The effect of Customer Satisfaction ratings on
shareholder value in European countries and
Singapore
Robert J. Bok1 1
University of Groningen, August 2012
Abstract
This study tests the effect of customer satisfaction ratings on shareholder value from the period of 2001 until 2012 for European countries and Singapore. The results show that high customer satisfaction ratings are not significantly related to higher shareholder value. However, when creating a back tested paper portfolio based on customer satisfaction ratings the portfolio of companies with the highest 20% ratings outperform the benchmark portfolio but consist of higher risk stocks. This study also investigates if an increase in customer satisfaction rating has more effect and takes more time to react than a decrease in shareholder value. The results for this part are not significant, however they indicate that a decreasing rating has an effect on the shareholder value, which is not the case for an increasing rating.
Key words: Customer satisfaction, shareholder value, risk, three-factor model, and portfolio study. JEL-classification: G-02, G-11
Customer satisfaction is a well-known notion for business people, however most of them remain skeptic towards customer satisfaction as a business performance metric. Most of the time customer satisfaction is related to marketing instead of financial
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
1 Robert Bok (s1605011) is a student at the University of Groningen. The author thanks Viola Angelina for all
! "!
performance, which makes it less important for senior management and shareholders. To convince senior management and shareholders of the importance of customer satisfaction, a theoretical and especially empirical support for the relationship between customer satisfaction and financial performance is essential. If this relation is not found, senior management and shareholders still remain ambivalent toward customer satisfaction and the direct relationship with business performance (Ambler (2000); Barwise, Marsh, and Wensley (1989); Day and Fahey (1988); Srivastava, Shervani, and Fahey (1998)). For this reason, this study investigates the relationship between customer satisfaction ratings of businesses and shareholder value in other countries than America, since many studies have already been performed for the American Customer Satisfaction Index (ACSI) (Fornell et al. (2006; 2009); Jacobson and Mizik (2009); Tuli and Bharadwaj (2009); O’Sulivan, Hutchinson, and O’Connell (2009)). The relationship will be investigated in the following countries: Denmark, Estonia, Finland, Latvia, Lithuania, Netherlands, Norway, Sweden, Singapore, Turkey, and the UK2. The countries are selected since the methodologies of the customer satisfaction ratings are the same or linked to the ACSI. The period for which the customer satisfaction ratings were available was different for each country, but the longest period ranges from 2001 until 2011. Singapore is added because it has the same methodology as the ASCI and increases the amount of data and therefore the validity of the results. Countries that are on the website of a customer satisfaction index but do not have specific customer satisfaction ratings for companies in these specific countries are excluded, for example Kazakhstan and Azerbaijan of the EPSI-rating website.
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
"!The following websites are used to gather the customer satisfaction ratings for each country:
! #!
This study builds on previous research by Fornell et al. (2006; 2009), who investigated this relationship from 1994 until 2008. This study especially contributes to the current field of research, since it extends the research to see whether this relationship is present in other parts of the world as well. To my knowledge this is the first study that investigates this relationship in most of these countries. Furthermore, this study investigates the risk profile of the businesses to see whether the possible relationship can be explained by higher risk or due to customer satisfaction, based on the previous study of Tuli and Bharadwaj (2009). In their research they use the Fama and French three-factor model (Fama and French (1993)), which is used in this study as well. Additionally, this research investigates if an increase in customer satisfaction ratings reward more than decreasing customer satisfaction ratings punish. Lastly, this study examines if there is a difference, between rewarding and punishing, in the time it takes to react to the new positive or negative customer satisfaction ratings.
! $!
20% of customer satisfaction ratings. However, this result can be explained by the fact this portfolio consists of more risky stocks. I did not perform the second portfolio study due to problems with the dataset, which I elaborate on in that specific section of the study.
