Was China the first domino? Assessing links between China and other Asian economies

Journal of International Money and Finance 18 (1999) 515–535 www.elsevier.com/locate/jimonfin

Was China the first domino? Assessing links between China and other Asian economies John Fernalda,*, Hali Edisona, Prakash Lounganib a

Board of Governors of the Federal Reserve, Division of International Finance, Washington, DC 20551, USA b International Monetary Fund, Asia and Pacific Department, Washington, DC 20431, USA

Abstract Several commentators have suggested that competition from China—in particular its 1994 ‘devaluation’ and strong export performance during 1994–95—may have contributed to the Asian crisis of 1997–98. We provide evidence against this view. We show that the devaluation was not important in economic terms, and China’s strong export performance in 1994–95 was matched by other Asian economies. We also show that the period 1993 and 1996 was marked by relative stability in the export shares of China and other Asian economies in various geographical regions and industries. The paper concludes by outlining channels through which the Asian crisis could affect China.  1999 Elsevier Science Ltd. All rights reserved. JEL classification: F31 Keywords: Asian Financial crisis; China; Contagion; Currency crises; Devaluation; Trade linkages

1. Introduction What are the links between China and other Asian economies?1 At the onset of the Asian crisis, press reports and several prominent economists speculated that com* Corresponding author. Tel: ⫹ 1-202-452-3889; fax: ⫹ 1-202-452-6424; e-mail: [email protected] 1 The term ‘country’ as used in this paper does not in all cases refer to a territorial entity that is a state as understood by international law and practice; the term also covers some territorial entities that are not states, but for which statistical data are maintained and provided internationally on a separate and independent basis. 0261-5606/99/$ - see front matter  1999 Elsevier Science Ltd. All rights reserved. PII: S 0 2 6 1 - 5 6 0 6 ( 9 9 ) 0 0 0 1 9 - 4

516

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

petition from China may have contributed to the crisis; more recently, there has been intense speculation on how the crisis has, and will, affect China. Motivated by these speculations, this paper tackles two questions. First, how has China’s emergence as a major exporter affected the export growth of other Asian economies? Second, through what channels might the current Asian financial crisis propagate to China? We begin our assessment of the impact of China’s export growth on other Asian economies by considering, in Section 2 of the paper, the effect of China’s exchange rate adjustment in January 1994. Numerous commentators have asserted that China’s devaluation of its official exchange rate, and a consequent surge in China’s exports since 1994, was the first domino in the Asian crisis. We argue, however, that this devaluation was not important in economic terms. The devaluation of the official rate unified that rate with a parallel floating rate which did not depreciate. The effective nominal depreciation relative to the dollar was relatively small and, in addition, took place at a time of high inflation. Hence, China’s real exchange rate did not depreciate, but instead appreciated fairly steadily from mid-1993 until 1998. In Section 2, we also show that China’s strong aggregate export performance in the aftermath of the 1994 ‘devaluation’ was mirrored in equally strong export among other Asian economies. We analyse next, in Section 3 of the paper, disaggregated trade data by region and industry to see if the aggregate data on exports are masking somewhat more intense competition between China and other Asian countries in more narrowly defined markets. We find that over the period 1989 to 1993, China’s export shares did increase in many markets, largely at the expense of the Asian NIEs (Korea, Singapore and Taiwan).2 In contrast, the period 1993 to 1996 is characterized by relative stability in market shares across China, the NIEs, and the major ASEAN economies. Hence, China’s robust export performance in 1994 and 1995, and the alleged devaluation of 1994, did not translate into major gains in market share. We also present in Section 3 correlations between Chinese real export growth and those of other Asian economies. The raw correlations between the two are strongly positive, reflecting the impacts of common factors, such as growth in industrialized countries. However, after estimating a standard trade model to control for obvious common global factors, we still cannot get a negative relationship between Chinese export growth and export growth of other Asian economies. While this may reflect in part our inability to control completely for global factors that drive trade, it does suggest that trade competition with China does not leap out as a major factor reducing Asian export growth. Thus, it is difficult to sustain the claim that China’s export performance ‘forced’ other Asian economies into an eventual competitive depreciation. Having argued, in essence, that China was not the first domino to fall in the Asian crisis, Section 4 then assesses the risks that it will be one of the later dominos to fall. In particular, we explore three channels through which the Asian crisis can

2

We use the label ‘Taiwan’ to refer to ‘Taiwan Province of China’.

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

517

affect, and propagate through, China’s economy: (1) reduced exports; (2) reduced capital inflows; and (3) concerns about financial sector soundness. 2. Chinese ‘devaluation’ of 1994 Fig. 1 provides a useful starting point for our analysis, showing two frequently noted, but highly misleading, stylized facts. Fig. 1(a) shows China’s official exchange rate, and its precipitous devaluation at the beginning of 1994; Fig. 1(b) shows a surge in Chinese exports in 1994 and 1995. Based on these observations, several commentators have argued that the Asian currency crisis originated, in part, from China’s 1994 currency devaluation; China, according to these analysts, was the first domino to fall in Asia.3 The devaluation supposedly gave China a competitive advan-

Fig. 1.

(a) Official exchange rate. (b) China’s reported annual exports.

