Journal Issue
Share
Article

Safe-Haven Korea? - Spillover Effects from UMPs

Author(s):
Jack Ree, and Seoeun Choi
Published Date:
April 2014
Share
  • ShareShare
Show Summary Details

I. Introduction

The discussion about spillovers from unconventional monetary policies (UMPs) has, until recently, focused on the impacts of quantitative easing (QE) of the United States through portfolio capital flows. Historically, capital flows tended to cause imbalance, at times severe ones, in the recipient economies—ranging from exchange rate overshooting and credit excess, to susceptibility to a sudden stop, and finally to a run-up to a balance of payments crisis. Worries about capital flow spillovers have thus prompted a global review of the policy tool kit available to respond to them.2 More recently, the focus of the spillover discussion has rapidly shifted to the risks associated with QE exit, particularly since May 2013 when these risks began to trigger capital outflows from a number of emerging markets (EMs). The Bank of Japan’s adoption of quantitative and qualitative easing (QQME) in April 2013 has also been an important recent development, the spillover of which is likely to add to the discussion.

In this paper, we examine how Korea’s capital flows and trade have been affected by UMPs, focusing on QE and QQME. Korea makes an intriguing case because of its borderline position between advanced and emerging market country groups and the common perception that Korea fiercely competes with Japan in the world market. Since the 2008-09 global financial crisis, Korea has experienced some recovery of capital flows during QE and QQME expansions, in line with EMs, but the size of the recovery was very moderate compared to Ems on average. This is an important stylized fact indicating the global investors’ growing tendency to differentiate Korea. Since May 2013, furthermore, the won has remained remarkably resilient despite the QE tapering turmoil. At the same time, The UMPs had also significantly affected Korea’s exchange rate. In particular the won has appreciated by 25 percent against the yen since May 2012. In view of Korea’s strong perceived competition in the third markets with Japan, the discussion on QQME spillovers to Korea has focused on the exchange rate channel. So far, however, the impact on Korea’s exports has remained muted. This could be merely a matter of time required for the exchange rate effects to eventually filter through. Or it may be a reflection of the strengthened resilience of Korea’s exports to price factors at play, which could have been also helped by other transitional factors (e.g., geopolitical tensions among countries in the region).

With the above-mentioned factors in mind, we examine the impact of UMPs on Korea through both financial and trade linkages. On the financial linkage, our main question would be whether QE and QQME have had a significant impact on capital flows to Korea, and thus the tapering of QE might cause a sudden stop of such flows. On the trade linkage, we ask (i) whether Korea’s sensitivity to the won-yen exchange rates is significant; and (ii) if Korea’s value chain link to imported inputs from Japan will help ameliorate negative competitiveness effects of the cheaper yen. We find that QE had little direct impact on capital flows to Korea, and the tapering is unlikely to cause capital outflows from it, owing partially to safe-haven behavior of capital flows to Korea (Section II). We also find that the exchange rate spillovers to Korea from QQME will be limited in the near term, although Korea’s newly gained non-price competitiveness may unwind, should the yen’s weakening cycle go deeper or become sufficiently prolonged (Section III).

II. Spillovers through the financial linkage

A. Korea—Capital Flow Trends

Early 2000s to 2008-09 Crisis

During the 2000s, Korea liberalized its capital account transactions, which introduced both new opportunities and risks. In the run- up to the 2008-09 crisis, expectations of trend appreciation in the won’s exchange rate created a large demand for hedging by exporters. In conjunction with the lack of commensurate demand for hedging by the importers (particularly refineries), such expectations led to a large buildup of liquidity mismatch in the banking system: banks (particularly foreign bank branches) relied heavily on external borrowing, particularly short-term one with low interest rates, in order to offset currency risks arising from relatively long-term forward contracts.3

Cumulative Capital Net Flows

(In billions of U.S. dollars)

Sources: CEIC Data Company; and IMF staff estimates.

2008-09 Crisis to Summer of 2013

In 2008Q4, global liquidity dry-ups caused a sudden outflow of about US$ 33 billion (14 percent of Korea’s gross international reserves) of banking capital from Korea. Stemming mainly from the unexpected repayment of short-term interbank FX loans—the bulk of which were interoffice loans—by the foreign bank branches, these outflows entailed the liquidation of local bond positions and conversion of the won proceeds to U.S. dollars, and caused a sharp depreciation of the won (to a maximum of 36 percent during the crisis). The outflows also caused Korea’s gross international reserves to decrease by 34 billion dollars that quarter.4 In response to the stress, the Korean authorities promptly put together a package of measures including monetary easing, fiscal stimulus, and the BOK’s dollar swap injection, as well as government guarantees on banks’ external debt. In addition, they also established bilateral currency swap facilities with the United States, as well as with Japan.

In 2009Q1, Korea was already able to stem the sudden stop in banking capital. The successful containment was not only due to the authorities’ own actions, but also to measures adopted by policymakers at the epicenters of the crisis (including the U.S. quantitative easing). Since then Korea has continued to restore the strength of its external position with the current account surplus and portfolio capital inflows more than offsetting a continued orderly reduction in banks’ external debt, which has allowed foreign direct investment (FDI) and reserve assets to build up over time. During this period, there were three waves of QE programs, Korea’s responses to which were visibly differentiated from EMs:

Korea - Net Capital Flows During U.S. Quantitative(In billions of dollars, monthly average)
QE1QE2QE3
Net Bonds1.70.20
Net Equity1.80.1−0.9
Net Other−0.50.7−2,4
Net FDI−1.1−1.6−1.3
Net Derivatives−0.40.20.5
Net Capital Flows1.5−0.3−4.1
Source: CEC data company Ltd

QE1 [11/2005 - 3/2010); QE2 (11/2010 - 6/2011); QE3 (9/2012 - 2/2013)

Source: CEC data company Ltd

QE1 [11/2005 - 3/2010); QE2 (11/2010 - 6/2011); QE3 (9/2012 - 2/2013)

EMs - EPFR Flows During U.S. Quantitative Easing(In billions of dollars, monthly average)
QE1QE2QE3
Bonds00.53.3
Equity4.53.710.5
Total4.54.113.8
Source: Haver Analytics

QE1 [2008.11-2010.3); QE2 (2010.11-2011.6); QE3 (2012.9-2013.2)

Source: Haver Analytics

QE1 [2008.11-2010.3); QE2 (2010.11-2011.6); QE3 (2012.9-2013.2)

  • Some push factor effects of the quantitative easing seem to have existed for Korea during QE1 (November 2008 to March 2010), but not so much during QE2 (November 2010 to June 2011) and QE3 (September 2012 to now) periods.5 During QE1, the monthly capital flows to Korea were 1.5 billion U.S. dollar net inflows on average (the average net flow was negative during the QE2 and QE3), with strong bond and equity flows (3.5 billion U.S. dollars) more than offsetting outflows in FDI, banking flows, and settlement payments related to derivative liabilities (i.e., valuation losses).

  • The impact of quantitative easing on bond and equity flows appears to be significantly milder for Korea than EMs. The global bond and equity fund flows to emerging market countries tracked by data provider EPFR Global rose to an average of 14 billion dollars per month during QE3 from of 4-5 billion dollars per month during QE1 and QE2. In the case of Korea, monthly gross bond and equity flows rose from 2.1 billion dollars during QE1 to 2.4 billion dollars during QE2, but they fell back to 2.1 billion dollars during QE3.

