• No results found

Banks' Customers Satisfaction and Stock's Returns : Banking Sector - Sweden, Norway, Denmark

N/A
N/A
Protected

Academic year: 2021

Share "Banks' Customers Satisfaction and Stock's Returns : Banking Sector - Sweden, Norway, Denmark"

Copied!
34
0
0

Loading.... (view fulltext now)

Full text

(1)

BANKS’ CUSTOMER SATISFACTION &

STOCK’S RETURNS

BANKING SECTOR – SWEDEN, DENMARK, NORWAY

Paper within Master Thesis in Finance Author: Konstantopoulou Nikoletta Tutor: Agostino Manduchi Jönköping May 21th 2012

(2)

i

ABSTRACT

Theoretical studies posit that marketing strategies increases customers’ satisfaction and loyalty and decreases the systematic risk of the company’s stock. Many variables such as size, book-to-market and others, which have no special standing in asset-pricing theory, show reli-able power to explain the cross-section of expected stock’s returns. By adding customers’ sat-isfaction to one of them, this research involves discovering the relationship between custom-ers’ satisfaction and stock’s returns systematic risk, if any, by conducting a panel data analy-sis of seven banks in Sweden, Denmark and Norway through the period of year 2002 – 2011. The results verify a significant negative relationship between customers’ satisfaction and stock’s returns systematic risk.

(3)

ii

Table of Contents

1

Introduction ... 1

2

Background ... 2

2.1 Market’s Up- and Downswings and Customers Satisfaction ... 2

2.2 Shareholder Value - Risk ... 4

2.3 Customer Satisfaction and Stock Returns Risk ... 7

3

Problem and Purpose ... 9

4

Research Questions ... 10

4.1 Customer Satisfaction and Systematic Risk ... 10

5

Methodology - Estimation Procedure ... 11

5.1 Dependent Variable ... 11

5.1.1 Systematic Risk ... 11

5.1.2 Robustness Check ... 12

5.2 Customer Satisfaction ... 13

5.3 Accounting Measurements ... 13

5.3.1 Size of the firm ... 13

5.3.2 Return on Assets (ROA) ... 13

5.4 Lag Dependent Variable ... 14

5.5 Final Model for systematic risk ... 14

6

Data Collection ... 16

7

Results ... 17

7.1 Final Model - Fama and French three factor model ... 17

7.2 CAPM ... 17

8

Conclusion and Limitations... 19

8.1 Conclusion ... 19 8.2 Limitations ... 19

9

References ... 20

Appendix... 23

Appendix 1 ... 23 Appendix 2 ... 24 Appendix 3 ... 25 Appendix 4 ... 27 Appendix 5 ... 28 Appendix 6 ... 29 Appendix 7 ... 31

(4)

1

1

Introduction

Nowadays, very often marketers are asked if the substantial resources they invest in market-ing strategies are financially efficient for the company. The complexity of answermarket-ing this question is that from one side marketing managers are translating their results into customers’ satisfaction and loyalty and from the other side investors and financial managers are interest-ed in increasing revenue, cash flow and shareholder value (Srivastava, Shervani and Fahey, 1998).

In finance, shareholder value is translated into high expected stock’s returns or low stock’s systematic risk (Brown, Martin and Gruber, 2010). Many variables, like size, book-to-market, momentum and others, which have no special standing in asset-pricing theory, show reliable power to explain the cross-section of expected stock returns (Fama and French, 1993). Con-sidering customers’ satisfaction as one of them, few researchers had shown a negative rela-tionship of customers’ satisfaction and stock’s systematic risk. Furthermore, they all focus on the U.S. market. After a comprehensive research, I found that no similar empirical research has been done so far specifically for the Scandinavian countries Sweden, Denmark and Nor-way.

This paper thus is based on the banking sector in Sweden, Denmark and Norway. It adds cus-tomers’ satisfaction to the variables used in asset pricing theory to explain the expected stock’s returns and examines if customers’ satisfaction is a factor which should be captured in asset pricing models and if it can be used as a proxy for the latent risk factor. By taking a sample including the leading banks of the Swedish, Danish and Norwegian market and the customers’ satisfaction scores from EPSI rating (2002 - 2011) and by conducting a panel data analysis, I am testing if customers’ satisfaction and stock’s returns systematic risk have a sig-nificant negative relationship (Epsi Rating, no year; Gujarati and Porter, 2009).

(5)

2

2

Background

2.1 Market’s Up- and Downswings and Customers Satisfaction

Market oriented economies are characterized by unpredictable fluctuations in aggregate eco-nomic activities, which are called market’s up- and downswings (Long and Plosser, 1983). Upswings and downswings involve fluctuations over time between periods of relatively rapid economic growth, booms, and periods of relative stagnation or decline, recessions (Long and Plosser, 1983). During booms, a shift in demand can raise markets output, lower its price and raise income (Rotemberg and Saloner, 1986). Price wars are common in booms. This can be explained by the fact that when demand is high, a firm can lower a bit its prices and take the whole demand (Rotemberg and Saloner, 1986). The benefit from this is higher than the pun-ishment of not collude1 (Rotemberg and Saloner, 1986). So, even though the total income ris-es, a frequent phenomenon during booms is the price wars. In recessions, on the contrary, a decline in demand can reduce the output. Demand outplaces supply and the prices remain constant, may even rise, but the total income declines and the market falls (Rotemberg and Saloner, 1986).

The effects of market’s up- and downswings are more intensive in the case of homogeneous products or services as in this case the customers are more price-sensitive (Singh and Vives, 1984). From the customers’ perspective, the product is a combination of value satisfactions (Levitt, 1980). When customers perceive a product or service as homogeneous the price be-comes the differentiated strategy (Levitt, 1980). A fractionally lower price then, gets the whole demand (Levitt, 1980). These fluctuations and their effects are especially relevant to up- and downswings in securities’ markets. Companies, therefore, need strategies in order to differentiate their product or service from the competitors, create value for the customer, fore-cast more accurate the demand and protect themselves from these fluctuations and their mac-roeconomic effects. In that way they can remain long term viable and profitable.

A definition of marketing suggests that based on the customers’ view, the marketing purpose is “to establish, develop and commercialise long term customer relationships, so that the

ob-jectives of the parties involved are met. This is done by mutual exchange and keeping of promises” (Grönroos, 1989). I am using this definition because by including the term “long

term customer relationships” it is more relevant to my topic. Based on marketing theories and researches, marketers are developing and implementing efficient marketing strategies in order

1

Game theory suggests that in equilibrium oligopolies collude under the threat of reverting to competitive be-havior whenever a single firm does not cooperate.

(6)

3

to increase customers’ satisfaction and to protect the company from the consequences of mar-ket’s up- and downswings. They support that by differentiating their product or service, the benefits and the satisfaction, that customers receive, increase. Therefore, by increasing cus-tomers’ satisfaction and loyalty and by strengthening the business and the product respective-ly service image, the company will be long-term profitable and will not be affected, or at least will be affected less, from market’s fluctuations (Yu and Dean, 2001).

