Journal Issue
Share
Article

Asian Flu or Wall Street Virus? Price and Volatility Spillovers of the Tech and Non-Tech Sectors in the United States and Asia

Author(s):
Jorge Chan-Lau, and Iryna Ivaschenko
Published Date:
September 2002
Share
  • ShareShare
Show Summary Details

I. Introduction

There is a widespread belief among the business and policymaking communities that the financial system has become an increasingly important mechanism for the transmission and amplification of local shocks to the international economy, especially in recent years. This belief is supported by the fact that cross-country correlations between financial variables – in particular, stock markets – increase during periods of market stress.2 Several arguments have been advanced to explain why stock prices have become more correlated across countries, including the diversification of production across countries by multinationals, easy access to information through the Internet, increased equity issuance and securitization, and the growing importance of the telecommunications, media, and technology (TMT) sector.3

Empirical research has found that the importance of global factors in determining stock returns across countries relative to country-specific factors has increased significantly in recent years.4 In particular, both “new” and “old” economy stocks became especially correlated during the so called tech bubble period—a synchronous rise of TMT stock prices across countries during the period between late 1998 and early 2001, followed by an abrupt correction—while economic fundamentals were not as synchronized. This made some observers to speculate that TMT stocks may have become a new channel of transmitting shocks throughout the world financial markets.

This paper attempts to shed light on the issue by analyzing whether transmission mechanisms differ in the TMT and non-TMT sectors, and whether there were structural breaks in their dynamics. It studies the transmission mechanisms of the conditional first and second moments of daily returns across stock markets in the United States and the Asia-Pacific region across three different periods. The periods under study include the period preceding the rapid increase in stock prices in the technology sector (or the “tech bubble”), the “tech bubble” period, and its aftermath or the price-correction period.5 We are specially interested in assessing whether price changes and volatility spillovers were generated mainly from the United States – the Wall Street Virus hypothesis – or from Asian countries - the Asian Flu hypothesis. The analysis is based on TGARCH models, which directly model the time-varying behavior of expected stock returns and volatility conditioning them on all information currently available to investors. These models can capture the asymmetric effect of negative and positive returns on the conditional variance of returns.6

The study finds that price change spillovers between the United States and the Asia-Pacific region are asymmetric. U.S. stock markets play an important role in determining the price dynamics in the Asian stock markets regardless of the sector analyzed. China is the only country not affected by the United States. Price spillovers from the Asia-Pacific countries have little or no effect on the U.S. stock markets, especially across TMT stocks. Only during the price correction period did spillovers from Japan, Singapore and Thailand became significant for the U.S. non-TMT stock prices. These findings support the Wall Street Virus hypothesis rather than the Asian Flu hypothesis. One explanation of these findings is that the process of price discovery in the United States is more efficient, and hence, closely monitored and followed in other countries. Another interpretation is that prices in other markets tend to mimic those of the United States, a symptom of herding behavior abroad.7

The results also show that at the regional level, the importance of price spillovers from Japan, Hong Kong SAR, and Singapore in determining price returns in other Asia-Pacific stock markets has increased during the price-correction period. Spillover patterns in the TMT and non-TMT sector show substantially different price dynamics with few markets exhibiting the same patterns across sectors. Finally, the relative importance of volatility spillovers compared with price change spillovers is quite small, and in many cases is not significant.

The structure of the rest of the paper is as follows. Section II reviews the related literature. Section III describes the data used in the study. Section IV explains the empirical methodology. Section V presents the results. Section VI concludes.

II. Related Literature

There is a substantial amount of theoretical and empirical work on documenting and understanding how stock returns and volatility are transmitted across countries.8 On the theoretical side, a number of explanations based on the “revision of expectations” have been advanced. For example, King and Wadhawani (1990) argue that mistakes could be transmitted between two markets as domestic traders have to infer information imperfectly from foreign prices. Kodres and Pritsker (1998) suggest that the existence of feedback traders and asymmetric information could lead to the propagation of shocks through portfolio rebalancing effects. Calvo (1999), and Calvo and Mendoza (2000) argue that comovements in stock markets are caused by herd behavior among portfolio managers. The importance of information asymmetries highlighted by the studies cited above is partly supported by survey studies such as Shiller, Konya, and Tsutsui (1991). Other studies argue that technical factors such as margin calls and convergence trades could lead to increased comovements in stock markets, as in Kyle and Xiong (2001), Lagunoff and Schreft (1999), and Schinasi and Smith (2000), among others.

Among empirical studies, the simplest way to evaluate the interrelation of stock returns across different markets is to compare simple correlations in different periods. Examples of this approach are numerous, as exemplified by Hilliard (1979), Eun and Shim (1989), Roll (1988,1989), Bertero and Mayer (1990), Baig and Goldfjan (1999), and Kumar, Chan-Lau, Richards, and Sloek (2001).9 This approach, however, has been questioned by Boyer, Gibson and Loretan (1999), Rigobón and Forbes (1998), and Rigobón (2000), among others. These authors point out that even if the data is generated from the same multivariate normal distribution, the sample correlation obtained from a high volatility subsample would be higher than that corresponding to a low volatility subsample. The use of simple correlations, then, would lead to identify- structural breaks in the transmission mechanism which are nonexistent.

Other studies have preferred a different empirical approach based on ARCH models and their variants. This approach is supported by the fact that stock price volatility is time-varying and that high volatility episodes are usually characterized by a high correlation of stock market returns. Among these studies, our work is closely related to that of Hamao, Masulis, and Ng (1990). These authors studied price changes and volatility spillovers across the New York, Tokyo, and London stock markets using various univariate GARCH models. They found volatility spillovers only in the period following the October 1987 crash and identified an asymmetry, since Tokyo was affected by London and New York but did not have any effect on the latter two stock markets. Lin, Engle, and Ito (1994) used a signal extraction model with GARCH processes to analyze spillovers between Tokyo and New York. Once the close-to-close returns were decomposed into daytime and overnight returns, spillovers from Tokyo to New York were detected.

The use of the ARCH methodology to study stock market interdependences has not been limited to univariate models. Booth, Martikainen, and Tse (1997) analyzed the price and volatility spillovers across Scandinavian stock markets using a multivariate EGARCH model and found that Sweden was a major spillover source. Theodossiou, Kahya, Koutmos, and Christophi (1997) estimated a multivariate GARCH model using weekly stock market data to analyze returns spillover in the United States, Japan, and the United Kingdom during the October 1987 stock market crash. In contrast to Hamao, et al., they found that spillovers from Japan to the United Kingdom were significant.

III. Data

The analysis in this paper uses daily close-of-day (business days) stock market prices in the United States and a number of markets of the Asia-Pacific region including Australia, China, Hong Kong SAR, India, Japan, Korea, Malaysia, New Zealand, Philippines, Singapore, Taiwan Province of China and Thailand. Stock market prices correspond to the proprietary indices compiled by Primark Data Stream for two sectors: the Telecommunications, Media, and Technology (TMT) sector and the Total Stock Market Excluding the TMT (TXT) sector or non-TMT sector.

