Developing the Government Bond Market in Korea after the Financial ...

5 downloads 121 Views 594KB Size Report
August 1998 Announcement of 'Government Bond Market Stimulus Plan'. March 1999 Establishment of inter-dealer market (IDM). July 1999. Enactment of ...
Developing the Government Bond Market in Korea after the Financial Crisis: Performance Evaluation Using Micro-data

Daekeun Parka, Changyong Rheeb and Sung Hwan Shinc,

August 2006

a) Department of Economics and Finance, Hanyang University, [email protected] b) Department of Economics, Seoul National University c) Department of Management, Hongik University

I. Introduction

Since the Asian financial crisis of 1997 the government bond market in Korea has achieved an outstanding improvement in terms of the size and the institutional arrangement.

The

government bond market has grown in size as the issuance of government bonds increased dramatically to finance the efforts of much needed financial restructuring and economic recovery in the aftermath of the financial crisis.

In addition, in response to the needs to reduce

the cost of issuing and servicing the enormous amount of government bonds, the Korean government took measures to enhance the institutional framework and the infrastructure for the government bond market.

These reform efforts resulted in the introduction of the Dutch

auction system to replace the old compulsory underwriting system, the primary dealer system, the DVP(delivery versus payment) settlement system, the reopening system, the marking-tomarket system, the repurchase agreement market, the KTB(Korea Treasury Bond) STRIPs and so on. One of the institutional reform measures was the introduction of the Korea Exchange Government Bond Market (hereafter the KRX market) in March 1999.

The KRX market,

which was based upon an electronic trading platform, was established with the intention of enhancing the transparency and the liquidity in the secondary market for government bonds. As a result of establishing the KRX Market, the secondary market for government bonds in Korea has been divided into two markets: the OTC market and the KRX market. In order to boost the trading activities in the KRX Market, the government also introduced the compulsory trading requirements in October 2002 mandating that the primary dealers (PDs) should trade benchmark issues of the governments bonds in the KRX market. Traditionally, the over-the-counter (OTC) trading where bonds are traded between dealers

2

through different forms of voice message system has been the dominant form of bond trading. The recent trend among the developed countries, however, is the one towards the electronic trading platform replacing the voice message system.

A good example of the electronic

trading platform for bond trading can be found in the EuroMTS and the MTS (Mercato dei Titoli di Stato) system of individual European countries.

It is widely understood that the

electronic trading system enhances efficiency of the secondary bond markets by reducing the transaction costs and by making the trading process transparent.

As a result, it is expected that

the imposition of the mandatory trading requirements to boost trading activities in the KRX market will be helpful in enhancing transparency and efficiency of the overall government bond market in Korea.

On the other hand, however, it is also argued that introduction of the KRX

Bond Market and imposition of the mandatory trading requirements may undermine the efficiency of the government bond market by restricting the trading activities of the primary dealers and by dividing market liquidity between the OTC market and the KRX market.

Thus,

whether the imposition of the exchange trading requirements has been beneficial to the entire government bond market in Korea remains an empirical question. In this paper we evaluate the effects of introducing the electronic trading system for government bonds and imposing the trading requirements on the KRX market and the OTC market using the intraday trading data for government bonds.

The main findings of this paper

are as follows. First, introduction of the KRX market and imposition of the mandatory trading requirements have been effective in increasing the trading volume of the KRX market and the trading share of the KRX market in the entire secondary market for government bonds.

The trading

requirements on benchmark issues have also boosted transactions of non-benchmark issues although these are not covered by the trading requirements.

3

In addition, the exchange trading

requirements on primary dealers have also helped to bring in other players into the exchange market. The bid-ask spreads on government bonds in the KRX market has also fallen after the measures to boost trading activities in the KRX market were taken. Second, the increase of the trading volume in the KRX market did not come at the expense of a lower trading volume in the OTC market.

In consequence, we can conclude that the

introduction of the exchange trading requirements for benchmark issues was helpful in improving the overall quality of the entire secondary government bond market in Korea. Third, the volatility of the KRX market as well as the OTC market decreased significantly since the introduction of the exchange trading requirements.

Yet, various measures of

volatility demonstrate that the OTC market is more volatile than the exchange market. In addition, the volatility measures such as the daily standard deviation of transaction prices and the intraday range of transaction prices displayed quite a few abnormal values in the OTC market.

Further analysis reveals that these abnormal values can be ascribed to lack of

transparency in the OTC market rather than to fundamental macroeconomic shocks. Fourth, the MEC (market efficiency coefficient) estimated from the intraday trading data and the execution costs derived from the estimates of the MEC demonstrate that the exchange trading requirement has not only improved the liquidity and the efficiency of the KRX market but added greatly to the efficiency and liquidity of the OTC market. This paper is organized as follows: Chapter II summarizes the changes in the government bond market in Korea after the financial crisis including the introduction of the KRX Government Bond Market and the imposition of the exchange trading requirements.