Moreover, as price wars are common in booms and as companies have realized that price is not an advantage weapon anymore, the companies which have loyal customers will be more profitable. A loyal customer will buy more, will make recurring purchases, will talk more about the firm (word of mouth strategy) and will not try to find a substitute (a competitor’s firm) (Berthon, Hulbert and Pitt, 1999). All these strategies will help the company to be fi-nancially improved without getting involved in the price war. Price wars are common in the business world because managers see them as an easy action to gain market share and tempo-rary profits. An increase in expected future profit reduces the initiatives for price-wars and markups staying in a higher level (Rotemberg and Woodford, 1991). In order to gain market share in the future without reducing the prices the companies can create efficient marketing strategies, differentiate their product or service, increase customers satisfaction and build “glamorous” brands and create a competitive advantage (Dutta, Zbaracki and Bergen, 2003). In downswings or recessions, when the total market declines, companies tend to compete more in order to keep their customers. In this case too, companies with loyal customers will have competitive advantages. Highly satisfied customers are more loyal and less likely to give in to other firms’ offers (Heide and Weiss, 1995). Consequently, remaining loyal to the firm makes the firm less vulnerable to market’s declines and more attractive to investors.

The value of customer retention is especially high in the banking sector (Reichheld and Ken-ny, 1990-1991). As they mention, the two main strategies of cost reduction or price increase, can cause only short-term profits. On the other hand, customers’ retention and loyalty cause growth and margins for several reasons. Balances raise through time as interest accrues, many accounts are consolidated and the economic situation of the customer is superior (Reichheld and Kenny, 1990-1991). The cost of keeping a customer is approximately fixed but the cost of attracting new customers is high and includes increased promotion and advertising expendi-tures and rate escalation (Reichheld and Kenny, 1990-1991). The cost of gaining a new cus-tomer is incurred only the first year, which means that the older the relationship is, the lower the amortised cost (Reichheld and Kenny, 1990-1991). Loyal customers could expand their purchasing manner also to other products/ services (Reichheld and Kenny, 1990-1991). A bank with loyal depositors has the advantage that its competitors will not react fast and this is because retention it is not easily measured (Reichheld and Kenny, 1990-1991). With all these arguments and some examples they are suggesting customers’ satisfaction as a comparative advantage against their competitors from one hand and market’s economic fluctuations from

(7)

4

the other. A measurement of the yearly market value is the Gross Domestic Product (GDP). The graph shows the gross domestic product of the countries and through the period tested, Sweden, Denmark and Norway. A downswing can be considered around 2008 and an up-swing around 2010.

graph 1

The question now is: Can marketing, by increasing customers’ satisfaction, protect the firm from the market fluctuations? Can we count this strategy in financial/ economical results? Some theoretical marketing studies posit that customers’ satisfaction lowers the stock’s re-turns risk (Srivastava, Shervani and Fahey, 1998). However, very little empirical research has verified this relationship (Tuli and Bharadwaj, 2009).

2.2 Shareholder Value - Risk

There are two ways to increase shareholder’s value: high expected stock’s returns or low stock’s returns systematic risk (Brown, Martin and Gruber, 2010).

From the asset pricing theory, the basic model for pricing risky securities, which describes the relationship between risk and expected return, is the Capital Asset Pricing Model (CAPM) (Fama and French, 2004). CAPM suggests that the equilibrium return on any risky security is equal to the sum of the risk-free rate of return and a risk premium measured by the product of the market price of risk and the security's systematic risk. The general idea behind CAPM is that investors need to be compensated in two ways: time value of money and risk (Fama and French, 2004). According to Brown, Martin, and Gruber (2010) the time value of money compensates the investors for placing money in any investment over a period of time and

rep-0 5E+11 1E+12 1,5E+12 2E+12 2,5E+12 3E+12 3,5E+12 2011 2010 2009 2008 2007 2006 2005 2004 2003 2002

GDP

(8)

5

resented with on the equation one below; (beta), risk premium, in the formula represents the systematic risk and calculates the amount of compensation the investor needs for taking on additional risk. Beta measures the sensitivity of the asset’s returns to variation in the mar-ket return (Fama and French, 2004).

(1) .

In simple words, when investors are evaluating a company, they are interested in the tradeoff between returns and systematic risk in minimum variance portfolios: the more systematic risk a security has, the higher the rate of return should be and the beta (systematic risk) is the only parameter which affects this relationship (graph two). Many researchers showed that some of the variation in expected returns is unrelated to market beta (Basu, 1977; Banz 1981). Thefore, accounting factors like earnings, size, etc add to the explanation of expected stock re-turns, explained by market beta and have been used as proxies for the latent risk factors. Therefore, market beta is not a complete description of asset’s systematic risk - high standard error (ei), not normally distributed.

graph 2

The risk of the asset can be divided in two different kinds of risks: systematic and idiosyncrat-ic (Brown, Martin, and Gruber, 2010). Systematidiosyncrat-ic risk is the market risk or the risk associated with the market movements and all securities are affected by the systematic risk (Brown, Martin, and Gruber, 2010). The companies which can protect themselves from the impact of market movements have lower systematic risk. Idiosyncratic risk, on the other hand, is the risk which is unique for every company and is affected mainly from company’s actions (Tuli and Bharadwaj 2009). Idiosyncratic risk can be eliminated in a well diversified portfolio. So, what matters to investors is only the systematic risk of stock’s returns.

(9)

6

As I already argued systematic risk is the risk associated with the market movements. Market is influenced by the different market’s up- and downswings (Long and Plosser, 1983). Mar-ket’s up- and downswings refers to the ups and downs seen somewhat unpredictably and sim-ultaneously in most parts of an economy. These fluctuations have an important effect on secu-rities’ markets. More detailed, in recessions, where the market declines, stocks with high ex-posure in market’s β (beta) will be more vulnerable and insecure. Even in the case of a boom, where the market rises, companies which can prevent to be involved in the price wars as-sumed as more stable in their future earnings without the risk to lose their market share in the future.

Stock’s returns systematic risk is a key component of shareholder value that affects the finan-cial markets (Barber and Odean, 2000). Thus, it makes sense that investors are trying to find factors that can reduce the systematic risk of the stock’s returns and its impact from the up- and downswings. Based on the financial literature, variables that are significant predictors of the systematic risk are: firm size, financial leverage, profitability, and earnings variability (Coles, Daniel and Naveen, 2006). McAlister, Srinivasan and Kim (2007) have shown that advertising/sales and R&D/sales are also important factors in reducing firm’s systematic risk. A model which adds more factors in order to explain the variability in returns is Fama- French three factors model (equation two) and it is the main model used in portfolio management to measure market returns (Fama and French, 1993; Lin, Wang and Cai, 2012).

Fama-French three factors model predicts company’s systematic risk by using the two non-market risk factors SMB (the difference between the return on a portfolio of small stocks and the return on a portfolio of large stocks) and HML (the difference between the return on a portfolio of high-book-to-market stocks and the return on a portfolio of low-book-to-market stocks) (Brown, Martin, and Gruber, 2010). Taking also these factors into account Fama-French three factors model is considered as more accurate than the CAPM (Fama and Fama-French, 1993). Nevertheless, it cannot be assumed as the complete description of the variability in re-turns (Fama and French, 1993).