3 For example, Makin (1997) claimed that China ‘devalued its currency by 35%’, and that its ‘preemptive devaluation…was the first in a number of events leading to the acute problems in Asia that surfaced

518

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

tage relative to other Asian economies, leading to the surge in Chinese exports in 1994 and thereafter. Other Asian countries kept their exchange rates pegged to the US dollar, and hence did not react immediately to China’s competitive devaluation. However, over time the effects of the loss in competitiveness accumulated, contributing to growing current account balances and eventually a currency crisis. We suggest that the argument outlined above is based on a faulty interpretation of Fig. 1. In terms of the exchange rate, China’s effective rate did not depreciate much at the beginning of 1994. Before 1994, China had a dual foreign exchange market. Fig. 2(a) shows the two main exchange rates: some transactions took place at the official rate (the thin dotted line), while others took place at a floating rate (often called the ‘swap-market rate’, shown with the thick solid line; [China continues to limit convertibility on the capital account, so at least officially, most transactions are trade-related]). On January 1, 1994, China unified the dual rates at the market rate, which required devaluing its official rate by about a third against the US dollar—the value of the renminbi fell from about 17 cents to about 11 cents.4 Before unification, Chinese exporters received an effective marginal rate that was a weighted average of the official and floating rates. IMF studies estimate that before unification, about 80% of the weight was on the floating rate, and 20% on the official rate.5 Hence, despite the enormous devaluation apparent in the official exchange series in January 1994, the effective nominal depreciation relative to the dollar was considerably smaller, about 7%. Moreover, on a real effective basis, the depreciation rapidly reversed itself. The unification of the two markets took place at a time of rapid, and rising, inflation in China, shown by the thin solid line in Fig. 2(a). Consumer price inflation, for example, was 19% in 1993 and 26% in 1994. Fig. 2(b) shows China’s exportweighted real effective exchange rate. The January 1994 depreciation is barely a blip in the upward trend from mid-1993 to early 1998, during which time the real exchange rate appreciated more than 60%. To focus on the competitive effects of China’s exchange rate, it may be better to study the 1989–1993 period, when China’s

this year’. Other recent references that link China’s 1994 devaluation to the Asian crisis include Bergsten (1997), and articles in the Financial Times (September 17, 1997, p. 29), Barron’s (October 27, 1997, p. 17), The Washington Post (November 27, 1997, p. B11; and February 4, 1998, p. A18), the Economist (November 22, 1997, p. 41), Euromoney (December, 1997, p. 38), and The New York Times (February 3, 1998, p. A31). Huh and Kasa (1997) also discuss the possible role of a Chinese devaluation as trigger for an Asian competitive devaluation (although their story could apply even without a Chinese devaluation in 1994). 4 The terms ‘renminbi’ and ‘yuan’ are generally used interchangeably to refer to China’s currency— the renminbi (literally, ‘people’s money’) is the currency, and the yuan is the unit of account. This exchange rate is more commonly quoted in terms of renminbi/dollar, so the devaluation represented an increase from 5.8 to 8.7 renminbi/dollar. 5 See, for example, Khor (1993); Tseng et al. (1994), p. 5), and the extended discussion of the exchange system in Mehran et al. (1996, pp. 55–61). Khor, in particular, argues that the 80% figure appropriately measures the marginal exchange-rate incentives faced by an exporter, since it represents the proportion of foreign exchange earnings that an exporter was allowed to retain for his or her own use, including selling on the floating-rate market.

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

519

Fig. 2. (a) Exchange rates and inflation. (b) Real renminbi exchange rates (export-weighted, chain-linked rate with 29 trading partners, using consumer price indices).

real exchange rate did, in fact, depreciate steadily. The focus on 1994 gets the timing wrong.6 Nevertheless, even if the 1994 devaluation was largely illusory, it may be that China’s strong export performance in 1994 and 1995 has contributed to the present crisis, to the extent that it has curtailed the export growth of other Asian countries. For example, China’s strong export performance may reflect economic reforms

6 The relative unimportance of China’s 1994 devaluation is well known to China scholars, as confirmed in numerous informal discussions in 1994, and again in 1997/98. Since writing the first draft of this paper in December 1997, we became aware of Liu et al (1998), who make a similar argument about the unimportance of the 1994 devaluation. See also the discussion in Noland et al. (1998).

520

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

implemented in the first half of the 1990s, and rapid investment in export-oriented industries, rather than a depreciation of the exchange rate.7 Fig. 3, however, shows striking evidence against this view. Fig. 3(a) shows export growth to the world from China (defined to include ‘Hong Kong’)8 and from the

Fig. 3. Export growth from China and developing Asia. (a) Total exports. (b) Exports to industrial countries. Exports are measured using data on trading partner imports from IMF Direction of Trade Statistics.

7 For a thorough discussion of the reforms of the foreign trade system in China, see Cerra and DayalGulati (1999). 8 We use the label ‘Hong Kong’ to refer to ‘People’s Republic of China–Hong Kong Special Administrative Region’.

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

521

rest of developing Asia; Fig. 3(b) shows export growth to industrial economies. Fig. 3 (a and b) use trading partner statistics. Fernald (1999) argues that it makes economic sense to combine data for China and Hong Kong even in the period preceding formal unification, since many goods use Chinese labor and Hong Kong management and distribution skills. It makes statistical sense to use trading-partner statistics, to avoid double-counting Chinese and Hong Kong exports.9 What is striking is the similarity in export growth between China and other Asian economies, including in the period 1994 to 1996: both show high growth in 1994 and 1995, and a slowdown in 1996. This suggests that common factors, such as growth in developed economies, movements in the world price of key exports such as semiconductors, and movements in the yen–dollar rate, were probably more important determinants for Asian exports than was competition with China. Discussions of China’s export performance tend to emphasize factors peculiar to China, such as economic reform initiatives, rapid investment, or tax incentives. However, these discussions appear to miss the common shocks that hit all Asian economies in 1996. Fig. 3(b) shows exports to industrial countries. Again, these two series move fairly closely together, and hence clearly seem to be reacting to common shocks, such as demand by industrial countries. However, the differences are considerably more pronounced than in the Fig. 3(a). The greatest differences came in early 1990s, particularly between 1990 and 1993, when import growth by industrial countries from China grew nearly 10% faster than from the rest of developing Asia. In sum, Fig. 3 tells a simple but fairly compelling graphical story. China’s export growth has a strong positive correlation with export growth of rest of Asia. Although export competition may well exist, the positive correlation raises the hurdle for demonstrating that export-competition had a first-order effect.

