• No results found

The occurrence of mispricing among Dutch, cross-listed companies – What are the determinants?

N/A
N/A
Protected

Academic year: 2021

Share "The occurrence of mispricing among Dutch, cross-listed companies – What are the determinants?"

Copied!
33
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

1

The occurrence of mispricing among Dutch, cross-listed

companies – What are the determinants?

Nanne J. Veenstra University of Groningen

Abstract

This study investigates mispricing among shares of stocks that trade simultaneously in different markets. Specifically, the prices of nine Dutch companies listed on the AEX-index are compared with the prices of their cross-listings in the period 1985-2007. The study provides empirical evidence that mispricing is present and that it is persistent during longer periods of time. Mispricing results from the fact that domestic and foreign variables influence the listings on their indices. The arbitrage opportunities appear not to be fully exploited.

Keywords: cross-listed companies, mispricing, determinants, arbitrage JEL Classification Codes: F30, G15, G32

I. Introduction

In 1871, the English economist and logician William Stanley Jevons introduced his law of indifference, which states that there cannot be more than one price for any one article in the same market at the same time (Jevons, 1871). Nowadays, this law has become a fundamental economic theory and is better known as the purchasing power parity (PPP) theory or the law of one price. It is stated as follows: “In an efficient market, all identical goods must have only one price”. The intuition behind this law is that sellers of the good will only accept the highest prevailing price, whereas the buyers will only give the lowest current market price. In an efficient market, the convergence to one price is instant.

An interesting example of identical goods would be the security of a multi-listed company. Though the company is listed on several stock exchanges in one or more markets, every security of that company represents a claim on the cash flows of one and the same entity and should, at least according to the law of one price, have the same monetary value as any other common security of that entity, regardless of the location it is being traded on. However, several empirical studies have shown that mispricing does actually occur among the many different forms in which the assets of multi-listed companies are traded.

(2)

Dual-2

listing involves two companies that have contractually agreed to operate their businesses as if they were a single enterprise, while distributing cash flows to their shareholders in a prescribed fashion and retaining their separate legal identity and separate stock exchange listings (De Jong et al., 2003). Their shares are not interchangeable, since one part of the securities represents a claim on the cash flows of the one firm, and the other part represents a claim on the cash flows of the other firm. This distinction has its implications for the possible set-up of an arbitrage strategy. In the case of cross-listing, investors can profit easily and fast from apparent mispricing by buying the security of a company in the market where it is cheapest, and selling it in another where it is most expensive. In the case of dual-listing, such a strategy cannot be pursued. Since the securities are not interchangeable, an investor has to buy the security of a company in the market where it is underpriced and go short in the market where it is overpriced, and wait till prices converge. This process is more costly and long-winded and might take up to eight years (De Jong, Rosenthal and Van Dijk, 2004).

As was stated in the beginning of this section, the easy and fast arbitrage strategy among cross-listed companies (henceforth CLCs) is not expected to last very long in efficient markets since the law of on price will make prices converge. Yet empirical research provides evidence of the contrary. The goal of this study is to investigate which determinants are responsible for preventing this from happening. Research will be conducted on a new, distinctive dataset when compared to those of prior studies: more companies and their cross-listings are investigated and the timeframe is broadened extensively.

This paper is built up as follows. In section II, prior literature on mispricing is discussed. The findings of research on mispricing and arbitrage are reported. Section 3 describes the dataset and methodology. The dataset under investigation contains nine companies that are from Dutch origin, together with their fourteen cross-listings. Section 4 presents the outcomes. Last is section 5 which contains the conclusion.

II. Literature

(3)

3

are penalized by the Dutch withholding tax, which is why shares of the Dutch companies are less valuable to them. Apart from this, no satisfactory explanations can be given by the authors.

Froot and Dabora (1999) also conducted a study about mispricing among these so-called “Siamese Twin” companies. They added the DLC SmithKline Beecham to their dataset. They found that the one-day Royal Dutch / Shell return differential yields coefficients of about 0.15 on the S&P500, -0.50 on the FTSE100, and 0.30 on the AEX-index. This implies that a 1 percent appreciation of the S&P500 increases the relative price of Royal Dutch over Shell by about 15 basis points. For the FTSE and AEX these values are -50 and 30 basis points respectively. The coefficients on the exchange rate changes are also large, at -0.10 and -0.50 for the guilder/dollar and guilder/pound exchange rates. A one-percent appreciation of the guilder against the dollar and pound, respectively, increases the relative price of Royal Dutch over Shell by about 10 and 50 basis points. These coefficient values also imply that a 1% appreciation of the dollar relative to the pound increases the relative price of Royal Dutch over Shell by about 40 basis points (Froot and Dabora, 1999). A similar story is revealed for Unilever NV / PLC and SmithKline Beecham. They exclude taxes as a possible determinant of this mispricing, and relate most of their findings to country-specific noise, explaining: “Suppose that a noise shock hitting, say, U.S. stocks, disproportionately affects the twin which trades relatively more in New York. In other words, stocks that trade more actively in the local market are more sensitive to local noise shocks and less sensitive to foreign noise shocks. This story has an interesting implication: the component of market movements explained by changes in twin's relative prices is likely to be noise. Twin price disparities, which are readily observable, may therefore be informative about market-wide noise shocks, which are not directly observable (Froot and Dabora, 1999).

Froot and Dabora refer to similar findings in a study that investigates the discount on country equity funds (Hardouvelis, La Porta, and Wizman, 1995). Country funds are publicly traded investment companies (closed-end funds) that trade on the open market and, unlike domestic equity funds, hold and manage portfolios concentrating in the equity markets of particular foreign countries (Hardouvelis et al., 1995). The authors found that the prices of these shares differ from the net asset values of the fund portfolios. In particular, it appears that closed-end fund share prices comove most strongly with the stock market on which they trade, while net asset values comove most strongly with their local stock markets. The average discount on these funds is almost seven percent, caused by noise trading. Though not exactly the same as stock listings, it is apparent that deviations of the prices from the underlying value also exist for closed-end funds, and that comovement with market indices occurs.

(4)

4

differentials that they identified in the Anglo-Dutch cases, so too do fundamentals-based explanations appear unable to explain substantial price differences between the Australian and UK arms of the three more recent DLCs. For example, the stocks in each of these DLCs are all actively traded and appear in benchmark market indices, suggesting that liquidity differences are unlikely to be able to explain substantial price differences. Furthermore, tax factors do not appear to be able to justify the differential, since investors from third countries do not obtain any major tax advantage from investing in a particular twin.” (Bedi et al., 2003).

De Jong, Rosenthal and Van Dijk (2004) underwrite earlier conclusions that it is hard to come up with fundamental determinants for mispricing, which is also obstinate in their dataset of thirteen DLCs. They found large deviations from theoretical price parity with average absolute price discrepancies for individual twins range from roughly 2.5 percent to almost 12 percent. The deviations from parity reach values of over 15 percent for every single DLC in the sample and occasionally exhibit levels of up to 50 percent (De Jong et al., 2004). They too find evidence that the relative return of a twin is strongly affected by fluctuations in domestic market indices. They interpret this as evidence of country-specific noise.