The financial theory requires that the systematic risks should be net priced. If the CAPM is not misspecified, omit some relevant explanatory variables, means that the expected return is completely captured by β (beta). If this is the case the standard error of the regression is white noise and normally distributed (Chen, 1983). Otherwise, the remaining part must be contained in the standard error (Chen, 1983). Similarly, Fama-French three factors model, although it includes also the size proxy, and the book-to-market equity proxy, possibly is not a complete description of the variability in returns. Thus, the part which is not explained must be

(10)

includ-7

ed in the standard error. In both cases, there must be some other explanatory factors which can price the remaining part of the expected returns (Chen, 1983). These factors can be used as proxies to the latent risk and should be net priced. Should we expect the “extra satisfac-tion” to be one of them? To verify this, the classical econometric tactic is to run the regression with the error term as dependent variable and the customers’ satisfaction scores as independ-ent and observe if some part of the error term can be priced by the customers’ satisfaction fac-tor.

2.3 Customer Satisfaction and Stock Returns Risk

Three main reasons are supporting the fact that firms with strong customers’ satisfaction scores should have lower systematic risk. First, by enabling rapid product identification and reducing consumer search costs, brands with high consumer based equity facilitate repeat-purchasing behavior (Berthon, Hulbert and Pitt, 1999). Second, customers’ satisfaction is re-lated to consumers’ emotional connection and strengthens the loyalty, which means they will rebuy the brand (Newman and Werbel, 1973). Last, but not the least, loyal consumers are less affected from the rivals (Rego, Billett and Morgan, 2009).

Few empirical researches have already been done about this topic, but they are based on US business world. For example, Tuli and Bharadwaj (2009) proved that customers’ satisfaction is a key component in reducing firm’s total risk (systematic, downside systematic, idiosyn-cratic, downside idiosyncratic). They conducted restrict sensitivity tests and their results are shown to be robust from financial and economical view.

Rego, Billett and Morgan (2009) found that a firm’s Consumers’ Based Brand Equity (CBBE) is associated with firm’s risk and explains variance in the risk measures. They also support the idea that CBBE has a stronger role in predicting firm-specific unsystematic risk than system-atic risk, but that it also has a particularly strong role in protecting equity holders from down-side systematic risk. The results of this research have clear economic significance (Rego, Billett, & Morgan, 2009) but they adopt conservative risk estimation and without a robustness test, they cannot be totally accepted from the financial researchers.

Fornell, Mithas, Morgeson III and Krishnan (2006) with their analyses concluded that firms with highly satisfied customers usually have positive abnormal returns. In the same time they posit that news about changes in customer satisfaction or the announcement of the scores does not have an effect on stock prices, so they cannot be used as for trading actions. The reasons for this are that maybe the market already expects satisfaction information, so there is no “surprise” from the announcement and, therefore, no reaction, or customer satisfaction is a new concept for analysts, consequently the market is slow to respond.

(11)

8

Aksoy, Cooil, Groening, Keiningham and Yalçın (2008) by creating four portfolios with dif-ferent customer satisfaction criteria, they found that a portfolio of firms with high and increas-ing customer satisfaction is much better to invest than a portfolio of firms with low and de-creasing customer satisfaction. Additionally, they suggest that even after we adjust for rele-vant risk factors, the results point out a positive risk-adjusted return for high customer satis-faction. Their results are robust and they have important implications for research analysts and portfolio managers alike.

Last but equal important, with their findings, Frieder and Subrahmanyam (2005) imply that individual’s decisions to hold stock are positive influenced from visible brand name stocks, “glamorous stocks”. With other words the investors prefer stocks of which they are cognizant and have more high quality information because they face lower parameter estimation risk.

(12)

9

3

Problem and Purpose

In this research I will examine the relationship between customers’ satisfaction and stock’s re-turns systematic risk in Sweden, Denmark and Norway market. More specific I will examine if customers’ satisfaction factor has some value which is not captured from the asset pricing models and if this factor should be used as a proxy of the latent risks factors and be priced. To the best of my knowledge, no empirical research is done on this topic specifically for these countries. Based on that, this research tries to close this gap by focusing only on the banking sector of these countries and will serve as a starting point for further and more extended stud-ies on this topic. The research is through the period 2002 – 2011. In this period a downswing can be considered around 2008 and an upswing around 2010 (graph one).

The banks which are taking part in this research are: 1. Danske Bank (Denmark)

2. Nordea Bank (Denmark - Sweden) 3. Swedbank (Sweden)

4. Sparebank1 (Norway) 5. Handelsbanken (Sweden) 6. Sydbank (Denmark) 7. Seb bank (Sweden)

(13)

10

4

Research Questions

The main argument that I am based on, is that satisfied customers are rewarding the company by being loyal and exhibiting recurrent purchasing behavior. Many empirical studies have proved this relationship (Hallowell, 1996). The company, then, is not affected, or affected less, from the market’s ups- and downswings. Consequently, the company’s stock’s returns are more secure, with lower systematic risk. Derived from this argument I am setting the hy-pothesis.

4.1 Customer Satisfaction and Systematic Risk

As I mentioned before stock’s returns systematic risk is the risk associated with the market. When a firm can protect itself from the exposure of the market movements, the stock returns systematic risk is lower.

Increase in customer satisfaction drives in increase in customer loyalty, in which the customer has greater commitment to the firm (Gustofsson, Johnson and Roos, 2005). When the market is falling the companies are becoming more competitive (offers, sales, promotion). Even though, highly satisfied customers are more loyal and less likely to give in to other firms’ of-fers. (Heide and Weiss, 1995). Consequently, remaining loyal to the firm, make the firm to forecast the demand more accurate, to be less vulnerable to market’s risk and more attractive to the investors. This means that customers’ satisfaction could be a factor that is not captured in asset pricing models and can be used as a proxy for the latent risk factors. The questions here are:

 Is there any significant relationship between customer’s satisfaction and bank’s stock’s returns systematic risk?

(14)

11

5

Methodology - Estimation Procedure

5.1 Dependent Variable

Firstly, I am using Fama-French three factors model to estimate banks’ systematic risk (equa-tion three).

5.1.1 Systematic Risk

As I argued before the beta may not be a complete description of asset’s risk - high standard error ( ), not normally distributed.

To estimate the systematic risk of each bank we run seventy regressions (one for every bank, every year).

Where:

 =

= monthly return on stock of bank i on month t, I calculate

= close price of stock of bank I on month t

= monthly risk-free return on month t,

 = monthly return on a value-weighted market portfolio on month t,

 = Fama–French size portfolio on month t,

= Fama–French market-to-book ratio portfolio on month t.

= random error term.

The coefficient represents the systematic risk of each bank and the coefficients and are the proxies for the latent risk factors SMB and HML respectively. For every bank I am

estimating ten , ten and ten (from 2002 - 2011).

By conducting the Jarque - Bera normality test, I find that the residuals are not white noise because they are not normally distributed, as they should be if the model captures all the nec-essary information. This means that there are some factors which explain the latent risk and are not net priced. No indications for autocorrelation or heteroscedasticity found.

(15)

12

Continuing, I replace , and in equation three and I calculate the expected excess re-turns for every year. I compare them with the actual excess returns . Their difference

- is the part of the systematic risk which is not explained from Fama and French three

factor asset pricing model and it represents my dependent variable . I estimate ten for

every bank (from 2002 - 2011).

5.1.2 Robustness Check

There are two main reasons which suggest also another way for calculating the different kind of systematic risks, in order to verify the robustness of the results.

Firstly, I am using monthly data for SMB and HML, Rmt and Rft which I have found at Ken-neth R. French bibliotheca (KenKen-neth R. French, 2012). Because these data are monthly I use monthly data to calculate also Rit. The sample is small because for every year I have only twelve observations, one for every month. The slope coefficients are not always significant and there is lack of more frequently data for Fama and French European factors.