3. Analysis of export-competition among Asian economies Fig. 3 showed the strong co-movement of China’s export growth and the export growth of other Asian economies. This co-movement suggests that during 1979–97 common factors, such as growth in industrial countries, drove the export performance of both China and other Asian countries. Nevertheless, the evidence presented in that figure may conceal competition among Asian countries in particular geographic markets or in particular industries. It may also be the case that once the influence of common factors is accounted for, the relationship between China’s export growth and export growth in other Asian countries is actually negative. This section investigates these possibilities. For the analysis in this section, the Asian countries considered here have been

9 See Arora and Kochhar (1995) for a comprehensive discussion of size and source discrepancies in bilateral trade statistics between China and its main industrial-country trading partners. See also Feenstra et al. (1998); Fung (1996); Lardy (1994).

522

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

classified into one of three groups: (1) China (China and HK); (2) the NIEs (Korea, Singapore, and Taiwan); and (3) the ASEAN-4 (Indonesia, Malaysia, Philippines and Thailand). 3.1. Competition in particular markets and industries We begin by examining the export performance of these three groups in different geographical regions and industries for evidence of ‘export competition’, defined as ‘shifts in market share’ across the three groups.10 In particular, we want to see if China’s market share has increased markedly in a particular region or industry. Note that by focusing on shares in particular markets we are stacking the deck in favor of the export-competition view. A country may have its share in a particular market decline without necessarily experiencing a decline in the level of its exports to that market. It may also be losing market share in one market but gaining it in another. Furthermore, some changes in shares may be deliberate. Several Asian economies have shifted production towards components that are then assembled in China for export. These shifts tend to increase China’s export shares and decrease the shares of other economies, without having an adverse effect on these other economies. These shifts were most pronounced in Hong Kong (whose statistics we have combined with China’s) and Taiwan, but to some extent affected the statistics of other Asian economies as well. 3.1.1. Competition in the US market Our analysis is based on three-digit industry level data published by the US Department of Commerce’s Bureau of Economic Analysis (BEA). The figures in Table 1 provide the ‘market shares’ of the three groups in the US market in 1989, 1993 and 1996–98. The total exports of these countries to the United States have been scaled up so that they add up to 100. As shown in column (1) below, in 1989 China and Hong Kong together accounted for about a quarter of total exports to the United States from the three groups. By 1993, China’s share had increased to a third. Mainland China alone nearly doubled its share of the US market, helped perhaps by the real depreciation of the renminbi over this period. The ASEAN-4 group also increased its market share, but by a smaller magnitude than the increase in Mainland China’s share. Correspondingly, the share of the NIEs fell from 59 to 44%. There is, therefore, some evidence of ‘competition’—shifts in market share—among the three groups over the period 1989 to 1993. By contrast, the period between 1993 and 1996 is one of virtual stability in market shares. The shares of China and ASEAN-4 inch up over this period at the expense of the NIEs. The evidence, therefore, does not point to Chinese export growth in

10 See Leamer and Stern (1970, Chapter 7) for a good discussion of export share analysis. The effects of export competition can be reflected not just in changes in export shares but also in export prices or profit margins. We intend to explore these other effects in future work.

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

523

Table 1 Export shares of selected Asian countries in the US market Country

China and HK China HK NIEs Korea Singapore Taiwan ASEAN-4 Indonesia Malaysia Philippines Thailand Total (China plus NIEs ⫹ ASEAN-4) Memo: Total in US$ (billions)

(1)

(2)

(3)

(4)

(5)

Memo: (6)

1989

1993

1996

1997

1998 Jan.–Aug.

1997 Jan.–Aug.

24 13 11 59 22 10 27 17 4 5 3 5 100

33 25 8 44 14 10 20 23 4 8 4 7 100

34 29 5 41 13 11 17 25 5 10 4 6 100

37 32 5 38 12 10 16 25 5 9 5 6 100

38 33 5 37 12 9 16 25 4 9 6 6 100

36 31 5 39 12 10 17 25 5 9 5 6 100

90

126

180

199

136

126

Note: Data are from Bureau of Economic Analysis, US Department of Commerce.

1994 and 1995 leading to a significant expansion of China’s market share in the US market. The Asian financial crisis, and the associated sharp real depreciations in the currencies of many Asian countries, has not so far led to any dramatic changes in market shares: the relative stability that characterized the period 1993 to 1996 has continued through August 1998. It may be argued that the evidence presented in Table 1, which was for an aggregate of all industries, masks changes in market shares in particular industries. Our analysis of data for the 48 industries that make up the aggregate shows that is not the case. In Tables 2 and 3 we show examples of our analysis for two key industries, industry 213 (computers, peripherals and semi-conductors) and industry 400 (apparel, footwear and household products). First consider the changes in industry 213 (Table 2). Here, the market share of China alone rose from essentially zero in 1989 to 5% in 1996; however, most of this increase appears to have come at the expense of Hong Kong. When the two are combined, their market share increases only slightly over the period. The share of ASEAN-4 increases somewhat more substantially, with a corresponding fall in the share of the NIEs. There is little evidence that the period 1993 to 1996 is characterized by greater shifts in market shares than the earlier period 1989 to 1993. In the period since the onset of the Asian financial crisis, both China and ASEAN-4 have continued to gain market share at the expense of the NIEs. The story in the case of industry 400 is a bit different (Table 3). Here, China does

524

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

Table 2 Export shares of selected Asian countries in the US market: data for industry 213 (computers, peripherals and semiconductors) Country

China and HK China HK NIEs Korea Singapore Taiwan ASEAN-4 Indonesia Malaysia Philippines Thailand Total

1989

1993

1996

1997

1998 Jan.–Aug.

Memo: 1997 Jan.–Aug.