In response to the resumption of capital flows, the authorities introduced macro-prudential regulations, including a ceiling on bank’s FX derivative position. The objective was to curtail systemic risks from short-term external debt, foreign currency denominated domestic lending, and foreign bond holding. Owing, at least partly, to these measures, banks’ external debt, led by short-term debt, sharply decreased from 220 billion dollars (short-term debt, 160 billion dollars) immediately prior to the crisis to 178 billion dollars (short-term debt, 77 billion dollars) in September 2013.6

Since May 2013

Since the 2008-09 crisis, the global financial market has continued to face recurring waves of volatility, and the current one is led by the fears about uncertainties related to the tapering of QE by the United States On May 22, 2013, the Federal Reserve Chairman Ben Bernanke made a public comment indicating a possibility that the Federal Reserve’s unconventional programs for monetary easing could be scaled back in the near future. From then on, spates of events, be they speeches or data release indicating the strength of the ongoing U.S. recovery, have increased financial market volatility, and triggered a significant amount of capital outflows from EMs. The latest wave of global financial volatility has taken harder hits on EMs with weaker fundamentals or policies, making them the most acute receivers of the negative spillovers from UMPs. As a means of defense, many EMs have strengthened capital flow measures, while striving to achieve an optimal balance between foreign exchange market intervention and allowing currency depreciation.

So far, however, Korea has done remarkably well, clearly differentiating itself, and, in fact, attracting significant inflows of both equity and bonds. This was partly because international investors have redirected their funds from emerging markets such as India, Indonesia, Brazil, and Turkey. This is an amazing departure from the past. Since the liberalization of the capital account in the early 2000s, Korea has been commonly labeled by market analysts as a high financial beta market, meaning that its asset prices are highly sensitive to fluctuations in global risk appetite. So why have global investors behaved differently this time? There seems to be a number of interconnected reasons:

First, the overall resilience of Korea’s financial system to external shocks has significantly increased since the 2008-09 global financial crisis, particularly because its banks have shed their external short-term debt, which used to be their key vulnerability to swings in capital flows.7

Second, Korea’s post-crisis rebound of equity prices has been very modest compared to many emerging markets that enjoyed a surge. And Korea’s specific factors that had dragged on equity began to fade in the summer of 2013. Three major factors that had weighed on Korean equity were (i) North Korea risks, (ii) benchmark index change by one large EM index fund (Vanguard EM ETF) from MSCI Global Equity indices(which include Korea as an EM) to FTSI Emerging Market indices (which does not), and (iii) the yen’s sharp depreciation since late 2012.

Third, bond flows are driven by increasingly diversifying the investor base, and hence are more stable. Moreover, our analysis presented later in the paper suggests that these investors have tended to increase exposure to Korea when the global market becomes more risk averse, at least up to a certain critical point, in a “safe-haven flow” sort of behavior.

Fourth, with a large current account surplus and the market’s perception that the Korean authorities are pacing the won’s appreciation, there seems to have been a one-sided expectation on the won’s exchange rates.8

Given all this, it is not unreasonable to speculate that if a QE tapering takes place in an orderly fashion, driven by the recovery of the U.S. growth outlook, portfolio capital flows are likely to continue as Korea’s economic outlook strengthens, as we will try to corroborate in the next section.

B. Effects of UMPs on Capital Flows: Estimation

Here, we construct a regression model of capital flows aimed at identifying both direct and indirect effects of UMPs on capital flows to Korea, and thus provide a basis to plausibly quantify the impacts of future tapering of QE, as well as of the continued rollouts of QQME by the Bank of Japan (BOJ). The model is a simple ordinary least-square model using Korean specific time series and global factors of the following form (See Appendix 1 for the list of variables):

(Yth: capital flows of type h, from an array comprising portfolio bonds, portfolio equity, and bank flows, Xth:m×1 vector of push factors, Zth:n×1 vector of pull factors, UMPt: n×1 vector of UMPs proxy variables).

Data

For portfolio (equity and bonds) flows, our baseline model uses weekly data from January 2008 to June 2013, using net purchases of listed Korean stocks and bonds by foreigners as the dependent variable. Previous studies on the impact of UMPs on capital flows have used EPFR Global data, which provide information on equity and bond fund flows to various countries by tracking country and regional weightings adopted by the global funds covered by the dataset. In the case of Korea, however, the EPFR Global data have been unable to correctly capture foreign portfolio capital flows, particularly for bonds, whose investor base is much broader than institutional investors (e.g., about 40 percent of the total foreign flows are generated by central banks). Fortunately, Korea’s major capital market institutions, such as the Korea Exchange, provide a detailed investor base breakdown of transaction flows on a daily basis.

For banking flows, however, the highest frequency data publicly available are the Bank of Korea’s monthly balance of payments statistics. Quarterly frequency options would include Bank for International Settlements (BIS) cross-border banking statistics and the Bank of Korea’s external debt statistics, both of which provide more delineated information than the standard balance of payments. However, given that quarterly frequency would smooth out the bulk of the most intriguing market dynamics, such as the impacts of market events and shifts in risk sentiment, which have been highly volatile, we used a monthly model as our baseline.

For proxies of UMP, we used the weekly change in the balance sheet of the U.S. Federal Reserve of the amount outstanding of unconventional operational measures, classifying the measures into Liquidity (LQ), Treasuries (TR), and MBS measures (mimicking the approach of Fratzscher and others (2013)). We also used an event dummy variable to capture the impact of QE announcements: a dummy variable equal to 1 (-1) in the two weeks following each announcement related to the expansion (reduction) of QE. For QQME, we were confined to the BOJ’s published balance sheet data—our models used the BOJ’s total asset stock (BOJ_ASST) and government securities holdings (BOJ_GS) as proxies for QQME—which is monthly frequency. Hence, we applied the QQME variable to the baseline bank flow model which is monthly. It is also applied to monthly equity and bond flow regressions.

Existing studies note that the expected signs of the direct impact of the QE measures and announcements can be equivocal. Because, depending on their nature and global financial setting (including risk appetite and perceived risk-adjusted return for EM assets relative to advance market (AM) assets), they may strengthen a safe-haven flow (negatively affecting equity and bond flows to many EMs) or crowd out private investors at the QE origin (positively affecting both equity and bond flows), exert push effects through a portfolio-rebalancing channel. Korean capital market’s borderline positioning between AM and EM asset classes and market indication that Korean government bonds are attracting safe-haven and reserve-diversification flows (including from central banks) add another layer of complication.

Summary of Capital Flow Regressions
Equity FlowBond FlowBank Flow
Direct effects from UMP (QE)− (small)+ (largel)
Direct effects from UMP (QQME)− (large)
Yield curve steepening++
Higher VIX+/-
Global liquidity stress (Libor-OIS spread)
Source: IMF staff estimates.Note: 1/ - denotes a negative impact, + a positive impact, and +/- a threshold behavior (all based on statistical significance)
Source: IMF staff estimates.Note: 1/ - denotes a negative impact, + a positive impact, and +/- a threshold behavior (all based on statistical significance)

Foreign Equity Flow Response to PBR

Equity regression

The result of the benchmark equity regression is reported in Appendix 2. Interestingly, none of the QE variables was significant in the baseline model. The model, however, suggests that QE could have supported equity capital flows to the extent that the measures and announcements helped alleviate risk aversion. The model also shows that QE could have affected the foreign investor response to stock fundamentals, particularly price-to-book ratio (PBR), through its interaction with risk aversion (VIX). It, too, shows that foreign equity flows to Korea may have reacted negatively on U.S. yield flattening.

  • Negative impact of VIX: Both the level (log) and the change (dlog) of VIX seem to have a negative association with equity flows. The equation indicates that a 1 percent increase in VIX can cause 0.6 billion dollars of outflows a week (which amounts to two-thirds of standard deviation from the average net flows over the sample period). Moreover, the impact of the change in VIX can be reinforced when the change is accelerated. For example, a one percentage point acceleration of sequential increases in VIX (e.g., additional 10 percent increase followed by a 9 percent increase in the previous week) can reduce the foreign equity flow response to PBR by 0.2 billion dollars for every one-tenth increase of PBR (e.g., from 1 to 1.1).