The selection of Primark Datastrream TMT and TXT indices is guided by the following considerations. First, the sectoral indices are constructed using the same methodology, which enhances the consistency of the results. Second, these indices are available as fixed-history indices. They are not recalculated historically when constituents change, which enables the effects of dead stocks to be seen on the index. This feature is especially important because the analysis of this paper is concerned with evaluating the aggregate impact of one equity market on the other. Finally, these data are available for a wide range of countries. With the exception of China and India, data is available for every country since January 2, 1990. All indices are expressed in local currency units since the goal of the paper is to analyze the transmission of stock market shocks across countries. Moreover, the use of local currency denominated indices helps to disentangle stock price effects from exchange rate effects.10

Data limitations force us to use close-to-close price returns that are calculated as changes in log closing prices, Rt =lnPt– ln Pt-1. Although the use of close-to-close returns may induce spurious cross-autocorrelation between stock markets because of asynchronous trading. However, results by Hamao et. al. (1990) show that correcting for this situation using open-to-close returns, not susceptible to the spurious cross-autocorrelation problem, do not change results significantly.

IV. Empirical Methodology

Generalized Autoregressive Conditional Heterokedastic (GARCH) models are used to analyze stock return and volatility spillovers across the different stock markets described in the previous section.11 In particular after some experimentation with model specifications using likelihood ratio (LR) statistics, the MA(1)-TGARCH(1,1) model was chosen because it provided a parsimonious specification that fitted the stock return series for all countries analyzed well.12 The choice of an asymmetric GARCH specification, TGARCH, was motivated by the apparent asymmetry in conditional volatility responses to negative and positive shocks, present in the stock return data.13 The MA(1) term was included to capture serial correlation present in stock index returns, following Bollerslev (1987) and French, Schwert, and Stambaugh (1987).14 While using a single model to fit all stock markets significantly facilitates cross-country comparisons, the results are generally robust to different model specifications that fitted domestic stock return data well.

The following TGARCH(1,1)-MA(1) model was used to test for spillovers in conditional mean and volatility across national stock markets:

where the price return in the stock market, R, is a linear function of its 1 day lagged-return, the lagged return on the foreign stock market which is assumed to be the source of spillovers (spillover market), Rt1f, and the error term, εt, that follows a MA(1) process. The foreign return lag is specified so that it accommodates the time differences between the United States and the Asia-Pacific region, and among the countries of the Asia-Pacific region. The conditional variance of the error term, σt2, is given by:

where the dummy variable, dt, is equal to 1 if the error term, εt, is positive, and zero otherwise.. The inclusion of the dummy variable allows capturing possible asymmetric effects of good news (positive εt) and bad news (negative εt). The regressors εt12,εt12dt1, and σt12 are commonly denominated as the GARCH, TARCH, and ARCH component respectively. Xtf is the volatility surprise from the spillover market, as in equation (1), and is lagged the same number of periods as the corresponding foreign stock return. Because the volatility surprise is an unobserved variable, econometric methods are needed to estimate it. In this paper, we follow the methodology first proposed by Hamao, Masulis, and Ng (1990).

Hamao et. al. (1990) suggest that the volatility surprise term correspond to the squared residual derived from a GARCH model estimated for spillover market without including other countries’ variables among the regressors. Hence, in this paper, the volatility surprise term was estimated as the squared residual derived from the following TGARCH(1,1)-M(1) model applied to the daily close-to-close return of the stock price index in the spillover market:

Equation (3) indicates that in any given period the price return, Rtf, is a function of its lag, the contemporaneous conditional variance, σt2, and a moving average error term. The dynamics of the conditional variance are given by equation (4). To simplify the description of the model, equations (1)-(2) are referred to as a spillover model, and equations (3)-(4), as a base model throughout the rest of the paper.

The base model was estimated for the three different periods, specified in the introduction: (i) the period prior to the boom in technology stock prices, from January 1, 1990 to August 31,1998 (called pre-bubble period); (ii) the technology boom period from September 1,1998 to March 27,2000 (bubble period); (iii) and the period of rapid technology stock price decline from March 28,2000 to May 1,2001 (post-bubble or price-correction period). The choice of the periods, although rather arbitrary, roughly reflects the pattern of stock price trends across the United States and the Asia-Pacific region.15 The estimation is performed for the TMT and TXT sectors for each country assumed to be a potential source of spillovers: the United States, Japan, Hong Kong SAR, Singapore, Malaysia, and Thailand. The United States and Japan were included as possible sources of spillovers to the Asian region because of their economic importance, Hong Kong SAR and Singapore because they are important regional financial centers, and Malaysia and Thailand because they were major sources of price and volatility spillovers during the 1997 Asian financial crisis.

For each sector and spillover market, volatility surprise terms were estimated from the base model, and used subsequently in the estimation of the spillover models for the remaining countries in the sample, for the corresponding sector. As in the case of the base model, the spillover models were estimated for the three different periods described above. Wald tests were conducted to evaluate the null hypothesis that spillover coefficients between the following two pairs of periods are the same: (i) technology bubble and pre-bubble periods, (ii) post-bubble and bubble periods, and (iii) post-bubble and pre-bubble. The main results are summarized in the next section.

V. Results

This section analyzes whether the United States, Japan, Hong Kong SAR, Singapore, Malaysia and Thailand were important sources of stock return and volatility spillovers in the Asian region, as well as whether spillovers from the five Asian stock markets above have had a significant impact in the United States stock markets.

Overall, the main results are: (ii) the United States has been the most important source of return spillovers to the Asia-Pacific region during the last ten years, but its relative importance during the post-bubble period has diminished as the importance of the stock markets in Japan, Hong Kong SAR, and Singapore has increased both in the TMT and non-TMT sectors; (iii) Asia-Pacific stock markets have had only a limited impact on the stock market price dynamics experienced in the United States, as there are neither large price changes nor volatility spillovers, regardless of the sector and period analyzed. In fact, the United States stock market has been one of the markets least affected by developments in foreign stock markets for the last ten years, both in the TMT and non-TMT sectors; (iv) generally, spillover patterns in the TMT and non-TMT sector differ across countries; (v) China and the United States are the least affected countries; in addition, Hong Kong SAR, and, especially, Korea have become the least affected countries since the tech-bubble period; and (vi) volatility spillovers are not as important as price change spillovers across all periods, irrespective of a spillover market chosen.

The results are summarized in Tables 1 to 3. Table 1 shows the three most and least affected countries for each of the spillover markets in the TMT and non-TMT sector, for each period. Table 2 presents the average spillover effects for all countries, as well as for the three most and least affected, for each period and for each sector of the stock market. Finally, Table 3 presents the spillover patterns. Table 4 to 8 present results for individual spillover markets. We proceed to review in detail the results obtained for the United States and the five Asian stock markets selected as a source of spillovers below.

Table 1:Countries Most And Least Affected By Price Spillover
Panel A: Most Affected Countries 1/
United StatesJapanHong Kong SARSingaporeMalaysiaThailand
TMTNon-TMTTMTNon-TMTTMTNon-TMTTMTNon-TMTTMTNon-TMTTMTNon-TMT
Australia1,212131
China
Hong Kong1,2,31,2,31,2,31,2,32,31,2,311,2,32,31,2,3
India1111,21,2
Indonesia3333
Japan32,33
Korea2,32,32,32,32,32,32,32,32,332,32,3
Malaysia111111
New Zealand1
Philippines3
Singapore211,2,31,22,3
Taiwan111
Thailand321,21,2,31,21,2,31,2,31,2
United States

1 = country affected in pre-bubble period, Jan. 2, 1990-Aug. 31, 1998; 2 = country affected in tech-bubble period, Sep. 1, 1998-Mar. 27, 2000; 3 = country affected in post-bubble period, Mar. 28, 2000-May 1, 2001.