7

7

8

10

11

10

0 7 72 21 31 20 21 0 12 4 5 100

3 4 68 16 29 23 25 0 15 4 6 100

5 3 65 18 28 19 27 1 15 6 5 100

7 3 60 16 24 20 29 1 15 5 8 100

9 2 57 14 23 20 32 1 15 6 10 100

7 3 61 16 25 20 29 1 15 5 8 100

Note: Data are from Bureau of Economic Analysis, US Department of Commerce.

Table 3 Export shares of selected Asian countries in the US market: data for industry 400 (apparel, footwear and household products) Country

China and HK China HK NIEs Korea Singapore Taiwan ASEAN-4 Indonesia Malaysia Philippines Thailand Total

1989

1993

1996

1997

1998 Jan.–Aug.

36

55

60

62

62

62

18 18 52 27 3 22 12 3 2 3 4 100

41 14 26 13 2 11 19 6 3 5 5 100

47 13 17 7 1 9 23 8 4 5 6 100

50 12 15 6 1 8 23 8 4 5 6 100

49 13 15 7 1 7 23 8 4 5 6 100

51 11 15 6 1 8 23 8 4 5 6 100

Note: Data are from Bureau of Economic Analysis, US Department of Commerce.

Memo: 1997 Jan.–Aug.

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

525

experience a big increase in market share between 1989 and 1996, from 36 to 60%. However, the bulk of this increase occurred over the period between 1989 and 1993. The share of the ASEAN-4 also increased over the period, with the change being more substantial in the earlier part of the period than the latter. Since 1996, there has been virtual constancy in market shares. In summary, our analysis of competition in the US market leads to four conclusions: 앫 Over the period 1989 to 1993, China did gain market share in many markets. 앫 In contrast, the period 1993 to 1996 is characterized by relative stability in market shares across the three groups. Hence, China’s robust export performance in 1994 and 1995 did not translate into major gains in market share. 앫 China’s gains have come largely at the expense of the NIEs rather than the ASEAN-4. 앫 The onset of the Asian financial crisis has not so far led to any dramatic changes in market shares across the three groups. 3.1.2. Competition in other markets We also considered the extent to which there were changes in market shares in Japan and important European markets. As shown in Table 4, China’s share of the Japanese market increased in the earlier period at the expense of the NIEs, while the ASEAN-4 share was essentially constant. In the latter period, Greater China’s share continues to increase, but the bulk of this comes from a decline in Indonesia’s share. In selected European markets shown in Table 5, the evidence once again suggests that: (1) there was greater export-competition in the period 1989 to 1993 Table 4 Export shares of selected Asian countries in the Japanese market Country

China China HK NIEs Korea Singapore Taiwan ASEAN-4 Indonesia Malaysia Philippines Thailand Total

(1)

(2)

(3)

1989

1993

1996

23 19 4 40 20 5 15 37 19 9 3 6 100

29 27 2 33 15 5 13 38 16 10 3 9 100

35 33 2 31 13 6 12 34 12 10 4 8 100

Note: Data are from the Direction of Trade Statistics, IMF.

526

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

Table 5 Export shares of selected Asian countries in European markets Country

China NIEs ASEAN-4 Total Country

China NIEs ASEAN-4 Total

France

Germany

1989

1993

1996

1989

1993

1996

31 46 23 100

38 36 26 100

39 35 26 100

36 42 22 100

40 37 23 100

40 35 25 100

Italy

United Kingdom

1989

1993

1996

1989

1993

1996

37 43 20 100

41 31 28 100

50 26 24 100

34 45 21 100

37 36 27 100

34 36 30 100

Note: Data are from the Direction of Trade Statistics, IMF.

than in the period since 1993; and (2) China’s gains have come at the expense of the NIEs rather than at the expense of the ASEAN-4. (The Italian market is an exception to this characterization of the evidence.) 3.2. Controlling for common factors In this sub-section, we go beyond the graphical evidence of Fig. 3 in two respects. First, while that figure was based on nominal export growth, the analysis here is based on real export growth. Second, and more important, we look at the relationship between Chinese real export growth and real export growth in other Asian economies after controlling for the effects of common factors. Our goal is to see if in contrast to the visual impression given by Fig. 3, the correlation between China’s export growth and export growth in other Asian countries is actually negative once we have controlled for the most important proximate determinants of Asian real export growth.11 Our basic finding is that the correlation either remains significantly positive or is reduced to zero; there is no evidence of the correlation turning significantly negative. To control for common factors, we estimate regressions of real export growth in a particular Asian economy on its proximate determinants, namely, the growth rate of foreign output and the (percentage) change in the country’s trade-weighted real exchange rate. (An increase in the real exchange rate indicates an appreciation of 11 In future work, we intend to augment the specification to include supply-side influences that are common across countries, such as coincident expansion in export capacity in these economies.