  • Nonlinear response to PBR: Foreign equity flows respond to PBR in a procyclical manner (i.e., a recent rise in valuation causes more inflows), until PBR reaches a certain valuation threshold (PBR*), beyond which its behavior turns countercyclical (i.e., a recent rise in valuation causes less inflows). The threshold PBR* decreases by about 0.36 (in ratio) with just a 1 percent increase in VIX. It collapses to zero when VIX increases by 2.7 percent in the current week, which implies that foreign equity flows would respond to PBR in a countercyclical manner regardless of levels of PBR during periods with a relatively small hiccup in VIX.

  • Positive impact of U.S. yield curve (noncrisis period): A 10 basis point acceleration of yield curve steepening (e.g., from 10 bps steepening to 20 bps steepening) is associated with about a 100 million dollar (or 1/10 of standard deviation over the sample period) increase in equity flows during the noncrisis period. During the crisis period, the positive correlation gets even stronger. This seems to be caused by the positive outlook for economic growth entailed by the yield curve.

Cumulative Effect of QQME on Equity Flows

Source: IMF staff calculations.

Our alternative monthly specification (Appendix 3) broadly confirms the main finding from the baseline weekly model, except that it shows a very significant negative association between QQME and equity flows to Korea.9 This result supports the view that foreign investors in the Korean equity market focused on expected negative exchange rate spillovers of QQME on Korean corporations, both as a driver of margin squeeze and as a potential game changer in the global competition with Japanese firms.

Bond regression

The result of the benchmark bond regression is reported in Appendix 4. The bond-flow regression model shows that an increase in the Federal Reserve’s bond-buying program (TR) can actually reduce foreign bond capital flows to Korea. It also shows that Korean bonds can benefit from “safe-haven” type flows when risk aversion remains below a certain threshold, which is positively associated with arbitrage gains, expected from currency-hedged trading positions on the Korea Treasury Bond (KTB) yield curve.

Foreign- Bond Flow Response to VIX

  • Effects of UMPs: The bond-flow regression establishes a highly significant and negative correlation between the Federal Reserve’s Treasury Purchase Program (TR) and foreign bond flows. The result, while not immediately intuitive, is consistent with previous studies that found the U.S. Treasury Purchase Program had caused investor shift away from bonds (Fratzscher and others, 2013). The model however indicates that UMPs could have positively affected bond flows by easing global dollar liquidity stress (captured by LIBOR-OIS spread), an important (negative) push factor for bond flows to Korea. It could also have affected bond flows positively by putting breaks on VIX particularly during periods of tail elevation.

  • Nonlinear response to VIX: Foreign bond flows positively respond to changes in VIX until it reaches a certain threshold (VIX*). In other words, larger VIX tends to bring more foreign investment in Korean bonds, in what appears as a search-for-safety. Such a behavioral pattern actually moves well with the continued upgrade of Korea’s sovereign risk ratings since the 2008-09 crisis (they began to exceed those of Japan in 2012) and a significant increase of regional central banks in the foreign investor base.10 However, once the threshold VIX is breached, the foreign investor reaction to VIX shifts to negative (i.e., larger VIX causes a decrease in flows).

    A decrease in the currency-swap basis spread (i.e., becoming more negative), however, will lower the threshold VIX except during the episodes of yield curve inversion, although it also increases arbitrage gains. This appears to reflect that a large prolonged decrease in the basis spread tended to occur during the episodes of dollar liquidity difficulties, during which foreign investors tended to sell off Korean bonds despite the spiking of arbitrage incentives.

    When the term premium (TB5YKRW-CD3MKRW) is zero, the threshold risk aversion (VIX*) above which bond flows start to respond negatively to VIX is 31.2 (close to one standard deviation above the median of VIX during the sample period). And yield curve steepening (by increasing arbitrage gains) will increase this threshold, implying that safe-haven flows would withstand higher levels of risk aversion.11

  • Interaction with domestic pull factors. Our model thus suggests that the underlying cause of the foreign investors’ nonlinear response to VIX is the interaction between VIX and domestic pull factors, in particular dollar funding conditions (captured by the cross currency swap basis spread) and term spread. These two variables determine the profitability of a typical arbitrage position (i.e., the term premium minus the basis spread, which is normally negative) in Korean government bonds.12

Cumulative Effect of QE on Bond Flows

Direct Effect of D(TR)

Source: IMF staff calculations.

Bank- Flow Regression

The result of the benchmark bank flow—short-term other-flow gross liabilities from the balance of payments statistics are used here—regression is reported in Appendix 6. Most importantly, the result shows that one particular type of the Federal Reserve’s QE operation (i.e., Mortgage-Backed Securities (MBS) Purchase Program) had a very large direct impact on short-term banking capital flows to Korea, significantly slowing its continued large net outflows since the 2008-09 crisis. This positive spillover suggests that securitization market freeze may have been a major supply-side bottleneck affecting the cross-border lending to Korea right after the GFC. However, if that is the case, it is not clear that QE tapering will cause an equally significant negative impact on banking flows, unless it brings the U.S. MBS market back to dislocation.

Cumulative Effect of QE on Bank Flows

Source: IMF staff estimates.

The indirect impact of UMPs seems to have worked in both directions. On the one hand, it exerted a positive impact on banking flows by alleviating global liquidity strains—this channel is captured by the negative coefficient of LIBOR-OIS spread. However, our regression also suggests that the impact of UMPs could have been negative to the extent that they contributed to flattening the U.S. yield curve (its coefficient is positive). Thus, a benign growth-led exit from UMPs that is likely to leave both the term premium and LIBOR-OIS spread anchored is not likely to much affect cross-border banking flows to Korea.

Implication for QE exit risks

A growth-driven smooth QE exit, which leaves long-term U.S. rates anchored and does not hurt investor confidence, is unlikely to cause capital outflows from Korea. In fact, our model even suggests that a growth-driven QE exit could cause inflows to Korean equities, as well as banking flows, through positive signaling effects associated with an orderly steepening of the yield curve. However, a disorderly QE exit can cause capital outflows, through VIX (both equity and bonds), and the LIBOR-OIS spread (bonds and banking flows). The negative effects would likely be exacerbated, particularly for bond flows, if the unwinding of bond positions causes a dislocation of the onshore dollar funding market.

VIX: Actual and Threshold

Source: IMF staff calculations.

An increase in VIX associated with a disorderly QE exit can, however, have different implications across asset classes.13 For equities, the impact would be unambiguously negative to capital flows. But for bonds, the impact can be positive (thus mitigating the negative impacts of dollar liquidity stress, particularly offshore) so long as it stays within about 1-1½ standard deviations from the historical median. A surge beyond these thresholds, however, can trigger a foreign investor sell-off.14

III. Spillovers Through The Trade Linkage

A. Korea—Trends in Korea’s Exports after the Global Financial Crisis

The share of Korea in the global export market has advanced significantly since the 2008-09 global financial crisis. Korea’s world market share rose from 2.7 percent in 2008 to 3.1 percent in 2010, and has so far remained unchanged. During this period, Korea’s immediate upper-level competitors, particularly Japan and France, saw a loss in their market shares. There are three widely quoted explanations for this phenomenon:

World Market Share

(In percent of total world exports)

Source: IMF Direction of Trade (DOT) Database.

  • Favorable Exchange Rate: The won’s depreciation in the wake of the 2008 crisis was sharper and more protracted than most competitor currencies. And this was particularly striking in the case of the won’s exchange rate with the yen, which had fluctuated in the 12-15 won per yen range until the third quarter of 2012, as opposed to a range of between 8-11 won per yen before the crisis. Given Korea’s strongly perceived export similarity with Japan,15 Korea’s exchange rate advantage has frequently been pointed to as the cause of Korea’s market share gain.