Chapter

III investigates the effects of imposing the exchange trading requirements on the KRX market and the OTC market using intraday trading data.

Chapter IV analyzes the effects of the

mandatory trading requirements on the liquidity and the transparency of the secondary

4

government bond markets in Korea by estimating the MEC and the execution cost using a micro data.

Chapter V presents our conclusions.

II. Changes in the Government Bond Market in Korea

Growth of the Government Bond Market after the Currency Crisis Before the currency crisis of 1997, the government bond market in Korea was in a rather stagnant state. Because of the emphasis on a healthy fiscal balance, the volume of bond issuances fell far short of the amount needed for an active secondary market to develop. The old regime of compulsory underwriting under which government bonds were issued at yields-tomaturity lower than the market interest rate provided little incentive for the secondary market to develop. After the financial crisis, however, issuance of government bonds rose dramatically boosted by the need to support the efforts of financial restructuring and economic recovery. see from Figure 1, the outstanding amount of government bonds1 which stood at the end of 1996 has increased almost eightfold to

As we can 25 trillion at

204.8 trillion by the end of June 2005. As

we can see in Figure 2, the government bond market was much smaller than the corporate bond market before the currency crisis with the outstanding volume of government bonds amounting to a third of that of corporate bonds.

After the currency crisis, however, the outstanding

amount of government bonds has grown continuously to surpass that of corporate bonds by the

1

Although twenty one kinds of government bonds have been issued since 1949, only three kinds of

bonds are currently issued including the Korea Treasury Bonds, the Korea Treasury Bills and the National Housing Bonds. The Foreign Exchange Stabilization Fund Bond has been consolidated into the Korea Treasury Bond since November 2003.

5

end 2003. Figure 2 shows that new issuance of government bonds in 2004 has grown sevenfold to

75.8 trillion compared to

11trillion in 1997.

Alongside the quantitative growth, the qualitative aspect of the government bond market has also improved significantly through institutional reforms and infrastructure build-ups. The efforts to develop government bond markets have been driven by the need to reduce the cost of issuing and servicing government bonds whose amount has grown dramatically to support the efforts of restructuring financial system in the aftermath of the financial crisis.

Table 1

summarizes some of the policy measures taken by the government to develop government bond markets in Korea

Table 1. Policy Measures to Develop Government Bond Markets in Korea

August 1998

Announcement of ‘Government Bond Market Stimulus Plan’

March 1999

Establishment of inter-dealer market (IDM)

July 1999

Enactment of primary dealer system

September 1999 Introduction of government bond futures November 1999 Introduction of DVP system February 2000 March 2000 May 2000

Introduction of inter-dealer brokers (IDB) Securities financing facilities for primary dealers Introduction of the reopening system

August 2000

Switch from multiple price auction to Dutch auction

October 2002

Introduction of exchange trading requirements for benchmark issues

January 2003

Strengthening obligations of primary dealers

2006

Introduction of KTB STRIPs and 20 year maturity government bonds

Introduction of the KRX Electronic Bond Trading System One of the reform measure taken by the Koran government to improve the government bond market was the establishment of the KRX Government Bond Market in 1999.

6

The KRX

market was initially set up exclusively for trading among government bond dealers. brokered trading through securities companies was allowed.

Later,

The secondary market for

government bonds in Korea comprises two markets: the OTC market operated by the Korea Security Dealers’ Association (KSDA) and the KRX market operated by the Korea Exchange.2 The KRX Government Bond Market adopted an electronic trading platform named the KRX Electronic Bond Trading System (the KTS).

In general, secondary markets for bonds have

developed in the form of an OTC market rather than a centralized exchange.

In the OTC

market, final investors who wish to trade bonds search for the best price quote by making calls to several dealers and make a deal with the dealer who offers the best price quote.

The

inefficiency and the obscurity that arise from the typical search process in the OTC market have lead to the recent trend in the developed markets such as the U.S. and Europe that more and more bond transactions are executed through electronic trading systems.

Introduction of the mandatory exchange trading requirements for benchmark issues When it was first established in 1999, the trading in the KRX market was so sluggish that the KRX market was not able to perform the function of price discovery properly.

In October

2002, the government imposed trading requirements on the primary dealers of government bonds to activate trading in the KRX government bond market making it compulsory for the primary dealers to make all of the trades of benchmark issues and at least 20% of the trades of government bonds in the KRX market.

In January 2003, the mandatory trading requirements

were further strengthened with the minimum trading proportion raised from 20% to 40%. The imposition of mandatory exchange trading requirements has been effective in

2

The exchange market for government bonds was initially operated by the Korea Securities Exchange,

which later merged with the Korea Futures Exchange to become the Korea Exchange.

7

activating transactions in the KRX market.

According to Figure 3, the transaction volume of

government bonds in the KRX market, which was nothing but negligible before October 2002, has increased greatly after the imposition of the trading requirements. Figure 4 shows the share of the KRX bond market in the secondary trading of the benchmark issues has also risen after the imposition of the trading requirements.