From the literature discussed so far it appears that many researchers relate part of the mispricing to comovement of one of the stocks with “habitat-based” factors such as a country-specific index or the national currency. Barberis, Schleifer and Wurgler (2002) have provided empirical evidence for these claims by showing that inclusion to or deletions from the S&P500 results in an increase in or decrease in comovement with the index itself.

(5)

5

To conclude the literature so far, mispricing exists among the stocks of DLCs, CLCs, closed-end funds, tech stock carve outs and possibly many others. It is present nationally as well as internationally, and short-term as well as long-term. Table 1 gives a summary of the most important findings of the discussed literature. As can be derived from the literature, no consensus has been reached about the actual determinants of mispricing, nor is it clear why the apparent arbitrage possibilities that result from mispricing are not fully exploited. If mispricing would be exploited, persistent mispricing would not occur due to converging prices. Taxes, currencies, noise and comovements together with numerous other variables have all been mentioned when trying to explain for the occurrence of mispricing, but the findings are not unanimous. Goal of this paper is to broaden the literature on this field by conducting research on mispricing on a whole new dataset, but it also hopes to make a contribution to the understanding of mispricing and to come up with new, valuable insights.

Now that prior literature on mispricing has been discussed, it is worthwhile to consider why companies in fact choose to cross-list in other countries. One study that has shed light on this subject is from Karolyi (1998). He comes forward with the following reasons for cross-listing:

- Share prices react favorably to cross-border listings initially;

- Post-listing trading volume increases on average, and, for many issues, home-market trading volume increases also;

- Share liquidity improves overall (but depends on the increase in total trading volume, the listing location and the scope of foreign ownership restrictions in the home-market);

- Exposure to domestic market risk is significantly reduced and is associated with only a small increase in global market risk and foreign exchange risk, which can result in a net reduction in the cost of equity of about 126 basis points;

An effective global diversification tool used for cross-listing is the American Depositary Receipt (ADR). ADRs were developed by JP Morgan in 1927 as a vehicle for investors to register and earn dividends on non-U.S. stock without direct access to the local market itself. They are more cost efficient to investors than are direct investments in home-market shares.

Another study on this subject is from Pagano, Röell and Zechner (1999), who add to the former enumeration the following reasons for cross-listing:

- Cross-listing appears to be driven by the need to raise equity to fund growth and foreign sales expansion;

- By listing abroad, firms may improve the terms on which they can raise capital or on which their shareholders can sell securities;

(6)

6

- Cross-listing reduces barriers for foreign investors (regulatory barriers, transaction costs and informational frictions);

- Cross-listing enables companies to profit from overvaluation in one country relative to another.

(7)

7 Table 1:

Summary of discussed literature

Authors Year of

publication

Period under investigation

Research Scope Findings Explanations

Rosenthal, L., and C. Young

1990 1979-1986 Can be profited from mispricing among DLCs?

Royal Dutch Petroleum vs. Shell Transport and Trading PLC, Unilever NV vs. Unilever PLC.

The stock prices exhibit persistent and strikingly large deviations from the ratio of adjusted cash flows.

The direction and magnitude of the mispricing indicate systematic rather than specific origins. Part of the early mispricing may be explained by the relative attractiveness of Shell and Unilever PLC over their Dutch counterparts to British institutional investors. Pension funds are able to obtain the full, pretax dividends from the British companies, but are penalized by the Dutch withholding tax, which is why shares of the Dutch companies are less valuable to them. Apart from this, no satisfactory explanations can be given. Despite pricing differences, no evidence of profitable intramarket trading rules has been found.

Hardouvelis, G.A., R. la Porta, and T.A. Wizman

1995 1985-1993 The discount on country equity funds.

Thirty-five single-country, publicly traded funds from all over the world.

Country funds trade at an average discount. Fund prices are “sticky” with respect to movements in the host country’s stock market and overly sensitive to variation in the U.S. and world stock markets.

The average discount for funds is largely attributable to noise trading. The large fluctuations in the discounts are the result of reversions in investor sentiment. Barriers to cross-border capital movements are determining the prevalence of discounts in the long-run: absence of such barriers make discount prevail over time. The oversensitivity to the world stock market index is partly driven by “world” fundamentals. Market frictions caused by informational factors and non-synchronous trading have no influence on the discount factor. Froot, K.A., and

E.M. Dabora

1999 1980-1994 Mispricing among DLCs.

Royal Dutch Petroleum vs. Shell Transport and Trading PLC, Unilever NV vs. Unilever PLC, SmithKline Beecham.

Stock prices are affected by the location of trade. The DLCs exhibit comovements between price differentials and market indexes.

Market-wide noise shocks from irrational traders can infect locally traded stocks more than foreign traded stocks. However, the source of noise or persistent irrationality is difficult to identify.

Institutional inefficiencies might explain comovements. By virtue of higher liquidity or inclusion in domestic-market indexes, one twin may be classified as a “domestic stock”. Such a classification appears to have contractual consequences.

Barberis, N., A. Schleifer, and J. Wurgler

2002 1976-2000 Comovement Inclusions to (590) and deletions from (565) the S&P500.

Stocks included into the index begin to comove more with other stocks in the index, and less with stocks out of the index. The converse holds for deletions.

(8)

8

Table 1, continued: Summary of discussed literature

Authors Year of

publication

Period under investigation

Research Scope Findings Explanations

Bedi, J., A. Richards, and R. Tennant

2003 1980 – 2002 The characteristics of DLCs.

Twelve DLCs from all over the world.

The results of this paper bolster the finds of Froot and Dabora (1999).

Price divergences between DLC twins and excess comovement in market valuation are pervasive phenomena in DLC cases, consistent with Froot and Dabora (1999) De Jong, A., L.

Rosenthal, and M. van Dijk

2004 1980-2002 Limited arbitrage among DLCs.

Thirteen DLCs from all over the world.

The relative prices of all twins exhibit statistically significant and economically substantial deviations from theoretical parity. Combined arbitrage strategies in all DLCs produce excess returns of up to 10% per annum on a risk-adjusted basis, after transaction costs and margin requirements.

The relative return of a twin is strongly affected by fluctuations in domestic market indices. This can be interpreted as evidence of country-specific noise.

The risks associated with arbitrage strategies form an important obstacle to DLC arbitrage.

Gagnon, L., and G.A. Karolyi

2004 1993-2002 Multi-market trading and arbitrage.

ADRs and other types of CLCs in U.S. markets relative to home-market shares, totaling 581 pairs of CLCs from 39 countries.

The return differentials on the CLCs exhibit excess comovements relative to market index returns. These excess comovements are related to the primary location of trading in the respective shares.

(9)

III. Data and methodology

The dataset under investigation contains Dutch, cross-listed companies that were listed on the AEX-index on the reference date December 31st, 2007. The AEX-index contains the 25 most actively traded securities in The Netherlands per annum. The particular dates have been chosen so as to conduct a study with an up-to-date dataset, yet large time frame. The base date of the dataset is March 4, 1983, the day on which the AEX-index was initiated. The total time span is thus 6.521 trading days. The initial dataset consisted of 22 companies (3 companies were already delisted in 2007 due to takeovers). Of those companies, 7 were excluded since they were not cross-listed. Another 4 companies were excluded since they could be referred to as dual-listed companies, and thus fall out of the scope of this research. Last, two companies were excluded since they were listed less than one year on the AEX-index. What is left is a dataset of 9 companies, which are presented in table 2 together with their time frame of listing in the AEX, the core-sector of their operations and their cross-listings. A precise realization of this dataset is presented in appendix I.