1 = country affected in pre-bubble period, Jan. 2, 1990-Aug. 31, 1998; 2 = country affected in tech-bubble period, Sep. 1, 1998-Mar. 27, 2000; 3 = country affected in post-bubble period, Mar. 28, 2000-May 1, 2001.

Panel B: Least Affected Countries 2/
United StatesJapanHong Kong SARSingaporeMalaysiaThailand
TMTNon-TMTTMTNon-TMTTMTNon-TMTTMTNon-TMTTMTNon-TMTTMTNon-TMT
Australia332323
China1,2,31,2,331,31,2,331,2,31,2,31,2,31,2,31,2,3
Hong Kong3
India331,333333
Indonesia11,211,21,21,21,32,312
Japan2
Korea11111
Malaysia2,3222
New Zealand2221,22111
Philippines3333
Singapore222
Taiwan22,3
Thailand2
United States1,2,31,2,31,2,31,2,31,2,31,21,2,31,2,31,2,31,2,3

1 = country affected in pre-bubble period, Jan. 2, 1990-Aug. 31, 1998, 2 = country affected in tech-bubble period, Sep. 1, 1998-Mar. 27, 2000; 3 = country affected in post-bubble period, Mar. 2 8, 2000-May 1, 2001; if more than three countries are reported for any given period, spillovers were insignificant.

1 = country affected in pre-bubble period, Jan. 2, 1990-Aug. 31, 1998, 2 = country affected in tech-bubble period, Sep. 1, 1998-Mar. 27, 2000; 3 = country affected in post-bubble period, Mar. 2 8, 2000-May 1, 2001; if more than three countries are reported for any given period, spillovers were insignificant.

Table 2.Price Spillovers in the TMT and non-TMT sectors
TMT Sector

(average price change spillover)
All countriesThree most affected countriesThree least affected countries
PeriodPeriodPeriod
Pre-bubbleTech-bubblePost-bubblePre-bubbleTech-bubblePost-bubblePre-bubbleTech-bubblePost-bubble
United States0.220.260.230.340.400.430.110.130.04
Japan0.140.160.270.200.360.570.070.000.03
Hong Kong SAR0.140.140.210.240.340.380.030.020.01
Singapore0.110.090.210.220.210.430.010.000.00
Malaysia0.090.080.070.180.210.200.010.000.00
Thailand0.060.120.260.120.260.490.000.020.08
Non-TMT Sector

(average price change spillover)
All countriesThree most affected countriesThree least affected countries
PeriodPeriodPeriod
Pre-bubbleTech-bubblePost-bubblePre-bubbleTech-bubblePost-bubblePre-bubbleTech-bubblePost-bubble
United States0.300.320.250.490.590.480.050.050.06
Japan0.150.310.240.240.640.540.040.050.16
Hong Kong SAR0.170.230.280.310.530.630.070.000.03
Singapore0.310.240.250.670.530.600.070.000.03
Malaysia0.170.100.200.450.230.480.030.000.00
Thailand0.110.130.140.190.240.300.030.000.01
1/Price spillovers measured as the percentage change in price return corresponding to a one percent change in the price return in the country originating the spillover.
1/Price spillovers measured as the percentage change in price return corresponding to a one percent change in the price return in the country originating the spillover.
Table 3.Spillover Patterns from the United States, Japan, Hong Kong SAR, Singapore, Malaysia and Thailand
United StatesJapanHong Kong
TMTnon-TMTTMTnon-TMTTMTnon-TMT
AustraliaΓ
China
Hong Kong SARΓ
India
IndonesiaΓ
JapanΓ
KoreaΓΓ
MalaysiaΛ
New Zealand
PhilippinesΓ
SingaporeΓ
Taiwan POC
Thailand
United States
SingaporeMalaysiaThailand
TMTnon-TMTTMTnon-TMTTMTnon-TMT
Australia
ChinaΓΛ
Hong Kong SAR
India
Indonesia
JapanΓ
Korea
Malaysia
New Zealand
Philippines
Singapore
Taiwan POC
Thailand
United States

Spillover patterns

inverted v-shapedone time decline in post-bubble period
v-shapedΓone time increase in tech-bubble period
increasingone time increase in post-bubble period
decreasingno change
one time decline in tech-bubble period

Spillover patterns

inverted v-shapedone time decline in post-bubble period
v-shapedΓone time increase in tech-bubble period
increasingone time increase in post-bubble period
decreasingno change
one time decline in tech-bubble period

Table 4.Price and Volatility Spillovers from the United States
TMT Sector
1/1/1990-8/31/19989/1/1998-3/27/20003/28/2000-5/1/2001
CoefficientP-valueCoefficientP-valueCoefficientP-value
Australia
Price0.4010.0000.3010.0000.3160.000
Volatility0.0070.6020.0490.1730.0350.281
China
Price0.3050.2430.0670.2220.0290.418
Volatility-0.1240.166-0.0070.7010.0790.006
Hong Kong SAR
Price0.3610.0000.4510.0000.3440.000
Volatility0.0530.072-0.0310.126-0.0020.696
India
Price0.2670.0000.3300.0000.1160.009
Volatility0.1180.283-0.1000.466-0.0030.931
Indonesia
Price0.2130.2140.2570.0000.2780.000
Volatility-0.3050.000-0.0090.622-0.0070.326
Japan
Price0.2580.0000.2600.0000.4050.000
Volatility0.0090.457-0.0030.8120.0480.117
Korea
Price0.1480.0000.3820.0000.3950.000
Volatility0.0520.082-0.1580.0000.0280.006
Malaysia
Price0.1930.0000.2090.0070.0420.401
Volatility0.0310.2620.0650.4080.1600.018
New Zealand
Price0.2680.0000.2060.0000.1680.000
Volatility0.0410.3650.0420.1920.0350.146
Philippines
Price0.2430.0000.3030.0000.1450.000
Volatility0.0320.219-0.0240.778-0.0370.000
Singapore
Price0.1810.0000.1620.0020.2400.000
Volatility0.0020.9080.0750.1140.0540.222
Taiwan POC
Price0.2380.0000.22S0.0000.2770.000
Volatility0.0050.863-0.0250.5890.0210.267
Thailand
Price0.1680.0010.3520.0000.1830.000
Volatility-0.0140.631-0.1870.0000.0170.073
1/The price coefficient corresponds to the foreign country’s stock market return; the volatility coefficient corresponds to the foreign country’s volatility surprise.
1/The price coefficient corresponds to the foreign country’s stock market return; the volatility coefficient corresponds to the foreign country’s volatility surprise.
Table 4 (concluded).Price and Volatility Spillovers from the United States 1/
Non-TMT Sector
1/1/1990-8/31/19989/1/1998-3/27/20003/28/2000-5/1/2001
CoefficientP-valueCoefficientP-valueCoefficientP-value
Australia
Price0.4960.0000.3130.0000.1660.000
Volatility0.0150.0360.0060.2970.0240.249
China
Price0.3330.000-0.0240.669-0.0290.517
Volatility-0.4160.0000.0040.4730.0270.592
Hong Kong SAR
Price0.5370.0000.6290.0000.5150.000
Volatility0.0220.3990.0070.4570.3990.001
India
Price0.2020.0000.3390.0000.1140.408
Volatility2.4790.0320.0950.2850.0300.318
Indonesia
Price-0.1440.0990.1010.1070.2200.019
Volatility-0.1260.0000.0040.9090.0370.529
Japan
Price0.3410.0000.3240.0000.1690.014
Volatility0.0100.4440.0270.0840.1870.003
Korea
Price0.1370.0010.6040.0000.4780.001
Volatility0.0300.2900.1640.5080.2500.326
Malaysia
Price0.3420.0000.2260.0000.2030.003
Volatility0.0110.4450.0130.7970.0670.490
New Zealand
Price0.4290.0000.2480.0000.1870.000
Volatility0.0300.1160.0180.1710.0790.035
Philippines
Price0.2590.0000.2900.0000.1880.646
Volatility0.0600.0200.0390.235-0.1110.575
Singapore
Price0.3140.0000.4910.0000.3240.000
Volatility0.0180.277-0.0050.6700.3270.002
Taiwan POC
Price0.2860.0000.1960.0050.2310.005
Volatility-0.0030.905-0.0210.157-0.0280.359
Thailand
Price0.3790.0000.4280.0000.3290.001
Volatility0.0780.0660.0460.027-0.0280.667

