(The) linkage between the depository receipts and the underlying stocks : an empirical analysis on the Korean companies listed in the United States and Great Britain

다국어 초록 (Multilingual Abstract)

Before the 1997 Asian Financial Crisis, Korea was still at an infancy level of financial development. Korea first issued its foreign listed shares in 1990 with the Global Depository receipt being issued by Samsung Corporations. However, since 1998, Korea's foreign listing has drastically increased, amongst all 1999 far stood out with US $6.7 billion, a 32% of the ten year foreign issuance. The DRs that are issued in foreign markets are based on the securities that are listed in Korea and traded through the Korea Stock Exchange, and recently, the DRs can be converted into shares, and the shares into DRs if some of the conditions are met.
This paper tries to understand the relationship between the underlying stock (UA) and the DRs (DR) that are issued to represent certain number of underlying stock, and examine whether a lead-lag relationship can be verified between the two pairs. If forecasting of the future stock movement of a security is possible using the VAR model, this method can be used for enabling the arbitrage trading, namely the famous and long run Wall Street trading technique, pairs trading. For this reason, the following null hypotheses have been tested on this paper:
[Hypothesis I] The daily return of the underlying stock (denoted as UA) and the Depository Receipt (denoted as DR) have a correlation of one.
[Hypothesis II] The daily return of UA at time t is only influenced and correlated with the daily return of DR adjusted by the F/X rate but it is not influenced by the historical return of DR.
[Hypothesis III] Daily return of DR and daily log return of UA have causal relationship.
3-1 The historic daily return of DR granger causes the daily return of UA. (If current time is t, the historic time is considered t-1, t-2, ..., 1)
3-2 The historic daily return of UA granger causes the daily return of DR.
The estimation results for the following hypotheses are as follows:
The cross correlation examination shows a close track record among the five ADRs vs UAs, however, the relationship was found to be in a much loose status in the case of GDR. The cross correlation in the ADR vs UA results between 0.55 and 0.71 and the cross correlation in the GDR vs UA, between 0.07 and 0.43.
The findings through the first hypothesis testing are the coefficients between the UAs and DRs that range between 0.53 and 0.88 in the case of ADR vs UA, and much lower result between 0.07 and 0.39 in the case of GDR vs UA. Through the Wald test, the null hypotheses that "coefficient between UA and DR is one" is rejected. However, the coefficients are found to be significant under the 95% confidence level with the t-statistics above two. This allows the coefficients to be accepted as it is found.
The Wald test conducted by restricting the coefficient between the historic return of the independent variable and the dependent variable, to zero, the following have been found. The ADR vs UA conclusion shows that historic return of ADR influences the return of UA, as the null hypothesis is rejected. In the case of GDR vs UA, the return of UA has been found to influence (i.e. granger cause) return of GDR. This finding shows that the ADR influences UA, which then influences the GDR. It is in accordance with the market opening sequence, with the US market opening first, followed by the Korean market, and then further passed on to the British market.
The Granger F-test and the VAR model results also show a similar findings as the results of the above tested hypotheses. ADRs are found to granger cause UAs, and UAs are found to granger cause GDR. The low r-squared are the only hindrance to a concrete forecasting of its paired security
Therefore, this paper concludes that a lead-lag relationship has been found between the DRs and UAs. Further research needs to be conducted, however to use the finding for an arbitrage trading, and more sophisticated models need to be invented for more accurate forecasting of the stock return.

목차 (Table of Contents)

• Acknowledgement = i
• Table of Contents = iii
• List of Tables = v
• List of Figures = vii
• List of Acronyms and Abbreviations = ix
• Abstract = x
• I. INTRODUCTION = 1
• 1.1. Research Purpose and Background = 1
• 1.2. Research Scope and Structure of the Paper = 2
• II. DEPOSITORY RECEIPT = 5
• 2.1. Definition of Depository Receipt = 5
• 2.1.1 Depository Receipt through Registered Offering = 7
• 2.1.2 Depository Receipt through Unregistered Offering = 10
• 2.2 The History of Korean Companies' Foreign Listing = 11
• 2.3 The Registration Process of DR and the Conversion Method = 14
• 2.3.1 Registration Process of DR = 15
• 2.3.2 Conversion Process between the DR and the Underlying Stock = 20
• III. ARBITRAGE AND ARBITRAGE TRADING = 22
• 3.1. Definition of Arbitrage = 22
• 3.2. Arbitrage Trading = 23
• 3.3. Short Selling and Stock Borrowing & Lending = 24
• IV. DATA AND SIMPLE DIAGNOSTIC ANALYSIS = 27
• 4.1. Sample Data and Sample Period = 27
• 4.2. Definition of Variables used in the analyses = 29
• 4.3. Explanation and Definition of the Hypotheses = 30
• 4.4. Contemporaneous Regression Analysis = 35
• 4.4.1 Cross Correlation Analysis = 35
• 4.4.2 Simple Linear Regression Analysis = 39
• 4.4.3 Multiple Regression Analysis = 53
• V. VECTOR AUTOREGRESSION MODEL = 61
• 5.1. VAR Estimation Results = 61
• 5.2. Impulse Response Function and Variance Decomposition Analysis = 64
• 5.3. Granger Causality Test = 77
• VI. IMPLICATIONS FROM THE VAR RESULTS = 79
• VII. CONCLUSION = 81
• BIBLIOGRAPHY = 84
• APPENDICES = 86