Table 2:

Nine Dutch Cross-listed companies in the time frame of March 4, 1983 – December 31, 2007

Company Base date End date ICB1 Sector Cross-listings

AEGON 01-03-1985 31-12-2007 Life insurance NYSE (ADR), TSE AHOLD KON 01-03-1985 31-12-2007 Food retailers and wholesalers Non-NASDAQ OTC (ADR) AKZO NOBEL 01-03-1985 31-12-2007 Specialty chemicals Non-NASDAQ OTC (ADR) ASML HOLDING 20-02-1998 31-12-2007 Semiconductors NASDAQ (ADR)

CORPORATE

EXPRESS 01-03-2001 31-12-2007 Industrial suppliers NYSE (ADR) ING GROEP 01-03-1991 31-12-2007 Life insurance NYSE (ADR)

KPN KON 17-02-1995 31-12-2007 Fixed line telecommunications FSE, NYSE (ADR) PHILIPS KON 03-01-1985 31-12-2007 Consumer electronics NYSE (ADR)

TNT 29-06-1998 31-12-2007 Delivery services NASDAQ OTC, Non-NASDAQ OTC (ADR)

The base and end date refer to the AEX-listing of each company. Appendix II provides an overview of the base and end dates of all the listings of each company. For the obtained listings go that data ranges may not coincide with the official listing data, which is due to the availability of the listing series in DataStream.

1

(10)

10

In line with the discussed literature, examination of the occurrence of mispricing and the occurrence of comovement with domestic as well as foreign variables will be performed by using the methodology of Froot and Dabora (1999). Their methodology is valued as being precise and appropriate since later literature also relies on and refers to the experimental design of Froot and Dabora.

Mispricing of the cross-listing relative to the home-listing is computed by taking the ratio of the former to the latter. The ratio gives information about the relative price premium (or discount) at which the cross-listing trades. The equation that will be estimated for every listing i is as follows:

(1) Pp, t = (Pf, t / Pd, t – 1) * 100%

where Pp, t is the price premium of the cross-listing relative to the home-listing on time t, Pf, t is the price of the cross-listing on time t, and Pd, t is the price of the home-listing on time t. The output of the equation when applied to all cross-listings for the complete dataset will be given in terms of descriptive statistics of the examined price premiums, as well as in graphical representations.

After having determined whether price premiums exist, possible comovement of all listings (home-listings and cross-listings) with their “habitat-based” variables such as the home-market, the exchange rate of the domestic currency, the domestic interest rate and the domestic inflation rate will be examined. In particular, an equation is applied that determines the slope coefficients of these “habitat-based variables” when trying to explain for the observed return of a listing. Examination of the slope coefficients will yield an answer to the question whether they are a determinant for the observed return. The equation, that will be estimated for every listing i, is as follows:

(2) Rt = β0 + β1Rc, t + β2Rm1, t + β3Rm2, t + β4Rm3, t + β5Rr, t + β6Rf r, t + µt

where Rt is the return of a specific listing on time t, Rc, t is the percentual change in the domestic exchange rate (relative to a foreign currency, which is discussed later) on time t, Rm1, t , Rm2, t and Rm3, t are the returns of the domestic and foreign market indices on time t respectively, Rr, t is the percentual change in the domestic interest rate on time t, and Rfr, t is the percentual change in the domestic inflation rate on time t. Since Rt reflects daily return data, a GARCH model is applied. The results of the regressions will be projected against the following hypotheses:

H0: All slope coefficients in (2) are equal to zero and thus the domestic variables have no explanatory power for the observed returns.

H1: All slope coefficients in (2) are not equal to zero and thus the domestic variables have explanatory power for the observed returns.

(11)

11

applied that determines the slope coefficients of the domestic and foreign variable differentials when trying to explain for the observed stock return differential. The equation is as follows:

(3) Rf, t - Rd, t = β0 + β1Rfc, t + β2Rfm1, t + β3Rfm2, t + β4Rdm, t + β5Rrr, t + β6Rfr, t + β7Rff, t + β8R fd, t + µt

where Rf, t is the return of a cross-listing on time t, Rd, tis the return of the home-listing on time t, Rfc, t is the percentual change in the foreign exchange rate on time t, Rdc, t is the percentual change in the domestic currency on time t, Rfm1, t and Rfm2, t are the returns of the first and second foreign market index on time t respectively, Rdm, t is the return of the domestic market index on time t, Rfr, t is the percentual change in the foreign interest rate on time t, Rrr, t is the percentual change in the domestic interest rate on time t, Rfr, tis the return of the foreign inflation rate on time t, and Rfd, t is the return of the domestic inflation rate on time t. Since Rf, t - Rd, t reflects daily return data, a GARCH model is applied. All returns are log returns and will be computed for every listing i with the following equation:

(4) Rt = log (Pt / Pt-1)

where Rt is the return on time t, Pt is the price on time t, and Pt-1 is the price on time t-1. The results of the regressions will be projected against the following hypotheses:

H0: All slope coefficients in (3) are equal to zero and thus the domestic and foreign variables do have explanatory power for the differentials in log returns.

H1: All slope coefficients in (3) are not equal to zero and thus domestic and foreign variables have no explanatory power for the differentials in log returns.

All coefficients, other relevant statistics and their implications will be discussed. Table 3 presents an overview of the exchange rates and foreign market indices that are used for equations (2) and (3). Since the listings trade in 4 different countries (The Netherlands, Germany, Japan and the United States of America), 4 exchange rates are used (this applies to the interest and inflation rates as well). The cross-listings trade on a total of 4 different stock exchanges (AEX, FSE, TSE and NYSE), but for some cross-listings more than one market index was included.

Table 3:

Exchange rates and foreign market indices

Country Exchange rate Market indices

The Netherlands Guilder-Dollar* and Euro-Dollar AEX

Germany Mark-Dollar** DAX30

Japan Yen-Dollar*** NIKKEI225

(12)

12

Stock prices, index prices, exchange rates, interest rates and inflation rates were all obtained from DataStream. For stock prices the total return index has been used, in order to adjust for stock-splits and dividend pay-outs. For the companies AEGON and KPN, the LSE-listings have been excluded from the dataset since DataStream could not provide for the proper equity series. The equity series were examined in EViews. The following section discusses the results.