  • Moving up the value chain: Another explanation emphasizes the interplay between Korea’s exchange rate gains and non-price competitiveness. Indeed, due credit should be given to Korean exporters’ efforts to lock in the exchange rate gains by moving up the value chain. For example, Samsung Electronics has successfully refocused its product lines toward high-value final products—which requires companies’ own branding powers (e.g., smart phones)—from its more conventional habitat of intermediate (e.g., semiconductor) products. The refocusing was accompanied by mutually reinforcing advances in branding, with Samsung moving up to the second place (next only to Apple, and followed by Google) in 2013 in the global brand ranking (it was ranked 43 in 2008).16

  • Offshore production: An increase in offshore production seems to have overall augmented Korea’s global export share and made it more diversified and resilient. Since the crisis, Korean companies have aggressively moved their production overseas to garner the benefits of vertical integration (e.g., Information and Communication Technology (ICT) products), or to move closer to the markets (e.g., automobiles). Most notable examples are smart phone makers, which increased the proportion of offshore production from 16 percent in 2008 to 80 percent in 2012; as well as the automobile industry that stepped up the same proportion to 73 percent in 2012 (38 percent in 2008). Although moving existing production lines overseas would undermine exports by substituting domestic production, an expansion of overall production capacity focusing on the offshore would substitute less from domestic production and boost exports of intermediary inputs.

Global Market Share in Smart Phones
20082009201112Q1
Samsung4319.930.1
Apple9.115.91923.8
Nokia40.133.415.88.1
Source: Strategy Analytics, May 2012,
Source: Strategy Analytics, May 2012,

B. Was Korea’s Gain Japan’s Loss?

Here we ask if Korea’s postcrisis market share gain was at Japan’s loss. This is an interesting question given the strikingly close magnitude between these two (about 0.4 percentage point for both), and Korea’s accentuated exchange rate advantage over Japan since 2009. A positive answer to the question would strengthen a prediction that Korea’s market share gains may unravel as the won-yen exchange rate moves back, in a lasting fashion, to the precrisis level or even lower owing to QQME.

A decomposition analysis, across both geographical and product markets, offers an intuitive initial path to explore the question. If Korea’s gains and Japan’s losses are indeed negatively correlated, then a dissection of the trade data into various market segments should lead to some evidence of demand substitution (e.g., demand in the U.S. automobile market shifting from Japan to Korea). On the contrary, if we fail to uncover any such evidence, the chance is higher that Korea’s gains and Japan’s losses may not be so closely connected, and the elevated levels of the won-to-yen exchange rate may have played a limited role in reshaping the market pie share.

The analysis is conducted based on the following accounting identity:

Decomposition of change in Korea’s global market share at time t (θt)

(Xt,hj: export of item h by country j at time t, Xtj: export of all items by country j at time t, Mt,h: world’s import of item h at time t, , Mt: world’s import of all items at time t)

Note that market segment h adds to (or subtracts from) the change in θ to the extent that its contribution to country j’s export growth ((XT,hjXo,hj)/Xo) exceeds its contribution to the growth of the aggregate world market ((MT,h−M0,h)/M0). For example, if the contribution of beer (or Asia) in country j’s exports was 1 percent, while its contribution to total global exports (=imports) was 2 percent, the beer (or Asia) would have contributed negatively in country j’s market share gains. Notice that market segment h can be either a geographic or product dimension subset of the global export market or an intersection of both.

The decomposition analysis highlights the following:

  • Global market gains by region: Korea’s recent global market share gains were led by China and ASEAN, which account for 73 percent of the 0.4 percent gain during 2008-12. These are geographic markets where Korea is known to have a relatively strong complementary relationship with Japan through the so-called Asian value chain, in which Japan and Korea both provide intermediate inputs (which tend to be differentiated particularly in sectors such as electronics and metals) to the local manufacturers of final products.

    • Exports to China: Sixty-three percent of Korea’s global market share gain owing to the Chinese market was made in professional and scientific instruments and electrical machinery, the same markets in which Japan has also either slightly gained or maintained its market share. In contrast, Korea’s market share remained flat or declined in areas in which Japan’s market share declined the most (e.g., petrochemicals and inorganic chemicals). Overall, Japan has lost only 0.04 percentage point of its global market share in China’s market, which is one-tenth of its overall market share loss, and one-fourth of Korea’s gain (+0.26 percentage point) in this specific market.

    • Exports to ASEAN: Korea’s global market share gain attributed to the ASEAN market was substantially broader based than in the case of China. However, the largest gain was made in petrochemical exports (+0.027 percentage point), where Japan has lost the most (-0.023 percentage point). This appears to reflect Japan’s energy supply shortage after the 2011 earthquake and tsunami. Both Korea and Japan increased their market share in road vehicles. Overall, Japan has lost only 0.01 percent of its global market share in the ASEAN market.

  • Global market gains by products: A decomposition by Korea’s export products indicates that a substantial portion of Korea’s market share gains during 2008-12 may indeed reflect demand substitution from Japan, and hence may be more prone to an unwinding in the future if the won/yen exchange rate permanently reverts to the precrisis levels.

    • About one-fifth of Korea’s market share gain was made in the global automobile market where Korean and Japanese firms fiercely compete under a limited degree of product differentiation (e.g., small passenger vehicles), and consumers tend to be sensitive to prices. Japan’s loss in this market closely mirrored Korea’s gain.

    • About one-fourth of Korea’s market share gain was made in electrical machinery and appliances—areas in which Japan also suffered a considerable market share loss (Japan’s loss in electrical machinery amounts to about 40 percent of Korea’s gain in the sector).

Global market Share Change Driver by Region

(Percent contribution, 2008-12)

Sources: IMF Direction of Trade; and IMF staff estimates.

Global market Share Change Driver by Product

(Percent contribution, 2008-12)

Sources: IMF Direction of Trade; and IMF staff estimates.

In sum, about 70 percent of Korea’s market share gains during 2008-2012 appear to have been achieved in geographical subsets (notably China and ASEAN) of the global market where Japan’s market share has not changed much. However, the dissection of Korea’s market share gains across product is suggestive of some demand shift from Japan.17

C. How Sensitive Are Korea’s Exports to the Japanese Yen?

In order to further corroborate the findings in the previous section, we examine the sensitivity of Korea’s exports to the yen using a simple econometric model. We focus our analysis on the following questions: (i) Have Korean exports become less sensitive to changes in the yen over time?; and (ii) How sensitive is Korea now, and how does it compare with other countries?

Accumulated Response of Korea’s Exports to Weakening Yen

(in percent)

Sources: CEIC Data Company; Bloomberg Ltd.; and IMF staff estimates.

To answer these questions, we ran a simple Vector Auto Regression (VAR) model, spanning 1998M1-2013M8 on the following three variables: (i) Korea’s export value in terms of U.S. dollars (seasonally adjusted); (ii) Korea’s import value in U.S. dollars (seasonally adjusted); and (iii) the won’s exchange rate against the Japanese yen18; in the same Cholesky ordering. In addition to endogenous variables, the model also included a dummy variable for financial crises (the 1997 Asian crisis and the 2008-09 global financial crisis), and CPB Netherlands Bureau for Economic Policy Analysis’s world trade value index (2005=100) as exogenous control variables.19 The VARs uses four lags following Akaike criterion.

Consistent with existing studies, the VAR estimation suggests that the sensitivity of Korea’s exports to the won/yen exchange rate has been reduced by about one-half since the late 1990s. The peak level of response of Korea’s exports to a 10 percent depreciation of the won/yen exchange rate fell to 0.7 percent (after five months of the shock) when fitting the model for the period 2006-13, compared to1.5 percent (after four months of the shock) estimated for the period 1998-2005. Our model, however, suggests that the spillover effect of QQME through the exchange rate channel to Korea is relatively larger for Korea than its Asian competitors.

Peak Impulse (cumulative) to a 10 Percent Depreciation Against the Yen
SingaporeMalaysiaThailandKorea
Peak−0.1−0.1−1.0−0.7
Lag (months]3635

D. Will Korea Benefit From a Weak Yen through Value Chain Effects?

The cross-border integration of production, including through direct investment and global sourcing, has added another layer of complication to the channel through which exchange rates affect international trade. Negative spillover effects on one country’s export price competitiveness from the currency depreciation of another, particularly a competitor in the global third markets, will be mitigated to the extent that the former sources intermediate inputs from the latter. Thus, countries that are more strongly integrated with the depreciating-currency country in the value chain would see their export cost competitiveness strengthened over third-party countries with less intense integration.