The price coefficient corresponds to the foreign country’s stock market return; the volatility coefficient corresponds to the foreign country’s volatility surprise.

The price coefficient corresponds to the foreign country’s stock market return; the volatility coefficient corresponds to the foreign country’s volatility surprise.

Table 5.Price and Volatility Spillovers from Japan 1/
TMT Sector
1/1/1090-8/31/19989/1/1998-3/27/20003/28/2000-5/1/2001
CoefficientP-valueCoefficientP-valueCoefficientP-value
United States
Price0.0570.0000.0160.7180.1390.036
Volatility-0.0010.784-0.0040.7250.0070.838
Australia
Price0.1750.0000.2790.0000.3580.000
Volatility0.0090.168-0.0020.9070.0700.071
China
Price0.1120.161-0.0540.3520.0360.371
Volatility0.0560.7400.0150.3420.0280.415
Hong Kong SAR
Price0.2470.0000.4100.0000.6160.000
Volatility0.0100.4640.0390.2070.0230.436
India
Price0.2130.0000.1300.0790.1890.482
Volatility0.1160.194-0.0510.159-0.0470.817
Indonesia
Price-0.0280.1960.1470.0320.5340.000
Volatility-0.1260.1770.0110.8740.0020.952
Korea
Price-0.0030.9110.3080.0020.5260.000
Volatility0.0050.6560.1260.1760.1570.038
Malaysia
Price0.1700.0000.1170.0280.1680.002
Volatility0.0030.7990.0070.8510.0550.599
New Zealand
Price0.1220.0000.1700.0000.2110.000
Volatility0.0050.6580.0020.8760.0780.124
Philippines
Price0.1340.0000.1530.0010.2140.001
Volatility0.0170.287-0.0140.7390.3130.021
Singapore
Price0.1310.0000.0790.1080.2500.000
Volatility0.0000.9380.0710.2390.1010.165
Taiwan POC
Price0.1550.0000.1990.0010.2300.000
Volatility0.0050.7090.0080.850-0.0210.395
Thailand
Price0.1630.0000.2560.0040.2850.000
Volatility0.0110.4460.1230.3140.0070.757
Non-TMT Sector
1/1/1990-8/31/19989/1/1998-3/27/20003/28/2000-5/1/2001
CoefficientP-valueCoefficientP-valueCoefficientP-value
United States
Price0.0540.0000.0470.3320.1160.048
Volatility0.0000.9310.0000.9830.0740.069
Australia
Price0.2080.0000.2700.0000.1460.000
Volatility0.0130.0370.0370.0360.0220.085
China
Price-0.0190.700-0.2270.006-0.0640.105
Volatility-0.0700.048-0.0450.0000.0420.299
Hong Kong SAR
Price0.2830.0000.6370.0000.4620.000
Volatility0.0170.2550.1070.1410.1380.147
India
Price0.1040.0150.2610.0060.2240.032
Volatility-0.0250.0000.2330.0920.0640.012
Indonesia
Price-0.0580.2050.0190.7600.3050.002
Volatility-0.0580.000-0.0440.2790.0800.341
Korea
Price0.0450.1140.7140.0000.8110.000
Volatility0.0320.0020.1260.4800.1110.579
Malaysia
Price0.2080.0000.3130.0000.2170.001
Volatility0.0030.3730.1180.1040.0450.513
New Zealand
Price0.1280.0000.1510.0000.0880.035
Volatility0.0050.167-0.0010.8480.0630.154
Philippines
Price0.0340.1080.2330.0000.4030.000
Volatility0.0020.5970.1150.003-0.0350.119
Singapore
Price0.1690.0000.5000.0000.2880.000
Volatility0.0120.0730.0450.2880.2520.031
Taiwan POC
Price0.1550.0000.2720.0000.1510.033
Volatility0.0190.204-0.0180.078-0.0270.083
Thailand
Price0.1450.0000.5760.0000.2310.021
Volatility0.0130.1920.3450.393-0.0220.575

The price coefficient corresponds to the foreign country’s stock market return; the volatility coefficient corresponds to the foreign country’s volatility surprise.

The price coefficient corresponds to the foreign country’s stock market return; the volatility coefficient corresponds to the foreign country’s volatility surprise.

