Catalog

The Uncoupling Myth: The G3 Slowdown and Developing Asia

The theme of uncoupling is one that has received considerable attention recently. It is certainly true that over a protracted period, rapid growth in developing Asia has had much more to do with Asian countries successfully exploiting opportunities for economic catch-up with richer industrial countries than with them hitching a ride on broader global growth (Figure 1.1.24). But it would be a mistake to assume that wider global developments do not matter for developing Asia (ADB 2007a, 2007b). Industrial-country growth exercises an influence on Asia, but the impact appears to be asymmetric (IMF 2007). When these countries are on the upswing of their business cycle, it is difficult to detect any impact on economic momentum in developing Asia, yet when their growth slows rapidly, Asia usually feels a downdraft. Box 1.1.2 presents one approach to quantifying these impacts using a global macroeconomic model. In the remainder of this section, ties between developing Asia and the G3 are viewed from a variety of other perspectives.

G3 non-oil import demand and Asia's exports

Previous issues of ADO have demonstrated the strong links between nonoil imports in the G3 economies and Asian exports (ADB 2007a). The high and rising correlations of Asian exports with non-oil imports in the G3 during the first few years of this century indicate that Asian exports are highly synchronized with G3 import demand (Figure 1.1.25).⁴

The links between US non-oil import demand and Asian exports are shown over a period of three decades. The relationship has become tighter in the most recent decade relative to the 1980s and 1990s (Figure 1.1.26).

In the case of the eurozone, the correlation has also strengthened in the current decade relative to the second half of the 1990s (Figure 1.1.27 below). The correlation is slightly less than in the US case but still quite robust, and indicates that Asian export activity has become more, not less, synchronized with external demand.

In the case of Japan (the largest national export market for some Asian economies such as Indonesia and the second largest for others), the correlation between non-oil imports and Asian exports has been positive over a longer period and has strengthened over time (Figure 1.1.28 below).

Therefore, even though intra-Asian trade has been expanding more rapidly than Asia's trade with the rest of the world, Asia has become ever more closely linked by globalization to the major global markets of the G3. This stems from the nature of Asian trade, with intra-Asian trade driven by vertically integrated Asian production chains and extra- Asian trade driven by G3 demand for the final goods produced in these networks (ADB 2007a).

The importance of the G3 markets for developing Asia is explored in the next section, with a focus on the linkage between the volume of retail sales in the G3 and the volume of Asian exports.

The uncoupling hypothesis: New evidence from G3 retail sales and Asian exports

Simple correlations between import growth in the G3 and Asia's export growth are revealing but other views of the data are possible. In this section, a statistical model is used to examine links between US, eurozone, and Japanese retail sales (as a proxy for demand) and export performance in nine Asian economies. Due to data limitations, only economies from East and Southeast Asia are included in the analysis.

The statistical model is estimated using monthly export data of the nine economies during 2000-2007, and retail sales of the G3. All data are suitably deflated and are measures of real economic activity. The approach chosen (Box 1.1.3 below) exploits the correlation between exports within Asia, as spillover effects along Asian supply chains are likely to be an important source of transmission of external shocks. Simple correlation analysis confirms that export performance across the nine Asian economies is highly synchronized. On average the correlation coefficient among them is around 0.8, with the highest correlations between the PRC; Hong Kong, China; Republic of Korea (hereafter Korea); Singapore; and Taipei,China.

1.1.3 The (unrestricted) vector auto-regression model Vector auto-regression (VAR) statistical methods are routinely used to capture the statistical relationships among a system of interrelated variables over time and to analyze the dynamic impact of random disturbances on the system of variables. The VAR approach sidesteps the need for structural modeling by treating every endogenous variable in the system as a function of the lagged values of all variables in the system. The standard representation of a VAR is as follows: (1) where e is a k vector of endogenous variables, X is a d vector of exogenous variables, A1, A2….., Ap and B are matrices of coefficients to be estimated, and ε is a vector of innovations that may be contemporaneously correlated but are uncorrelated with their own lagged values and uncorrelated with all of the right-hand side variables. In examining the impact of the G3 slowdown on exports in nine Asian economies, a VAR model is used because exports in these economies tend to be correlated. In particular, the emergence of product fragmentation and trade in parts and components has increased intraregional trade interdependence so that exports in the region have become highly correlated. Equation (1) can be rewritten in terms of exports and the G3 demand as follows: (2) where ex is the vector of (real) exports in nine Asian economies, US, EU, and Japan are the demand, represented by (real) retail sales, of the US, eurozone, and Japan, respectively. Since only lagged values of the endogenous variables appear on the right-hand side of the equations, simultaneity is not an issue and ordinary least square estimation yields consistent estimates. The Schwarz information criterion is used to select the lag length of the unrestricted VAR model. According to this criterion, the first lag of unrestricted VAR is chosen. In fact, there are other criteria for choosing the lag order of VAR models and these suggest third and seventh order lags. However, almost all variables become insignificant when the lagged values of all the endogenous variables are expanded to the third and seventh order. Diagnostic tests do not significantly improve relative to the first lag order of the VAR model. Meanwhile, all models (the first, third, and seventh lag orders) suggest the stability (stationary character) of the estimated VAR. Based on estimation results and diagnostic tests, the first lag model of VAR is preferred.