IV. Results

Table 4 presents the descriptive statistics of the computed price premiums resulting from equation (1). For all distributions of the price premiums go that they are not symmetric, given their respective values for skewness and kurtosis. The values resulting from the Jarque-Bera test underwrite the observation of non-normality. Graphical representations of the price premiums are presented in appendix III. It is immediately clear that persistent mispricing is present for all cross-listings. This first observation proofs already that mispricing is more rule rather than exception, as was also the conclusion of the literature review. The extent to which mispricing among cross-listings occurs, differs however. From the figures in appendix III can also be seen that outliers and price shocks are present. These are not representative of the distribution, since they represent mispricings that can be related to non-systematic occurrences. Table 5 gives an enumeration of occurring events at the time these outliers and shocks are identifiable and which may have been the cause of them being present. A clear example is the correction of AHOLD’s stock price on the OTC-listing after the announcement of a rights offering of new common shares in November 2003: mispricing relative to the AEX-listing reduced from more than 30% to approximately 12% in only a few trading days. Yet what remains striking is the fact that the apparent mispricing was not ruled out at once: it took a reverse stock split almost four years later to put an end to the arbitrage opportunity.

Now that the outliers and shocks have been explained by events that probably can be held accountable for their occurrence, “regular” mispricing is discussed. For seven out of twelve cross-listings the sign of mispricing fluctuates regularly: at some periods in time the foreign listing sells at a premium, whereas at other the AEX listing is more expensive again. For the cross-listings of that group of companies goes that the mean and median of their deviations from parity are close to zero. These cross-listings are ING’s NYSE-listing, KPN’s FSE-listing and KPN’s NYSE listing. Profiting from arbitrage is not worthwhile in these cases, apart from incidental somewhat larger deviations, since the spread is too small. The economic law of one price somehow holds for these cross-listings when compared to the AEX-listings of their company.

(13)

13

TNT’s OTC- and OTC (ADR)-listing. Arbitrage opportunities may be worthwhile to exploit since the apparent gains can be significant, but corrections follow up very quickly. This implies that profiting from arbitrage may be a risky business, especially when fluctuating signs of mispricing follow up on one another regularly. This was also concluded by De Jong, Rosenthal and Van Dijk (2004).

Cross-listings that deviate from the average mispricing being near to zero are the following: AEGON’s NYSE-listing that sells at a discount up until early 1995; AHOLD’s OTC-listing that sells at a striking premium of approximately 30% up until mid 1994 and thereafter at a 12% premium before the reverse stock split put the deviations from parity to an end in mid 2007; AKZO’s OTC-listing selling from a 10% discount and gradually declining to zero during its first three years; CORPORATE EXPRESS’s NYSE-listing that traded between a 5% and 10% discount during its first three year. It is clear that the exploitation of the arbitrage possibilities is very profitable: deviations from parity are persistent over larger periods of time and easily exceed +/- 20%. In some instances, deviations of +30% or even +40% occurred. These figures are similar to those of Froot and Dabora (1999) and De Jong et al. (2004). Deviations also often gradually decline to zero. However, it is striking that these possibilities were not used to the fullest, otherwise such long and extreme mispricing would not have occurred: prices would have converged faster then.

(14)

14 Table 4:

Descriptive statistics of the computed price premiums resulting from equation (1)

Mean Median Max. Min. Std. Dev. Skewn. Kurt. Jarque-Bera Sum Sum Sq. Dev. Obs.

AEGON

NYSE / AEX -0.028 -0.005 0.153 -0.184 0.048 -0.523 1.944 540.858 *** -163.146 13.324 5,872

TSE / AEX 0.005 -0.007 0.664 -0.269 0.108 1.518 8.262 6,847.964 *** 24.460 46.091 4,453

AHOLD KON OTC / AEX 0.262 0.314 0.468 -0.018 0.098 -0.888 2.488 622.127 *** 1,144.886 42.308 4,366

AKZO NOBEL OTC / AEX -0.016 -0.001 0.079 -0.143 0.042 -1.222 3.466 1,255.495 *** -75.671 8.558 4,865

AMSL HOLDING NASDAQ / AEX -0.046 -0.048 0.175 -0.159 0.025 1.172 10.254 8,080.777 *** -153.799 2.061 3,338

CORPORATE EXPRESS NYSE / AEX 0.037 0.041 0.178 -0.048 0.039 0.426 2.260 80.796 *** 56.779 2.372 1,522

ING GROEP NYSE / AEX 0.005 0.002 0.150 -0.139 0.016 1.905 12.944 15,937.61 *** 17.964 0.885 3,373

KPN KON

FSE / AEX 0.002 0 0.134 -0.096 0.013 0.517 11.033 8,130.733 *** 4.709 0.474 2.975

NYSE / AEX 0.006 0.003 0.148 -0.132 0.020 0.607 5.812 1,243.862 *** 19.900 1.230 3.181

PHILIPS NYSE / AEX 0.012 0.017 0.365 -0.242 0.070 -0.179 3.526 100.431 *** 72.797 29.070 5,957

TNT

OTC / AEX -0.003 -0.002 0.139 -0.111 0.028 0.389 6.812 413.003 *** -2.105 0.512 655

OTC (ADR) / AEX 0.001 0.001 0.132 -0.096 0.021 0.117 6.056 971.044 *** 3.060 1.059 2.481

(15)

15

Table 5:

Enumeration of occurring events that may have led to the outliers and shocks that are present in appendix III

Company and listing relative to AEX Date Occurring event

AEGON

NYSE / AEX October 19, 1987 Black Monday: stock market crash

September 11, 2001 Terrorist attack upon the United States of America July 22, 2002 Court filing for violating the Securities Exchange Act

TSE / AEX

May 19, 1998 No attributable cause found October 9, 1998 No attributable cause found

March 26, 1999 Announcement of the opening of an Aegon Square September 11, 2001 Terrorist attack upon the United States of America September 23, 2002 Restructuring of the equity in place

AHOLD KON OTC / AEX November 26, 2003 “Two for three” rights offering of new common shares August 22, 2007 Reverse stock split

AKZO NOBEL OTC / AEX April 18, 1994 Possible aftermath of the IPO with unfounded trading September 14, 1992 No attributable cause found

AMSL HOLDING NASDAQ / AEX January 3, 2001 April 24, 1995 Possible aftermath of the IPO with unfounded trading No attributable cause found

September 11, 2001 Terrorist attack upon the United States of America

CORPORATE EXPRESS NYSE / AEX

August 9, 2002 Publication of 2nd quarter results

November 1, 2002 Publication about appointment new CEO March 18, 2003 No attributable cause found

March 30, 2005 Successful completion of shares repurchase

ING GROEP NYSE / AEX

August 31, 1998 Publication of 2nd quarter results

October 1, 1998 No attributable cause found

September 11, 2001 Terrorist attack upon the United States of America July 24, 2002 Publication about resigning CFO

KPN KON

FSE / AEX

March 11, 1997 No attributable cause found January 4, 1999 No attributable cause found May 1, 2000 No attributable cause found

September 11, 2001 Terrorist attack upon the United States of America

NYSE / AEX

August 31, 1998 No attributable cause found

September 11, 2001 Terrorist attack upon the United States of America November 22, 2001 Extraordinary layoffs on UMTS

July 24, 2002 Filed request of approval for call rates with OPTA

PHILIPS NYSE / AEX May 10, 2001 No attributable cause found

TNT OTC / AEX August 10, 2007 Repurchase of own shares

OTC (ADR) / AEX January 1, 1999 No attributable cause found

(16)

16

goodness of fit coefficient and an insignificant F-value. Their results may thus be considered not representative.