Table 6.Price and Volatility Spillovers from Hong Kong SAR 1/
TMT Sector
1/1/1990-8/31/19989/1/1998-3/27/20003/28/2000-5/1/2001
CoefficientP-valueCoefficientP-valueCoefficientP-value
United States
Price0.0420.0020.0470.1390.1170.022
Volatility0.0020.4200.0150.0580.0180.233
Australia
Price0.1880.0000.0900.0420.2550.000
Volatility0.0050.1140.0140.3350.0230.038
China
Price-0.0220.6860.0620.2180.0870.002
Volatility0.0320.4190.1070.053-0.0090.028
India
Price0.2470.0000.1510.0100.1410.009
Volatility0.2720.0000.0020.934-0.0090.697
Indonesia
Price0.0480.1130.1490.0060.2660.000
Volatility-0.0010.602-0.0120.5400.2950.001
Japan
Price0.1460.0000.2070.0000.3410.000
Volatility0.0050.0970.0240.0530.0130.141
Korea
Price0.0650.0040.4000.0000.5130.000
Volatility0.0090.3510.1180.0470.0670.006
Malaysia
Price0.2340.0000.1510.0810.1590.001
Volatility0.0600.004-0.0200.0450.1180.023
New Zealand
Price0.0910.0000.0840.0070.1260.001
Volatility0.0540.0030.0040.7140.0560.013
Philippines
Price0.1840.0000.1620.0000.1650.006
Volatility0.0170.0120.0340.319-0.0150.411
Singapore
Price0.1390.0000.1040.0100.3050.000
Volatility0.0140.0080.0560.0140.0710.005
Taiwan
Price0.1850.0000.0960.0240.1390.010
Volatility0.0180.1050.0050.6830.0090.641
Thailand
Price0.2350.0000.3830.0000.2410.000
Volatility0.0580.0000.1430.0810.0070.357
Non-TMT Sector
1/1/1990-8/31/19989/1/1998-3/27/20003/28/2000-5/1/2001
CoefficientP-valueCoefficientP-valueCoefficientP-value
United`States
Price0.0540.0000.0170.6350.0160.754
Volatility0.0010.573-0.0030.5250.0060.782
Australia
Price0.2040.0000.1700.0000.1080.000
Volatility0.0230.0000.0050.0000.0170.261
China
Price0.0400.2270.0230.6350.0140.702
Volatility-0.0520.0000.1390.1490.0360.101
India
Price0.3780.0000.3560.0000.1480.104
Volatility0.0240.8530.0600.2790.1850.177
Indonesia
Price0.0010.7900.0950.0290.2780.000
Volatility-0.0110.000-0.0140.166-0.0290.407
Japan
Price0.1640.0000.2810.0000.2870.000
Volatility0.0070.0040.0080.3060.0220.202
Korea
Price0.0940.0000.5740.0000.8650.000
Volatility0.0100.0870.0320.0000.2940.027
Malaysia
Price0.3200.0000.1630.0010.2870.000
Volatility0.0550.0000.0560.2160.0310.323
New Zealand
Price0.1120.0000.1180.0000.1520.000
Volatility0.0060.0570.0080.2090.0070.314
Philippines
Price0.1610.0000.1640.0000.5570.000
Volatility0.0260.0010.0950.013-0.0180.381
Singapore
Price0.2650.0000.5240.0000.5340.000
Volatility0.0580.0000.0630.0210.0310.283
Taiwan
Price0.1850.0000.1910.0000.3130.000
Volatility-0.0010.715-0.0090.097-0.0480.156
Thailand
Price0.3320.0000.5600.0000.4460.000
Volatility0.0530.0010.3060.0910.0120.832

The price coefficient corresponds to the foreign country’s stock market return; the volatility coefficient corresponds to the foreign country’s volatility surprise.

The price coefficient corresponds to the foreign country’s stock market return; the volatility coefficient corresponds to the foreign country’s volatility surprise.

Table 7.Price and Volatility Spillovers from Singapore 1/
TMT Sector
1/1/1990-8/31/19989/1/1998-3/27/20003/28/2000-5/1/2001
CoefficientP-valueCoefficientP-valueCoefficientP-value
United States
Price0.0370.0190.0200.6450.0370.019
Volatility0.0070.0030.0270.1580.0070.003
Australia
Price0.1090.0000.1130.2220.1390.003
Volatility0.0000.995-0.0360.3730.0950.028
China
Price-0.0940.0970.0090.8570.0010.983
Volatility0.0900.305-0.0060.7390.0100.253
Hong Kong SAR
Price0.1860.0000.2120.0000.2920.000
Volatility0.0250.0730.0410.3100.0710.021
India
Price0.1440.0000.1880.0240.1110.092
Volatility0.1090.0090.5560.003-0.0030.791
Indonesia
Price-0.0310.344-0.0140.8100.4460.000
Volatility-0.0080.0790.0580.3250.2160.016
Japan
Price0.0740.0000.0840.0200.2330.000
Volatility0.0040.3510.0030.7980.1270.067
Korea
Price0.0250.4020.1520.0800.5250.000
Volatility0.0180.0770.1330.3700.6260.000
Malaysia
Price0.1820.0000.0160.8080.2420.000
Volatility-0.0010.8780.1660.0080.1240.058
New Zealand
Price0.0420.0150.0570.1420.1450.001
Volatility-0.0090.1930.0280.3280.0390.097
Philippines
Price0.1280.0000.1030.0300.0150.730
Volatility0.0060.3460.0750.3410.0530.038
Taiwan POC
Price0.1120.0010.0880.1380.2440.000
Volatility-0.0010.9200.0760.3830.0040.709
Thailand
Price0.1790.0000.1930.0400.2520.000
Volatility0.0020.8770.8290.0310.1430.011
Non-TMT Sector
1/1/1990-8/31/19989/1/1998-3/27/20003/28/2000-5/1/2001
CoefficientP-valueCoefficientP-valueCoefficientP-value
United States
Price0.0690.0000.0230.4780.1430.008
Volatility0.0010.3660.0000.8850.0590.063
Australia
Price0.2890.0000.1490.0000.0870.001
Volatility0.0580.000-0.0010.3560.0290.025
China
Price-0.0520.4020.0390.3260.0040.928
Volatility-0.0140.094-0.0280.0000.0340.015
Hong Kong SAR
Price0.6600.0000.5490.0000.5770.000
Volatility0.0590.0040.0020.7400.0760.001
India
Price0.5190.0000.3780.0000.2310.014
Volatility0.0200.8600.2250.0130.0490.304
Indonesia
Price0.1230.0580.0530.1240.2600.001
Volatility-0.0270.000-0.0180.007-0.0090.760
Japan
Price0.2750.0000.2250.0000.2110.000
Volatility0.0200.0140.0050.5170.0480.057
Korea
Price0.1900.0000.5210.0000.7770.000
Volatility0.0290.0210.2190.0750.1770.152
Malaysia
Price0.7580.0000.2280.0000.2290.000
Volatility0.0660.0000.0530.0830.0290.001
New Zealand
Price0.2020.0000.1150.0000.1280.000
Volatility0.0190.0290.0050.3200.0680.003
Philippines
Price0.3300.0000.2060.0000.0290.673
Volatility0.0430.0010.0870.0040.3460.410
Taiwan POC
Price0.2760.0000.2260.0000.2360.000
Volatility-0.0010.4320.0000.935-0.0140.242
Thailand
Price0.6120.0000.6370.0000.4790.000
Volatility0.1040.0030.2460.036-0.0150.620

The price coefficient corresponds to the foreign country’s stock market return; the volatility coefficient corresponds to the foreign country’s volatility surprise.

The price coefficient corresponds to the foreign country’s stock market return; the volatility coefficient corresponds to the foreign country’s volatility surprise.