The KRX bond market

used to have a mere 10% share in the secondary trading of benchmark issues but the share has risen to 50% after the government introduced the mandatory exchange trading requirements. The mandatory exchange trading requirements have also been effective in increasing the trading volume of non-benchmark issues in the KRX market.

Figure 4 shows that the share of

the KRX bond market in the secondary trading of non-benchmark issues also rose from 10% in September 2002 to 25% in January 2003 although non-benchmark issues were not subject to the mandatory trading requirements. Imposition of the trading requirements on the primary dealers of government bonds has also contributed greatly to attracting players other than primary dealers to the KRX bond market. According to Figure 5, the share of the KRX bond market in the trades of government bonds of primary dealers rose from 30% in April 2002 to approximately 50% in January 2003.

Figure 5

also shows that the share of the KRX market in the trades of government bonds of dealers who are not primary dealers has also grown from 3% in April 2002 to 20% in April 2005.

Such a

phenomenon well demonstrates the characteristic of bond markets that “liquidity attracts liquidity.” That is, the higher liquidity in the KRX bond market resulting from increased participation of primary dealers has attracted other dealers to the exchange market. The fact that trading requirements were able to attract other participants to the exchange market can be also verified from Figure 6 that shows the trading volumes in the KRX market by different types of dealers.

As we can see from the figure, it is not the transaction between

8

primary dealers but the transaction between a primary dealer and a non-primary dealer that has the largest trading volume in the exchange market.

Reduction in the bid-ask spreads The benefit of developing the electronic trading system for government bonds can be also confirmed by taking a look at the bid-ask spreads in the KRX market. As we can see from Figure 7, the bid-ask spreads that used be over 18 bps before the trading requirements were introduced has been reduced to 3.5 bps in 2005.

Normally, the bid-ask spreads are used as an

indicator of market liquidity.

III. The Trading Requirements and the Government Bond Market

Chapter II demonstrated that the introduction of the electronic trading system and the imposition of the mandatory trading requirements were effective in enhancing the trading activity and the liquidity of the exchange market.

However, if these benefits were available at

the cost of the OTC market in terms of a reduced trading volume or a lower level of liquidity, then we cannot conclude that the introduction of the electronic trading system was a definite benefit to the government bond market as a whole. As mentioned in the introduction, one of the most important objectives of introducing the electronic trading system was to enhance the efficiency and the transparency of the OTC market as well as the exchange market by ensuring more transparent dissemination of the transaction prices.

In this chapter, we present a variety of measures of market performance calculated

using the intraday trading data from the KRX market and the OTC market to assess the effects

9

of the electronic trading system on the transparency and the efficiency of these two markets.

Transaction Volume Figure 8 shows the monthly transaction volume of the Korea Treasury Bonds (KTBs) and the Foreign Exchange Stabilization Fund Bonds (FESFBs)3 in the KRX market and the OTC market for the period from January 2000 to April 2005.

In addition, the solid line in the figure

shows the share of the KRX market in the monthly turnover of the KTBs and the FESFBs in the entire secondary market.

Table 2 presents the annual turnover of the KTBs and the FESFBs in

the KRX market and the OTC market together with the share of the KRX market.

Table 2. Yearly Turnover of KTBs and FESFBs

Year

The KRX Market

The OTC Market

The Share of

Turnover (billion won)

Turnover(billion won)

the KRX Market

2000

21,644

2,513,045

0.09

2001

10,100

4,430,658

0.02

2002

42,600

3,432,220

0.12

2003

207,930

4,539,385

0.46

358,400

7,077,729

0.51

110,630

2,501,905

0.44

2004 2005

*

Source: Financial Supervisory Service, Monthly Bulletin, various issues Note: The numbers for 2005 are based on the data from January to April only.

If we take a look at Figure 8, we can observe a dramatic increase in the transaction volumes of the KTBs and the FESFBs in the KRX market after the introduction of the exchange

3

Although the Foreign Exchange Stabilization Fund Bond is no longer issued since it has been

consolidated into the Korean Treasury Bond in November 2003, there still remain some outstanding issues.

10

trading requirements.

The annual transaction volume of the KTBs and the FESFBs in the

KRX market amounted to 2003.

42.6 trillion in 2002 but it has skyrocketed to

207.9 trillion in

As a result of the dramatic increase in the transaction volume, the share of the KRX

market in the entire secondary market turnover has also risen from 12% in 2002 to 46% in 2003. Figure 8 and Table 2 also demonstrate that the monthly turnover as well as the yearly turnover in the OTC market has also increased noticeably after the imposition of the trading requirements.

The fact that not only the transaction volume of government bonds in the KRX

market but that in the OTC market has also increased after the imposition of the trading requirements supports to the argument that the introduction of the electronic trading system and the imposition of the trading requirements have contributed to enhancing the trading activities in the entire secondary market for government bonds.