In general, it can be said that the listings of the examined companies are influenced by domestic market variables to a greater or lesser degree. A variable whose coefficient is mostly positive and highly significant is the home-market index. For almost all listings goes that this variable is highly influential. Foreign market indices are also quite often deterministic, though be it to a lesser degree. The Dow Jones hardly shows signs of comovement overall. This may result from the fact that inclusion of the S&P500 already takes this into account. For cross-listings goes that the foreign market index is mostly influential: the AEX-index often exhibits a significant relationship. This implies that the cross-listing comoves with the home-market index. Only for the cross-listings of AEGON, PHILIPS and TNT this was not the case, but it was stated earlier that these results are not representative.

The exchange rate also exhibits significant influence, though not as influential as the market index. This result may stem from the fact that the degree to which companies rely on foreign turnover differs. It is known that AEGON’s and AKZO NOBEL’s financial results are for a large part attributable to their success on the U.S. market. It is therefore not surprising that the Euro-Dollar exchange rate is highly significant. Other firms show a less significant relation with the exchange rate, because of their lower degree of internalization. It can thus be stated that the influence of the exchange rate is present, though be it case-dependable. It can be seen that the domestic interest rate and inflation rate appear to have minor influence. In the rare cases that they are significant, it is only marginal. Based on the results, these variables can not be attributed an influential role.

(17)

17 Table 6:

Coefficients and relevant statistics resulting from equation (2)

β0 β1 Β2 β3 β4 β5 β6 R2 Adj. R2 Durbin Watson F-statistic

Exchange rate Market index Market index Market index Interest rate Inflation rate

AEGON AEX 0 (-1.227) 0.135 (0.864) 0.988 *** (14.807) -0.077 (-0.455) 0.104 (0.553) 0.037 0.484) 0.008 (1.468) 0.633 0.617 1.905 39.803 *** NYSE (ADR) - 0.001 (-1.256) 0.555 *** (3.910) 1.115 *** (5.957) 0.040 (0.193) 0.122 ** (2.101) -0.080 (-1.120) 0.015 (1.432) 0.693 0.679 1.865 52.063 *** TSE 0 (0.248) 0.153 (-0.481) 0.100 (0.531) -0.066 (-0.364) - -0.090 (-0.857) -0.001 (-0.752) 0.017 -0.021 2.038 0.446 AHOLD KON AEX 0 (1.240) -0.138 (-0.817) 0.897 *** (12.453) 0.426 ** (1.985) -0.497 ** (-2.339) -0.022 (-0.269) 0.023 (1.380) 0.369 0.341 2.355 13.496 *** OTC 0 (0.229) 0.378 (1.451) 0.355 ** (2.092) 0.309 (1.509) 0.250 ** (2.315) 0.137 (1.431) -0.009 (-0.543) 0.264 0.232 2.488 8.272 *** AKZO NOBEL AEX 0 (-0.314) 0.140 (0.713) 0.833 *** (10.091) -0.483 ** (-2.072) 0.403 (1.611) 0.146 (1.262) -0.003 (-0.234) 0.273 0.242 1.738 8.679 *** OTC -0.001 (-0.980) 1.016 *** (5.051) 0.349 *** (2.669) 0.184 (1.528) 0.402 *** (5.570) 0.103 * (1.745) 0.049 (1.390) 0.303 0.273 1.947 10.045 *** ASML HOLDING AEX 0 (0.139) 0.431 (1.445) 0.856 *** (6.704) -1.039 *** (-3.005) 0.910 *** (2.847) -0.106 (-0.711) 0.007 (0.295) 0.259 0.225 2.062 7.607 *** NASDAQ 0 (0.201) 0.254 (0.833) 0.513 *** (2.807) 0.432 * (1.889) 0.153 (1.066) -0.281 * (-1.953) 0.024 (1.361) 0.357 0.328 1.931 12.102 *** CORPORATE EXPRESS AEX -0.003 * (-1.742) 0.155 (0.455) 0.914 *** (6.182) -0.409 (-0.911) 0.271 (0.542) 0.083 (0.353) 0.023 (0.924) 0.145 0.108 1.807 3.909 *** NYSE -0.003 * (-1.895) 0.656 * (1.857) 0.656 (1.312) 0.131 (0.241) 0.184 * (1.670) 0.117 (0.745) -0.004 (-0.112) 0.171 0.135 1.912 4.769 *** ING GROEP AEX -0.001 ** (-2.111) 0.130 (0.943) 1.091 *** (19.472) -0.068 (-0.425) 0.056 (0.369) 0.006 (0.085) -0.006 (-0.641) 0.738 0.726 1.747 64.968 *** NYSE 0 (-0.825) 0.628 *** (3.820) 1.142 *** (6.798) -0.002 (-0.012) 0.195 *** (3.838) -0.005 (-0.091) 0.007 (0.717) 0.734 0.722 2.208 63.648 ***

(18)

18 Table 6, continued:

Coefficients and relevant statistics resulting from equation (2)

β0i β1 Β2 β3 β4 β5 β6 R2 Adj. R2 Durbin Watson F-statistic

Exchange rate Market index Market index Market index Interest rate Inflation rate

KPN KON AEX 0 (0.109) -0.233 (-0.946) 0.907 *** (8.897) -0.113 (-0.359) 0.013 (0.044) -0.076 (-0.620) -0.016 (-1.080) 0.390 0.364 2.024 14.776 *** FSE 0 (0.130) - 0.167 (0.833) 0.533 *** (2.809) - 0.115 * (1.887) -0.007 (-0.543) 0.222 0.196 2.330 8.561 *** NYSE 0 (0.188) 1.076 *** (4.137) 0.666 * (1.787) 0.061 (0.179) 0.298 *** (2.759) 0.059 (0.607) -0.014 (-0.907) 0.446 0.422 2.030 18.637 *** PHILIPS KON AEX -0.001 (0.968) -0.624 * (-1.766) 0.120 (0.728) -0.274 (-0.637) -0.145 (-0.340) -0.105 (-0.534) -0.023 (-0.972) 0.057 0.016 2.102 1.400 NYSE 0 (-0.197) 0.470 (1.274) -0.247 (-0.562) -0.176 (-0.404) 0.114 (0.689) 0.069 (0.349) -0.051 * (-1.988) 0.069 0.029 2.232 1.714 * TNT AEX -0.001 (-0.570) 0.561 ** (2.288) 0.817 *** (5.814) -0.358 (-1.175) 0.575 ** (2.053) -0.200 (-1.088) 0.017 (0.748) 0.375 0.329 2.153 8.066 *** OTC 0 (-0.066) 0.756 (1.324) 0.124 (0.333) 0.205 (0.443) -0.093 (-0.300) 0.101 (0.598) 0.004 (0.093) 0.095 0.027 2.055 1.405 OTC (ADR) 0 (0.062) -0.233 (-0.797) 0.259 (1.073) -0.414 (-1.550) 0.001 (0.009) -0.071 (-0.750) 0.025 (1.418) 0.039 -0.003 2.086 0.933

(19)

19

Table 7 presents coefficients and relevant statistics resulting from equation (3). Appendix VI gives the denotations of all variables for every specific listing that are used when applying equation (3), and appendix VII presents the variance equation variables for equation (3). For the results goes that the explanatory power of most regressions is high. The significant F-statistics imply that all slope coefficients of the regressions together significantly differ from zero. As a result, the null hypothesis which states that all slope coefficients are equal to zero should be rejected. The exceptions are KPN (FSE-listing), PHILIPS (NYSE-listing) and TNT (OTC (ADR)-listing, but based on the coefficients of determination their results may be considered not representative. In none of the regressions autocorrelation in the residuals is present, since the Durbin Watson statistics are all close to 2.