In this section, we study the positive effects from currency appreciation of this sort, particularly relative to the Japanese yen. We focus specifically on Korea’s vertical production links within Asia and the rest of the world.

Korean manufacturers traditionally have developed strong upstream links to Japanese companies as suppliers of intermediate inputs. For example, Korean ICT companies, the most successful of which have now risen to the ranks of global leaders, continue to rely heavily on Japanese high-precision equipment, including to produce semiconductors (the share of Japan in imported inputs was 30 percent in 2010 compared to 42 percent in 2000) and Liquid Crystal Displays (LCDs; the same share of Japan was 81 percent in 2010 compared to 83 percent in 2000). They also use Japanese parts and components (e.g., system semiconductors) in products such as smart phones. Hence, Korea and Japan are usually viewed more as partners than competitors in ICT, which is buttressed by the fact that Korea’s leading product spectrum (e.g., smart phones, memory chips, and LCDs) is differentiated from Japan’s (e.g., system chips and game stations). However, the importance of China is rapidly growing in production linkages both to Korea and Japan, while Korea’s reliance on Japan as the supplier of intermediary inputs has been steadily declining.

With these factors in mind, we examined the OECD’s latest value-added- based trade statistics, which allows one to disentangle gross export value, first into domestic and foreign value added, and then, within the foreign value added, value components generated from some 40 source countries. Detailed data are reported in Appendices 7 and 8.

Korea’s Foreign Value-Added (FVA) Exports

(In percent of gross exports, 2009)

Source: OECD.

  • High degree of production linkage. Foreign value added (FVA) accounts for 40 percent of Korea’s gross export value, which makes Korea’s FVA to gross export ratio the fifth highest among the OECD countries and also substantially higher than its major competitors including Japan (15 percent) and China (33 percent). A high FVA share implies less exposure to exchange rate shocks for the exporters with larger offsets provided by prices of imported inputs.

  • Modest overall production link to Japan. Yet, Korean exporters’ dependence on Japan is modest with the share in gross export value of the FVA originating in Japan standing at 5.1 percent. The data show Korean-Chinese two-way (i.e., both up and downstream production links) value chain linkage may have already become stronger than Korea-Japan links (Appendix 7), likely reflecting vibrant and growing direct investment, mainly by Korean companies in China. Moreover, the domestic value-added ratio of Japan’s exports (85 percent) is significantly higher than the ratios for most of the countries covered by the data. Hence, most of the competitive gains or losses from changes in the yen’s exchange rate will fall on Japanese exporters, although Korean exporters will also receive some trickled-down gains.

  • But Korea’s link to Japan is still larger than third-country competitors. Although the positive spillover from the yen’s depreciation through the production links to Japan would be modest, it can still help strengthen Korea’s competitive edge against third -country competitors, which tend to gain even less from the yen’s depreciation. For example, Korean transportation equipment manufacturers’ FVA share owing to Japan (6.8 percent) is substantially higher than their competitors in the United States (2.5 percent) and Germany (1.2 percent).

In sum, while value chain links could mitigate negative spillovers from the depreciation of one country’s currency to another, the mitigation effect would likely be relatively small in the case of Korea. Nonetheless, a depreciation of the Japanese yen would likely strengthen Korea’s competitive edge against third-country competitors.

Comparison of Japan Contents: Korea’s Top Export Sectors(In percent; shares of FVA to gross exports)
KoreaJapanChinaIndonesiaU.S.GermanyFrance
Overall5.185.24.41.40.90.81.8
Electrical6.882.26.34.11.61.71.1
Transport6.885.96.93.02.51.20.2
Chemical4.178.94.00.90.80.70.9
Logistics2.493.21.51.50.30.62.0
Metal6.180.53.51.61.00.60.8
Sources: OECD; and IMF staff estimates.
Sources: OECD; and IMF staff estimates.

IV. Conclusion

In this paper, we assessed the impacts of UMPs on Korea through both financial and trade linkages. On the financial linkage, our question boils down to whether QE and QQME have had, so far, a significant impact on capital flows to Korea, and thus whether the tapering of QE will cause a sudden stop in such flows. Our capital flow regressions have identified a few intriguing observations that draw a clear distinction between Korea and EMs:

First, QE and QQME operations or announcements were found to have had little or no direct influence on capital flows to Korea. The only significant exception was the cross-border banking capital flows, whose sustained outflow was substantially offset by the positive impulse generated by the Federal Reserve’s MBS purchase programs. This impulse likely would have stemmed from its role in thawing the U.S. MBS market, which was a very important source of bank liquidity.

Second, while the indirect effects of QE, through financial prices, seem to have been significant, impulse responses indicated the existence of an important nonlinear behavior. In particular, capital flows to Korean bonds have demonstrated a safe-haven behavior (i.e., flows increase with higher VIX), until the risk aversion has reached certain threshold levels. Overall, the indirect effects point to a largely positive impact of QE through the alleviation of risk aversion (captured by VIX) and global dollar liquidity stress (represented by the LIBOROIS spread).

In light of this, a growth-driven smooth QE exit, which leaves long-term U.S. rates anchored and does not hurt investor confidence, is unlikely to cause capital outflows from Korea.20 A growth-driven QE exit could even cause inflows to Korean equity and bank debt through positive signaling effects associated with an orderly steepening of the yield curve. However, a disorderly QE exit can cause capital outflows, through VIX (equities and bonds), and the LIBOR- OIS spread (bonds and banking flows). The negative effects would likely be exacerbated, particularly for bond flows, if the unwinding of bond positions causes dislocation of the onshore dollar funding market.

An increase in VIX associated with a disorderly QE exit can, however, have different implications across asset classes. For equities, the impact would be unambiguously negative. But for bonds, the impact can be positive, except in the event of a tail eruption of VIX.

These analyses point to Korea’s possible graduation from a high-capital-flow beta country, as corroborated by the impressive degree of resilience of the won and asset prices to recent QE tapering-related global turmoil. While there is a need for continued vigilance, Korea’s sound macroeconomic fundamentals and policies should enable the country to weather external shocks much better than in the past and other EMs.

The findings are consistent with IMF’s recent policy papers21 on the challenges of UMPs. One of the policy papers concludes that the effects of an exit from UMPs on non-UMP countries will depend on their exposure (likelihood of a shock caused by a sudden stop related to an exit) and resilience (ability to absorb that shock), and identifies Korea as an example of a low-exposure country. This paper reinforces this through an in-depth examination of Korea’s exposure and adds some insights on Korea’s safe- haven characteristics. The other paper found significant financial spillovers of the UMPs through both asset price and capital flow channels, which have, overall, been benign for the recipient countries. Our case study of Korea too observes financial spillovers significant. However, we also find that the direction of the spillover can change depending on the levels of market stress against which UMPs are deployed.

On the trade side, we find that (i) the bulk of the increase in Korea’s global market share since the crisis appears not to have been at Japan’s loss, which coincidentally was in a very similar magnitude to Korea’s gain; (ii) Korean exports are no longer as sensitive to the won/yen exchange rate as before although the sensitivity is larger than other Asian competitors; and (iii) expected positive spillovers from the yen’s depreciation through value chain effects would also be small in Korea’s case.

However, this paper’s benign assessment of the spillovers from QQME through the trade linkage will need to be read cautiously. Given the margin compression suffered by large Japanese companies since the 2008-09 crisis, largely owing to the then-strong yen, it would be natural for them to first decompress margins as the yen weakens. Moreover, some Japanese companies have experienced the negative consequences of too much and too rapid expansion of their market shares (e.g., massive recalls on Japanese automobile manufacturers in 2010), and may not yet be ready to expand them back.