This paper is organized as follows. Firstly, in section I the previous literature will be discussed. Section II explains the hypotheses that will be tested in this study. Section III presents the data. In section IV, the methodology will be discussed. Section V summarizes the results of the study. Section VI elaborates on the two portfolios. Lastly, section VII presents the conclusions based on the results and presents recommendations for future research.
I. Literature review
! %!
and Sasser (1990); Srivastava, Shervani, and Fahey (1998)). Furthermore, higher customer retention will probably result in more stable future revenue, lower volatility of anticipated future revenue, since customers will come back and will be less influenced by external factors such as competition (Anderson and Sullivan (1993); Narayandas (1998)).
! &!
Figure 1 - The model of the relationship between customer satisfaction and shareholder value (Matzler et al. (2005)).
A. Repurchase and shareholder value
It is generally agreed that customer satisfaction results in repurchases (e.g. Anderson and Sullivan (1993); Bolton (1998); Bolton, Kannan, and Bramlett (2000); Fornell (1992); O’Sulivan, Hutchinson, and O’Connell (2009); Verhoef, Franses, and Hoekstra (2001)). Srivastava, Shervani, and Fahey (1998) mention that the stable relationship with customers will increase the shareholder value in various ways. As a consequence of the repurchasing of the products/services a stable relationship between the customer and the company is established. Through this relationship the company can generate important information about the customers. This results in the ability to lower relationship costs, lowering the costs for acquiring new customers, and reducing the volatility of the incoming cash flows. As a consequence of the decreasing costs the shareholder value will increase. Furthermore, the decrease in the volatility of the incoming cash flows result in a lower cost of capital. Lastly, customer
results) which are related to outcomes (second level results) that directly influence the drivers of shareholder value.
Repurchase and Shareholder Value
Customer satisfaction leads to repurchases (e.g. Anderson & Sullivan, 1993; Fornell, 1992; Rust et al., 1995). The continuous repurchase of a company’s product results in a stable relationship between customer and supplier, which allow a firm to generate meaningful knowl-edge about the customers. Through experience curve effects and economies of scale, a company is able significantly to lower its relationship costs. Furthermore, costs for acquiring new customers decrease. As a result, shareholder value will be enhanced. In addition, the stable customer base can enhance a firm’s shareholder value in multiple ways (Srivastava et al., 1998). The faster acceptance of new products by loyal customers accelerates market pen-etration and therefore also cash flows. A large stable customer base reduces the volatility of the cash flows. The lower volatility of the cash flows also leads to a lower cost of capital and therefore to an enhancement of cash flows. Finally, customers’ loyalty enhances the residual value of the firm through size and quality of the customer base.
Cross Selling and Shareholder Value
Customer satisfaction also leads to cross-selling (e.g. Hallowell, 1996; Homburg & Scha¨fer, 2002; Reichheld & Sasser, 1990). Enhanced cross-selling has two effects. First, the total sales of the company grow and markets can be penetrated faster because customers who have become loyal are responding better to a firm’s marketing efforts (Srivastava et al., 1998). The additional sales increase cash flows and reduce their vola-tility by means of diversification beyond the core business (cross selling). A faster market penetration accelerates cash flows and therefore also enhances shareholder value. Figure 1. A conceptual model of the relationship between customer satisfaction and shareholder
value (adapted from Matzler & Stahl, 2000)
! '!
satisfaction increases the quality of the customer base and increases the residual value of the business.
B. Cross-selling and Shareholder value
Previous research states that customer satisfaction is related to cross selling (e.g. Hallowell (1996); Homburg and Schäfer (2002); Reichheld and Sasser (1990)). Customer satisfaction results in, faster penetration of new markets since satisfied customer’s respond better to marketing efforts, e.g. extension of products, of the particular company. Therefore, an increase in cross selling results in an increase in the sales of the company and accelerates the timing of new cash flows (Srivastava, Shervani, and Fahey (1998)); cross selling can increase shareholder value.