(20)

20 Table 7:

Coefficients and relevant statistics resulting from equation (3)

β0 β1 Β2 β3 β4 β5 β6 β7 β8 R2 Adj. R2 Durbin Watson F-statistic

Exchange rate Market index Market index Market index Interest rate Interest rate Inflation rate Inflation rate

AEGON NYSE (ADR)-AEX 0 (0.298) -0.230 ** (-2.325) 1.194 *** (8.758) -0.075 (-0.569) -0.857 *** (-17.116) 0.003 (0.092) -0.065 (-0.996) 0.005 (1.138) -0.001 (-0.284) 0.794 0.783 2.386 72.110 *** TSE-AEX 0.001 (1.003) -0.693 *** (-3.558) 0.108 (0.713) - -1.197 *** (-6.913) -0.061 (-0.616) -0.120 (-0.866) -0.006 (-0.863) -0.002 (-0.118) 0.345 0.313 2.061 10.882 *** AHOLD KON OTC-AEX 0 (-1.426) -0.001 (-0.011) 0.036 (0.316) 0.706 *** (5.129) -0.632 *** (-8.572) 0.009 (0.172) -0.129 (-1.567) -0.003 (-0.940) 0.003 (0.724) 0.321 0.284 2.542 8.834 *** AKZO NOBEL OTC-AEX 0 (1.391) -0.196 ** (-2.193) -0.089 (-1.255) 0.901 *** (11.422) -0.634 *** (-14.776) 0.008 (0.247) 0.001 (0.026) 0.004 * (1.820) -0.003 (-1.010) 0.615 0.594 2.700 29.873 *** ASML HOLD. NASDAQ-AEX 0 (0.274) -0.456 ** (-2.459) 0.300 *** (2.984) 0.733 *** (6.083) -0.751 *** (-8.333) 0.040 (0.961) 0.012 (0.112) 0.017 *** (5.418) -0.007 (-1.450) 0.553 0.528 2.727 21.828 *** CORP. EXPR. NYSE-AEX 0 (1.132) -0.061 (-0.442) 1.048 *** (4.622) -0.180 (-0.816) -0.652 *** (-8.786) -0.085 (-1.105) -0.144 (-1.592) 0.005 (1.175) 0.003 (0.449) 0.427 0.396 2.847 13.953 *** ING GROEP NYSE-AEX 0 (1.053) -0.338 *** (-3.004) 1.071 *** (9.102) -0.154 (-1.140) -0.638 *** (-13.785) -0.047 (-1.133) -0.149 (-1.171) 0 (-0.051) 0.001 (0.349) 0.682 0.665 2.679 40.150 *** KPN KON FSE-AEX 0 (-0.261) - -0.019 (-0.186) -0.159 * (-1.668) 0.050 (1.087) -0.013 (-0.185) 0.001 (0.232) -0.001 (-0.270) 0.037 0.002 2.939 1.063 NYSE-AEX 0 (-0.288) 0.019 (0.120) 0.726 *** (4.153) -0.046 (-0.262) -0.713 *** (-6.399) 0.062 (0.862) 0.020 (0.202) 0.010 (0.897) -0.007 (-1.238) 0.742 0.666 2.214 9.682 *** PHILIPS NYSE-AEX 0 (0.877) -1.014 *** (-4.520) 0.184 (0.863) -0.210 (-1.025) 0.130 (1.380) 0.010 (0.135) -0.012 (-0.088) -0.013 (-1.240) 0.007 (1.104) 0.086 0.037 2.965 1.768 * TNT OTC-AEX 0 (-0.135) -0.040 (-0.110) 0.330 (1.339) -0.195 (-0.697) -0.616 *** (-3.397) 0.059 (0.451) 0.052 (0.243) -0.006 (-0.447) -0.002 (-0.246) 0.118 0.075 2.312 2.728 *** OTC (ADR)-AEX 0.001 (1.177) -0.749 ** (-2.488) 0.083 (0.397) -0.112 (-0.463) -0.980 *** (-6.766) -0.117 (-1.323) 0.104 (0.802) 0.005 (0.727) -0.011 (-1.422) 0.227 0.190 2.829 5.995 ***

(21)

21

V. Conclusion

This study provides evidence of significant differences in the prices of shares of stock that trade simultaneously in different markets around the world. For nine Dutch, cross-listed companies with a total of twelve cross-listings, synchronous and intraday prices of these cross-listings relative to the home-listings were compared. The magnitude of the price deviations from parity, their persistence over time and the systematic comovement with domestic and foreign variables like market indices, exchange rates, interest rates and inflation rates were examined. The foremost finding is that mispricing occurs among cross-listings as much as it does among other types of derivatives and listings. The effect may be brief and fluctuate highly, but it may also be long-lasting and significantly one-sided. The foremost example of the latter is AHOLD’s OTC/AEX deviation from price parity for more than 16 years. The conclusion from this study is very clear: domestic and foreign market indices as well as the domestic exchange rate are responsible for the observed mispricing: they influence the returns of the listings under investigation and result in deviations from price parity.

The advantage of the dataset used in this study in comparison to those of the discussed literature is that it contains more companies and their cross-listings (21 listings in total) over a larger and more recent timeframe. The findings lead to the observation that the puzzle is a lot bigger than expected: with a more elaborate dataset and when applied to cross-listings instead of dual-listings, the expectation would be that mispricing should be less apparent for cross-listings, yet excessive mispricing is still found. There appear to be no differences among “normal” shares and ADRs.

(22)

22

the home-country bias. This phenomenon states that investors suffering from this bias are more comfortable investing in firms within their own country, for the simple fact that these are more familiar to them. In doing so, mispricing holds.

(23)

23

References

Barberis, N., A. Shleifer, and J. Wurgler, 2003, Comovement, working paper, University of Chicago. Bedi, J., A. Richards, and P. Tennant, 2003, The characteristics and trading behavior of dual-listed

companies, working paper, Reserve Bank of Australia.

De Jong, A., L. Rosenthal, and M. van Dijk, 2003, The limits of arbitrage: Evidence from dual-listed companies, working paper, Erasmus University.

De Jong, A., L. Rosenthal, and M. van Dijk, 2004, The risk and return of arbitrage in dual-listed companies, working paper, Erasmus University.

Froot, K.A., and E.M. Dabora, 1999, How are stock prices affected by the location of trade?, Journal of Financial Economics, 53(2), pp 189–216.

Gagnon, L., and G.A. Karolyi, 2003, Multi-market trading and arbitrage, working paper, Ohio State University.

Hardouvelis, G., R. la Porta, and T. Wizman, 1994, What moves the discount on country equity funds?, The Internationalization of Equity Market, The University of Chicago Press.