Table.3. Yearly Turnover Ratio of KTBs and FESFBs Year

KRX Market

OTC Market

Overall Market

2000

41.2%

478.0%

519.1%

2001

16.9%

743.0%

759.9%

2002

59.6%

480.2%

539.8%

2003

197.8%

431.7%

629.5%

2004

246.7%

487.2%

734.0%

2005

305.8%

414.7%

720.4%

Source: Financial Supervisory Service, Monthly Bulletin, various issues Note: The numbers for 2005 are based on the data from January to April only.

Turnover Ratio The fact that the imposition of the exchange trading requirements on benchmark issues has been effective in invigorating the trading activities in entire secondary market for government bonds can be also verified from Figure 9 that shows the monthly turnover ratio of the KTBs and

11

the FESFBs in the KRX market and the OTC market and from Table 3 that presents the yearly turnover ratios.4 Both of Figure 9 and Table 3 clearly demonstrate that the turnover ratio of the KTBs and FESFBs in the KRX market has increased significantly after the imposition of the exchange trading requirements.

For example, the yearly turnover ratio for government bonds in the

KRX market has risen from 59.6% in 2002 to 246.7% in 2004 and is still on the rising trend. Meanwhile, the turnover ratio in the OTC market decreased only slightly even though the exchange trading requirements for benchmark issues have been put in place so that the overall turnover ratio in the entire secondary market for government bonds has also increased.

Once

again, such findings can be interpreted to indicate that strengthening the trading activities in the KRX market was achieved without sacrificing the trading activities in the OTC market.

Standard Deviation of Transaction Prices In order to assess the effect of the exchange trading requirements on market volatility, we calculated the standard deviation of the transaction prices5 using the intraday transaction data available from the KRX and the KSDA.

To calculate the standard deviation, we constructed a

bond portfolio consisting of all 3-year KTBs with a remaining term to maturity of no less than 30 months6 and calculated the standard deviation of the transaction price of the bond portfolio Figure 10 shows the estimated daily standard deviation of the transaction prices quoted in terms

4

Turnover ratios are calculated as the market turnover divided by the outstanding amount at the end of

each period. 5

Transaction prices are quoted in terms of the yields to maturities in the KRX market as well as in the

OTC market in Korea. Although they are quoted in yields, we will still use the term “transaction prices” in this paper. 6

We used bonds with a remaining maturity of less than 30 months because bonds with a short remaining

term to maturity are not liquid.

12

of the yields to maturity for the 3-year KTBs for the period between May 2000 and February 2005. Figure 10 clearly demonstrates a clear difference in the volatility of the transaction prices in the KRX market between the two periods divided by the introduction of the trading requirements: the volatility of the KRX market as measured by the standard deviation of the transaction prices fell precipitously after the trading requirements were put in place. Figure 10 shows that the volatility of the transaction prices in the OTC market has decreased as well after the imposition of the trading requirements.

It should be clear why the

OTC market has become less volatile when the trading requirements were put in place in the KRX market and not in the OTC market.

It is because the implementation of the trading

requirements provided the participants in the OTC market with a reliable reference price for government bonds.

Although the transaction prices in the KRX market had been available to

participants in the OTC market before the implementation of the trading requirements the KRX government bond market had been too thin and too volatile for market participants to accept the transaction prices in the KRX market as the equilibrium prices that correctly reflect the market demand and supply.

After the implementation of the trading requirements, however, the KRX

market has become more active and less volatile so that participants in the secondary market began accepting the transaction prices in the KRX market as the market price that accurately reflected the demand and supply condition and began using them as a benchmark for their transactions.

Use of a common reference price for the deals in the OTC market naturally

resulted in convergence of the transaction prices in the OTC market.

This demonstrates how

the introduction of the trading requirements did not only help improve the liquidity and the efficiency of the exchange market but also enhance the stability and the transparency of the OTC market.

13

Although the OTC market has become less volatile, we can still find quite a few extreme values for the standard deviation in the OTC market.

To make this point, an enlarged version

of the graph of the daily standard deviations for the period after the imposition of the trading requirements is presented in Figure 11. is more volatile than the KRX market.

Looking at Figure 11, we can tell that the OTC market In addition, we can observe quite a few excessive

values in the standard deviation of the transaction prices in the OTC market.

An excessive

value of the standard deviation indicates that at least one of the transactions was conducted at an abnormal price during that day.

Existence of such outliers in the intraday transaction prices

may be interpreted as a piece of evidence of the shortage of transparency in the OTC market. It is still possible to argue that these outliers occurred as a result of new information shocks creating turbulence in the bond market.