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

527

the Asian currency relative to that of its trading partners.) We add China’s real export growth as a regressor to these standard export equations. The data used in the estimation are annual, and extend from 1973 to 1996. To obtain sufficient degrees of freedom, we pool the data for the three NIEs, for ASEAN-4 members, and also for all seven economies. Country fixed effects are included in all regressions, but their estimates are not reported. The coefficient estimates are shown in Table 6; numbers in parentheses are standard errors. The first column presents the results of a regression of real export growth in the NIEs on (1) country fixed effects, (2) a lagged dependent variable12 and (3) China’s real export growth. As shown, the coefficient estimate on the last of these variables is positive, quite large in magnitude, 0.58, and significantly different from zero. In the second column, the growth rate of foreign output and the change in the real exchange rate Table 6 Conditional correlations between China’s real export growth and real export growth in other Asian countries Independent variable

China’s real exports Lag 1

NIEs (Korea, Singapore, Taiwan)

ASEAN-4 (Indonesia, Malaysia, Philippines, Thailand)

All seven countries (NIEs plus ASEAN-4)

(1)

(4)

(7)

0.58 (0.18)

Lag 2 Foreign demand Lag 1 Lag 2 Real exchange rate Lag 1 Lag 2 Lagged dependent variable Adj. R2

0.13 (0.12) 0.11

(2)

(3)

0.13 (0.17)

0.11 0.56 (0.25) (0.14) 0.00 (0.19) ⫺0.05 (0.22) 3.27 2.63 (0.67) (0.77) ⫺1.67 (0.82) ⫺0.58 (0.90) ⫺0.40 ⫺0.20 (0.20) (0.18) ⫺0.08 (0.15) ⫺0.04 (0.12) 0.04 0.10 0.09 (0.11) 0.43

(0.14) (0.09) 0.36 0.24

(5)

(6)

0.35 (0.14)

0.40 0.56 (0.15) (0.11) 0.22 (0.22) 0.004 (0.15) 1.45 1.46 (0.52) (0.74) ⫺0.18 (0.14) ⫺0.84 (0.61) ⫺0.11 ⫺0.21 (0.08) (0.09) 0.07 (0.08) 0.14 (0.12) 0.10 0.08 0.11 (0.08) 0.29

(0.11) (0.07) 0.30 0.21

(8)

(9)

0.23 (0.11)

0.26 (0.14) 0.12 (0.15) ⫺0.04 (0.12) 2.34 2.01 (0.45) (0.52) ⫺1.05 (0.57) ⫺0.67 (0.51) ⫺0.17 ⫺0.18 (0.08) (0.08) 0.01 (0.07) 0.06 (0.08) 0.07 0.10 (0.06) 0.36

(0.08) 0.35

Note: IFS and National income accounts data from country sources.

12 The estimates of the other coefficients are not very sensitive to whether or not a lagged dependent variable is included in the regression.

528

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

are included as regressors. The coefficient estimates of these two variables have the expected signs and, statistically, are significantly different from zero. For present purposes, the key result is that the coefficient on Chinese real export growth now drops to 0.11 (we cannot reject the null that the correlation is zero). Adding in lags of the independent variables, as is done in column (3), does not materially affect the conclusion that the coefficient estimate is essentially zero. A similar set of regressions for the ASEAN-4 group is presented in columns (4) to (6). In this case, the conditional correlations are always positive, large in magnitude, and significantly different from zero. When data for all seven economies are pooled, the conditional correlations generally lie in between the correlations for the two groups, as shown in columns (7) to (9). Since our sample period has been characterized by many structural changes (e.g. the ‘opening-up’ of China) and changes in exchange rate regimes, there may be concern that the parameters of the export equations may not be stable over time. While a full investigation of this issue is beyond the scope of the paper, we did reestimate the equations over two sub-sample periods, (i) 1973 to 1985 and (ii) 1986 to 1996. The conclusions were fairly similar to those for the full sample period.13

4. The impact of the Asian crisis on China How does the Asian crisis affect China? In this section we highlight the adverse effects of reduced exports and capital inflows on China’s growth; in particular, we argue that slowing capital inflows would make economic restructuring more difficult for China. We also discuss the short-term risks from China’s financial market weaknesses.14 As of February 1999, China had largely been spared the worst effects of the region’s financial crisis. The currency remained stable against the dollar and foreignexchange reserves increased. Though output growth slowed from an average annual growth of over 11% between 1991 and 1996 to under 8% in 1998, most observers attributed this to domestic rather than external economic developments. In part, the relative immunity may reflect the lack of convertibility of China’s currency, which makes a speculative attack difficult.15 In addition, China has a large current account surplus, large foreign-exchange reserves, a relatively low ratio of debt-to-GDP, and relatively little portfolio investment. Although these features tend to weaken some of the usual channels for contagion, several channels remain at work, including trade linkages, reassessments of fundamentals, and perceptions of risk. Hence, to the extent speculation simply accelerates adjustments to changes in fundamentals or perceptions, the lack of capital account convertibility suggests that China’s adjustments 13 The referee suggested that including China’s real exchange rate in the equations (instead of China’s real export growth) would be an alternate test of the export competition view. 14 Fernald and Babson (1999) examine in greater detail why China has survived the crisis so well and what risks remain. 15 See Edison and Reinhart (1999) for a discussion of the role of capital controls in stopping crises.