Additionally and equally important, financial analysis suggests that conclusions based on the analysis risk measures is possible to change when the method of calculating returns or risk will change (Fama, 1998).

In order to prove the robustness of the results I am calculating the systematic risk again with the CAPM, presented before, but by using daily stock prices this time.

Where:

=

= daily return on stock of bank i on day t, I calculate

= close price of stock of bank I on day t

= daily risk-free return on day t,

= daily return on a value-weighted market portfolio on day t,

= random error term.

The coefficient represents the systematic risk of each company. For every bank I estimate

(16)

13

By conducting the Jarque-Bera normality test, I find that the residuals are not white noise be-cause they are not normally distributed, as they should be if the model captures all the neces-sary information. This means that there are some factors which explain the latent risk and are not net priced. No indications for autocorrelation or heteroscedasticity found.

And by following the same procedure I replace in equation four and I calculate the ex-pected excess returns for every year. Then, I compare them with the actual excess returns

. Their difference - is the part of the systematic risk which is not explained from

CAPM and it represents my dependent variable I estimate ten for every bank (from 2002 - 2011).

5.2 Customer Satisfaction

I collect these data from EPSI rating based on Swedish, Danish and Norwegian quality index. Swedish quality index for the banking sector exists from 1989 and in 2011 they collected data from more than 300.000 customers through telephone interviews (Epsi Rating, no year). I am using data from 2002. The scores are scaled from 0 to 100. The questions are concern the im-age of the company, the expectations, the quality of service, the value and the loyalty (Epsi Rating, no year).

5.3 Accounting Measurements

Firm’s systematic risk is also influenced by fundamental factors such as size, leverage, ROA and others. The final model will include some of these accounting measurements, in order to be more precise (Brown, Martin and Gruber, 2010).

5.3.1 Size of the firm

I measure the size of the firm by using the natural logarithm of the Total Assets (appendix seven) of each bank (Brown, Martin and Gruber, 2010). Large firms should have lower sys-tematic risk and the reason for that is that they have the ability to lesser the effect of economic changes (Brown, Martin and Gruber, 2010).

(17)

14

Return on assets is a measurement of bank’s performance (European Central Bank, 2010). I collect these data from Amadeus Database or from the annual reports of the banks (appendix seven). ROA is calculated as the ratio of operating income to total assets and from the Capital Asset Pricing Model is known that it has a negative correlation with the systematic risk of a stock. However, in some cases, for example financial institutions, the relationship is positive and this is because financial institutions become more profitable when they take more risk.

5.4 Lag Dependent Variable

Because stock’s returns systematic risk predicts the future systematic risk, I am included in our model a lag variable of the dependent difference variable. By using this lagged variable I prevent the misspecification which may exist if I omit some relevant variables. The inclusion of the lagged dependent variable also controls for inertia, persistence, and different initial conditions (Mizik and Jacobson, 2004)

5.5 Final Model for systematic risk

I am using a panel data analysis with fixed or random cross section effects. The final model for systematic risk is:

Where:

 = –

 = - of bank i for year (t), equation (3), (4)

 = log of customer satisfaction score of bank i for year (t),  = total assets of bank i for year (t),

 = return on assets of bank i for year (t),

 = random error term.

I am using the first difference of the variables in the final model for systematic Risk, because I found the existence of unit root in some of the variables.

By running the OLS regressions for each bank separately, I observe if the intercept or any of the slope coefficients can be assumed as constant (fixed or random) through different banks.

(18)

15 The assumptions made, in both cases, are:

 The intercepts are assumed as constant among the different banks (fixed or random),

 The slope coefficients of customer satisfaction scores are assumed as constant among the different banks (fixed or random),

 The slope coefficients of the lag value of the dependent variable are assumed as con-stant among the different banks (fixed or random),

 The slope coefficients of the ROA of bank are assumed as constant among the differ-ent banks (fixed or random),

 The slope coefficients of the total assets of bank are different for different banks. In order to choose between Random and Fixed effects model I conduct the Hausman’s test (Gujarati and Porter, 2009). With p-value = 1 > 0.05 in both cases, the random effect model is more appropriate (see Appendix 3 and Appendix 6). The test assumes that the intercept and the slope coefficients are not constant for each bank but random and it uses their mean values (Gujarati and Porter, 2009).

(19)

16

6

Data Collection

The data used are secondary and are taken from different sources. I obtain yearly data for the customers’ satisfaction from EPSI rating (Epsi Rating, no year). Epsi Rating is an independ-ent organization that conducts researches of non financial performance indicators. It is devel-oped a lot in Scandinavian countries, especially in Sweden. EPSI Rating is collecting and ana-lyzing customer satisfaction in different industries across Europe and focus on both on the business to consumer (B2C) and business to business (B2B) segments. It has a verified meth-odology that can give organizations insight into their customer base, and benchmark it to the main competition. I use the reports for the banking sector conducting in October 2011, 2010. Because the final model includes the lagged values of the dependent variable I am using com-panies with more than 8 years of data are available (Epsi Rating, no year).

The data for firms’ monthly and daily stocks prices comes from yahoo finance and the Euro-pean Fama–French size and market-to-book ratio monthly factors from the data library by Kenneth French (Yahoo, 2012), (Kenneth R. French, 2012). The accounting measurements such as total assets and return on assets are obtained from Amadeus database provided by university’s library, from Morningstar database, from yahoo finance and from the annual re-ports of the tested banks (Amadeus, 2012), (DanskeBank, 2012), (Handelsbanken, no year), (Morningstar, 2012), (Nordea, 2012), (SEB, no year), ( SpareBank 1, no year), (Swedbank, no year), (Sydbank, no year), ( Yahoo, 2012). As a market proxy I used the OMX Stockholm in-dex.

Important factor which determines the risk of stock’s returns of a company is the company’s R&D investments (Tuli and Bharadwaj, 2009). I could not find R&D data for none of these banks.

(20)

17

7

Results

7.1 Final Model - Fama and French three factor model

Table 1 below summarizes the results of the final model, with the dependent variable

be-ing calculated with the Fama and French three factor model (see appendix 2).

= -

-0.346408***

(cust. satisfaction) -4.021079***

-0.007023***

Different for every bank

Table 1 *p < .10. **p < .05. ***p < .01

Source: own calculations

The results support that the = - (the part of the systematic risk which is not com-

pletely captured by β (beta)) of the bank has a significant negative relationship with the cus-tomers’ satisfaction of the same bank. If cuscus-tomers’ satisfaction increases by 1 percent,

decreases -4.021079 units with p=0.003. With significance level 1%, the result is significant. The lag value of the dependent variable and the return on assets (ROA) have a significant negative relationship with the customers’ satisfaction of the same bank, as I expected.

7.2 CAPM

For the robustness check, table 2 below outlines the results with the dependent variable

being calculated with the CAPM (see appendix 4).

(21)

18

-0.479601***

(cust. satisfaction) -2.585262*

0.002612

Different for every bank

Table 2 *p < .10. **p < .05. ***p < .01

Source: own calculations

The results support that the = - (the part of the systematic risk which is not com-

pletely captured by β (beta)) of the bank has a significant negative relationship with the cus-tomers’ satisfaction of the same bank. If cuscus-tomers’ satisfaction increases by 1 percent, decreases -2.585262 units with p=0.0993. With 10% significance level, the result is signifi-cant.

