be tried, when more data become available, in order to separate the components of the import elasticities.

The export price elasticities may overestimate the responsiveness of FTCs and enterprises to price changes. The assignment of overall export value targets to local governments 5/ allowed them to shift export targets to enterprises whose products enjoyed an increase in world prices. The extent to which local governments used this ability increased the estimated export price elasticity even if there has been no change in the price responsiveness of FTCs and enterprises.

Another issue is the choice of the price variable. As discussed in Chapter II.c, the relevant prices for the trading entities in China may differ from the dollar unit values and even from the real yuan prices in Table 3. 1/ However, while transparency of world prices in China is far from perfect, the use of retention quotas for FTCs, local governments, and enterprises that are set as a percentage of total revenue creates a positive correlation between world and domestic prices. 2/ The effects of retention quotas may be especially important because they were very high at the margin (i.e., for exports above the target). 3/

An additional factor that may affect the results follows from the way in which the UVI is calculated. Dividing the trade value by the quantity for each commodity creates a negative correlation between the volume index and the UVI if there is a measurement error in the data on quantities. 4/ A possible solution for this problem is to use world prices of the commodities as substitutes for the calculated unit values. However, since products, especially in the manufacturing sector, are not homogeneous, this method is unreliable. The comparison in Chapter III between the estimated UVI for China’s textile and apparel exports and the prices reported by Peltzman (1990) shows that the two sets of prices do not differ significantly, at least for these commodities.

2. Income elasticities

The income elasticities for imports, like the price elasticities, have two components. One is the income elasticity of the economic entities’ demand, and the other is the willingness of the government to allow imports. As mentioned above, only a system of structural equations would permit the separate identification of these effects.

The effect of domestic production in the export equations can be positive or negative. As Bond (1987) has shown, the sign may be positive because an increase in production represents a greater ability to export, or negative if it results from an exogenous increase in domestic demand for exported or non tradable products. In the single commodity equations it may be more likely that the coefficient will reflect income effects unless production of the commodity is highly correlated with the total size of production.

3. Estimation

Regression equations were estimated in order to obtain the required elasticities. For the export elasticities, two sets of equations were used: including and excluding petroleum.

The dependent variable in the export equations was the log of the seasonally adjusted volume index. The right-hand side variables (in logs) were the comparable real unit value index, a seasonally adjusted index of real industrial production, 1/ and an index of world imports as a proxy for world demand. Slope dummy variables were used to test for changes in the price elasticity after the reforms in 1985 and 1988 (as discussed in Chapter II). In addition, lagged values of the price variables were included in order to test for the length of the adjustment process.

In the import equations, two different dependent variables were used to explore the possible effects of sampling bias on the results. These variables were (1) the log of seasonally adjusted imports for the UVI sample commodities and (2) the log of the seasonally adjusted index of total import volume. 2/ The right-hand side variables were the log of the UVI, log of the index of real industrial output, and slope dummy variables to test for changes in the price elasticity since 1985 and 1988. The import coverage of reserves (in quarters), the level of export value, and the terms of trade in current values and with lags were also included. The latter variables were introduced as possible components that affect government restrictions on imports. A complete list of the variables and their definitions appears in Annex III. 3/ However, in the future, it may be useful to experiment with other variables that may affect the government’s policies.

The regressions were estimated for 40 quarters between the first quarter of 1981 and the last quarter of 1990. In order to correct the observed serial correlation in the data, corrections for moving average and autoregressive processes were implemented using the Cochrane-Orcutt (1949) method. 4/

4. Results

a. Imports

The equations for the two different import volume indices are reported in Table 11 and reveal different estimates for the income elasticity-- an expected result owing to the different commodity content of the indices.

Table 10.

China: Import Regressions (OLS)

article image

For complete definitions of variables see Annex III.

* - Significant at the 10 percent level. ** - Significant at the 5 percent level.

Absolute values of t statistics in parentheses. All the regressions are corrected for moving average serial correlation. The degrees of freedom are adjusted to account for seasonal adjustments of the data.