Table 8.Price and Volatility Spillovers from Thailand 1/
TMT Sector
1/1/1990-8/31/19989/1/1998-3/27/20003/28/2000-5/1/2001
CoefficientP-valueCoefficientP-valueCoefficientP-value
United States
Price0.0100.2030.0190.4920.0920.185
Volatility0.0010.1120.0130.0030.0020.893
Australia
Price0.0680.0000.0480.0140.1650.000
Volatility0.0000.6220.0100.3000.0860.009
China
Price-0.0100.2550.0500.0540.0090.834
Volatility-0.0120.000-0.0020.6420.0160.015
Hong Kong SAR
Price0.0990.0000.2210.0000.4120.000
Volatility0.0320.0000.0010.8580.0330.035
India
Price0.1500.0000.2600.0000.2100.006
Volatility0.0610.0010.1620.0090.0650.379
Indonesia
Price00120.5240.0870.0050.4330.000
Volatility0.0050.191-0.0260.000-0.0250.072
Japan
Price0.0520.0000.0640.0360.3130.000
Volatility0.0040.0100.0020.7860.0030.839
Korea
Price0.0510.0010.3060.0000.5440.000
Volatility-0.0010.6620.0500.2890.1070.257
Malaysia
Price0.1170.0000.1190.0010.2120.003
Volatility0.0130.0210.0230.2780.0170.042
New Zealand
Price0.0200.1050.0700.0070.1880.000
Volatility0.0120.0230.0150.0770.0410.257
Philippines
Price0.0970.0000.1510.0000.2630.018
Volatility0.0510.0010.0180.2160.1950.127
Singapore
Price0.0620.0000.0930.0020.2030.008
Volatility0.0000.7740.0310.0090.2850.112
Taiwan POC
Price0.0720.0000.0970.0170.3650.000
Volatility0.0040.1690.0220.0520.0220.340
Non-TMT Sector
1/1/1990-8/31/19989/1/1998-3/27/20003/28/2000-5/1/2001
CoefficientP-valueCoefficientP-valueCoefficientP-value
United States
Price0.0330.0000.0230.2540.0140.620
Volatility0.0010.1110.0030.219-0.0050.320
Australia
Price0.0840.0000.0980.0000.0360.041
Volatility0.0170.0000.0040.0620.0110.452
China
Price0.0210.5540.0210.3890.0090.752
Volatility-0.0860.000-0.0020.4930.0250.202
Hong Kong SAR
Price0.2110.0000.2460.0000.2540.000
Volatility0.0510.000-0.0020.5090.0680.113
India
Price0.3290.0000.2040.0000.0740.230
Volatility0.0050.9340.0450.1550.0350.000
Indonesia
Price0.0980.0030.0170.5630.0460.379
Volatility-0.0190.0000.0070.4640.0000.996
Japan
Price0.0740.0000.1180.0000.0820.010
Volatility0.0060.0070.0160.048-0.0020.674
Korea
Price0.0920.0000.3590.0000.4500.000
Volatility0.0100.0120.0230.5100.2470.020
Malaysia
Price0.1850.0000.1060.0010.1460.000
Volatility0.0180.0480.0170.2150.0320.562
New Zealand
Price0.0540.0000.0850.0000.0530.007
Volatility0.0020.1120.0040.006-0.0020.375
Philippines
Price0.1090.0000.1200.0000.2530.005
Volatility0.0420.0000.0310.0460.0530.312
Singapore
Price0.1580.0000.2620.0000.2370.000
Volatility0.0300.0000.0410.0120.0710.158
Taiwan POC
Price0.1030.0000.0590.0550.1810.000
Volatility-0.0010.3930.0010.787-0.0070.706

The price coefficient corresponds to the foreign country’s stock market return; the volatility coefficient corresponds to the foreign country’s volatility surprise.

The price coefficient corresponds to the foreign country’s stock market return; the volatility coefficient corresponds to the foreign country’s volatility surprise.

A. United States

Overall, return spillovers in the United States, as measured by the impact of the previous day price returns in the United States on the price return of Asia-Pacific stock markets, are significant in both the TMT and non-TMT sectors (Table 4). The sole exception in the Asia-Pacific region is China, which is not affected by spillovers from the United Sates, for both sectors. In the TMT sector, a one percent change in the stock return in the United States translated in an average 0.25 percent change in returns in the Asian region during the pre-bubble period, a 0.30 percent change in returns during the bubble period, and a 0.23 percent change in returns during the post-bubble period. The corresponding stock return changes for the non-TMT sector are somewhat higher – 0.35 percent, 0.34 percent, and 0.27 percent in the pre-bubble, bubble, and post-bubble periods respectively. On average, the magnitude of spillovers increases significantly during the bubble period in both sectors.

There is significant variation in the magnitude of spillovers across countries. In the TMT sector, average price spillovers in the three most affected countries are 0.34 percent, 0.40 percent, and 0.43 percent in the pre-bubble, tech-bubble and post-bubble periods, respectively. In the three least affected countries they are considerably smaller – 0.11, 0.13, and 0.04 percent in the same periods. The same large variation of spillovers is also observed in the non-TMT sector, where the average price spillovers in the three most affected countries are 0.49, 0.59, and 0.48 percent in the pre-bubble, tech-bubble and post-bubble periods, respectively. In the three least affected countries the spillovers are just about 0.05 percent during all three periods.

Spillovers pattern are also significantly different in the TMT and non-TMT sector in each country. The only exceptions are China, where spillovers were insignificant; New Zealand, where spillovers in the TMT and non-TMT sector decline continuously from the pre-bubble period to the post-bubble period; and Taiwan Province of China, where spillovers in the TMT and non-TMT sector decline during the tech-bubble period but recover in the post-bubble period (see Table 3). In addition, Wald test results indicate that price spillovers return to their pre-bubble levels in Singapore for the TMT sector; and in Malaysia and Thailand for the non-TMT sector.

With respect to volatility spillovers, the results indicate that they become significant in the TMT sector during the tech-bubble period. However, their magnitude is small, and only in Korea they are significantly different from zero. In the non-TMT sector, volatility spillovers tend to be significant in the pre-bubble period but their importance diminishes in the subsequent periods. No country shows significant volatility spillovers from the United States in more than one period. Finally, the results also indicate that volatility in both the TMT and non-TMT sectors in the United States have been negligibly affected by volatility spillovers from other countries’ TMT stock markets.

B. Japan

In the TMT sector, average price spillovers from Japan to other countries in the Asia-Pacific region are lower than those from the United States in the pre-bubble and tech-bubble periods, but become higher in the post-bubble period (see Tables 2 and 5). In the non-TMT sector, average price spillovers are lower than those from the United States only in the pre-bubble period, and roughly the same in the subsequent two periods. However, price spillovers in the three most affected countries are higher than those corresponding to the United States during the bubble and post-bubble period, amounting to 0.64 and 0.54 percent in those periods respectively, compared with only 0.59 and 0.48 percent in the United States,

In contrast to the United States, TMT spillovers from Japan did not increase significantly during the bubble period (i.e. the inverted V-shaped spillovers are not observed, see Table 3). The opposite is true for the non-TMT sector, where inverted V-shaped patterns predominate in 9 out of 13 countries, suggesting a surge in spillovers from the Japanese non-TMT sector during the bubble period. As in the case of the United States, spillover patterns were different between TMT and non-TMT sectors in each country, with the exception of Indonesia, which shows a one time-increase in the bubble period, and the United States, which shows a common V-shaped pattern. Spillovers to China are negative in both sectors, so that positive price changes in Japan are reflected in negative price changes in China.

It is worth noting that while before and during the bubble spillovers from Japan in the non-TMT sectors are higher than in the TMT sector, after the bubble spillovers in the TMT sector become higher than in the other sector.