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

529

will be slower but that China will eventually feel the effects of the crisis. The primary mechanisms for these channels to affect China is through reduced exports, and reduced capital inflows. 4.1. The export channel About 40% of the exports of China and Hong Kong go to Asia, including Japan (see Fernald, 1999). Given the weakness in these economies, their demand for imports from China fell sharply in late 1997 and 1998. In addition to this income effect on the demand for Chinese exports, there is a relative price effect working through the real depreciations of Asian (especially ASEAN-4) currencies. The magnitude of the relative price effect is difficult to gauge. To the extent that China produces different goods than other Asian economies, it is less sensitive to the depreciations of other Asian economies. Furthermore, more than half of China’s export volume (56% in 1996)16 involves processing trade, in which China imports partially assembled goods, and completes labor-intensive assembly; the relative appreciation of the renminbi reduces the renminbi-cost of these imported components, partially offsetting the competitive effects of the Asian crisis. On the other hand, the Asian depreciations reduce the incentives to locate labor-intensive assembly in China, so at the margin export growth will slow. In addition, the changes in relative prices can also lead to a shift away from goods in which China has a comparative advantage towards goods where other Asian economies have a comparative advantage. 4.2. The capital inflows channel A likely reason for inflows to slow is that the Asian crisis has already caused investors to increase their China risk premium. Fig. 4(a) shows the widening of the yield spread between Chinese sovereign debt and US Treasuries (using a dollardenominated Chinese government bond due in February 2004) from an average of about 75 basis points until early September 1997 to over 200 basis points in late 1998. Even more striking is the offshore renminbi forward exchange rate, shown in Fig. 4(b).17 The forward rate remained relatively stable until late October 1997; when the Hong Kong dollar came under speculative pressure on October 24, the renminbi forward rate also spiked upwards. These rates imply an interest-rate differential (using covered interest parity) of more than 20 percentage points. These forward rates, although somewhat volatile, have remained considerably above the spot rate. The forward rate primarily reflects exchange-rate risk, and incorporates both expectations about the expected future spot rate, and a currency risk premium. The risk premium is likely to be particularly important, since the main participants in the 16

From Zhonguo Xinwen She (1997). There is no formal forward market for renminbi in China, but there is an offshore non-deliverableforward market, where all contracting and delivery is done in foreign currency, but based on the value of the renminbi. 17

530

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

Fig. 4. (a) International yield spreads (government bonds relative to US treasuries). (b) Non-deliverable forwards (rates from offshore forward market, where all transactions are settled in US dollars based on the value of the renminbi).

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

531

market are probably investors who have a large renminbi exposure over the next year, and are willing to pay a sizeable premium to insure against the possibility of a large devaluation. Whether it reflects a change in exchange rate expectations or a change in the risk premium, the forward rate suggests a substantial shift in perceptions by financial market participants.18 A higher risk premium is likely to show up in reduced capital inflows. Of course, more than 85% of China’s capital inflows take the form of foreign direct investment (FDI)19—amounting to US$38 billion in 1995, US$42 billion in 1996, and US$45 billion in 1997—which is usually presumed to be less subject to reversal than other forms of capital inflows.20 Nevertheless, even if FDI does not reverse (with investors dismantling a factory and shipping it to, say, Korea), the level of inflows can fall substantially. Indeed, the value of new contracts signed for FDI fell sharply in 1997 from 1996 (down 29%); although the relationship between China’s contracted FDI and actual future inflows is not clear, the figures already suggest a likely slowdown. The slowdown in contracted FDI appeared in the first half of 1997, even before the Asian crisis. The crisis in the second half of 1997 only makes things worse. Even apart from the evidence on contracted FDI, there are a number of reasons to expect the Asian crisis to contribute to a sharp decline in actual FDI. First, more than 80% of Chinese FDI comes from other Asian economies, and investors from many of these economies are in no position to continue rapid investment into China. For example, between 1992 and 1995, Hong Kong accounted for about 60% of FDI,21 Taiwan 10%, and Japan 7%. (The United States, by comparison, accounted for a little over 7%.)22 Second, even Western firms that undertake FDI in China may well 18

For comparison, the Consensus Economics (1999) survey of forecasters for January 11, 1999 shows an expected renminbi depreciation of about 1% over the next year; by contrast, if the forward rate were an unbiased estimate of the future spot rate, the January 11 one-year forward rate of about 8.9 renminbi/dollar would imply roughly a 7% depreciation. 19 Figures from ‘China: Joint-Stock System Development in Perspective’, FBIS-CHI-98-022, January 22, 1998. See also China Statistical Yearbook (1996), Table 16-15, p. 598. 20 Anecdotal evidence suggests that some of China’s FDI may be disguised lending. First, some FDI contracts specify fixed dividends that look a lot like interest payments. Second, Chinese firms sometimes circumvent the severe restrictions on foreign borrowing by establishing subsidiaries in Hong Kong, who borrow from international markets and then undertake FDI in China. Conceptually, both of these examples are closer to being debt than to being FDI. 21 In recent years, investors from other Asian economies have also been investing increasing amounts in China. In 1995, for example, the ASEAN-4 economies, plus Korea, accounted for about US$2 billion in FDI (5% of the total), and Singapore accounted for an additional US$2 billion. Data on the regional sources of FDI are from China Statistical Yearbook, various years. 22 The ultimate sources of financing for Hong Kong FDI into China remains unclear. One source is Chinese money, illegally diverted to Hong Kong and then repatriated in order to take advantage of foreign investment incentives. In addition, Chan and Chow (1997) report that in a sample of audited Chinese joint ventures, only about half of Hong Kong-incorporated enterprises were owned primarily by Hong Kong residents. Many multinationals invest in China through Hong Kong subsidiaries, and smaller Western investors often form joint ventures with Hong Kong firms in order to benefit from their Chinese expertise; many Taiwan residents invest in China through Hong Kong for political reasons. In addition, much of the ultimate funding for FDI may take the form of borrowing (e.g. from Japanese banks) by Hong Kong firms, who then undertake FDI. If this bank lending dries up, as seems likely, Hong Kong investment will also decline.