More sophisticated, model-based estimates of the strength of the relationship between retail sales in the G3 and Asian exports are presented in Table 1.1.2 below. The coefficients associated with G3 retail sales in all nine Asian economies are almost all positive, and most are statistically significant. For example, 3 months after a 1% rise in US retail volumes, the PRC's exports rise by 0.79%. For the eurozone, the lag is longer (5 months) and the impact smaller (0.35%). Though the estimates suggest that the PRC's exports respond to a rise in Japanese retail volumes after just 2 months, this particular estimate is not statistically reliable, and should be interpreted cautiously. The single largest response is for Indonesian exports to Japan, which 2 months after a 1% rise in retail volumes jump by 2.62%. Japan is Indonesia's single biggest export market, accounting for 21% of its total exports in 2002-2006.

These estimates suggest that the ongoing slowdown in the G3 could have a substantial cumulative effect on Asian exports if it were to persist. The Asian economies and export sectors likely to be hardest hit by the G3 demand slowdown vary depending on the share and composition of exports into the individual G3 markets and the severity of the slowdown in each of the G3 economies.

One advantage of the model is that by including indirect demands it allows estimation of the contagion (multiplier) effects across the nine Asian economies from the G3 demand slowdown (Figure 1.1.29).⁵ Inclusion of the induced impacts that travel down Asia's supply chains is likely to amplify the direct impacts shown in Table 1.1.2, and it is probable that these effects will depend on the strength of intraregional links through the PRC to the G3.

The model results suggest that the PRC is likely to suffer the most from falling consumer demand in the US. (The shares of the G3 economies in Asia's exports are shown in Figure 1.1.30.) In 2006, the latest year for which data are available, the US market accounted for more than 20% of the PRC's exports, up from 16% in 1995. For the other Asian economies, the US market accounted for about 15% of exports on average in 2006, down from an average of 21% in 1995. Shipments from the PRC and from Hong Kong, China are weighted more toward consumer products that are sensitive to consumer discretionary spending, including garments and miscellaneous manufactures, than are shipments from Korea, Malaysia, Philippines, Singapore, and Thailand which export machinery. Indonesia is less exposed to slowing US demand because it exports relatively greater amounts of food and raw materials than the other Asian economies.

The PRC also has the greatest exposure to a slowdown in the eurozone among the Asian economies, with about 18% of exports destined for that market. The eurozone is a significant destination for low-cost consumer goods exports from the PRC; Korea; and Taipei,China as well as from Malaysia, Singapore, and Thailand. There is some impact on the Philippines, but it dissipates quickly. Again, Indonesia's exposure to a slowdown in the eurozone is cushioned by the high proportion of food and raw materials in its exports. The smaller impacts stemming from the eurozone are probably a reflection of the fact that machinery exports dominate Asia's sales to the eurozone. This may make Asia less vulnerable to a short-term decrease in retail sales but if the slowdown persists, durable goods purchases may also fall and cut into Asian exports.

Indonesia is most exposed to a slowdown in the Japanese market. As it exports a large volume of primary products directly to Japan, indirect effects coming through a reduction in exports of third countries are unlikely to be important. For the PRC, the statistical estimate of the impact of Japanese retail volumes is not reliable. The PRC exports a large volume of garments, textiles, and footwear to Japan (but more so to the US) and these are highly cyclically sensitive product groups. But the PRC also exports a comparatively large volume of food to Japan, which tends to be less sensitive to income, and this may be influencing the net result. Backward linkages from the PRC to other Asian countries from its exports of garments and food are likely to be weaker than for other goods. Korea's exports are also sensitive to a decline in Japanese retail volumes, despite a modest share of Japan in Korea's total exports.