We conclude that such an argument cannot explain the

existence of these outliers for the following reasons. First, we calculated the daily standard deviation for each individual bond included in the bond portfolio and identified that dates on which the standard deviation has excessively large values. Examining these dates, we found that the dates on which each individual bond has extremely large values of standard deviation do not coincide with the dates on which other bonds have extremely large values of standard deviation. Therefore, it is difficult to conclude that these outliers were caused by common economic shocks that affected the entire bond market simultaneously. Second, we investigated whether there were any macroeconomic events that would have had a significant effect on the government bond market on the dates when the outliers for each individual bond occurred but failed to find any meaningful macroeconomic events that could explain the occurrence of the outliers. Third, we also examined whether the transaction price of the benchmark issues showed

14

abnormal movements on the dates when the outliers for each individual bond occurred but failed to find any significant movement in the benchmark issues.

For example, the daily

standard deviation of transaction prices of the 3-year KTB, KR1035017P32, was 0.27 on 17 August 2004, which is almost ten times the average daily standard deviation.

To see whether

there had been any macroeconomic shock on that particular day, we investigated the movement of the transaction price of the benchmark issue for the 3-year KTB on that day.

During that

particular day, the transaction price of the benchmark issue changed only by 3 bps, which by no means can be regarded as a dramatic change. a

In addition, we found that on that day there was

22.6 billion worth of bloc trading of the KR1035017P32 at the yield to maturity of 4.5%

which was far above the average transaction yield on that day. As is illustrated by the case explained above, the majority of the outliers cannot be attributed to macroeconomic shocks that affected the entire government bond market simultaneously.

On the contrary, it is lack of transparency in the OTC market that is

responsible for most of the outliers.

Although various reform efforts have been made to

improve the transparency and the efficiency of the government bond market including the OTC market, existence of the outliers in the OTC market indicates that there still is a lot of room for improvement.

Range of the Intraday Transaction Prices Besides the standard deviation of transaction prices, market volatility can be assessed by the range of transaction prices as well.

Figure 12 shows the range of the intraday transaction

prices calculated as the difference between the maximum and the minimum of the intraday transaction prices for the 3-year KTBs issued in 2002.

As was the case with the standard

deviation of the intraday transaction prices, the intraday range of the transaction prices in the

15

KRX market has decreased significantly when the exchange trading requirements were put in place.

Figure 12 also shows that the mandatory trading requirements ware effective in

reducing the volatility in the OTC market as well.

Yet as we can see from Figure 12, the OTC

market is much more volatile than the KRX market and produces quite a few outliers.

The Intraday Day Pattern of Transaction Volume and Market Volatility Figure 13 shows the transaction volume of the KTBs accumulated for the entire year of 2004 for each 20 minute interval.

The transaction price is calculated as the transaction price of

the portfolio of the entire KTBs traded during 2004.

According to Figure 13, the intraday

pattern of the trading volume has a U-shape with the volume decreasing noticeably around the lunch time of 12:00-13:00. Figure 14 shows the standard deviation of the transaction prices of the KTBS for each 20 minute interval. The standard deviation was calculated as the standard deviation of the transaction prices during each 20 minute interval in each day averaged for the entire year of 2004.

The standard deviation of the transaction prices in the KRX market as presented in

Figure 14 does not show any particular pattern of intraday fluctuation.

On the other hand, the

standard deviation of the transaction prices in the OTC market shows abnormally high values during the lunchtime of 12:00-13:00.

Coincidence of the low trading volume and the high

volatility of transaction prices in the OTC market could be the result of some market participants conducting abnormal transactions by taking advantage of the low trading activities in the OTC market during the lunchtime.

16

IV. MEC and the Execution Cost

Liquidity can be broadly defined as the degree of easiness with which an asset can be converted into cash.

It can be measured by how long it takes to convert an asset into cash

through a sale at a reasonable price or how much it costs to convert an asset into cash through a sale within a short amount of time. execution cost.

The cost of turning an asset into cash is called the

The execution cost not only includes the obvious cost like commissions and

taxes but the cost arising from the market impact effect in the case of large orders.

In order to

measure the market execution cost rather than the mere cost of individual transaction, all these implicit costs have to be included in the calculation of the execution cost. Various measures have been adopted to assess the liquidity of an asset or an asset market including the bid ask spread, the quote size, the transaction volume, the transaction frequency, the frequency of price quotes, the price impact coefficient (sometimes known as the Kyle Lambda), and the yield spreads between on-the-run issues and off-the-run issues.

In this paper,

we use the market efficiency coefficient (MEC) developed by Hasbrouck and Schwarz (1988) to estimate the execution cost in the government bond market and evaluate the effect of developing the KRX bond market on the liquidity of the secondary market for government bonds in Korea. The MEC is defined as the ratio between the variance of the long run rate of return and the time adjusted variance of the short run rate of return.

MEC =

var( R L ) , q × var( RS )

Specifically, the MEC is defined as:

(1)

where R L and RS denote the long-run rate of return and the short-run rate of return

17

respectively and q stands for the number of short-run periods comprising the long-run period. If we denote the closing transaction price of each short-run period as P0 , P1 , L, Pt , L, Pq , we can get the following identity:

Pq PT P P = 1 × 2 ×L× P0 P0 P1 Pq −1

(2)

Taking the natural logarithm on both sides of Equation (2), we can derive the following relation between the long-run return and the short-run returns:

q

R L = ∑ R S ,t

(3)

t =1

In general, if the market is efficient, the short-run rates of return will follow a random walk process with independent and identical probability distribution and as a result the value of the MEC will be equal to one.