C. Hong Kong SAR

On average, price change spillovers in both the TMT and non-TMT sectors increase throughout the three periods analyzed (see Table 2). Compared to Japan and the United States, TMT spillovers from Hong Kong SAR have been lower, especially in the price-correction period (see Table 6). Spillovers from the non-TMT sector increased almost twofold on average, from 0.15 percent in the pre-bubble period to 0.28 in post-bubble period. On average, spillovers from the Hong Kong SAR’s non-TMT sector to other Asia-Pacific countries have become more important relative to those from the United States and Japan in the post-bubble period.

Price spillover patterns in the TMT and non-TMT sectors in each country are the same only for Korea, which shows an increasing pattern, and Malaysia, which shows a V-shaped pattern (see Table 3). Interestingly, the V-shaped pattern predominated in six of the markets under study, pointing towards a declining influence from the Hong Kong SAR TMT sector during the bubble period. In contrast, there is no predominant spillover pattern in the non-TMT sector.

D. Singapore

The importance of Singapore as a source of price spillovers in the TMT sector is similar to that of Hong Kong SAR (see Table 2 and Table 7). It should be noted, though, that the magnitude of the spillovers to the three most affected countries have increased relative to that of Hong Kong SAR since the bubble period, and matches the magnitude of spillovers from the United States. In the non-TMT sector, the average price spillover effect from Singapore declined from 0.31 percent in the pre-bubble period to 0.25 percent in the post-bubble period. However, for the three most affected countries, Singapore is the most important source of spillovers in the pre-bubble period, and the second most important one in the bubble and post-bubble periods.

Among the affected countries, only Malaysia showed the same spillover pattern in the TMT and non-TMT sector (increasing pattern, see Table 3). The predominant increasing spillover pattern in the TMT sector suggests a more influential role of Singapore as a regional spillover source that may be explained by the increased importance of information-and computer-related manufacturing activities as a share of Singapore’s GDP. The evidence in the non-TMT sector is mixed, as four markets exhibit continuing spillover declines (Australia, Japan, Malaysia and the Philippines) and other five exhibit a V-shaped pattern (Hong Kong SAR, Indonesia, New Zealand, Taiwan Province of China and the United States).

E. Malaysia

The Malaysian stock market is the least important of all the possible TMT spillover sources analyzed in this study so only average results are reported in this paper. Relative to the United States and the Asian markets described above, spillovers effects from the Malaysian TMT sector are roughly equal to one third of the effects from the countries listed above (see Table 2). Spillovers from the TMT sector to the three most affected countries have remained constant during the last ten years, and equal to 0.20 percent. In the non-TMT sector, spillovers recovered to 0.48 percent in the post-bubble period after experiencing a decline from 0.45 percent in the pre-bubble period to 0.23 percent in the bubble period.

Spillovers from the TMT and non-TMT sectors exhibited the same pattern in three countries: Korea (increasing), New Zealand and Singapore (V-shaped) (see Table 3). There was no predominant pattern in the TMT sector, while the predominant spillover pattern in the non-TMT sector is V-shaped (6 countries). This suggests an increasing importance of the Malaysian non-TMT sector in the Asian region, though the magnitude of the spillovers remain low.

F. Thailand

TMT spillovers from Thailand surged dramatically during the last ten years: they doubled from 0.06 percent in the pre-bubble period to 0.12 percent in the bubble period, and then doubled again to 0.26 percent in the post-bubble period (see Table 2 and Table 8). It is worth noting that in the post-bubble period, the magnitude of the Thai TMT spillovers is only surpassed by those from Japan.

Regarding non-TMT sector spillovers, Thailand seems to be one of the least important player. On average, the spillovers never surpassed 0.15 percent. The numbers are two to three times as high when only the three most affected countries are considered. However, the magnitude of the spillovers is still about one half of the other countries analyzed.

The spillovers patterns across the TMT and non-TMT sectors were the same for Hong Kong SAR and Korea (increasing), and India (inverted-V shaped) (see Table 3). The increasing importance of Thai TMT spillovers described in the previous paragraph is also supported by the fact that 9 out of 13 countries exhibited an increasing pattern. In the non-TMT sector, four countries exhibit a V-shaped pattern (Indonesia, Malaysia, Taiwan Province of China, and the United States) while other five countries exhibit an inverted V-shaped pattern (Australia, India, New Zealand, the Philippines and Singapore).

VI. Conclusions

This study contributes to the knowledge of price dynamics in the technology and non-technology sectors of the stock market in the United States and a number of Asian countries, by focusing on price changes and volatility spillovers across countries in both sectors prior to, during, and after the run-up in tech stock prices.

The empirical findings suggests that, when it comes to price change spillovers between the United States and the Asian region, the Wall Street virus is the most important carrier of the “disease” rather than the Asian flu, regardless of stock market sector. The results also point out that spillover mechanisms have experienced structural changes during the periods analyzed, and that spillover patterns in the TMT and non-TMT sectors in each country differ significantly.

In sharp contrast to results in some earlier papers, volatility spillovers do not appear to play an important role, as the volatility surprise term was insignificant in most of the cases. This result is likely driven by allowing for asymmetric effects in the conditional variance Some caution is warranted in deriving policy conclusions from the results, though, as the analysis did not look in detail into particular crisis episodes, such as the 1997 Asian financial crisis or the 1998 Long-Term Capital Management debacle, which took place during the period analyzed.

References

    BacaS.P.B.Garbe and R.A.Weiss2000The Rise of Sector Effects in Major Equity MarketsFinancial Analysts JournalSeptember/October35-40.

    • Search Google Scholar
    • Export Citation

    BaigT. and I.Goldfajn1999Financial Market Contagion in the Asian CrisisIMF Staff Papers46167-95.

    BerteroE. and C.Mayer1990Structure and Performance: Global Interdependence of Stock Markets Around the Crash of October 1987European Economic Review341155-80.

    • Search Google Scholar
    • Export Citation

    BlackF1976Studies of Stock Price Volatility Changes” in Proceedings of the 1976 Meetings of the Business and Economic Statistics Section American Statistical Association177-81.

    • Search Google Scholar
    • Export Citation

    BollerslevT.1987A Conditionally Heteroskedastic Time Series Model for Speculative Prices and Rates of ReturnReview of Economics and Statistics69 pp. 542-47.

    • Search Google Scholar
    • Export Citation

    BollerslevT.R.Chou and K.Kroner1992ARCH Modeling in Finance: A Review of the Theory and Empricial EvidenceJournal of Econometrics525-60.

    • Search Google Scholar
    • Export Citation

    BollerslevT.R.F.Engle and D.Nelson1994ARCH ModelsChapter 49in R.F.Engle and D.McFadden(eds.) Handbook of Econometrics (Elsevier, New York).

    • Search Google Scholar
    • Export Citation

    BoothG.G.T.Martikainend and Y.Tse1997Price and Volatility Spillovers in Scandinavian Stock MarketsJournal of Banking and Finance21811-23.

    • Search Google Scholar
    • Export Citation

    BoyerB.H.M.S.Gibson and M.Loretan1997Pitfalls in Tests for Changes in CorrelationsInternational Finance Discussion Paper 597 (Washington, DC: Board of Governors of the Federal Reserve).

    • Search Google Scholar
    • Export Citation

    BrooksR. and L.Catao2000The New Economy and Global Stock ReturnsIMF Working Paper WP/00/216International Monetary FundWashington, DC.