Evidence from cyclically sensitive manufactured exports

The aggregate analysis of imports and exports and the country breakdown within the G3 and East and Southeast Asia suggest the likelihood of significant impacts on Asian exports of a G3 slowdown. In this section, disaggregated trade data are examined to see if such effects can be traced in the recent evolution of shipments of manufactured goods. Specifically, the impact of weaker consumer demand in the US and Japan on key Asian manufactured exports is probed using quarterly data for the former and 6-monthly data for the latter. An attempt is also made to see if the weakness has spilled over into the EU and if this is also translating itself into more sluggish shipments of Asian products. The knock-on effects on commodity shipments through weakened exports between G3 members as a result largely of the US economy's weakness are also considered. South Asian as well as East Asian and Southeast Asian countries are covered in this part of the analysis.

United States: Evidence of consumer demand contraction and implications for Asian exports of manufactures

US consumer demand accounts for approximately 70% of the $14 trillion US economy and this represents a huge slice of world effective demand. The US consumer has in part fueled the Asian boom over the past 5 years by providing a strongly growing market for Asian manufactures. US retail sales are essential to growth in important labor-intensive manufactures as well as in more technology- and capital-intensive products. US import demand for garments is a bell-wether for how US consumer weakness is likely to impact developing Asia—particularly countries that rely heavily upon garments as their chief export product (such as Bangladesh, Cambodia, Lao PDR, Nepal, Pakistan, and Sri Lanka).

Clothing and footwear are sensitive to consumer sentiment and imports of these products account for almost all US consumption. Unlike durable consumer goods, purchases of garments and footwear are highly discretionary.⁶ Moreover, orders for clothing are placed frequently and must be met by just-in-time delivery so that a downturn in one month's orders will be reflected in the next month's shipments (Rosen 2002, pp. 180-182). If this is the case, examination of quarterly clothing shipments to the US over the course of 2007 and year-on-year growth rates in each quarter may be particularly revealing. Clothing shipments are closely monitored by the US Department of Commerce, which has established an Office for Textiles and Apparel for the purpose of providing real-time data on the volume and value of imports from all world suppliers of textiles and garments.

For descriptive simplicity, data are presented for three Asian groupings: PRC, the Association of Southeast Asian Nations (ASEAN), and the South Asian Association for Regional Cooperation (SAARC)— combinations that encompass the vast bulk of clothing shipments from Asia to world markets, including the US. Volumes are measured in million square meter equivalents.

The overall picture that emerges is stark. Garment exports to the US from developing Asia, whether measured in volume or value terms, weakened steadily over the course of 2007. First quarter growth is robust, second quarter growth is about in line with overall US demand growth, but the third quarter slips and the fourth is flat (Figures 1.1.31 and 1.1.32). Exports from the PRC have grown more slowly in each successive quarter of 2007 and were virtually flat in the fourth. In contrast, ASEAN managed to increase its growth in the fourth quarter relative to earlier quarters by taking advantage of the constraints facing PRC shippers under US product-specific safeguards.⁷

SAARC suppliers, which include garments giants such as Bangladesh, India, and Pakistan, failed to take advantage of the constraints on the PRC and actually saw their volume growth rates crumble after the second quarter, falling to one third of the previous quarter's growth in the third and then retreating absolutely in the fourth. In value terms, the third quarter's performance was half that of the previous quarter, and there was virtually no growth in the fourth quarter. World suppliers as a whole saw growth steadily erode as the year wore on; the impact of the subprime crisis and a loss of confidence are clearly indicated in the much slower growth in world shipments in the third quarter and the contraction in the fourth for both volumes and values.

If retail sales of garments can be taken as an indicator of consumer appetites more generally, the picture that emerges is one of deepening gloom—and gloom that is touching Asian suppliers.

A second example of recent trends in US consumer demand is that of footwear shipments—a product group that is complementary to clothing and would be expected to show similarity in patterns of retail sales and of import orders and deliveries. Imports of footwear from leading Asian suppliers, including the dominant PRC, and from the world as a whole are shown in Figure 1.1.33.