Jevons, W.S., 1871, The Theory of Political Economy, Reprint of 1931 edition, Charlottesville, Virginia, pp 90-5.

Karolyi, G.A., 1998, Why do companies list shares abroad: A survey of the evidence and its managerial implications, Financial Markets, Institutions & Instruments, 7(1), pp 1–60.

Pagano, M., A.A. Röell, and J. Zechner, 1999, The geography of equity listing: Why do European companies list abroad?, CSEF Working Paper no. 28, University of Salerno.

Rosenthal L., and C. Young, 1990, The seemingly anomalous price behavior of Royal Dutch/Shell and Unilever N.V./PLC, Journal of Financial Economics, 26(1), pp 123–141.

Shleifer, A., and R. Vishny, 1997, The limits of arbitrage, Journal of Finance, 53, pp 35–55. Young, A.A., 1912, Jevons' "Theory of Political Economy", American Economic Review, Vol. 2, No.

(24)

24

Appendix I: Realization of AEX dataset

Deviations from the original 2007 AEX-index compilation (base date: March 2, 2007)

Initial number of firms in sample: 25

Deleted due to mergers and acquisitions: 3 (ABN AMRO, HAGEMEYER, NUMICO)

Deleted because of no cross-listing: 7 (ARCELOR MITTAL, DSM, HEINEKEN, SBM OFFSHORE, TOMTOM, VEDIOR, WOLTERS KLUWER)

Deleted because of dual-listing: 4 (FORTIS, ROYAL DUTCH SHELL, REED

ELSEVIER, UNILEVER)

Deleted because listing less than one year: 2 (RANDSTAND, UNIBAIL RODAMCO)

Final dataset: 9 (AEGON, AHOLD KON, AKZO NOBEL, ASML

HOLDING, CORPORATE EXPRESS, ING

(25)

25

Appendix II: Data range for each listing in the respective stock exchange

Company Cross-listings Base date End date

AEGON AEX 01-03-1985 31-12-2007

NYSE 28-06-1985 31-12-2007

TSE 06-12-1990 31-12-2007

AHOLD KON AEX 01-03-1985 31-12-2007

OTC 08-04-1991 31-12-2007

AKZO NOBEL AEX 01-03-1985 31-12-2007

OTC 09-05-1989 31-12-2007

ASML HOLDING AEX 15-03-1995 31-12-2007

NASDAQ 15-03-1995 31-12-2007

CORPORATE EXPRESS

AEX 01-03-2001 31-12-2007

NYSE 14-03-2002 31-12-2007

ING GROEP AEX 01-03-1991 31-12-2007

NYSE 26-01-1995 31-12-2007

KPN KON AEX 17-02-1995 31-12-2007

FSE 06-08-1996 31-12-2007

NYSE 23-10-1995 31-12-2007

PHILIPS KON AEX 03-01-1985 31-12-2007

NYSE 03-01-1985 31-12-2007

TNT AEX 29-06-1998 31-12-2007

OTC 28-06-2005 31-12-2007

(26)

26

Appendix III: Graphical representations of price premiums resulting from equation (1)

-25% -20% -15% -10% -5% 0% 5% 10% 15% 20% Ju n-85 Dec -87 Ju n-90 Dec -92 Ju n-95 Dec -97 Ju n-00 Dec -02 Ju n-05 Dec -07 P r ic e d e v ia ti o n

Figure A-1: Log deviations from AEGON’s NYSE/AEX parity during the period of 28/06/1985 up until 12/31/2007. This figure shows on a percentage basis the deviations from theoretical parity of AEGON shares traded on the NYSE and AEX.

-40% -20% 0% 20% 40% 60% 80% Dec -90 Ju n-93 Dec -95 Ju n-98 Dec -00 Ju n-03 Dec -05 P r ic e d e v ia ti o n

Figure A-2: Log deviations from AEGON’s TSE/AEX parity during the period of 06/12/1990 up until 12/31/2007. This figure shows on a percentage basis the deviations from theoretical parity of AEGON shares traded on the TSE and AEX.

-10% 0% 10% 20% 30% 40% 50% Ap r-91 Oct -92 Ap r-94 Oct -95 Ap r-97 Oct -98 Ap r-00 Oct -01 Ap r-03 Oct -04 Ap r-06 Oct -07 P r ic e d e v ia ti o n

(27)

27 -20% -15% -10% -5% 0% 5% 10% Ap r-91 Oct -92 Ap r-94 Oct -95 Ap r-97 Oct -98 Ap r-00 Oct -01 Ap r-03 Oct -04 Ap r-06 Oct -07 P r ic e d e v ia ti o n

Figure C-1: Log deviations from AKZO’s OTC/AEX parity during the period of 09/05/1989 up until 12/31/2007. This figure shows on a percentage basis the deviations from theoretical parity of AKZO shares traded OTC and on the AEX.

-20% -15% -10% -5% 0% 5% 10% 15% 20% Mar -95 Se p-96 Mar -98 Se p-99 Mar -01 Se p-02 Mar -04 Se p-05 Mar -07 P r ic e d e v ia ti o n

Figure D-1: Log deviations from ASML’s NASDAQ/AEX parity during the period of 15/03/1995 up until 12/31/2007. This figure shows on a percentage basis the deviations from theoretical parity of ASML shares traded on the NASDAQ and AEX.

-10% -5% 0% 5% 10% 15% 20% Mar -02 Se p-03 Mar -05 Se p-06 P r ic e d e v ia ti o n

Figure E-1: Log deviations from CORPORATE EXPRESS’s NYSE/AEX parity during the period of 14/03/2002 up until 12/31/2007. This figure shows on a percentage basis the deviations

(28)

28 -20% -15% -10% -5% 0% 5% 10% 15% 20% Ja n-95 Jul-9 6 Ja n-98 Jul-9 9 Ja n-01 Jul-0 2 Ja n-04 Jul-0 5 Ja n-07 P r ic e d e v ia ti o n

Figure F-1: Log deviations from ING’s NYSE/AEX parity during the period of 26/01/1995 up until 12/31/2007. This figure shows on a percentage basis the deviations from theoretical parity of ING shares traded on the NYSE and AEX.

-15% -10% -5% 0% 5% 10% 15% Au g-96 Fe b-98 Au g-99 Fe b-01 Au g-02 Fe b-04 Au g-05 Fe b-07 P r ic e d e v ia ti o n

Figure G-1: Log deviations from KPN’s FSE/AEX parity during the period of 06/08/1996 up until 12/31/2007. This figure shows on a percentage basis the deviations from theoretical parity of KPN shares traded on the FSE and AEX.

-15% -10% -5% 0% 5% 10% 15% 20% Oct -95 Ap r-97 Oct -98 Ap r-00 Oct -01 Ap r-03 Oct -04 Ap r-06 Oct -07 P r ic e d e v ia ti o n

(29)

29 -30% -20% -10% 0% 10% 20% 30% 40% Mar -85 Se p-86 Mar -88 Se p-89 Mar -91 Se p-92 Mar -94 Se p-95 Mar -97 Se p-98 Mar -00 Se p-01 Mar -03 Se p-04 Mar -06 Se p-07 P r ic e d e v ia ti o n

Figure H-1: Log deviations from PHILIPS’s NYSE/AEX parity during the period of 01/03/1995 up until 12/31/2007. This figure shows on a percentage basis the deviations from theoretical parity of PHILIPS shares traded on the NYSE and AEX.