    • Search Google Scholar
    • Export Citation

    CalvoG.A.1999Contagion in Emerging Markets: When Wall Street is a Carrier” (unpublished; College Park, Maryland: University of Maryland).

    • Search Google Scholar
    • Export Citation

    CalvoG.A. and E.G.Mendoza2000Rational Contagion and the Globalization of Securities MarketsJournal of International Economics5179-113.

    • Search Google Scholar
    • Export Citation

    CohenK.G.HawawiniS.MalerR.Schwartz and D.Whitcomb1980Implications of Microstructure Theory for Empirical Research on Stock Market BehaviorJournal of Finance35 pp. 249-57.

    • Search Google Scholar
    • Export Citation

    De BandtO. and P.Hartmann2000Systemic Risk: A SurveyECB Working Paper No. 35 (Frankfurt, Germany: European Central Bank).

    • Search Google Scholar
    • Export Citation

    EunC. and S.Shim1989International Transmission of Stock Market MovementsJournal of Financial and Quantitative Analysis24241-56.

    • Search Google Scholar
    • Export Citation

    FrenchK.G.Schwert and R.Stambaugh1987Expected Stock Returns and VolatilityJournal of Financial Economics19 pp. 3-29.

    FloodR. and P.Garber1980Market Fundamentals versus Price-Level Bubbles: The First TestsJournal of Political Economy88 pp. 745-70

    • Search Google Scholar
    • Export Citation

    GlostenL.R.R.Jaganathan and D.Runkle1993On the Relation Between the Expected Value and the Volatility of the Normal Excess Return on StocksJournal of Finance481779-1801.

    • Search Google Scholar
    • Export Citation

    HamaoY.R.W.Masulis and V.Ng1990Correlations in Price Changes and Volatility Across International Stock MarketsReview of Financial Studies3281-307.

    • Search Google Scholar
    • Export Citation

    HentschelL.1995All in the Family: Nesting Symmetric and Asymmetric GARCH ModelsJournal of Financial Economics3971-104.

    HilliardJ.1979The Relationship Between Equity Indices on World ExchangesJournal of Finance34103-14.

    International Monetary Fund2001World Economic Outlook (Washington, DC: International Monetary Fund).

    JarrowR.(editor) Volatility: New Estimation Techniques for Pricing Derivatives (LondonRisk Books).

    KingM. and S.Wadhwani1990Transmission of Volatility Between Stock MarketsReview of Financial Studies35-33.

    KingM.A.Sentana and S.Wadhwani1994Volatility and Links Between National Stock MarketsEconometrica62901-33.

    KodresL.E. and M.Pritsker1998A Rational Expectations Model of Financial ContagionFinance and Economics Discussion Paper Series 98/48 (Washington, DC: Board of Governors of the Federal Reserve System); forthcoming in Journal of Finance.

    • Search Google Scholar
    • Export Citation

    KumarM.J. A.Chan-LauA.Richards and T.Sloek2001Turbulence in Global Stock Markets: Trends and Correlations”(unpublishedWashington, DC: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    KyleA.S. and W.Xiong2001Contagion as a Wealth EffectJournal of Finance561401-40.

    LagunoffR. and S.L.Schreft1999Financial Fragility with Rational and Irrational ExuberanceJournal of Money Credit and Banking31531-60.

    • Search Google Scholar
    • Export Citation

    LinW.-L.R.F.Engle and T.Ito1994Do Bulls and Bears Move Across Borders? International Transmission of Stock Returns and VolatilityReview of Financial Studies7507-38.

    • Search Google Scholar
    • Export Citation

    RigobónR.2000Contagion: How to Measure ItmimeoMIT.

    RigobónR. and K.Forbes1998No Contagion, Only Interdependence: Measuring Stock Market ComovementsNBER Working Paper 7267 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation

    RollR.1988The International Crash of October 1987Financial Analysts JournalSeptember/October19-35.

    RollR.1989Price Volatility, International Market Links, and Their Implications for Regulatory PoliciesJournal of Financial Services Research3211-46.

    • Search Google Scholar
    • Export Citation

    SchinasiG.J. and R.T.Smith2000Portfolio Diversification, Leverage, and Financial ContagionIMF Staff Papers47159-76.

    ScholesM. and J.Williams1977Estimating Betas from Nonsynchronous DataJournal of Financial Economics17 pp. 197-210.

    ShillerR.J.F.Knoya and Y.Tsutsui1991Investor Behavior in the October 1987 Stock Market Crash: The Case of JapanJournal of the Japanese and International Economies51-13.

    • Search Google Scholar
    • Export Citation

    ZakoianJ.M.1994Threshold Heteroskedastic ModelsJournal of Economic Dynamics and Control18931-944.

1We would like to thank Yasushi Hamao, Donald Mathieson, Paul Soderlind, and Christopher Towe for their comments. All errors and omissions remain our sole responsibility.
2For example, King, Sentana, and Wadhwani (1994) and King and Wadhwani (1990) found that sharp changes in one stock market transmitted quickly to other stock markets during the October 1987 crash.
3See Business Week, September 11, 2000, and The Economist, “Rise and Fall: A Survey of Global Equity Markets,” May 5, 2001.
5The definition of a bubble in this paper is rather loose and refers to a period characterized by a surge in equity prices. For example, the period from the fourth quarter of 1998 to the first quarter to the first quarter of 200 has been commonly referred to as the “tech bubble” period by the media. For a formal definition of bubble, which is not used here, see Flood and Garber(1982).
6Bollerslev, Chou and Kroner (1992) and Bollerslev, Engle and Nelson (1994) provide two excellent surveys on the application of ARCH models in the area of financial econometrics.
7This behavior is exemplified by the following quotation: “Asian shares posted broad-based gains… after Wall Street posted its biggest rise in more than six weeks.” (Reuters World service, August 27,2001).
8De Bandt and Hartmann (2000) provide a useful survey of the literature, emphasizing the relationship between increased stock market correlation and systemic risk. Jarrow (1998) provides a selected survey of multivariate GARCH models.
9The correlations can be obtained either as sample correlations, using simple OLS regressions, or using the exponential smoothing method popularized by RiskMetrics.
10Hamao, et. al. (1990) report that currency denomination of stock prices would not change results substantially.
11The GARCH model was first suggested by Bollerslev (1986) as a generalization of the ARCH model developed by Engle (1982).
12The TGARCH family of statistical models was first introduced by Zakoian (1994) and Glosten, Jaganathan and Runkle (1993), and is a special case of the Box-Cox transformation of the absolute GARCH (AGARCH) model, as shown by Hentschel (1995).
13Black (1976) showed the existence of asymmetric effects (so called leverage effects) on the conditional variance, as negative equity returns are usually followed by larger increases in volatility than equally large positive returns.
14Scholes and Williams (1977) and Cohen et. al. (1980) show that nonsynchronous trading in individual stocks, bid-ask spreads, and minimum-price changes can cause serial correlation in stock returns. Hamao et. al. (1990) show that the inclusion of a MA(1) term is sufficient to extract serial correlation from the first moments of stock returns, finding no support for higher-order MA terms.
15The results are generally robust to minor changes in the periods definition.

Other Resources Citing This Publication