532

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

reduce investments. If firms invest as an export platform, other Asian economies now appear more competitive. If firms invest in order to get access to China’s domestic market, a possible slowdown in China makes that less enticing. Third, firms may perceive Asia generally as a less desirable place to invest and produce, consistent with a higher regional risk premium. One important effect of a slowdown of capital inflows is that it makes enterprise restructuring more difficult. In 1997, China announced plans to restructure stateowned enterprises (SOEs) and reduce their role in the economy. This major shift requires a massive reallocation of labor and capital, which in turn requires substantial investments in new firms. But FDI is a major source of financing for the non-state sector, so any decline in FDI is likely to make this reallocation more difficult. In other words, downsizing SOEs requires destroying existing jobs and laying off workers, a politically difficult step. This downsizing is easier if new jobs are being created, and the pace of job creation is likely to slow because of reductions in the financing available to the non-state sector. The main argument for continued enterprise and banking reforms is that without them, growth will slow over time and perhaps stagnate. However, radical reforms are costly in the short run, and could easily generate a sharp contraction. The substantial enterprise restructuring of the sort discussed following China’s 15th Party Congress would likely entail a short-run contraction.23 A concern is that Chinese leaders may find the political risks of a short-term contraction greater than the risks from a longterm slowdown in growth. Unemployment is already a concern of Chinese policymakers, and press commentary suggests concerns about social stability.24 4.3. Soundness of the financial system After Thailand’s currency crisis in July 1997, investors appear to have reassessed the risks associated with weak financial systems in Asian economies. Some commentators argue that China’s financial system is at least as weak, and perhaps far weaker,

23 There evidence on the extent to which restructuring is associated with output declines is not clearcut. Restructuring appears to have been enormously costly in terms of output in Eastern Europe and the former Soviet Union. For industrial countries, there is a substantial body of macroeconomic literature that explores the importance of sectoral shifts, suggesting that reallocations are costly in the short-term. The controversy in this literature is not whether reallocations/restructuring are costly in principle, but whether they are important in practice (they might not turn out to be a significant source of shocks to the United States, say). See Loungani (1996) for a survey of the literature on sectoral shifts. 24 The official urban unemployment rate was only 4% at the end of September 1997, but is misleadingly low for at least two reasons. First, it excludes workers who have been sent home on partial pay. Second, it includes only official urban residents (29% of the population); hence, it does not include the presumably substantial unemployment among the ‘floating population’ of generally unskilled rural workers who have moved to the cities in search of jobs. At least 30 million surplus rural workers (and according to some estimates, perhaps well over 100 million, or around 9% of the population) have moved to urban areas but do not have official residency status.

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

533

than those of other regional economies.25 Most notably, China’s banking system has a significant overhang of bad loans.26 In most Asian economies, policymakers’ post-crisis concern with banking-sector health reflects both short- and long-term concerns. In the short-term, poor bank health can lead to a ‘credit crunch’, as banks reduce lending even to viable non-financial firms, thus exacerbating the real effects of the crisis. This short-term concern is probably not as relevant in China as in the other Asian economies. In particular, it is unlikely that in order to restore their ‘profitability’, Chinese banks will be forced to cut back on other loans. First, it is fairly clear that the Chinese government continues to guarantee bank deposits, which are, after all, primarily held in state banks. Hence, depositors continue to have faith in the banking system. Second, if a severe credit crunch begins to impinge on the real economy, Chinese authorities can induce banks to lend. In other words, despite substantial moves in recent years, the banking system is not fully commercially driven and does not operate independent of the government. The problems of China’s banking system are primarily a fiscal issue, and as long as government guarantees are credible, the financial sector weaknesses are unlikely to lead to a short-term credit crunch. Nevertheless, even in the absence of an immediate credit-crunch, one short-term risk is that a slowdown in output and increases in bankruptcies might cause some of banks to become illiquid, when interest income cannot cover normal deposit withdrawals. Alternatively, if depositors decide to increase their withdrawals—perhaps because they lose faith in the banking system, or perhaps because they simply decide to reduce their (currently large) holdings of broad money-banks may find themselves illiquid. If the government is then forced to rescue the banks, the most accessible source of funding is the central bank. Then the government may face the undesirable choice of seeing an increase in inflation, or a substantial slowdown in growth. In the long-term, China’s financial market weaknesses raise severe concerns about whether resources are channelled to their most productive uses. Because of these long-term concerns, Chinese policymakers need to continue progress in limiting policy loans (through enterprise reform), providing banks with experience and skill in assessing loans on commercial grounds, and improving the transparency and accountability of banks. These are necessary steps before Chinese policymakers can successfully recapitalize the banks or otherwise try to solve the underlying problems of inherited bad loans of the banking system. It is noteworthy that in early 1998, policymakers announced numerous plans to accelerate financial market reforms, explicitly citing a desire to try avoiding an Asian-style crisis. Policies included central bank reforms, in order to improve local supervision, and bank reforms to improve the ability of banks to operate on commercial grounds, with greater independence from government-directed quotas. Nevertheless, as shown by the apparent reversals 25

See, for example, Cashmore (1997); Melloan (1997). In January 1998, for example, People’s Bank of China Governor Dai stated that more than 20% of bank loans are non-performing, although he argued that only 5–6% of loans are unrecoverable. Western observers generally estimate that the proportion of non-performing and unrecoverable loans may be far higher; see, for example, Roell (1997). 26

534

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

of reforms in late 1998 (with increased bank lending to state enterprises), it is clear that these steps are unlikely to proceed smoothly.