The lag value of the dependent variable has a significant negative relationship with the cus-tomers’ satisfaction of the same bank, as I expected. Return on Assets has a positive but not significant relationship with the customers’ satisfaction of the same bank.

The test’s results seem quite robust across different specification of . They are also show-ing similar results with the previous researchers as well. Customers ‘satisfaction is an im-portant factor which is not captured from the asset pricing models. It can be used as a proxy of the latent risk and be net priced. Customers’ satisfaction has a significant negative relation-ship with the expected excess rate of return. From the asset pricing theory, lower excess re-turns mean lower systematic risk. Consequently, customers’ satisfaction has a significant neg-ative relationship with the systematic risk. The banks with high customers’ satisfaction scores can be thought as “glamorous” assets. In simple words famous brands are preferred from the investors, even though they may not present high rates of returns. Investors used to overesti-mate these assets by seeing them as safe assets, with less systematic risk. They prefer to invest to them even if they perform less excess returns.

(22)

19

8

Conclusion and Limitations

8.1 Conclusion

In this paper, by using the main asset pricing models, Capital Asset Pricing Model and Fama – French three factor model, I am examining if customers’ satisfaction is a factor which ex-plains the expected stock’s returns and is not captured by them. The result of the research suggests that customers’ satisfaction can be used as a proxy for the latent systematic risk and it should be net priced. More specific customers’ satisfaction has a significant negative rela-tionship with bank’s stock’s returns systematic risk. Therefore, a bank which implements effi-cient marketing strategies and pose high customers’ satisfaction scores are less vulnerable to market’s up- and downswings observed consequently to market’s risk. These stocks are pre-ferred from the investors. They are considering them as “glamorous” assets and they invest to them even though they do not perform high excess returns. Considering specifically the bank-ing sector, the results are quite important. Banks, especially in the period tested, which in-volves the financial crisis, have a higher risk than the other sectors. Consequently, banks with high customers’ satisfaction scores can be considered as more safe in the eyes of investors and create a competitive advantage. This paper is the first step in this area in Scandinavian countries, Sweden, Denmark and Norway. It contributes in marketing – finance interface and it can be used as a starting point for further investigation in other sectors and countries.

8.2 Limitations

In considering the results of this research, the reader must always keep in mind several limita-tions. First, an important factor that determines the risk of stock’s returns of a company is the company’s R&D investments (Tuli and Bharadwaj, 2009). I could not find R&D data for none of these banks. Second, the data used are secondary and are taken from different sources. Third and most important, because of data source unavailability, the sample used is small and includes only banks. With only seven banks in my dataset and ten years time peri-od, the final observations, after the adjustments, are marginally acceptable in terms of robust-ness. Additionally, the fact that only banks are participating in the research and by consider-ing that each business sector has particular characteristics, the results may not be equally gen-eralizable for other sectors. However, by including the leading banks, I expect these results to have real economic significance for the banking sector in Sweden, Norway and Denmark.

(23)

20

9

References

Aksoy, L., Cooil, B., Groening, C., Keiningham, T. L. and Yalçın A. (2008). The Long-Term Stock Market Valuation of Customer Satisfaction. Journal of Marketing, 72(Jul), pp. 105-122.

Amadeus. (2012). Amadeus Companies. Retrieved on Feb. 25, 2012 from http://hj.se/bibl/en.html

Banz, R. W. (1981). The Relationship between Return and Market Value of Common Stocks. Journal of Fmanctal Economics, (9)(1), pp. 3-18

Barber, B. M. and Odean, T. (2000). Trading is Hazardous to Your Wealth: The Common Stock Investment Performance of Individual Investors. Journal of Finance, (55)(2), pp. 773–806

Basu, S. (1977). Investment Performance of Common Stocks in Relation to their Price-Earnings Ratios: A Test of the Efficient Market. The Journal of Finance, (12)(3), pp. 129-56

Berthon, P., Hulbert, J. and Pitt, L.F. (1999) M., Brand Management Prognostications. MIT

Sloan Management Review, (40)(2), pp. 53-65

Brown, S. J., Martin, W. and Gruber, J. (2010). Modern Portfolio Theory and Investmnet

Analysis, 8th Int. student ed. JohnWiley & Sons Ltd.

Chen, N. F. (1983). Some Empirical Tests of the Theory of Arbitrage Pricing. The Journal of

Finance, (38)(5), pp. 1393-1414

Coles, J. L., Naveen D. D. and Naveen L. (2006). Managerial Incentives and Risk-Taking.

Journal of Financial Economics, (79)(2) pp. 431–68.

DanskeBank. (2012). Financial Reports. Retrieved on Feb. 20, 2012, from http://www.danskebank.com/EN-UK/IR/REPORTS/Pages/financial-reports.aspx

Dutta, S., Zbaracki, M. J. and Bergen, M. (2003). Pricing Process as a Capability: A resource-based Perspective. Strategic Management Journal, (24)(7), pp. 615-630

Epsi Rating. (no year). Customer Satisfaction. Retrieved on Feb. 8, 2012, from http://www.epsi-rating.com/expertise/customer-satisfaction.htm

European Central Bank. (2010). Beyond ROE – How to Measure Bank Performance.

Appen-dix to the report on EU banking Structures.

Fama, E. F. (1998). Market efficiency, long-term returns, and behavioral finance. Journal of

Financial Economic, (49)(3), pp. 283-306

Fama, E. F. and French, K. R. (2004). The Capital Asset Pricing Model: Theory and Evi-dence. The Journal of Economic Perspectives, (18)(3), pp. 25-46

Fama, E. F. and French, K. R. (1993). Common risk factors in the returns on stocks and bonds. Journal of Financial Economic, (33)(1), pp. 3-56

Fornell, C., Mithas, S., Morgeson III, F. V., and Krishnan M.S. (2006). Customer Satisfaction and Stock Prices: High Returns, Low Risk. Journal of Marketing, (70)(Jan), pp. 3-14. Frieder, L. and Subrahmanyam, A. (2005). Brand Perceptions and the Market for Common

(24)

21

Stock. Journal of Financial & Quantitative Analysis, (40)(1), pp. 57-85

Grönroos C., (1989),Defining Marketing: A Market-Oriented Approach. European Journal of

Marketing, (23)(1), pp. 52 - 60

Gujarati D. N. and Porter D. C. (2009). Basic Econometrics, 5th ed. McGraw-Hill: New York,

USA.

Gustafsson, A., Johnson, M. D. and Roos, I. (2005). The Effects of Customer Satisfaction, Relationship Commitment Dimensions, and Triggers on Customer Retention. Journal of

Marketing, (69)(4), pp. 210-218

Hallowell, R. (1996). The Relationships of Customer Satisfaction, Customer Loyalty and Profitability: an Empirical Study. International Journal of Service Industry Management.

(7)(4), pp. 27-42

Handelsbanken. (no year). Investor Relations. Retrieved on Feb. 15, 2012, from http://www.handelsbanken.se/ireng

Heide, J. B. and Weiss, A. M. (1995). Vendor Consideration and Switching Behavior for Buyers in High-Technology Markets. Journal of Marketing, (59)(3), pp. 30-43

Kenneth R. French. (2012). International Research Returns Data. Retrieved on Feb. 10, 2012, from http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html Lin, J., Wang, M. and Cai, L. (2012). Are the Fama–French factors good proxies for latent

risk factors? Evidence from the data of SHSE in China. Economics Letters, (116)(2), pp. 265-268

Levitt, T. (1980). Marketing success through differentiation - of anything

.