Variables List: TOIVOS: Total import volume; IVOSES: Import volume for UVI sample; SINOUT: Industrial output; IRDUP: Import prices in real dollars; RDUP85: Slope dummy for IRDUP since 1985; RESQUA: Value of foreign exchange reserves in terms of quarters of imports; IUP: Import prices in dollars; DUP85, DUP88: slope dummies for IUP since 1985 and 1988 respectively; IRYUPI: Import prices in real yuans; DYUP85: Slope dummy for IRYUPI since 1985; IEYUPI: Imports prices in real yuan using the secondary exchange rate until 1984; TERMTR: China’s terms of trade, calculated by dividing imports prices by exports prices.

Numbers in parentheses after variable names indicate lagged values of the variable.

Table 11.

China: Export Regressions (OLS): Dependent Variable: Seasonally Adjusted Export Volume Excluding Petroleum (SEVNOP) 1/, 2/

article image

For complete definitions of variables, see Annex III.

* - Significant at the 10 percent level. ** - Significant at the 5 percent level.

Absolute values of t statistics in parentheses. All the regressions are corrected for moving average serial correlation. The degrees of freedom are adjusted to account for seasonal adjustments of the data.

Variables List: SEVNOP: Exports volume excluding petroleum; SINOUT: Industrial output; ERDUPI: Exports prices in real dollars; SWORTR: Value of world imports, RDUP85; RDUP88: Slope dummy for ERDUPI since 1985 and 1988 respectively; RYUPNP: Exports prices in real yuans; RYUP85: Slope dummy for RYUPNP since 1985.

Numbers in parentheses after variable names indicate lagged values of the variables.

The first volume index includes only the UVI sample commodities. This sample excludes most of the advanced technology imports into China and represents the more basic commodities. The income elasticity for these commodities is 0.30 when we use real dollar unit values in the equation and increases to 0.47 when we use real yuan prices. 1/ In other specifications the value increased to as much as 0.71 but statistically, in all cases, the elasticity is significantly lower than 1. When one uses dollar prices, the inclusion of the last period’s terms of trade or lags of the foreign exchange reserves drives the value of the estimated coefficients close to zero. 2/

This suggests that the demand for standardized goods did not increase as fast as income. This may be due to the substitution of local production for imports of these goods, as shown in Appendix Table 15, and the increase in demand for other goods; that is, technologically more advanced goods. The negative values of income elasticity estimated for many commodities in Appendix Table 17 also suggest that import substitution was an important factor in China’s trade.

The income elasticity for total imports is much larger than that estimated for the UVI sample. The elasticity estimates ranged from 1.22 to 1.47 1/ (Table 10). Thus, the income elasticity for total imports shows a tendency for the share of imported goods in domestic expenditure to increase with income. This increase applies mostly to imports of mechanical and electronic goods other than consumer goods. For most of these products the industrial production index reflects only income effects because there are no domestically produced substitutes, a factor that creates the large difference in income elasticities between total imports and the UVI sample.

Attempts to estimate the effect of foreign exchange availability on the Government’s tendency to allow increased imports yielded significant results only with regard to imports of goods in the UVI sample. Those imports depend positively on the value of foreign reserves with a lag of two quarters, and the terms of trade with a lag of one quarter. No effects were found on the volume of total imports or even on the size of the income elasticity in their equations. The coefficients were insignificantly different from zero or had the wrong signs. While more work is required in order to estimate the Government’s reaction function, these results may indicate that the Government’s determination of the volume of technology imports is based primarily on its perception of domestic needs, while the imports of other inputs and consumer goods are determined also in conjunction with balance of payments considerations. Indeed, most of the controls over trade during the 1980s were on imports of consumer goods and basic inputs.

The estimated price elasticities of imports are significantly different from zero and very close to -1. The price elasticities for the UVI sample goods were between -0.79 and -1.02 regardless of the price variables we chose. Statistically, in all cases, the price elasticity estimator was insignificantly different from -1. A price elasticity of -1 reflects fixed-budget importers who have to adjust the volume of their purchases exactly in an opposite magnitude to changes in the price. This is consistent with the trade system in China where import quotas are allocated to local governments and enterprises. The price elasticity of those products became significantly smaller (in absolute values) after the liberalization in 1985, reflecting more discretion for the importing entities in determining the value of trade.