C. Low price sensitivity and shareholder value
Customer satisfaction decreases price sensitivity which results in the ability for a company to charge a higher price for their product/service and that customers are less sensitive for a decrease in the price of the product/service of the competitors (e.g. Anderson (1996); Krishnamurthi and Raij (1991); Narayandas (1998); Reichheld and Sasser (1990); Stock (2003)). Therefore, a decrease in the price sensitivity of customers can increase the cash flow and enhance shareholder value.
D. Word-of-mouth and shareholder value
! (!
company’s reputation improves by recommendations and positive word-of-mouth, which result in an enhancement of the residual value of the company (Anderson, Fornell, and Mazvancheryl (2004); Matzler et al. (2005)).
E. Customer satisfaction and competitive forces
Anderson, Fornell, and Mazvancheryl (2004) argue that customer satisfaction influences the competitive forces of the company and therefore enhances the shareholder value. A large and satisfied customer base increases the bargaining power of the company towards suppliers, partners, other members of the firms’ value chain, and other channels who wants to serve the same customers. For instance, suppliers want to establish and sustain a good relationship with the particular company and therefore the company should be able to obtain greater value from for example the suppliers. This can result in lower costs, higher prices, higher volumes, and faster market penetration due to the bargaining position the particular company has since it possesses a valuable “asset”. Therefore, customer satisfaction can enhance shareholder value due to the improvement of the bargaining position of the company.
! )!
II. Hypotheses
This section elaborates on the hypotheses that are tested in this research. As mentioned previously there are a lot of studies that find that there is a positive relationship between customer satisfaction ratings and shareholder value (Anderson, Fornell, and Mazvancheryl (2004); Fornell et al. (2006; 2009); Matzler et al. (2005)). However, Jacobson and Mizik (2009) find that the positive relationship is limited to firms in the computer and Internet sector. All the previous studies use the ACSI ratings for their research; therefore in this research customer satisfaction indices from other countries are used to examine if there also exists a positive relationship between customer satisfaction ratings and shareholder value for companies in these countries as well. Hence, the first hypothesis that will be tested in this research is:
H1: There is a positive relation between customer satisfaction ratings and shareholder return.
Fornell et al. (2006) find that the higher returns on stocks of firms with high customer satisfaction ratings are not caused by an increase in risk. On the contrary they find that satisfied customers are assets with high returns and low risk. Additionally, O’sullivan et al. (2009) find above market returns with low systematic risk and lower volatility than the market. Furthermore, Tuli and Bharadwaj (2009) find that a positive change in customer satisfaction ratings results in a declining of the overall- and downside-systematic and idiosyncratic risk. Therefore, this study tests if this is also true for companies in other countries than America, which results in the following hypothesis:
! *+!
As mentioned previously, studies find a cause-and-effect relationship between customer satisfactions ratings and shareholder value. However, to my knowledge no study made a distinction between the strength of the relationship between decreasing and increasing customer satisfaction ratings. To see if there exists a difference in the strength of the relationship this is tested in this research. Additionally, due to the fact that there is a cause-and-effect relationship points to the assumption that a time lag in the relationship could exist. Matzler et al. (2005) tested if there was a time lag between changes in customer satisfaction ratings and the effect of it on shareholder value. Their results show that overall the impact of the changes in customer satisfaction ratings is the largest after three quarters. However, they do not make the distinction between increasing or decreasing ratings. Therefore, this research examines if there is a difference in the time lag between a decrease or increase in the customer satisfaction rating and the effect on the shareholder value. To test the two relationships described, the following hypotheses are used:
H3a: Decreasing customer satisfaction ratings punish more than increasing customer satisfaction ratings reward shareholder value.
H3b: It takes more time to react to an increasing customer satisfaction rating than to a decreasing customer satisfaction rating for shareholder value.
III. Data
! **!
! *"!
Figure 2 - The model of the American Customer Satisfaction Index3.