But just as Roll (1984) had shown, if there exist some execution

costs, successive price changes will have a negative serial correlation and as a result the MEC will be smaller than one even if the market is efficient.

Therefore, assuming that the market is

efficient, we can evaluate the size of the execution cost by calculating the value of MEC.

In

practice, Hasbrouck and Schwarz (1988) showed that the execution cost can be derived from the MEC using the following equations:

C = [0.5 × var(RS ) × (1 − MEC )]1 / 2 , if MEC ≤ 1 and C = −[0.5 × var(RS ) × ( MEC − 1)]1 / 2 , otherwise

18

(4)

In this paper, we use the intraday trading data from the KRX Government Bond Market and the OTC market to estimate the MEC and the execution cost in each market on the condition that each market is efficient.

In order to calculate the MEC, we chose one hour as

the length of the short-run period and took the closing transaction price of each one hour interval as the transaction price of that period. The KRX Government Bond Market is in session from 9 am to 3 pm.

Since the KRX

market is based on an electronic trading system, the transaction details including the transaction time are recorded with accuracy. In the OTC market, dealers are required to report the specifics of each transaction to the KSDA through computer terminals within 15 minutes after the transaction is conducted. However, since the transaction time is not included as a part of the report, we can only identify the time when each transaction is reported. off 15 minutes from the reporting time.

Thus, we estimated the transaction time by taking

Yet, since the regulation allows quite a few exceptions

to the 15 minute reporting requirement, quite a number of transactions in the OTC market are reported after 3 pm even if the transactions have been conducted between 9 am and 3 pm.

For

these transactions, there is no way of estimating the transaction time and as a result we excluded the transactions reported after 3 pm from the analysis. In order to calculate the MEC it is necessary to choose the length of the long run and the short run.

The rate of return for one hour has been chosen as the short run rate of return whilst

the rates of return rates for four days has been chosen as the long term rates of return. Given that only the transactions that took place between 9 am and 3 pm are included in the analysis the long run corresponds to 24 hours (4 days multiplied by 6 hours a day) implying that the value of

19

q in Equation (1) should be 24.7 We estimated the MEC and the execution cost for twelve benchmark issues of the 3-year KTBs and twelve benchmark issues of the 5-year KTBs.

The 3-year KTBs and the 5-year

KTBs issued between May 2000 and January 2003 maintained the benchmark status for three months after the initial offering and those issued after January 2003 maintained the benchmark status for six months.

Table 4. Estimates of the MEC and the Execution Cost (3-year Bonds) Item

Benchmark Period

KRX MEC OTC MEC KRX C(%) OTC C(%)

. KTB00-10

2000.5.12~2000.8.1

0.296***

0.113***

0.105

0.254

KTB00-12

2000.8.16~2000.11.6

0.042***

0.282**

0.266

0.075

KTB01-1

2001.1.10~2001.2.12

0.063*

0.165*

0.475

0.099

KTB01-6

2001.7.5~2001.10.4

0.044***

0.192***

0.310

0.113

KTB0644-0504

2002.4.3~2002.7.1

0.164***

0.375**

0.105

0.063

KTB0562-0507

2002.7.11~2002.9.25

0.082***

0.138***

0.128

0.094

KTB0520-0510

2002.10.2~2003.1.3

0.520*

0.471**

0.052

0.063

KTB0510-0601

2003.1.8~2003.2.28

0.684

0.515

0.023

0.034

KTB0450-0603

2003.3.5~2003.8.28

0.552**

0.630*

0.042

0.040

KTB0450-0609

2003.9.3~2004.2.26

0.509**

0.384***

0.051

0.065

KTB0475-0703

2004.3.3~2004.9.8

0.940

0.645*

0.016

0.044

KTB0375-0709

2004.9.10~2005.3.9

0.753

0.596**

0.034

0.050

Note: *, ** and *** denote that the null hypothesis that the successive short-run returns follow a random walk process can be rejected at the significance level of 10%, 5% and 1% respectively.

Table 4 and Figure 15 present the estimates of the MEC and the execution cost for twelve benchmark issues of the 3-year KTBs.

Table 5 and Figure 16 present the estimates for twelve

benchmark issues of the 5-year KTBs.

For most of the issues, the estimated value for the MEC

7

The choice of 4 days for the long run is somewhat arbitrary. We also calculated the MEC and the

execution cost by setting q to be equal to 12 (two days) instead of 24 (four days) and got similar results.

20

is less than one.

Table 4 and Table 5 also present the result of testing the null hypothesis that

the successive short-run rates follow a random walk process.