In view of this, if the yen’s weakening cycle deepens or becomes sufficiently prolonged, Japanese companies will likely take strategic advantage, for example by foraying back into top-notch product streams that were sidelined (e.g., smart phones). This could lead to an unwinding of the product differentiation that Korea has managed to achieve so far. Korean companies should turn the challenge into an opportunity and continue to find ways to step up non-price competitiveness.

Appendix 1. Regression Variables
VariablesDescription
BOJ_GSBank of Japan’s holding of government securities (in billions of yen)
BONDFLOW_ALLNet foreign bond inflows (in millions of U.S. dollars)
CCR1Y_RATE1-year cross country swap spreads (in basis points)
CCR5Y_RATE5-year cross country swap spreads (in basis points)
CD3MKRWKorea’s 3 month CD interest rate (in percent)
CRISISDummy variable that captures the existence of a financial crisis
DC_HHDomestic claim by banks to households (in billions of won)
DC_PRVNFDomestic claim by banks to nonfinancial private sector (in billions of won)
EQUITYFLOWNet foreign equity inflows (in millions of U.S. dollars)
IPIKorean industrial production index (2010=100)
KOSPI_PBRPrice-to-book ratio for KOSPI index
TB5YKRWKorea 5-year government bond yield (in percent)
LIBOIS_SPREAD_USDLIBOR-OIS spread for U.S. dollars (in percent)
LQOutstanding amount of the U.S. Federal Reserve’s liquidity facility (weekly average, in billions of U.S. dollars).
MBSOutstanding amount of the Federal Reserve’s MBS facility (weekly average, in billions of U.S. dollars).
OIL_STLKorean balance of payments-short-term other inflows (in millions of U.S. dollars)
OIS3M_USDU.S. dollar-OIS spread (in percent)
PINL_DEBTKorean balance of payments-Portfolio bond inflows (in millions of U.S. dollars)
PINL_EQKorean balance of payments-Portfolio bond inflows (in millions of U.S. dollars)
SOV_CDS_KRKorean sovereign credit default spread (in basis points)
TB5YKRW5-year Korean government Bond
TRENDTime trend
USGB10YRU.S. 10-year government bond yield (in percent)
VIXThe Chicago Board Options Exchange Market Volatility Index
Source: Authors.
Source: Authors.
Appendix 2. Equity Regression (Baseline)
Dependent Variable: EQUITYFLOW
Metdod: Least Squares
Sample (adjusted): 1/15/2008 - 6/25/2013
Included observations: 285 after adjustments
VariableCoefficientStd. Errort-StatisticProb.
CRISIS−2313.369356.6128−6.4870620
LOG(VIX)−614.2112198.5488−3.0935030.0022
DLOG(VIX)*KOSPI_PBR(-1)−2378.536413.8241−5.7476980
D(OIS3M_USD)13905.347513.3011.8507630.0653
DLOG(KOSPI(-1))9075.0351674.6815.4189640
D(USGB10YR-OIS3M_USD)1090.736423.13692.5777380.0105
(USGB10YR-OIS3M_USD)*CRISIS695.0036113.57196.1195010
D(OIS3M_USD)*CRISIS−13186.187628.783−1.7284780.085
KOSPI_PBR(-1)6357.2981673.3613.799120.0002
KOSPI_PBR(-1)^ 2−3218.339873.1274−3.6859910.0003
@TREND−5.2356761.300901−4.0246530.0001
R-squared0.437907Mean dependent var29.16361
Adjusted R-squared0.417393S.D. dependent var947.5578
S.E. of regression723.2586Akaike info criterion16.04324
Sum squared resid1.43E+08Schwarz criterion16.18422
Log likelihood−2275.162Hannan-Quinn criter.16.09976
Durbin-Watson stat1.688911
Source: Authors’ estimates.
Source: Authors’ estimates.
Appendix 3. Equity Regression (Alternative)
Dependent Variable: PINL_EQ
Metdod: Least Squares
Sample (adjusted): 2008M03 - 2013M06
Included observations: 64 after adjustments
VariableCoefficientStd. Errort-StatisticProb.
D(BOJ_GS)−0.3299630.122261−2.6988370.0092
LOG(VIX)−4098.3541132.019−3.6203940.0006
DLOG(VIX)*KOSPI_PBR(-1)−6559.3391447.157−4.5325710
D(USGB10YR-OIS3M_USD)−1659.9911149.549−1.4440370.1544
DLOG(KOSPI(-1))15252.786356.2792.399640.0198
KOSPI_PBR(-1)56054.8521596.362.5955690.0121
KOSPI_PBR(-1)^2−27735.999704.225−2.8581360.006
D(IPI)119.9348148.02790.8102180.4213
@TREND−34.9201325.78497−1.3542820.1812
R-squared0.652723Mean dependent var659.9547
Adjusted R-squared0.60221S.D. dependent var3607.325
S.E. of regression2275.16Akaike info criterion18.42719
Sum squared resid2.85E+08Schwarz criterion18.73078
Log likelihood−580.67Hannan-Quinn criter18.54679
Durbin-Watson stat2.109061
Source: Authors’ estimates.
Source: Authors’ estimates.
Appendix 4. Bond Regression (Baseline)
Dependent Variable: BONDFLOW_ALL
Metdod: Least Squares
Sample (adjusted): 1/08/2008 - 6/18/2013
Included observations: 285 after adjustments
VariableCoefficientStd. Errort-StatisticProb.
C1728.232168.734610.242310
CRISIS−2162.402483.8886−4.4688010
D(TR)−0.0178790.005791−3.0875330.0022
LIBOIS_SPREAD_USD−802.0172235.4032−3.4069940.0008
SOV_CDS_KR−5.2049421.304547−3.9898470.0001
CCR1Y_RATE*CRISIS−9.372972.345515−3.9961240.0001
CCR5Y_RATE2.4099071.1180842.1553910.032
CCR1Y_RATE*VIX−0.1802550.064917−2.7766770.0059
CCR1Y_RATE*VIX ^ 20.0028860.0009253.1214890.002
100*(TB5YKRW-CD3MKRW)*VIX0.0633130.0254712.4857380.0135
R-squared0.319349Mean dependent var869.4192
Adjusted R-squared0.297073S.D. dependent var933.2361
S.E. of regression782.4322Akaike info criterion16.19715
Sum squared resid1.68E+08Schwarz criterion16.32531
Log likelihood−2298.094Hannan-Quinn criter.16.24852
F-statistic14.33608Durbin-Watson stat1.421518
Prob(F-statistic)0
Source: Authors’ estimates.
Source: Authors’ estimates.
Appendix 5. Bond Regression (Alternative)
Dependent Variable: PINL_DEBT
Metdod: Least Squares
Sample (adjusted): 2008M02 - 2013M06
Included observations: 65 after adjustments
VariableCoefficientStd. Errort-StatisticProb.
D(LQ)−0.0228220.010156−2.247070.0286
D(BOJ_GS)0.1142670.119970.9524590.345
VIX155.355276.27242.0368470.0464
USGB10YR-OIS3M_USD871.2663582.26961.4963280.1402
D(KOSPI)13.123214.5398522.890670.0055
SOV_CDS_KR−20.5809311.32564−1.8171970.0745
CCR1Y_RATE*VIX^20.0039780.002491.5973960.1158
TB5YKRW-CD3MKRW−1449.414688.0612−2.1065190.0397
D(IPI)84.38745187.67410.4496490.6547
R-squared0.526079Mean dependent var1558.149
Adjusted R-squared0.458376S.D. dependent var3522.924
S.E. of regression2592.7Akaike info criterion18.68667
Sum squared resid3.76E+08Schwarz criterion18.98774
Log likelihood−598.3169Hannan-Quinn criter.18.80547
Durbin-Watson stat2.154972
Source: Authors’ estimates.
Source: Authors’ estimates.
Appendix 6. Bank Flow Regression (Baseline)
Dependent Variable: OIL_STL
Metdod: Least Squares
Sample (adjusted): 2008M02 - 2013M06
Included observations: 65 after adjustments
VariableCoefficientStd. Errort-StatisticProb.
D(MBS)0.0269340.0113712.3685750.0215
D(BOJ_GS)−0.0105580.159223−0.066310.9474
D(LIBOIS_SPREAD_USD)−3191.421560.517−2.0451050.0458
USGB10YR-OIS3M_USD2681.553681.32723.9357780.0002
OIS3M_USD3452.462726.35194.7531540
D(KOSPI)−15.245525.156699−2.9564490.0046
CCR5Y_RATE−34.5765210.72605−3.2236040.0022
CCR1Y_RATE*VIX0.7769770.1049847.4009050
DLOG(IPI)−10194.1321038.99−0.4845350.63
D(DC_PRVNF)0.0081160.030170.2690160.789
D(DC_HH)−0.5248860.135891−3.8625550.0003
@TREND−23.591346.745452−3.4973690.001
R-squared0.726795Mean dependent var−891.0846
Adjusted R-squared0.670092S.D. dependent var4958.822
S.E. of regression2848.229Akaike info criterion18.91192
Sum squared resid4.30E+08Schwarz criterion19.31334
Log likelihood−602.6373Hannan-Quinn criter.19.07031
Durbin-Watson stat2.359909
Source: Authors’ estimates.
Source: Authors’ estimates.
Appendix 7. Value Chain Links to Japanese Origin Inputs