For total imports, the dollar price elasticity is also very close to -1 and became more negative after 1985. The yuan price elasticity is smaller (between -0.36 and -0.60). The reason may be that for technology imports, which constitute most of the difference between the two volume indices, there are no local substitutes and therefore their import volume is not affected by the yuan price. The difference between the dollar price elasticities of the two volume indices is statistically insignificant.

Price and income elasticities for imported commodities are reported in Appendix Table 17. Equations that also include slope dummies for changes in the price elasticity since 1985 and/or the effect of foreign exchange reserves on imports are reported only if the inclusion of these variables reduced the standard error of the regression.

b. Exports

The export equations are reported in Table 11 and reveal the strong reaction of export volume to output. The coefficients, including or excluding petroleum, 1/ are between 0.62 and 0.89. Without seasonal adjustment or when the value of world trade is excluded from the equations, the elasticity is significantly larger than 1. 1/ As production grows the tendency to export increases. One of the reasons may be that exports react to increases in the demand for imports as income increases. The increasing demand for imports requires more exports in order to finance the increased foreign currency expenditure. Future estimation of a complete macroeconomic model of the Chinese economy can help to sort out these effects. 2/

The volume of China’s exports is also responsive to the size of world trade. As world imports expand, China increases its exports. The elasticity estimates ranged from 0.62 to 1.04 and statistically are always significantly different from zero. In most specifications the coefficient is smaller than 1. 3/ This may indicate that the direct reaction of China’s exports to world demand would have led to a decreasing or constant share of China in world trade. Only the additional effect of growth in production caused the increase in China’s share.

The price elasticity of total export volume is not significantly different from zero. For export volume excluding petroleum, we observe a price elasticity of about -0.3 (Table 11). There had been a small but statistically significant change in this elasticity in 1985, and again in 1988 as reflected by the slope dummies coefficients. The negative supply price elasticity may mean that the phenomenon of export value quotas is significant and strongly affects the responsiveness of China’s exports to their unit values. As unit values decrease, exporters are required to export larger quantities in order to achieve their value quota. As unit values increase, the trading entities have more slack and they can reduce their efforts to export larger volumes. As there had been a shift toward more value targets, rather than volume targets, the elasticity became more negative after 1985. The increase in retention quotas in 1988 may be the cause for the correction in the opposite direction in 1988-90.

The positive lagged price elasticity during 1985-90 may suggest a mechanism in which enterprises and FTCs reduce their export volume as prices increase to avoid larger value targets in the future (the ratchet effect). 4/ When the authorities (usually the local governments) observe the price increase they indeed raise the export target, and this causes the lagged price elasticity to be positive. 1/ This may be a result of the separation between management and ownership. Managers received more control over operations during the 1980s, but they have little financial incentive to increase profits beyond the contract level and may, therefore, prefer to minimize the risk of failure to meet contract targets.

Another possibility is that for China’s exports the “small country” assumption is incorrect and the observed price elasticity is dominated by the demand price elasticity. To correct this problem, two-stage least squares (TSLS) models for China’s exports were estimated. In this method only price changes that result from demand variables are included in the second stage equation and therefore, the estimated price elasticity reflects only the supply reaction. 2/ The estimation of the TSLS equations also yields negative supply price elasticities, although the equation with dollar prices was very inefficient (Table 12).

Table 12.

Two-Stage Least Squares Dependent Variable: Seasonally Adjusted Export Volume

article image
Source: Organization for Economic Cooperation and Development (OECD), (1991).Note: The price variables are instrumented by consumption in the OECD countries. The numbers in parentheses are t values.

Another explanation for the negative export price elasticities may be that even if China is a “small economy” its supply curve for exports may shift simultaneously with that of other countries because of changes in input prices. Controlling for this possibility in addition to the TSLS process is difficult because it requires a knowledge of the inputs that are used for the exported commodities. However, oil and cotton are two important inputs in China’s production and therefore their world prices were used to control for shifts in world supply. This procedure maintained the negative supply price elasticities and increased their t values.