Harvard Business Review, (58)(1), pp. 83-91

Long, J. B., Plosser, Jr. and Plosser, Ch. I. (1983). Real Business Cycles. Journal of Political

Economy, (91)(1), pp. 36-69

McAlister, L., Srinivasan, R. & Kim, M. (2007). Advertising, Research and Development, and Systematic Risk of the Firm. Journal of Marketing, (71), pp. 35-48

Mizik, N. and Jacobson, R. (2004). Are Physicians ‘Easy Marks? Quantifying the Effects of Detailing and Sampling on New Prescriptions. Management Science, (50)(12), pp. 1704– 1715

Morningstar. (2012). Morningstar Quote. Retrieved on Mar. 10, 2012, from http://www.morningstar.com/

Newman, J. W. and Werbel, R. A. (1973). Multivariate Analysis of Brand Loyalty for Major Household Appliances. Journal of Marketing Research, (10)(May), pp. 363-389

Nordea. (2012). Annual Reports. Retrieved on Feb. 20, 2012, from http://www.nordea.com/Investor+Relations/Reports/Annual+reports/804982.html

Rego, L. L., Billet, M. T. and Morgan, N. A. (2009). Consumer-Based Brand Equity and Firm Risk. Journal of Marketing, (73)(6), pp. 47-60.

Reichheld, F. F. and Kenny, D. W. (1990-1991). The Hidden Advantages of Customer Reten-tion. Journal of Retail Banking Services, 12(4), pp. 19

Rotemberg, J. J. and Saloner, G. (1986). A Supergame-Theoretic Model of Price Wars during Booms. The American Economic Review, (76)(3) pp. 390-407

(25)

22

Rotemberg, J. J. and Woodford, M. (1991). Markups and the Business Cycle. NBER

Macroe-conomics Annual, (6), pp. 63-129

SEB. (no year). Financial Reports. Retrieved on Feb. 21, 2012, from http://sebgroup.com/pow/wcp/sebgroup.asp?website=TAB3&lang=en

SpareBank 1. (no year). Investorblog. Retrieved on Feb. 21, 2012, from http://investor.sparebank1.no/in-english/

Srivastava, R. K., Shervani, T. A. and Fahey L. (1998). Market-Based Assets and Shareholder Value: A Framework for Analysis. Journal of Marketing, (62), pp. 2 - 18

Singh, N. and Vives, X. (1984). Price and Quantity Competition in a Differentiated Duopoly.

The RAND Journal of Economics, (15)(4), pp. 546-554

Swedbank. (no year). Financial Information and Publication. Retrieved on Feb. 20, 2012, from http://www.swedbank.com/investor-relations/financial-information-and-publications/risk-report/index.htm

Sydbank. (no year). Financial Reports. Retrieved on Feb. 23, 2012, from http://sydbank.com/sydbankcom/about/ir/finreports

Tuli, K. R. and Bharadwaj, S. G. (2009). Customer Satisfaction and Stock Returns Risk.

Journal of Marketing, (73)(6), pp. 184-197.

Yahoo! (2012). Finance. Retrieved on Jan. 20, 2012, from http://finance.yahoo.com/

Yu, Y. T. and Dean, A. (2001). The Contribution of Emotional Satisfaction to Consumer Loyalty. International Journal of Service Industry Management, (12)(3), pp. 234 – 250

(26)

23

Appendix

Appendix 1

Final Model – Fama and French 3 factor model Panel Data – fixed effects

Dependent Variable: DY_? Method: Pooled Least Squares Date: 05/07/12 Time: 21:02 Sample (adjusted): 2004 2011

Included observations: 56 after adjustments Cross-sections included: 7

Total pool (balanced) observations: 392

Variable Coefficient Std. Error t-Statistic Prob.

DROA_? -0.007822 0.005435 -1.439116 0.1510 C 0.188239 0.063845 2.948358 0.0034 LAGDY_? -0.345574 0.094245 -3.666780 0.0003 DLOGX_? -4.229528 1.613628 -2.621130 0.0091 DANSKE--DLOGTOTASS_DANSKE -2.236977 1.076859 -2.077315 0.0385 NORDEA--DLOGTOTASS_NORDEA -6.114819 1.922516 -3.180634 0.0016 SEB--DLOGTOTASS_SEB 0.426883 1.011529 0.422017 0.6733 HANDEL--DLOGTOTASS_HANDEL 6.698514 1.633132 4.101637 0.0001 SPARE--DLOGTOTASS_SPARE -0.235636 0.545318 -0.432107 0.6659 SWEDBANK--DLOGTOTASS_SWEDBANK -1.157432 1.368906 -0.845516 0.3984 SYDBANK--DLOGTOTASS_SYDBANK -1.615936 1.332568 -1.212648 0.2260

Fixed Effects (Cross)

DANSKE--C -0.066445 NORDEA--C 0.651383 SEB--C -0.268292 HANDEL--C 0.014371 SPARE--C -0.189316 SWEDBANK--C -0.004164 SYDBANK--C -0.137537 Effects Specification Cross-section fixed (dummy variables)

R-squared 0.213128 Mean dependent var 0.072754

Adjusted R-squared 0.179555 S.D. dependent var 0.840886

S.E. of regression 0.761662 Akaike info criterion 2.335771

Sum squared resid 217.5483 Schwarz criterion 2.507994

Log likelihood -440.8111 Hannan-Quinn criter. 2.404027

F-statistic 6.348147 Durbin-Watson stat 1.452946

(27)

24

Appendix 2

Panel Data – random effects Dependent Variable: DY_?

Method: Pooled EGLS (Cross-section random effects) Date: 05/07/12 Time: 20:59

Sample (adjusted): 2004 2011

Included observations: 56 after adjustments Cross-sections included: 7

Total pool (balanced) observations: 392

Wansbeek and Kapteyn estimator of component variances

Cross-section weights (PCSE) standard errors & covariance (d.f. corrected)

Variable Coefficient Std. Error t-Statistic Prob.

C 0.150861 0.112126 1.345459 0.1793 LAGDY_? -0.346408 0.093916 -3.688494 0.0003 DROA_? -0.007023 0.001848 -3.799919 0.0002 DLOGX_? -4.021079 1.053561 -3.816656 0.0002 DANSKE--DLOGTOTASS_DANSKE -2.276512 0.530770 -4.289071 0.0000 NORDEA--DLOGTOTASS_NORDEA -3.735151 0.580120 -6.438588 0.0000 SEB--DLOGTOTASS_SEB 0.211335 0.728048 0.290275 0.7718 HANDEL--DLOGTOTASS_HANDEL 6.820924 3.378075 2.019175 0.0442 SPARE--DLOGTOTASS_SPARE -0.326492 0.143059 -2.282216 0.0230 SWEDBANK--DLOGTOTASS_SWEDBANK -1.118124 0.558010 -2.003772 0.0458 SYDBANK--DLOGTOTASS_SYDBANK -1.782781 0.997785 -1.786738 0.0748

Random Effects (Cross)

DANSKE--C -0.022218 NORDEA--C 0.337586 SEB--C -0.186141 HANDEL--C 0.034023 SPARE--C -0.116043 SWEDBANK--C 0.025063 SYDBANK--C -0.072270 Effects Specification S.D. Rho Cross-section random 0.253909 0.1000 Idiosyncratic random 0.761662 0.9000 Weighted Statistics

R-squared 0.131089 Mean dependent var 0.027070

Adjusted R-squared 0.108283 S.D. dependent var 0.805215

S.E. of regression 0.760371 Sum squared resid 220.2803

F-statistic 5.747979 Durbin-Watson stat 1.444916

Prob(F-statistic) 0.000000

Unweighted Statistics

R-squared 0.163378 Mean dependent var 0.072754

(28)

25

Appendix 3

Hausman’s Test

Correlated Random Effects - Hausman Test Pool: Untitled

Test cross-section random effects

Test Summary

Chi-Sq.