5. Conclusions We provide evidence against the view that China’s emergence as a trading nation and its 1994 ‘devaluation’ had a significant adverse effect on the export performance of other Asian economies. Between 1993 and 1996, China’s exchange rate appreciated, its export growth was strikingly similar to that of other Asian economies, and the composition of Asian exports (measured by export shares to various regions and industries) was relatively stable. To the extent there is evidence of export competition, it is during the period from about 1989 to 1993 when China’s exchange rate depreciated sharply, its export growth exceeded export growth of other Asian economies, and export shares changed substantially. We also discuss three channels for the propagation of the Asian crisis to China. First, China faces reduced demand for its exports because of the large real depreciations of many Asian currencies and the economic slowdown in the region. Second, China is also likely to face a slowdown in capital inflows, which will lead to reduced domestic investment and make state-enterprise reform more difficult by reducing growth and job-creation in the non-state sector. Third, the crisis focuses increased attention on the vulnerabilities of China’s financial sector. In sum, although China has remained insulated from the crisis so far, it remains susceptible in the future.

Acknowledgements We acknowledge useful comments from seminar participants at the International Monetary Fund, the October 1998 Journal of International Money and Finance conference on ‘Perspectives on the Asian Crisis’, the January 1998 American Economic Association meetings and the Federal Reserve Board. We thank, in particular, an anonymous referee, Lewis Alexander, Vivek Arora, Tom Dorsey, Caroline Freund, Dale Henderson, Steve Kamin, Marcus Noland, Steve Saeger, Nathan Sheets, Chris Towe, and Ted Truman for comments on earlier drafts. We thank Oliver Babson, Jennifer Cascone, Adam LaVier, Michael Sharkey, and Katherine Vanderhook for help with data and charts. The views in this paper are solely the responsibility of the authors and should not be interpreted as reflecting the views of the Federal Reserve System or the International Monetary Fund.

References Arora, V., Kochhar, K., 1995. Discrepancies in bilateral trade statistics: the case of China. IMF Paper on Policy Analysis and Assessment (PPAA/95/10), June.

J. Fernald et al. / Journal of International Money and Finance 18 (1999) 515–535

535

Bergsten, F., 1997. The Asian monetary crisis: proposed remedies. Prepared remarks to the US House of Representatives Committee on Banking and Financial Services, November 13. Cashmore, N., 1997. Work in progress. Asiamoney June, 51–55. Cerra, V., Dayal-Gulati, A., 1999. China’s trade flows: changing price sensitivities and the reform process. IMF Working Paper (WP/99/1), January. Chan, K., Chow, L., 1997. An empirical study of tax audits in China on international transfer pricing. Journal of Accounting and Economics 23, 83–112. China Statistical Yearbook, 1996. China Statistical Publishing House. Consensus Economics, 1999. Asia Pacific consensus forecasts. 11 January. Edison, H., Reinhart, C., 1999. Stopping hot money: on the use of capital controls during crises. Mimeo. Feenstra, R., Hai, W., Woo, W.T., Sachs, J. and Yao, S., 1998. The US–China bilateral trade balance: its size and determinants. Manuscript prepared for the UNDP-HIID Conference on China’s Integration into the Global Economy, January. Fernald, J., 1999. China and Hong Kong: reconciling bilateral trade statistics. Manuscript. Fernald, J., Babson, O., 1999. Why has China survived the crisis so well? What risks remain? International Finance Discussion Paper No. 633, Federal Reserve Board. Fung K., 1996. Accounting for Chinese trade: some national and regional considerations. NBER Working Paper No. 5595. Huh C., Kasa K., 1997. A dynamic model of export competition, policy coordination, and simultaneous currency collapse. Manuscript, Center for Pacific Basin Monetary and Economic Studies, Economic Research Department, Federal Reserve Bank of San Francisco WP PB97-08. Khor H.E., 1993. China’s foreign currency swap market. IMF Paper on Policy Analysis and Assessment (PPAA/94/1), December. Lardy N., 1994, China and the world economy. Institute for International Economics, Washington, DC. Leamer, E., Stern, R., 1970. Quantitative international economics. Allyn and Bacon, Boston. Liu, L., Noland, M., Robinson, S., Wang, Z., 1998. Asian competitive devaluations. Manuscript prepared for the UNDP-HIID Conference on China’s Integration into the Global Economy, January. Loungani, P., 1996. Macroeconomic effects of labor reallocation: a comment. Carnegie–Rochester Conference Series on Public Policy, vol. 44, June 1996, pp. 117–124. Makin J., 1997. Two new paradigms. American Enterprise Institute, October. Mehran, H., Quintyn, M., Nordman, T., Laurens, B., 1996. Monetary and exchange system reforms in China: an experiment in gradualism. IMF Occasional Paper 141. Melloan, G., 1997, The stakes are rising on China’s banking reforms. Wall Street Journal, 14 July, p. A15. Noland, M., Liu, L., Robinson, S., Wang, Z., 1998. Global economic effects of the Asian currency devaluations. Institute of International Economics. Roell, S., 1997. It’s your move, Mr Zhu. Euromoney February, 52–56. Tseng, W., Khor, H.E., Kochhar, K., Mihaljek, D., Burton, D., 1994. Economic reform in China: a new phase. International Monetary Fund Occasional Paper 114. Zhonguo Xinwen She, 1997. Foreign trade approaches $300 billion in 1996. 10 January. Reported as FBIS-CHI-97-007.