Lo and MacKinlay (1988)

showed the following test statistic asymptotically has a standard normal distribution when the null hypothesis is true:

z (q) =

nqM (q) 2(2q − 1)(q − 1) / 3q

where M (q ) denotes the estimator of the MEC.

,

(5)

The test result shows that in all but a few

cases, the null hypothesis of random walk can be rejected.

Table 5. Estimates of the MEC and the Execution Cost (5-year Bonds) Item

Benchmark Period

KRX MEC OTC MEC KRX C(%) OTC C(%)

KTB0820-0508

2000.8.23~2000.10.12

0.215**

0.079**

0.192

0.170

KTB0715-0604

2001.4.11~2001.4.20

0.129

0.016

0.275

0.286

KTB0625-0607

2001.8.17~2001.10.5

0.004*** 0.211**

0.739

0.174

KTB0690-0701

2002.3.6~2002.3.21

0.004*

0.887

0.160

KTB0717-0704

2002.4.10~2002.7.5

0.072*** 0.138*** 0.329

0.149

KTB0615-0707

2002.7.19~2002.10.4

0.146*** 0.416*

0.206

0.108

KTB0577-0710

2002.10.9~2003.3.7

0.329*** 0.228*** 0.112

0.134

KTB0475-0803

2003.3.12~2003.9.4

0.528**

0.295*** 0.084

0.138

KTB0450-0809

2003.9.9~2004.3.4

0.416*** 0.154*** 0.087

0.206

KTB0500-0903

2004.3.10~2004.6.11

0.307**

0.204*** 0.075

0.102

KTB0450-0906

2004.6.15~2004.12.9

1.109

0.619*

-0.024

0.059

. KTB0350-0912

2004.12.14~2005.3.25

1.124

0.554*

-0.036

0.093

0.330

Note: *, ** and *** denote that the null hypothesis that the successive short-run returns follow a random walk process can be rejected at the significance level of 10%, 5% and 1% respectively.

As we can see from Figure 15 and 16, the values of the MEC in the KRX market rose dramatically after the imposition of the trading requirements.

As a matter of fact, we can find

a few cases where the null hypothesis of random walk cannot be rejected in the KRX market

21

after the imposition of the trading requirements.

Table 6 and Table 7 compare the averages of

the MEC and the execution cost for the period before the trading requirements were imposed and the period after.

As we can see from these tables, the KRX market had values of the MEC

lower than those of the OTC market and values of the execution cost higher than the OTC market before the imposition of the trading requirements.

After the imposition of the trading

requirements, however, the values of the MEC have risen to become larger than those of the OTC market and the values of the execution cost have fallen to become smaller than those of the OTC market.

Therefore, we can conclude that the mandatory trading requirements were

effective in enhancing the liquidity and the efficiency of the KRX market and the OTC market but that they were more effective in enhancing the liquidity and the efficiency in the KRX market. We can also observe from Figure 15 and 16 that the MEC of the OTC market have also risen significantly after the imposition of the trading requirements.

The increase in the MEC

and the decrease in the execution cost in the OTC market can be also confirmed from Table 6 and Table 7 that compare the averages of the MEC and the execution cost for the period before the introduction of the trading requirements and the period after the introduction of the trading requirements.

These findings confirm our earlier proposition that the trading requirements

have contributed to enhancing not only the liquidity and the transparency of the KRX bond market but those of the OTC market.

That is, the imposition of the trading requirements

enabled the transaction prices in the KRX market to reflect the supply and demand condition of the government bond market more accurately and thus contributed to enhancing the liquidity and the transparency of the OTC market as the participants in the OTC market began using the transaction prices in the KRX market as the reference prices for their own deals.

22

Table 6. Averages of MEC and Execution Cost of 3-year KTBs Period Before

Variable MEC

imposing the trading requirement After imposing the trading requirement

C(%) MEC C(%)

Entire Market

Exchange Market

OTC Market

0.235

0.193

0.277

(0.153)

(0.187)

(0.109)

0.135

0.165

0.106

(0.085)

(0.093)

(0.070)

0.693

0.782

0.604

(0.144)

(0.107)

(0.124)

0.033

0.026

0.040

(0.010)

(0.006)

(0.007)

Table 7. Averages of MEC and Execution Cost of 5-year KTBs Period Before

Variable MEC

imposing the trading requirement After imposing the trading requirement

C(%) MEC C(%)

Entire Market

Exchange Market

OTC Market

0.206

0.118

0.294

(0.173)

(0.079)

(0.201)

0.300

0.433

0.167

(0.243)

(0.289)

(0.064)

0.567

0.755

0.379

(0.372)

(0.426)

(0.196)

0.074

0.030

0.117

(0.071)

(0.065)

(0.049)

23

V. Conclusion

To develop the secondary market for government bonds, the Korean government has taken a unique strategy of developing an exchange market in addition to the OTC market even though there was the risk of dividing the liquidity between two markets.