Statistic Chi-Sq. d.f. Prob.

Cross-section random 0.000000 10 1.0000

* Cross-section test variance is invalid. Hausman statistic set to zero. ** WARNING: robust standard errors may not be consistent with assumptions of Hausman test variance calculation. Cross-section random effects test comparisons:

Variable Fixed Random Var(Diff.) Prob.

LAGDY_? -0.345574 -0.346408 0.008858 0.9929 DROA_? -0.007822 -0.007023 -0.000002 NA DLOGX_? -4.229528 -4.021079 0.562956 0.7812 DANSKE--DLOGTOTASS_DANSKE -2.236977 -2.276512 -0.206147 NA NORDEA--DLOGTOTASS_NORDEA -6.114819 -3.735151 -0.275246 NA SEB--DLOGTOTASS_SEB 0.426883 0.211335 -0.231188 NA HANDEL--DLOGTOTASS_HANDEL 6.698514 6.820924 31.467045 0.9826 SPARE--DLOGTOTASS_SPARE -0.235636 -0.326492 -0.014462 NA SWEDBANK--DLOGTOTASS_SWEDBANK -1.157432 -1.118124 -0.249569 NA SYDBANK--DLOGTOTASS_SYDBANK -1.615936 -1.782781 -0.498930 NA

Cross-section random effects test equation: Dependent Variable: DY_?

Method: Panel Least Squares Date: 05/07/12 Time: 21:01 Sample (adjusted): 2004 2011

Included observations: 56 after adjustments Cross-sections included: 7

Total pool (balanced) observations: 392

Cross-section weights (PCSE) standard errors & covariance (d.f. corrected)

Variable Coefficient Std. Error t-Statistic Prob.

C 0.188239 0.098635 1.908440 0.0571 LAGDY_? -0.345574 0.132958 -2.599122 0.0097 DROA_? -0.007822 0.001294 -6.044297 0.0000 DLOGX_? -4.229528 1.293424 -3.270024 0.0012 DANSKE--DLOGTOTASS_DANSKE -2.236977 0.274901 -8.137389 0.0000 NORDEA--DLOGTOTASS_NORDEA -6.114819 0.247575 -24.69887 0.0000 SEB--DLOGTOTASS_SEB 0.426883 0.546687 0.780854 0.4354 HANDEL--DLOGTOTASS_HANDEL 6.698514 6.548163 1.022961 0.3070 SPARE--DLOGTOTASS_SPARE -0.235636 0.077482 -3.041168 0.0025 SWEDBANK-- -1.157432 0.248608 -4.655644 0.0000

(29)

26

DLOGTOTASS_SWEDBANK

SYDBANK--DLOGTOTASS_SYDBANK -1.615936 0.704730 -2.292985 0.0224

Effects Specification Cross-section fixed (dummy variables)

R-squared 0.213128 Mean dependent var 0.072754

Adjusted R-squared 0.179555 S.D. dependent var 0.840886

S.E. of regression 0.761662 Akaike info criterion 2.335771

Sum squared resid 217.5483 Schwarz criterion 2.507994

Log likelihood -440.8111 Hannan-Quinn criter. 2.404027

F-statistic 6.348147 Durbin-Watson stat 1.452946

(30)

27

Appendix 4

Final Model – CAPM Panel Data – fixed effects

Dependent Variable: DY_? Method: Pooled Least Squares Date: 05/02/12 Time: 17:59 Sample (adjusted): 2004 2011

Included observations: 56 after adjustments Cross-sections included: 7

Total pool (balanced) observations: 392

Variable Coefficient Std. Error t-Statistic Prob.

DLOGX_? -2.573671 1.300105 -1.979587 0.0485 LAGDY_? -0.475583 0.042903 -11.08501 0.0000 DROA_? 0.002392 0.004380 0.546221 0.5852 C 0.028155 0.052095 0.540459 0.5892 DANSKE--DLOGTOTASS_DANSKE -3.263824 0.875932 -3.726115 0.0002 NORDEA--DLOGTOTASS_NORDEA -3.279839 1.570856 -2.087931 0.0375 SEB--DLOGTOTASS_SEB 0.825703 0.782467 1.055256 0.2920 HANDEL--DLOGTOTASS_HANDEL -3.505073 1.335616 -2.624312 0.0090 SPARE--DLOGTOTASS_SPARE 2.795505 0.440678 6.343650 0.0000 SWEDBANK--DLOGTOTASS_SWEDBANK -2.153573 1.103087 -1.952314 0.0516 SYDBANK--DLOGTOTASS_SYDBANK 0.144693 1.063672 0.136031 0.8919

Fixed Effects (Cross)

DANSKE--C 0.159317 NORDEA--C 0.406450 SEB--C -0.105213 HANDEL--C 0.266986 SPARE--C -0.807064 SWEDBANK--C 0.128189 SYDBANK--C -0.048664 Effects Specification Cross-section fixed (dummy variables)

R-squared 0.400335 Mean dependent var -0.061817

Adjusted R-squared 0.374749 S.D. dependent var 0.778872

S.E. of regression 0.615876 Akaike info criterion 1.910857

Sum squared resid 142.2387 Schwarz criterion 2.083080

Log likelihood -357.5279 Hannan-Quinn criter. 1.979113

F-statistic 15.64680 Durbin-Watson stat 2.364803

(31)

28

Appendix 5

Panel Data – random effects Dependent Variable: DY_?

Method: Pooled EGLS (Cross-section random effects) Date: 05/02/12 Time: 18:02

Sample (adjusted): 2004 2011

Included observations: 56 after adjustments Cross-sections included: 7

Total pool (balanced) observations: 392

Wansbeek and Kapteyn estimator of component variances

Cross-section weights (PCSE) standard errors & covariance (d.f. corrected)

Variable Coefficient Std. Error t-Statistic Prob.