As in the case of garments, footwear for US consumption is almost all imported.⁸ Footwear is unregulated by safeguard quotas and is protected mainly by tariffs that average about 10% in the case of most Asian suppliers. Footwear imports unambiguously reflect a sharp contraction in consumer demand and retail sales in the fourth quarter of 2007 from which even the PRC is unable to escape. The quarterly pattern observed in US consumer demand and imports from Asia clearly reflects a sharp slowdown with shipments from all suppliers (including the PRC) falling by about 5% year on year. Indonesia, a fairly large exporter of shoes, takes a brutal hit during 2007 with annualized shipments contracting by 19%.

A third example of a product group for which US consumption is heavily import dependent—toys, games, and sports equipment—is also worth examining. Imports and sales of these items are likely to be more strongly influenced by seasonal demand than clothing and footwear as they are extremely popular as gift items in the winter holiday season. Hence, one would expect third and fourth quarter imports to rise strongly relative to the first and second quarters. The growth of this product group is also likely to be strongly influenced by consumer sentiment and discretionary income. Again, as with footwear, the PRC is the dominant supplier of these items to the US import market, with a share of over 80% in 2006 and 2007 (Figure 1.1.34).

This product group is important for many other exporters in Asia. The pattern of growth over the course of 2007 clearly reflects a slowdown year on year as each successive quarter shows a reduced, albeit strongly positive, growth rate. The fourth quarter's outturn of single-digit growth after starting the first and second quarters with growth in excess of 30% is sobering—and for most Asian suppliers fourth quarter contraction has the same feel of export stagnation that is seen in clothing and footwear. Countries whose currencies strongly appreciated against the dollar in 2007, such as Korea and Thailand, have seen their growth rates fall the most.

A cursory examination of shipments of consumer durables does not show the patterns seen in more sensitive semi-durable goods for which US consumption is heavily import dependent, but this is likely to change if the downturn endures over the first half of 2008 or longer. Already reports of steep cuts in sales of automobiles in early 2008 are coming in, along with further evident weakness in consumer sentiment and in retail sales more generally.⁹

A broader look at the value of developing Asia's exports to the US, including mineral fuels, when deflated by the US import price index shows a clear pattern of deceleration in the second half of 2007 relative to the same period in 2006 (Figure 1.1.35).

Japan's faltering recovery and impact of consumer anxiety on Asian exports

Japanese consumption is not as large as that of the US but the country is still one of the premier markets for exports of manufactured goods from the rest of Asia. It remains the second-largest national economy worldwide and still outweighs the PRC in terms of overall purchasing power with a real GDP of $3.8 trillion versus $1.9 trillion for the PRC in 2006 in 1990 constant US dollars, or $4.4 trillion versus $2.7 trillion in current dollars (UNSD 2008). Japan also has a higher consumption-to- GDP ratio than the PRC, although it is less import dependent. In the case of Japan, data are examined over two half-years in 2007 compared with the same periods in 2006 (that is, January-June and July-December).

Imports of clothing and accessories of clothing from all suppliers were growing at a healthy 4% in yen value terms in the first half of 2007 but then contracted in the second (Figure 1.1.36).

In particular, imports from the dominant supplier, the PRC, underwent a sharp reversal from growth to contraction in the second half. Imports from ASEAN also slowed in the second half relative to the first but remained in positive territory. Imports from the newly industrialized economies (NIEs) of Hong Kong, China; Korea; Singapore; and Taipei,China experienced very sharp reductions of 19% in the first half and almost 26% in the second. Thus, led by the PRC, the whole of Asia experienced a switch to contraction in the second half and this ensured that imports from all suppliers declined. Perhaps surprisingly, given developments in the foreign exchange markets, imports to Japan from the EU remained positive while those from the US were down sharply, by over 20% in both the first and second half. Conversion of these imports into US dollars flattens out growth over the two periods to an average of 0.5% but still leaves imports from the PRC in negative territory in the second half (Figure 1.1.37).

This impression of weakness in consumer purchases appears to apply as broadly to Japan as to the US. Japanese imports of computers and computer parts (Figure 1.1.38) fell in both half-years (and by double-digits in the second).

Electrical machinery growth also weakened as the year progressed with growth of about 10% in the first half, declining to about 6% in the second.

The pattern of growth over the 4 quarters varies according to whether one uses yen or US dollars. Looking at Japan's general merchandise imports from developing Asia, on the basis of growth in yen values, the impression of weaker rather than stronger growth in the fourth quarter is reinforced. This is likely to be reflected in weaker export growth in developing Asia at the start of 2008.