(In percent)

KoreaJapan 1/ChinaIndonesiaUnited StatesGermanyFrance
Value Chain - Snap Shot
FVA/EXPORT (all country i and sector j)40.614.832.614.411.326.673.6
Of which: largest contributor countryJapanUnited StatesJapanUnited StatesCanadaUnited StatesGermany
(contribution: FVAi/EXPORT)5.12.24.41.61.52.78.2
Of which: largest contributor sectorElectricalElectricalElectricalChemicalChemicalChemicalBusi servcs
(contribution: FVAj/EXPORT)13.64.314.32.73.36.116.9
Significance of Japan
Japan’s contribution (FVAjp/EXPORT)5.185.24.41.40.90.81.8
Max Japan content sector: MAXj(FVAjp, j/EXPORTj) 2/ElectricalFinancialTransportMachineryTransportElectricalBusi servcs
(FVA ratio of that sector)6.896.86.96.62.51.77.0
Significance of Asia
Asia 6’s contribution (FVAasia6/EXPORT)13.54.09.84.32.52.77.5
Top 5 FVA contributors (country)Japan 5.0United States 2.1Japan 4.3United States 1.5Canada 1.4United States 2.7Germany 8.2
China 4.7China 1.6United States 3.6Japan 1.4China 0.9France 1.8United States 7.8
United States 4.5Australia 0.9Korea 2.9China 1.3Japan 0.8United Kingdom 1.7United Kingdom 5.8
Australia 1.8Korea 0.6Germany 1.5Korea 0.6Mexico 0.7Italy 1.6Italy 5.7
Germany 1.5Indonesia 0.5Australia 1.2Germany 0.5Germany 0.5Russian Federation 1.3Spain 4.4
Top 5 Japan content sectors (FVAjp, j/EXPORTj) 1/Electrical 6.8Financial 96.Transport 6.9Machinery 6.6Transport 2.4Electrical 1.6Busi servcs 7.0
Transport 6.7Business services 96.Electrical 6.3Electrical 4.1Electrical 1.5Transport 1.2Financial 5.3
Metal 6.0Trade, Hotel, Restaurant 95.Machinary 5.9Transport 2.9Machinary 1.5Machinary 1.0Oth servcs 4.0
Machinery 5.1Oth servcs 94.Utility 3.9Oth servcs 1.6Textile 1.1Manufacturing nec 0.7Utility 3.9
Chemical 4.1Logistics 93.Chemical 3.9Manufacturing nec 1.5Metal 0.9Textile 0.7Trade, Hotel, Restaurant 3.7
Sources: OECD; and IMF staff estimates.

For Japan, Japanease contents refers to domestic value added to export in each sector.

FVAi, j represents foreign value added for product group j originating from country i, For each product group j, FVA captures value-added contents of a country’s gross export that is attributed to country i which is supplying inputs for those exports.

Sources: OECD; and IMF staff estimates.

For Japan, Japanease contents refers to domestic value added to export in each sector.

FVAi, j represents foreign value added for product group j originating from country i, For each product group j, FVA captures value-added contents of a country’s gross export that is attributed to country i which is supplying inputs for those exports.

Appendix 8. Value Chain Links for Korea’s Top Export Products: Comparision with Its Global Competitors

(In percent)

KoreaJapanChinaIndonesiaUnited StatesGermanyFrance
Value Chain - Snap Shot
ElectricalChina 8.80China 2.88Japan 6.32Japan 4.11China 1.70United States 2.94United States 4.58
Japan 6.84Korea 1.05Korea 5.39United States 3.11Japan 1.58China 1.72Germany 4.23
United States 5.74Germany 0.71United States 4.51China 2.70Mexico 1.02Japan 1.65Italy 3.78
Germany 1.82Indonesia 0.59Germany 1.74Germany 1.42Canada 0.97France 1.58United Kingdom 2.87
Australia 0.98Australia 0.57Australia 0.93Korea 1.15Korea 0.67Italy 1.44Spain 2.24
TransportJapan 6.75Australia 0.62United States 1.93United States 1.88Canada 1.02United States 2.11United States 11.4
United States 4.77China 0.39Japan 1.46Japan 1.47Russian Federation 0.48United Kingdom 1.40Germany 9.65
China 4.40Indonesia 0.27Australia 1.04China 1.00Mexico 0.43Russian Federation 1.34United Kingdom 8.08
Germany 2.21Russian Federation 0.22Russian Federation 0.83Australia 0.50Japan 0.29France 1.26Italy 6.52
Australia 1.42Germany 0.14Korea 0.78Korea 0.43United Kingdom 0.28Spain 1.07Belgium 5.24
ChemicalAustralia 4.80Australia 2.14United States 4.37United States 1.18Canada 3.50United States 4.00Germany 5.46
Japan 4.14China 1.19Japan 3.97Japan 0.92Mexico 1.36Russian Federation 3.20United States 4.82
United States 3.96Indonesia 0.90Australia 2.87China 0.88Japan 0.83United Kingdom 2.99Italy 3.56
Indonesia 2.90Russian Federation 0.67Korea 2.10Korea 0.40Russian Federation 0.82Netherlands 2.03Spain 3.36
Russian Federation 2.56Germany 0.66Russian Federation 1.86Germany 0.37United Kingdom 0.82France 1.86United Kingdom 3.33
LogisticsUnited States 5.88Australia 0.62United States 1.93United States 1.88Canada 1.02United States 2.11United States 11.4
Japan 2.39China 0.39Japan 1.46Japan 1.47Russian Federation 0.48United Kingdom 1.40Germany 9.65
China 2.26Indonesia 0.27Australia 1.04China 1.00Mexico 0.43Russian Federation 1.34United Kingdom 8.08
Australia 1.69Russian Federation 0.22Russian Federation 0.83Australia 0.50Japan 0.29France 1.26Italy 6.52
Russian Federation 1.27Germany 0.14Korea 0.78Korea 0.43United Kingdom 0.28Spain 1.07Belgium 5.24
MetalJapan 6.07United States 1.29Japan 3.50Japan 1.55Canada 1.88Russian Federation 2.74Germany 6.70
China 3.74Indonesia 1.20Australia 2.98China 1.27China 1.22France 2.62United States 4.71
Australia 3.09China 0.98United States 2.23United States 1.27Mexico 1.07Italy 2.47Italy 4.16
United States 2.81Russian Federation 0.98Russian Federation 1.91Australia 0.62Japan 0.97United States 2.32Spain 3.78
Russian Federation 2.03Korea 0.64Brazil 1.50Korea 0.57Germany 0.67United Kingdom 2.31United Kingdom 3.21
Sources: OECD; and IMF staff estimates.