In order to control for the possibility that the negative supply price elasticity results from the effect of a few commodities that have large market shares in their export markets, or from errors in the calculation of the aggregate unit value index, separate equations were estimated for each commodity that had sufficient data. The results are presented in Appendix Table 16 and discussed in the next section.

c. Export price elasticities by commodity

Estimation of the price elasticities reveals that most of the commodities have negative export price elasticities. This occurs even for commodities for which China’s world market share is negligible. Only in the case of a few agricultural products is the export price elasticity positive. This supports the assumption that the observed export price elasticities are indeed supply elasticities and supports the notion that the negative export price elasticity is a result of the quota system. In addition, since aggregate export value targets are assigned to local governments which have to allocate them among the products they export, this is a test for efficient substitution between commodities at the local government level. The results indicate that such substitution effects are not very strong in China.

The estimated values for the price elasticity are usually negative and in most cases significantly different from zero. In some cases, we cannot reject the hypothesis that the elasticity is -1, but the elasticities are generally between zero and -1. In the garment industry, the elasticities became significantly negative only after the reforms in 1984-85. The result is that, for most commodities, the value of exports is positively correlated with unit values. As unit values decrease, some increase in export volume takes place in order to reach the targeted quota, but export supply does not adjust fully to the price change. As unit values increase, some of the increase is used to reduce the export effort. This is true at the commodity level as well as at the national level.

V. Conclusion

The developments In China’s external sector during 1981-90 were examined in this paper in light of the reforms in the external and internal sectors of the economy. Close links were found between the implementation of different stages in the reforms and trade developments. The institutional environment seemed to have an important effect on the enterprises and corporations involved in China’s trade.

China’s trade grew rapidly during the decade--faster than world trade, and also faster than the rest of the economy. This rapid growth was achieved in spite of a decline in the terms of trade attributable to the decline in world oil prices and an increase in import prices. Export unit values (excluding petroleum) remained almost unchanged while import unit values increased by 12 percent between 1981 and 1990 (37 percent between 1984 and 1990). The export unit values of manufactures increased, but, at least for the textile and apparel industries, this increase is shown to be a result of general price increases in the market and not a result of higher relative prices for Chinese products.

Growth in China’s imports took place in large discrete steps following import liberalization, possibly reflecting excess demand. The analysis of import volumes also suggests that imports were used to release bottlenecks in the economy when the demand for certain commodities exceeded supply. The analysis of export and import price elasticities shows the strong effect of trade restrictions and the management system on the development of trade. It seems that many of the enterprises and FTCs involved in China’s trade are still oriented toward meeting administrative targets rather than exploiting opportunities in the market to increase profits. This results from the way in which local governments implement the diversification of microeconomic decisions and may offer important lessons for other countries engaged in efforts to decentralize decision making. While each country’s circumstances differ, the Chinese experience may suggest that privatization, rather than delegation of power to local governments, would be a more effective way to achieve market-oriented behavior and efficiency at the microeconomic level in formerly centrally planned economies.

The availability of quarterly data on China’s trade, including unit values, should facilitate future work in developing a more complete model of the country’s trade. However, development of a complete macroeconomic model that will be able to forecast trade developments would require information (on a quarterly basis) about macroeconomic policy variables, national accounts, and the factors that underlie the Government’s policy toward trade. The large degree of government involvement in the Chinese economy, even after the reforms, makes such an analysis an important component of any trade model for China.

Analysis at the industry and commodity level may also be useful especially for business planning purposes both in China and in its trade partners. In particular, it may be useful to relate production data by commodity to the trade data.

In order to further investigate the accuracy of China’s negative export price elasticities as estimated in this paper, more work is required using world unit value data instead of the prices calculated for China’s export commodities. The use of disaggregated data on trade of other developing economies can also be very useful. Further development of a theoretical framework to explain the existence of negative export supply price elasticities for the entities involved in China’s trade is needed, and is