DLOGX_? -2.585262 1.564587 -1.652361 0.0993 LAGDY_? -0.479601 0.057051 -8.406557 0.0000 DROA_? 0.002612 0.003393 0.769609 0.4420 C 0.016859 0.151501 0.111276 0.9115 DANSKE--DLOGTOTASS_DANSKE -3.183842 1.428607 -2.228635 0.0264 NORDEA--DLOGTOTASS_NORDEA -2.586217 1.089187 -2.374447 0.0181 SEB--DLOGTOTASS_SEB 0.802008 0.521720 1.537238 0.1251 HANDEL--DLOGTOTASS_HANDEL -3.231048 0.811614 -3.981017 0.0001 SPARE--DLOGTOTASS_SPARE 2.602110 0.623630 4.172524 0.0000 SWEDBANK--DLOGTOTASS_SWEDBANK -2.075622 0.612111 -3.390923 0.0008 SYDBANK--DLOGTOTASS_SYDBANK 0.122787 0.731927 0.167758 0.8669

Random Effects (Cross)

DANSKE--C 0.156773 NORDEA--C 0.316613 SEB--C -0.087515 HANDEL--C 0.243765 SPARE--C -0.723286 SWEDBANK--C 0.127678 SYDBANK--C -0.034028 Effects Specification S.D. Rho Cross-section random 0.372489 0.2678 Idiosyncratic random 0.615876 0.7322 Weighted Statistics

R-squared 0.386130 Mean dependent var -0.013337

Adjusted R-squared 0.370018 S.D. dependent var 0.775922

S.E. of regression 0.615860 Sum squared resid 144.5069

F-statistic 23.96527 Durbin-Watson stat 2.347400

Prob(F-statistic) 0.000000

Unweighted Statistics

R-squared 0.209392 Mean dependent var -0.061817

(32)

29

Appendix 6

Hausman’s test

Correlated Random Effects - Hausman Test Pool: Untitled

Test cross-section random effects

Test Summary

Chi-Sq.

Statistic Chi-Sq. d.f. Prob.

Cross-section random 0.000000 10 1.0000

* Cross-section test variance is invalid. Hausman statistic set to zero. ** WARNING: robust standard errors may not be consistent with assumptions of Hausman test variance calculation.

Cross-section random effects test comparisons:

Variable Fixed Random Var(Diff.) Prob.

LAGDY_? -0.475583 -0.479601 -0.000458 NA DLOGX_? -2.573671 -2.585262 -0.660385 NA DROA_? 0.002392 0.002612 -0.000008 NA NORDEA--DLOGTOTASS_NORDEA -3.279839 -2.586217 -0.805000 NA HANDEL--DLOGTOTASS_HANDEL -3.505073 -3.231048 -0.513743 NA SEB--DLOGTOTASS_SEB 0.825703 0.802008 -0.219927 NA DANSKE--DLOGTOTASS_DANSKE -3.263824 -3.183842 0.135969 0.8283 SWEDBANK--DLOGTOTASS_SWEDBANK -2.153573 -2.075622 -0.314554 NA SYDBANK--DLOGTOTASS_SYDBANK 0.144693 0.122787 -0.426962 NA SPARE--DLOGTOTASS_SPARE 2.795505 2.602110 -0.091173 NA

Cross-section random effects test equation: Dependent Variable: DY_?

Method: Panel Least Squares Date: 05/07/12 Time: 20:57 Sample (adjusted): 2004 2011

Included observations: 56 after adjustments Cross-sections included: 7

Total pool (balanced) observations: 392

Cross-section weights (PCSE) standard errors & covariance (d.f. corrected)

Variable Coefficient Std. Error t-Statistic Prob.

C 0.028155 0.038235 0.736370 0.4620 LAGDY_? -0.475583 0.052886 -8.992644 0.0000 DLOGX_? -2.573671 1.336992 -1.924971 0.0550 DROA_? 0.002392 0.001876 1.275380 0.2030 NORDEA--DLOGTOTASS_NORDEA -3.279839 0.617518 -5.311329 0.0000 HANDEL--DLOGTOTASS_HANDEL -3.505073 0.380754 -9.205611 0.0000

(33)

30 SEB--DLOGTOTASS_SEB 0.825703 0.228615 3.611756 0.0003 DANSKE--DLOGTOTASS_DANSKE -3.263824 1.475427 -2.212121 0.0276 SWEDBANK--DLOGTOTASS_SWEDBANK -2.153573 0.245206 -8.782696 0.0000 SYDBANK--DLOGTOTASS_SYDBANK 0.144693 0.329779 0.438756 0.6611 SPARE--DLOGTOTASS_SPARE 2.795505 0.545656 5.123197 0.0000 Effects Specification Cross-section fixed (dummy variables)

R-squared 0.400335 Mean dependent var -0.061817

Adjusted R-squared 0.374749 S.D. dependent var 0.778872

S.E. of regression 0.615876 Akaike info criterion 1.910857

Sum squared resid 142.2387 Schwarz criterion 2.083080

Log likelihood -357.5279 Hannan-Quinn criter. 1.979113

F-statistic 15.64680 Durbin-Watson stat 2.364803

(34)

31

Appendix 7

Total Assets 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Nordea Bank 249.619 262.190 280.074 325.549 346.890 389.054 474.074 507.544 580.839 716.204 Danske Bank 1.751.553 1.826.134 2.078.497 2.431.988 2.739.361 3.349.530 3.543.974 3.098.477 3.213.886 3.424.403 Seb 1.241.112 1.279.393 1.606.551 1.889.738 1.934.441 2.344.462 2.510.702 2.308.227 2.179.821 2.362.653 Handelsbanken 1.277.514 1.260.454 1.349.090 1.582.907 1.790.008 1.859.382 2.158.784 2.122.843 2.153.530 2.454.366 Sparebank1 23.923 21.561 37.552 43.631 49.166 52.833 53.250 84.509 97.997 101.455 Swedbank 957.503 1.022.281 1.032.038 1.197.283 1.352.989 1.607.984 1.811.690 1.794.687 1.715.681 1.857.065 Sydbank 66.857 73.565 85.258 98.913 114.758 132.323 155.975 157.821 150.843 153.441

Return on Assets (ROA)

2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Nordea Bank 3,61 41,33 66,24 15,2 6,31 4,53 5,92 0,98 2,75 3,5 Danske Bank 1.751.553 1.826.134 2.078.497 2.431.988 2.739.361 3.349.530 3.543.974 3.098.477 3.213.886 3.424.403 Seb 0,51 0,52 0,51 0,48 0,64 0,63 0,42 0,05 0,3 0,5 Handelsbanken 0,768 0,4299 0,5071 0,6929 0,5242 0,3511 0,1988 0,2936 0,6276 0,699 Sparebank1 23.923 21.561 37.552 43.631 49.166 52.833 53.250 84.509 97.997 101.455 Swedbank 0,69 0,94 1 1,33 1,1 1,02 0,64 -0,58 0,4 0,65 Sydbank 2 2 1,7 1,7 1,4 0,4 0,5 0,3 0,12

References

Related documents

As for disparities in conclusions around the market size link to market efficiency, we also examine if there is any variation between large-cap, mid-cap and small-cap Swedish

In pI and PII the portfolios are evenly weighted, in vwpI and vwpII the portfolio are value weighted The portfolios have been regressed against different risk premiums, rp10 being

and relatively high returns on earning assets of the Norwegian banks may be explained by their seemingly higher credit risk-taking in terms of a larger lending share of earning

For each announcement, the 180 trading days prior to the event window will be the estimation window, which will be used in a regression to estimate a normal return (MacKinlay,

The first search generated 176 articles, which were further limited to those directly focus on investigating online customer satisfaction in Internet banking, and other themes

Re-examination of the actual 2 ♀♀ (ZML) revealed that they are Andrena labialis (det.. Andrena jacobi Perkins: Paxton &amp; al. -Species synonymy- Schwarz &amp; al. scotica while

Even if UN’s 17 SDGs do not affect credit rating directly, banks assume that manufacturing industries, such as steel production, work and include sustainable aspects in

A study was conducted to investigate how the local branch of HB in Luleå, market itself to meet banking needs of the young people between 13-18 years and how did they