For each country in column i, cell (i, j) shows top five input source countries in FVA terms for its export product j (in row j). For example, China contrubutes 14.2 percent of total value added of Korea’s gross electrical exports.

For Japan, Japaneses contents refers to domestic value added to export in each sector

Sources: OECD; and IMF staff estimates.

For each country in column i, cell (i, j) shows top five input source countries in FVA terms for its export product j (in row j). For example, China contrubutes 14.2 percent of total value added of Korea’s gross electrical exports.

For Japan, Japaneses contents refers to domestic value added to export in each sector

References

    AroraVivekKarlHabermeierJonathan D.Ostry and RhodaWeeks-Brown (coordinators)2012The Liberalization and Management of Capital Flows: An Institutional ViewIMF Policy Paper (Washington: International Monetary Fund)

    FratzscherMarcelMarco LoDuca and RolandStraub2013On the International Spillovers of U.S. Quantitative EasingECB Working Paper Series 1557 (Frankfurt: European Central Bank).

    International Monetary Fund2013Global Financial Stability Report, April 2013, Chapter 1 (Washington: International Monetary Fund)

    International Monetary Fund2013aSpillover Report (Washington: International Monetary Fund).

    International Monetary Fund2013bGlobal Impact and Challenges of Unconventional Monetary Policies”; available via the Internet at: www.imf.org/external/np/pp/eng/2013/090313.pdf (Washington: International Monetary Fund).

    International Monetary Fund2013cUnconventional Monetary Policies—Recent Experience and Prospects”; available via the Internet at: https://www.imf.org/external/np/pp/eng/2013/041813a.pdf (Washington: International Monetary Fund).

    LagardeChristine2013The Global Calculus of Unconventional Monetary Policies”; available via the Internet at: http://www.imf.org/external/np/speeches/2013/082313.htm.

    LeeJisun2013Impact of Weak Yen in Exports, Expected to Increase in the Second Half of 2013LG Business Insight Weekly Focus Vol. 3 (July) (Seoul: LG Economic Research Institute).

    OhSekyung2013Analysis of the Impact of the Weak Yen on Korea’s Exports and ImportsWorld Economy Update Vol. 13 No. 3 (Seoul: Korea Institute for International Economic Policy).

    ReeJooK.Kyoungsoo Yoon and HailPark2012FX Funding Risks and Exchange Rate Volatility–Korea’s CaseIMF Working Paper 12/268 (Washington: International Monetary Fund).

    ShinoharaNaoyuki2014Unconventional Monetary Policies: Looking Ahead”; available via the Internet at: http://www.imf.org/external/np/speeches/2014/012314.htm.

This paper was prepared as part of research projects to underpin the 2013 Article IV Consultation with Korea. The authors would like to thank Isabelle Mateos y Lago for her insightful guidance and the seminar participants at the IMF APD and the Bank of Korea Research Institute Joint Seminar (Seoul, October 23, 2013).

IMF, “The Liberalization and Management of Capital Flows: An Institutional View,” November 14, 2012 (www.imf.org/external/np/pp/eng/2012/111412.pdf).

The model estimation suggests that a one percentage point increase in the net forward position (scaled by the total asset size) of a foreign bank branch is associated with a 0.3 percentage point increase in short-term external debt (scaled the same way) in the same quarter, and 0.2 percentage point increases in the subsequent quarters (see Jack Joo K. Ree, Kyoungsoo Yoon, and Hail Park, “FX Funding Risks and Exchange Rate Volatility-Korea’s Case,” IMF Working Paper, WP/12/268, November 2012).

This was a decline by 16 percent; the maximum decline during the 2008 global financial crisis (GFC) amounted to US$ 65 billion dollars or a peak- to- trough fall of 24 percent.

The regression analysis presented later does not allow us to decompose push and pull factors because of interaction variables. However, the IMF’s 2013 Spillover Report finds Korea at the borderline between emerging and advanced countries in terms of the significance of the global or regional common factors—which is larger for EMs—driving capital flows.

However, an exact quantification of the affects of the measures is not easy because they were introduced during a period when both demand (reflecting a sharp decline in demand for FX hedging from shipbuilders owing to decreases in new orders) and supply (owing to deleveraging of global banks) of external short-term debt were declining.

In concluding 2013 Article IV Consultation with Korea, the Executive Board of the International Monetary Fund “noted the widening of Korea’s external imbalance and the increased pressure on the exchange rate, which the staff assessed as being moderately undervalued in real effective terms.” (http://www.imf.org/external/np/sec/pr/2014/pr1420.htm).

The weekly model does not include QQME variables because of a lack of availability.

See Ree and others (2012), page 19, Figure 10.

In the graph, yield curve steepening is captured by movement along the VIX* curve to the left because CCR1Y_RATE is negative.

The interaction complicates the nonlinear response of bond flows to risk sentiments. For example, a QE expansion during a period in which VIX is below the threshold would increase the KTB term spread by dampening bond flows. However, the ensuing steeper yield curve will increase the threshold VIX*, which would reinforce the robustness of safe-haven behavior of bond flows.

The latest global financial turmoil indicates that a disorderly QE exit can indeed steeply increase VIX, which rose from 50 on May 22 to 62 on June 21, 2013. Then VIX rapidly backed down to about 40 through early August, since when fluctuating in a 20-45 range,

However, these findings should be read with due caution, bearing in mind that the behavior of capital flows during the periods of QE expansion (which occupy most of the sample period) may be different from one that would pan out during the periods of QE tapering, especially if it involves abrupt and nonlinear dynamics often seen during the bursting of financial bubbles. The non-linearity may set off tail risks, which could be chaotic and difficult to predict. That being said, Korea had not really had a capital flow-related credit bubble during the periods of QE expansion.

Existing studies show somewhat mixed pictures on Korea-Japan export similarity, with some finding substantially lower export similarity under finer product category classifications but others finding high proximity in product spectrum across both product functions and price ranges.

The two findings from the geographical and product dimension analyses may appear contradictory, but they are not. As an illustration, suppose two extreme assumptions: (i) none of Korea’s gains in China and ASEAN markets was at Japan’s loss, and (ii) all of Japan’s loss in automobile and electrical machinery was captured by Korea. In this case, the two major contributors to Korea’s market share gains (i.e., China and ASEAN markets on the one hand, and auto and electrical on the other) will be mutually exclusive. Even so, the sum of these two segments would not exceed 100 percent. In reality, there will be a considerably large intersection between the two sets, which would free space for other market segments to contribute to Korea’s market share gains.

A model using real effective exchange rates yielded a similar result with the peak level of response of Korea’s exports to a 10 percent depreciation declining from 1.4 percent (after four months of the shocks) for the sample period 1998-2005 to zero for 2006-13.

The specification is similar to recent studies done in Korea, whose findings are broadly consistent with ours. See Lee (2013) and KIEP (2013).

This is consistent with the IMF’s 2013 Spillover Report.

IMF, “Global Impact and Challenges of Unconventional Monetary Policies,” October 7, 2013b (www.imf.org/external/np/pp/eng/2013/090313.pdf) and IMF, “Unconventional Monetary Policies—Recent Experience and Prospects,” April 18, 2013c (https://www.imf.org/external/np/pp/eng/2013/041813a.pdf). The first paper also finds that an exit from UMPs, even a well-managed one, can cause disruptive spillovers to non-UMP countries, and both exiting too soon and too early can hurt all. It stresses the need to make use of the space provided by UMP for reforms (UMP countries), to carefully manage communication (UMP central banks), and strengthen resilience to spillover shocks (non-UMP countries). “The Global Calculus of Unconventional Monetary Policies,” a speech by IMF Managing Director also summarizes these points (http://www.imf.org/external/np/speeches/2013/082313.htm). Another helpful reference is a speech by IMF Deputy Managing Director Naoyuki Shinohara. “Unconventional Monetary Policies: Looking Ahead (http://www.imf.org/external/np/speeches/2014/012314.htm).

Other Resources Citing This Publication