The rationale for taking such

a strategy is that the KRX market based on the electronic trading platform will be able to facilitate the price discovery process of the secondary market and thereby enhance the transparency of secondary market transactions.

In addition to adopting the most advanced

electronic trading system, the government introduced and strengthened the exchange trading obligations on the primary dealers to boost trading in the KRX market. Our analysis using the intraday trading data in the KRX market and the OTC market demonstrate that the Korean government took the right strategy. Various measures of market performance confirm that the liquidity of the KRX market has improved significantly after the imposition of the trading requirements. of the OTC market.

What is more, these benefits did not come at the cost

On the contrary, our analysis using the intraday trading data shows that

the liquidity and the transparency of the OTC market has improve as well after the imposition of the trading requirements though there is still some room for improvement in the transparency of the OTC market. The Korean experience will be useful for Asian countries that try to develop their own domestic bond markets.

Recently, there has been a great deal of discussion and effort to

develop regional bond markets in Asia. (Ito, 2004, Park and Park, 2004 and Oh, Park, Park and Rhee, 2003)

The necessity of developing regional bond markets in Asia originates from the

realization that the Asian financial crisis of 1997 could have been avoided or at least mitigated had there been well developed bond markets in the bank dominated Asian countries.

24

As is

pointed out by many including ADB(2003) and Eichengreen(2004), bond markets in Asia are small and underdeveloped except for a few countries. One of the lessons from the Korean experience is that these Asian countries should build up and develop the secondary market for government bonds based on an electronic trading system in the beginning rather than trying the develop the OTC market.

25

Reference

Asian Development Bank (2000), “Government Bond Market Development in Asia,” March. BearingPoint (2005), The Electronic Bond Market 2005. Christodoulopoulos, Thanasis N. and Ioulia Grigoratou (2005), "Measuring Liquidity in the Greek Government Securities Market," Working Paper No. 23, Bank of Greece. Downing, Chris and Frank Zhang (2004), “Trading Activity and Price Volatility in Municipal Bond Market," Journal of Finance, pp. 899-938 Eichengreen, Barry and Pipat Luengnaruemitchai (2004), “Doesn't Asia Have Bigger Bond Markets?" Paper presented at the BIS/KU conference on Asian Bond Market: Issues and Prospects on March 22-23.. Hasbrouck, Joel and Robert A. Schwartz (1988), "Liquidity and Execution Costs in Equity Markets," Journal of Portfolio Management 14, 10-16. Ito, Takatoshi (2004) “Promoting Currency Basket Bonds,” T. Ito and Y.C. Park (eds.) Developing Asian Bondmarkets, Asia Pacific Press, 67-89. Kang, Kenneth, Geena Kim and Changyong Rhee (2005) "Developing the Government Bond Market in Korea: History, Challenges and Implications for Asian Countries," paper presented at the AEP, October 7-8, 2004, Columbia University. Lo, Andrew W. and Craig A. MacKinlay (1988), “Stock Market Prices Do Not Follow Random Walks: Evidence from a Simple Specification Test.", the Review of Financial Studies, spring, pp. 41-63. Oh, Gyutaeg, Daekeun Park, Jaeha Park and Changyong Rhee (2003), “Building a Settlement Infrastructure for the Asian Bond Markets: AsiaSettle,” paper presented at “Workshop on East Asia’s Strategy for Regional and Global Financial Cooperation” on 11 October, 2003, Seoul.

26

Park, Daekeun and Yung Chul Park (2004), “Creating Regional Bond Markets in East Asia,” T. Ito and Y.C. Park (eds.) Developing Asian Bondmarkets, Asia Pacific Press, 16-66. Roll, Richard (1984) "A Simple Emplicit Measure of the Effective Bid-Ask Spread in an Efficient Market," Journal of Finance, 1127-1139.

27

Figure 1. Outstanding Amount of Government Bonds

Figure 2. Ratio of the Outstanding Amount to GDP

28

Figure 3. Trading Volume in the KRX Market and the OTC Market

Figure 4. Share of the KRX Market for Benchmark and Non-Benchmark Issues

29

Figure 5. Share of the KRX Market Trading by PDs and non-PDs

Figure 6. Types of Trading by Trading Partners (Benchmark Issues)

30

Figure 7. Bid-Ask Spreads of Benchmark Issues

Figure 8. Monthly Transaction Volume of Government Bonds

31

Figure 9. Monthly Turnover Ratio of Government Bonds

Figure 10. Standard Deviation of Transaction Price

32

Figure 11. Daily Standard Deviation after September 2003

Figure 12. Range of Intraday Transaction Prices

33

Figure 13. Transaction Volume of KTBs for Each 20 Minute Interval

Figure 14. Standard Deviation of Transaction Price for Each 20 Minute Interval

34

Figure 15. Estimates of the MEC and the Execution Cost for 3-year KTBs

Figure 16. Estimates of the MEC and the Execution Cost for 5-year KTBs

35