-15% -10% -5% 0% 5% 10% 15% 20% Ju n-05 Dec -05 Ju n-06 Dec -06 Ju n-07 Dec -07 P r ic e d e v ia ti o n

Figure I-1: Log deviations from TNT’s OTC/AEX parity during the period of 28/06/2005 up until 12/31/2007. This figure shows on a percentage basis the deviations from theoretical parity of TNT shares traded OTC and on the AEX.

-15% -10% -5% 0% 5% 10% 15% Ju n-98 Dec -99 Ju n-01 Dec -02 Ju n-04 Dec -05 Ju n-07 P r ic e d e v ia ti o n

(30)

30

Appendix IV: Denotation of each variable in equation (2)

Rc, t Rm1, t Rm2, t Rm3, t Rr, t Rf r, t

AEGON

AEX Euro-Dollar AEX S&P500 Dow Jones Dutch interest rate Dutch inflation rate NYSE Dollar-Euro S&P500 Dow Jones AEX U.S. interest rate U.S. inflation rate

TSE Yen-Euro NIKKEI225 AEX - Japanese interest rate Japanese inflation rate AHOLD KON AEX Euro-Dollar AEX S&P500 Dow Jones Dutch interest rate Dutch inflation rate

OTC Dollar-Euro NASDAQ100 S&P500 AEX U.S. interest rate U.S. inflation rate AKZO NOBEL AEX Euro-Dollar AEX S&P500 Dow Jones Dutch interest rate Dutch inflation rate

OTC Dollar-Euro NASDAQ100 S&P500 AEX U.S. interest rate U.S. inflation rate ASML

HOLDING

AEX Euro-Dollar AEX S&P500 Dow Jones Dutch interest rate Dutch inflation rate NASDAQ Dollar-Euro NASDAQ100 S&P500 AEX U.S. interest rate U.S. inflation rate CORPORATE

EXPRESS

AEX Euro-Dollar AEX S&P500 Dow Jones Dutch interest rate Dutch inflation rate NYSE Dollar-Euro S&P500 Dow Jones AEX U.S. interest rate U.S. inflation rate ING GROEP AEX Euro-Dollar AEX S&P500 Dow Jones Dutch interest rate Dutch inflation rate

NYSE Dollar-Euro S&P500 Dow Jones AEX U.S. interest rate U.S. inflation rate KPN KON

AEX Euro-Dollar AEX S&P500 Dow Jones Dutch interest rate Dutch inflation rate FSE - DAX30 AEX - German interest rate German inflation rate NYSE Dollar-Euro S&P500 Dow Jones AEX U.S. interest rate U.S. inflation rate| PHILIPS KON AEX Euro-Dollar AEX S&P500 Dow Jones Dutch interest rate Dutch inflation rate|

NYSE Dollar-Euro S&P500 Dow Jones AEX U.S. interest rate U.S. inflation rate| TNT

(31)

31

Appendix V: Variance equation variables for equation (2)

(32)

32

Appendix VI: Denotation of each variable in equation (3)

Rfc, t Rfm1, t Rfm2, t Rdm, t Rfr, t Rrr, t R fr, t R fd, t

AEGON

NYSE-AEX Dollar-Euro S&P500 Dow Jones AEX U.S. int.rate Dutch int.rate U.S. infl.rate Dutch infl.rate TSE-AEX Yen-Euro NIKKEI225 - AEX Japanese

int.rate Dutch int.rate Japanese infl.rate Dutch infl.rate AHOLD KON OTC- AEX Dollar-Euro NASDAQ100 S&P500 AEX U.S.

int.rate Dutch int.rate U.S. infl.rate Dutch infl.rate AKZO NOBEL OTC- AEX Dollar-Euro NASDAQ100 S&P500 AEX U.S.

int.rate Dutch int.rate U.S. infl.rate Dutch infl.rate ASML NASDAQ-AEX

Dollar-Euro NASDAQ100 S&P500 AEX U.S. int.rate Dutch int.rate U.S. infl.rate Dutch infl.rate CORP. EXPRESS

NYSE- AEX Dollar-Euro S&P500 Dow Jones AEX U.S. int.rate Dutch int.rate U.S. infl.rate Dutch infl.rate ING GROEP NYSE- AEX Dollar-Euro S&P500 Dow Jones AEX U.S.

int.rate Dutch int.rate U.S. infl.rate Dutch infl.rate KPN KON

FSE- AEX - DAX30 - AEX German

int.rate Dutch int.rate German infl.rate Dutch infl.rate NYSE- AEX Dollar-Euro S&P500 Dow Jones AEX U.S.

int.rate Dutch int.rate U.S. infl.rate Dutch infl.rate PHILIPS KON NYSE- AEX Dollar-Euro S&P500 Dow Jones AEX U.S.

int.rate Dutch int.rate U.S. infl.rate Dutch infl.rate TNT

OTC- AEX Dollar-Euro NASDAQ100 S&P500 AEX U.S. int.rate Dutch int.rate U.S. infl.rate Dutch infl.rate OTC (ADR)- AEX

(33)

33

Appendix VII: Variance equation variables for equation (3)

β0 Residual(-1)² GARCH(-1) AEGON NYSE (ADR)-AEX 0 (1.428) 0.145 ** (2.032) 0.738 *** (6.515) TSE-AEX 0 *** (6.616) -0.039 *** (-4.339) 0.908 *** (44.440)

AHOLD KON OTC-AEX 0 ***

(7.198)

1.005 *** (4.664)

-0.004 (-0.406)

AKZO NOBEL OTC-AEX 0 **

(2.460)

0.660 *** (3.523)

0.333 *** (3.092)

ASML HOLD. NASDAQ-AEX 0 **

(2.082) 0.231 ** (2.572) 0.655 *** (5.392) CORP. EXPR. NYSE-AEX 0 ** (2.545) 0.297 *** (3.413) 0.641 *** (9.158)

ING GROEP NYSE-AEX 0 ***

Referenties

GERELATEERDE DOCUMENTEN

The following subjects are discussed during the interviews: the process concerning choosing the appropriate study, more specific the wants and needs of people concerning

So far, UTAUT2 has been used to study online subjects similar to social media platforms, such as consumer acceptance behavior of mobile shopping (Marriott & Williams,

Variables include: fraction of cash (FC), Intra- industry acquisitions (IntraInd), pre-deal acquirers firm size (Size), cash holdings of acquirer (CashHol), relative size of

Investment is derived by dividing capital expenditures (CAPEX) by total net Property, Plant & Equipment (PPE). Tobin's Q equals the market value of equity plus total assets

Next to that, due to a higher percentage of companies that are exposed when stock returns are lagged with one period these levels of exposure will be applied in the

In the case of foreign competition it is expected the inverted U shape has a short but steep positive effect, meaning for relatively low levels of competition innovation

The results in this model indicate that an appreciation of the local currency against the US Dollar has a more significant effect on domestic credit growth

Although this study does not find significant evidence that differences among cross-border and domestic M&As exist, it does find significant differences