EU Trade: A last bastion of G3 demand for Asian exports—or the beginning of the end for the Asian export boom?

The availability of data limits the extent to which the impact of demand changes in the EU on developing Asia can be assessed. However, the data that run through the third quarter of 2007 indicate that consumer demand in the EU held up rather well compared with the US market. For example, clothing imports from developing Asia continued to grow by about the same pace in the third quarter as the first half of 2007 (Figure 1.1.39). Growth in the euro value of imports of clothing showed slower growth for the PRC in the third quarter but still ran at doubledigit rates. In contrast, growth from ASEAN and SAARC (as groups) was negative over the first 3 quarters of 2007. The complex systems of preferential trade that the EU maintains explain this performance.¹⁰ The individual performances by Asian suppliers, aside from the PRC, are influenced by the varying extent to which they enjoy preferential access to that market.

Looking at general imports (Figure 1.1.40), it appears that demand was strong in the fourth quarter but this may mask the underlying situation. Unfortunately, no data on volumes were available as of March 2008. Another way to assess the situation is to examine export data from developing Asian sources. Again these are very limited but as the figures indicate, early 2008 data suggest that growth is decelerating in most cases relative to 2007 (Figure 1.1.41).

The impact of the slowdown in the G3 on their trade is likely to become more severe as the slowdown deepens. Preliminary data for 2008 are compared with annual data for 2007 in the cases of the US and Japan, and they show that the slowdown is indeed reducing the growth of exports between the two largest national economies in the world. Imports of the US and Japan from the EU in 2008 are showing signs of weakness compared with annual imports in 2007. The growth rates for shipments from the EU slip from 6.4% to 4.2% in the US and from 9.4% to 4.6% in Japan.¹¹

This section has until now focused on the G3 as a source of demand for developing Asia's exports. Conversely though, robust growth in developing Asia may provide a cushion as a source of demand for G3 exports. Equally, it is possible that strong final demand in Asia may benefit intra-Asian exports. In the next section, the role of the PRC in propagating the external downturn is examined through its role in generating intraregional trade in final consumer and investment goods.

Is the PRC uncoupling from developing Asia?

The results from the economic models presented above indicate that intraregional trade linkages are likely to transmit a G3 demand slowdown within Asia. In particular, the results confirm an important role of the PRC in intraregional trade and linking that trade with global markets. As illustrated in ADB (2007a, 2007b), the increasing importance of intraregional trade is attributed mainly to the parts and components trade, with the PRC functioning as an assembly hub for final products in Asian production networks. Recent studies by Cui and Syed (2007) and Albaladejo and Lall (2004), however, suggest that trade patterns in Asia are now changing.

In particular, the PRC's demand for imports of intermediate products from Asian economies has been declining as a share of its imports, and the domestic content of its exports has risen. If this is indeed true, it would imply that the PRC is shifting from being an export hub for the rest of Asia and is deepening backward linkages domestically. Such a structural shift would seem to imply that the PRC would be even more exposed to an economic slowdown in the G3, and that other economies in East and Southeast Asia may be more sheltered. The data presented here, however, suggest that these shifts may be more apparent than real.

Over the past 5 years, the PRC's trade surplus has grown over fivefold in US dollar terms and it has also risen hugely as a share of its GDP. The rise in the trade surplus over the past 3 years stems from a slowdown in import growth relative to export growth. Import growth declined from more than 35% a year in 2004 to less than 20% a year by mid-2007 while export growth decelerated from 34% to 27% over the same period. Does this dramatic slowing of import growth suggest substitution toward intermediate goods in the PRC's production of exports?

Looking at intraregional trade balances, after 4 years of being in deficit with the remainder of Asia, the PRC has been running surpluses since early 2006. Cui and Syed (2007) believe that the slowdown in the PRC's import growth reflects important structural changes. Their thesis is that large investments and technological upgrading have boosted domestic production capacity, especially for intermediate products, leading to a decline in the import content of exports. These trends would seem most apparent for home electrical appliances, ordinary machinery, and high-tech products.

