RDP 2004-05: Big Fish in Small Ponds: The Trading Behaviour and Price Impact of Foreign Investors in Asian Emerging Equity Markets 2. Data

2.1 Basic Data

The six Asian markets studied in this paper are the Jakarta Stock Exchange (JSX), Korea Stock Exchange (KSE), Philippine Stock Exchange (PSE), Stock Exchange of Thailand (SET), Taiwan Stock Exchange (TWSE), and Kosdaq Stock Market.^[3] The first five are ‘main boards’, while the sixth, which focuses on Korean start-up and technology-related companies, is a ‘second board’, but nonetheless has a larger market capitalisation than many main boards in other emerging markets. Data on daily net purchases were obtained from the exchanges and from CEIC and Bloomberg, two secondary providers.^[4] Other data used in the study are taken from Bloomberg and include data for the capitalisation-weighted price index and market capitalisation of each local market, as well as data for various mature market equity price indices.

Foreign investors in these markets must register with the local exchange or regulator, and brokers must report the nationality of the buyer and seller in each transaction that occurs.^[5] The resulting data capture the trading of all registered foreign investors. One possible shortcoming with the data is that they do not capture net purchases by foreigners of ADRs or country funds in foreign markets, or equity futures trading in the domestic market. In the first two cases, the omission is unlikely to be serious, since trading in these is likely to be largely between foreigners, and is unlikely to result in substantial net purchases or sales by foreigners. The omission of futures (and other derivatives) trading might be more serious. Fortunately, daily data on the net purchases of foreigners are available for the Korea Stock Exchange's equity futures contract. Thus in one case the data capture essentially all changes in foreign investors' equity exposures.

The study uses data for January 1999 to September 2002, except for the Philippines which begins in March 1999. The sample corresponds fortuitously to the period after which trading on the two Korean exchanges switched (in December 1998) from six- to five-day trading.^[6] Trading on all six markets is order-driven and fully computerised. To facilitate price discovery, all six markets have call auctions to determine opening and closing prices. The trading hours of the six exchanges all correspond to periods when US markets are closed.

Summary data for each market and the role of foreigners are provided in Table 1. The markets include two exchanges, the KSE and TWSE, which are among the largest of all emerging markets and are comparable in capitalisation to some mid-sized mature equity markets. Trading on these two markets is also highly active, with 2001 annual turnover ratios well above most mature markets (including the New York Stock Exchange's 2001 turnover ratio of 0.89). The Kosdaq market is even more active, with annual turnover equivalent to about ten times market capitalisation, making it the most active exchange in the world. Turnover ratios for the JSX, and especially the PSE, are quite low. The latter two exchanges are also those where there is least variation in the daily net purchases of foreign investors (normalised by dividing by the previous day's market capitalisation). In those cases where a comparison is feasible, the share of foreign investors in total trading is lower than their ownership share, suggesting that foreigners trade less actively than domestic investors. Moreover, with the development of institutional investors still at a relatively early stage in these markets, trading tends to be dominated by individual investors: the share of individual investors in total trading in 2001 was 73 per cent on the KSE, 77 per cent in Thailand, 84 per cent in Taiwan, and an amazing 94 per cent on the Kosdaq.^[7]

2.2 Descriptive Statistics for Net Purchases of Foreign Investors

Data on the properties of daily net inflows (or ‘flows’) are shown in Table 2. Here and subsequently, flows are expressed as a percentage of the previous day's market capitalisation. The data in the top panel show substantial positive autocorrelation in daily inflows, consistent with Froot et al (2001), with a median autocorrelation of 0.47. Daily returns in these markets are far less autocorrelated, with a median autocorrelation of 0.09. This positive autocorrelation in flows could be due to particular investors establishing positions slowly (perhaps to reduce market impact), or to investors of similar types responding in the same direction – but with different speeds – to new information. In those four markets where net purchases data are available separately for domestic institutions and individuals, the net flows of these groups are not surprisingly also highly autocorrelated.

Within each market, there is a strong positive same-day correlation between net foreign inflows and returns, with a median correlation coefficient of 0.33 (middle panel). By contrast, the net purchases of domestic individuals are strongly negatively correlated with returns, while the pattern is more mixed for domestic institutions, with median correlations of −0.40 and 0.11, respectively. In most cases there is also strong positive correlation between net inflows across different exchanges (bottom panel), although it is not as strong as the cross-exchange correlations in returns, with median correlation coefficients of 0.17 and 0.22, respectively. The positive correlations in net inflows suggest that there are common or related factors influencing flows, as will be confirmed in Section 3.2.

Finally, some preliminary data for the relationship between physical and futures trading on the KSE are shown in Table 3. In contrast to the results in Table 2, the data show significant negative first-order autocorrelation in the net purchases of foreigners in the Kospi 200 contract on the KSE futures market. Somewhat surprisingly, there is only weak positive correlation between net purchases in the physical market and same-day net purchases on the futures market. However, there is a strongly significant positive correlation between net purchases in the physical market and previous-day net purchases on the futures market. Together, these correlations are highly suggestive of a pattern whereby some foreign investors wishing to effect a change in their exposure do so by first taking a short-term position in the futures market, and then unwinding the futures market position on the next day as they carry out the desired change in their longer-term position in the physical market. Foreign investors trading on the futures market is significant; the standard deviation of daily futures market net purchases is 0.044 per cent of market capitalisation, only modestly smaller than the equivalent figure for the physical market. Accordingly, and given the different trading behaviour in the two markets, it follows that concentrating only on physical market transactions in this market may give an incomplete picture of the timing and impact of foreign investors' trading. To conserve space, the results presented in the remainder of the paper for the KSE are based on the sum of the net purchases of each investor group on both the physical and futures markets, although results for the physical market alone are discussed at a few points, and are available upon request.

Footnotes

Some other studies have also used data from these exchanges. For example, Choe, Kho, and Stulz (2004) have used the KSE data at the individual stock level, Seasholes (2001) has used the Taiwanese and Thai data, Bonser-Neal et al (2002) have used the Indonesian data, and Griffin, Nardari and Stulz (2003) have studied data for five of the exchanges, but not the Kosdaq. [3]

Net purchases data from different sources were checked against each other, and numerous errors were corrected. A few potential outliers remained for Indonesia and Philippines, so eight observations (all cases of apparent very large net inflows) were omitted because they appeared to be data errors. In several of these cases, the observation coincided with a large privatisation sale, suggesting that large off-market privatisation transactions had somehow shown up in the trading data. [4]

There was no general limit on foreign investment for four of the six markets (the KSE, Kosdaq, JSX, and PSE) during the period of the study. Taiwan saw a substantial relaxation of foreign ownership limits during the sample period, with the limit for each firm being increased from 30 to 50 per cent in March 1999, then to 75 per cent in October 2000, before its removal at the end of 2000. Thailand had a general limit of 49 per cent throughout the sample period. In addition, all markets had specific foreign ownership limits for a few particular firms or industries. [5]

However, Saturday trading continued in Taiwan in 1999 and 2000 on the first, third and fifth Saturdays of each month. On those 51 occasions when there was Saturday trading, the data for Saturday were merged into the following Monday, with Monday returns being measured relative to Friday close, and Saturday net inflows included in Monday's flows. [6]

Data on the average size of individual trades of foreigners were not available, although data on the trades of foreign investors in Froot et al (2001) indicate an average trade size of about $200,000. This suggests that foreign investors are indeed ‘big fish’ in these markets even if they trade less actively than domestic investors. [7]