But there is no consensus on these points. Some pundits (as for example Gilboy 2004, Athukorala 2007) have observed that the slowdown in the PRC's import growth may be a reflection of temporary overinvestment in certain sectors and is likely to be short-lived. They argue that supply-side complementarities between the PRC and its East and Southeast Asian neighbors remain significant and that the likelihood of export crowding-out by domestic producers in the PRC tends to be vastly exaggerated.¹²

A closer look at the data shows that, overall, the share of parts imports in the PRC import basket is still increasing. Moreover, export data do not suggest a growing role for the PRC as an intermediate parts supplier. The share of parts in total manufacturing exports did gradually rise over 1992-2002 but has since stabilized at around 15%. In contrast, the share of parts in total manufacturing imports has risen steadily, more than doubling in the past 15 years. In 2006, imports of parts accounted for almost 37% of total manufacturing imports, up from 16% in 1992 and 29% in 2000 (Figure 1.1.42).

In East Asian economies, the share of parts in total manufacturing exports has grown strongly, nearly doubling between 1992 and 2006, to 33%. The share of parts in total manufacturing imports has also risen. These patterns suggest growing specialization in intermediate goods production, with a high degree of vertical integration among these economies. For Southeast Asian economies, the shares of parts in total manufacturing exports and imports have also steadily risen. Malaysia, Philippines, and Singapore stand out as being heavily specialized in parts trade. Only Thailand gives any sign of a decline in such trade. The establishment of Thailand as a hub for the automotive industry would appear to have curbed its cross-border parts trade. The share of parts exports declined from 30% in 2000 to 21% in 2006, and the import share from 35% to 27% over this period.

In South Asia, too (albeit from a small base), intermediate goods trade is on the rise. The share of parts in total exports has increased from 2% in 1998 to around 6% in 2006. The share of parts imports has been relatively stable at around 12% over the past 15 years.

The largely unskilled labor-intensive manufacturing industry is the only sector in which trade in parts/intermediate products is in decline. In the PRC, such a decline strongly reflects a reduction in intra-industry trade in textiles (ADB 2007a, p. 90). The share of fabric imports in total manufacturing imports in the PRC has declined significantly over the past 15 years, from almost 7% in 1992 to only 1% in 2006, mirroring a huge investment in textile machinery and production capacity in the textile and clothing industries there in anticipation of the PRC's membership in the World Trade Organization (WTO) (ADB 2007a, p. 93).

In Figure 1.1.43 the composition of Asia's exports is shown for parts and components for different geographic groupings within developing Asia. These data confirm the importance of parts trade within the region: the export share of parts rises within the region and the PRC becomes an important export destination for parts from, particularly, East and Southeast Asia. Importantly, the underlying data do not provide any strong indication that other Asian final goods exports to the PRC are rising strongly.

The reliance of industrial production in the PRC on imported parts and components has risen over the past decade (Figure 1.1.44). The share of parts imports in gross industrial output rose from only 1.7% in 1992 to almost 6% in 2006. To date, PRC industrial investment, and that of multinational enterprises in particular, has not been involved significantly with parts production. Naughton (2007) observes that foreign-invested enterprises are overwhelmingly concentrated in the final assembly stage of production, which is the most labor-intensive layer in production processes spread over many countries. Basic research, product design, and physical capital- and human capital-intensive stages of the production process tend to be carried out in the home countries of multinational enterprises or in other Asian countries that are in a more advanced stage of industrial development than the PRC.¹³ A slight decline in the share of parts imports in gross industrial output over the past few years is likely to be a reflection of reduced import content for the garment industry. The import value of fabric to gross output in that industry has continuously declined, to 10% in 2006 from 26% in 2000.¹⁴

In summary, the data show that the PRC has not become a significant producer or exporter of parts and components. In contrast, parts trade in other subregions (East Asia, Southeast Asia, and South Asia) is generally on an upward trajectory, with rising shares for exports and imports. The parts share of PRC imports has also risen. The idea that the PRC has been able to replace imported supplies with domestic components is not supported by these trends. The claim that the PRC now provides an important source of demand in final goods markets for other Asian countries also sits ill with the data. The base share of final demand is small (ADB 2007a, p. 95) and virtually all growth in intraregional trade in recent years has been attributable to parts and components. Therefore, although the characteristics of intra-Asian trade may have started to change, it will be some time yet before they fundamentally alter the nature of the transmission of external shocks.

So far, the center of focus has been on the transmission of the slowdown in the G3 to developing Asia through its exports and vertically integrated trade channels, particularly within East and Southeast Asia. But there are other possible channels through which impacts may be felt. In the next section, the potential for financial contagion is examined.

Financial integration and contagion

Each global slowdown is different. Current difficulties have followed an extraordinary period for global financial markets—tremendous growth, deepening integration, and rapid financial innovation. The potential for financial contagion in an increasingly borderless world of international capital movements has clearly risen.

McKinsey's annual report, Mapping global capital markets (McKinsey Global Institute 2008) observes that at the end of 2006, the major economies of developing Asia held assets to the value of $14.2 trillion, equivalent to 250% of combined GDP. Developing Asia's financial asset holdings are dominated by the PRC, which possesses a bit over a half of them, with Korea and India together accounting for 40%. In the PRC, financial assets are over 300% of GDP, whereas in India the corresponding ratio is just over 200%. Although the PRC's financial system is still bank dominated, asset distribution in other markets is more evenly balanced among equities, debt, and deposits. However, in some countries such as India, Indonesia, and Philippines, government has a large profile in debt markets, with only a small private sector presence. In 1990-2006, the stock of financial assets in developing Asia grew at an average rate of 15.5% a year, easily outrunning growth of nominal income.

Financial deepening in developing Asia has occurred during a period in which the "home bias" in financial asset investment has been weakening. In 1990, cross-border capital flows amounted to just 5.2% of global GDP. By 2006, this figure had more than tripled to 17.2%. As a share of total global assets, financial investments overseas have increased from 28.9% of total assets in 1995 to 44.6% by 2006 (Figure 1.1.45). Yet, though cross-border capital flows have been growing much faster than trade, trade flows still dominate in absolute size. In 2006, total crossborder capital flows amounted to 62.1% of global imports.

For developing Asia, gross capital inflows in 2006 were a shade below $300 billion, equivalent to 4.6% of GDP or 11.6% of its total exports. Inflows have climbed steeply since 2001 when they were just $39 billion, but have also more than doubled from their precrisis peak of $125 billion in 1996. In addition, developing Asia is also a large net lender to the rest of the world, with net capital outflows from East Asia alone reaching $539 billion by 2006. Notably, the PRC was the world's single largest exporter of capital that year. Data from IMF's Coordinated Portfolio Investment Survey suggests that while the share of East Asia's crossborder capital investments in other East Asian countries has grown quickly and that the share of investments in the US has declined, this picture changes if the definition of East Asia is widened to include Japan. Japan still invests heavily in the US and invests little in East Asia. Moreover, a major source of portfolio investment in East Asia is the US. Kim and Lee (forthcoming) estimate that in 2006 the US accounted for about 38% of the total portfolio liabilities of East Asia (excluding Japan), little changed from its 37% share in 2001.

Cross-border capital flows exhibit a pronounced upward trend, but with periodic sharp breaks from trend. In the past decade, flows contracted in absolute value in 1997-1998 and in 2001-2002. Recent reversals coincided with the Asian crisis of 1997-98 and with the bursting of the dot-com bubble and the US recession of 2001. Although from a much smaller base, a contraction also occurred in the early 1990s when there was a global recession. The most volatile component of crossborder capital flows has been cross-border lending and deposits, with foreign direct investment being the most stable. Surprisingly perhaps, cross-border investments in equity markets have been quite stable, with investments in debt securities being comparatively volatile (Figure 1.1.46).

The rapid expansion of global capital markets and the even faster growth of cross-border capital flows suggest that the potential for the financial transmission of global shocks is likely to have increased. Direct evidence on this point from Asia is fragmentary. Kim and Lee (forthcoming) show that East and Southeast Asian equity markets now track the US equity market more closely than before the Asian crisis and that within East and Southeast Asia, markets are also more closely synchronized with each other than a decade ago (Figure 1.1.47). Chai and Ree (2005) note that although regional factors have played an increasingly important role in explaining equity market movements in East Asia, markets in the region have closer ties with the US than with one another. Another measure of financial integration—the covered interest rate differential—does not show much change in 2000-2007, but suggests closer integration of East Asian markets with the US than with Japan (Kim and Lee forthcoming).

Kim and Lee (forthcoming) also look at the extent to which East Asian economies engage in "risk sharing" with each other and with the rest of the world. In circumstances where risk sharing among countries is high—shocks in one country are dispersed and are partially absorbed by others—close financial integration can be expected to align consumption paths across countries. The empirical evidence points to significant risk sharing between East Asia and the global economy (suggesting that in response to a shock to global consumption, consumption in East Asia would change by between 30-50%) but not within East Asia.

Given Asia's presence and growing participation in international financial markets, it would be surprising if a global slowdown were not to make itself felt through asset market adjustments. Credit spreads have widened for Asian borrowers in offshore markets and this should deter debt issuance. Equity markets in Asia have also moved closely in step with those of the US, and the erosion of wealth can be expected to lead to some belt-tightening by households. Investors may also find it more difficult to raise capital in volatile domestic equity markets. Cross-border lending might also be expected to slow and even dip, if previous episodes of financial distress provide a reliable indication. But where interest rates are expected to rise and there is an expectation that domestic currencies will appreciate, this may attract capital inflows. Also, some Asian debt may look like a good bet alongside fast degrading debt in mature markets.

Evidence on these points is still fragmentary. But data from the Institute for International Finance (2008) show that both equity and syndicated loan issuance by "Emerging Asia" held up well comparing the September to February period 2006-2007 with the same period in 2007-2008, but that bond issuance fell sharply, by 31%. Developing Asia's dependence on the offshore funding market is, however, small.

There is little hint of significant vulnerability within Asia's bankdominated financial systems. Despite close financial coupling, the direct exposure of banks in developing Asia to US subprime mortgage debt or to other "unsafe" assets is believed to be small. The region's banks are, in general, reasonably well capitalized and profitable and appear not to have indulged in leveraged investment activities on a large scale. Recent data on real loan growth show that it is increasing in many countries, including Indonesia, Korea, and Singapore (Figure 1.1.48). Only in countries where there is significant domestic monetary tightening, such as the PRC, is there evidence that real loan growth is decelerating. These data do not suggest that a credit crunch is imminent. Nevertheless, in some countries banks may be exposed to risks of rising bad debts, especially where there is significant lending to support investment in inflated equity or property markets. Currency mismatches on bank balance sheets have recently surfaced as a source of risk in a few economies.

Global financial turbulence may even have a silver lining for some large surplus-savings countries in developing Asia. A reduction in net private capital inflows, as forecast by the Institute for International Finance (2008), may ease pressures on domestic liquidity. Financial trauma in more mature markets has also created a gap between the global demand for equity capital and its ready availability. This offers investment opportunities for countries with foreign exchange assets in excess of their reserve needs.

Summary

Evidence suggesting that Asia has "uncoupled" from the global economy is scant. The relationships embedded in large empirical models of the global economy, such as the Oxford Economics model (Box 1.1.2 above), are broadly confirmed by fresh evidence drawing on recent high-frequency time-series analysis of the links between retail demand in the G3 economies and exports in developing Asia (see The uncoupling hypothesis: New evidence from G3 retail sales and Asian exports, above). Both suggest that developing Asia's exports do respond quickly and in some cases strongly to variations in G3 demand. Precise impacts differ depending on the source of the demand shock and trade structure. Close inspection of data for cyclically sensitive manufacturing exports confirms slowing across developing Asia in precisely those sectors that would be expected to feel the effects of a global downturn first. Evidence from durable goods exports and from the EU will have to await the release of data.

The PRC's role in the transmission and buffering of shocks could be pivotal. Yet the evidence marshaled above (Is the PRC uncoupling from developing Asia?) suggests that strong growth of domestic demand in the PRC is unlikely to provide a buffer for other countries because their export presence in the PRC's internal final goods market is still limited. Moreover, through their ties along vertically integrated supply chains, East and Southeast Asian economies will be hit by any slowing in demand for the PRC's exports to G3 markets.

Developing Asia's ties to global financial markets have greatly strengthened in the past decade. The stock of financial assets as a proportion of GDP in developing Asia has risen quickly, as has crossborder ownership of assets and liabilities. Despite the unmistakable trend toward stronger financial coupling, the impacts of global credit market difficulties on developing Asia are expected to be limited. Domestic banking systems, which still play a significant intermediation role, generally appear well positioned to weather the global storm.

Finally, the point bears repeating: though developing Asia's economy is not immune to the vicissitudes of global demand, its longer-run growth trajectory will be much more a function of structural and supply-side dynamics. To maintain momentum, countries will have to address and overcome a variety of constraints (see Part 3 of ADO 2008). In the short run, the impact of the global slowdown is likely to be modest: even a highly unfavorable global scenario that dents growth in developing Asia by more than forecast in the ADO 2008 baseline should not leave lasting scars. The main risks to future growth lie elsewhere: in reversals of market access and of trade liberalization, and in the failure to meet domestic challenges.