The ACSI model measures the overall customer satisfaction for a company by using customer interviews as input to a multi-equation econometric model (Fornell et al. (1996)). The customer satisfaction rating for a company represents it’s served markets overall evaluation of total purchase and consumption experience (Anderson, fornell, and Lehman (1994); Fornell (1992); Johnson and Fornell (1991)). The factors “Perceived quality”, “Customer expectations”, and “ Perceived value” in figure 2 are the drivers of customer satisfaction in the center, and the outcomes of customer satisfaction are “Customer complaints” and “Customer loyalty”. The indexes in the boxes are multivariable components measured by several questions that are weighted within the model. The customer satisfaction index is scaled on a 0 to 100 basis. The ACSI model is self-weighting to maximize the explanation of customer satisfaction on customer loyalty. This means that when looking at the components and relations, the users of the model can conclude for themselves, which drivers of satisfaction would have the most effect on customer loyalty if those drivers were improved. The definitions of the indexes in the boxes speak for themselves, therefore will not be explained here but the descriptions are in Appendix A, which are copied from the ACSI website4.
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
#!http://www.theacsi.org/index.php?option=com_content&view=article&id=48&Itemid=122
! *#!
I obtained the customer satisfaction ratings for companies from 2001 until 2011, however the period for each country is different, since I use the EPSI ratings for different European companies, the DCPI, the NCSI-UK, the CSISG, and the TMME. In this research companies are included that satisfy the following criteria in the dataset:
i) Only individual company ratings (industry averages and a sum of small companies in the industry are excluded);
ii) Only companies with a continuous rating set of three years or more;
iii) Customer satisfaction ratings for the same company in different countries are averaged (based on the assumption that the turnover is equally divided);
iiii) Only companies where the additional data can be collected from DataStream.
! *$!
Furthermore, I collect the market value of equity, book value of total assets, book value of liabilities, market capitalization, book-to-market ratio, stock prices, and the return indexes (RI) for the companies for the same period from DataStream. The RI of the companies is used since it provides the value of an investment in a stock, while re-investing any cash dividends in the same stocks. Hence, if you calculate the return it will include dividends. The market value of equity is the share price multiplied by all the company's outstanding shares. Moreover, I collect the RI of the MSCI world index and the yield of the 10-year German government bond. The MSCI World index is used, since Singapore is not located on the same continent as the other countries. For the total of 207 companies, I am able to collect the data from DataStream. However for some companies I only can gather the data of the mother company instead of the data for the daughter company itself. For example, there are customer satisfaction ratings for daughter companies of “Telia Sonera” but for those specific companies the data from DataStream is unavailable since the data is only available for the mother company, in this case “Telia Sonera”. Therefore, the data from DataStream of the mother company is used for the daughter companies. This results in 139 (mother) companies for which I am able to gather the data from DataStream and there are 75 daughter companies. The separate customer satisfaction ratings for the mother company are used instead of using the average customer satisfaction ratings of all the daughter companies.
! *%!
IV. Methodology
The method used to examine if there is a positive relationship between customer satisfaction and shareholder return is based on the method used by Fornell et al. (2006). To see whether there is a significant relationship between customer satisfaction ratings and shareholder return, I follow the standard practice to define an equation, which estimates the effect of an identifiable asset on market capitalization in a combined longitudinal cross-sectional research, though controlling for accounting book values (Barth and McNichols (1994); Ittner and Larcker (1998); Landsman (1986)). In this research, equation (1) estimates whether customer satisfaction rating has a significant effect on the market value of equity and controls for the influence of the recorded assets and liabilities:
!"#$%! !!!!!!!!!"#$%!! !!!"#$!!! !!!"#$!! !!, (1)
where MVE is the yearly market value of equity, BVA is the yearly book value of total assets, BVL is the yearly book value of liabilities, CS is the yearly customer satisfaction score for the specific company, !!is the constant, and !! is the disturbance
term.
To test if companies with high customer satisfaction ratings have lower risk than companies with low customer satisfaction ratings the Fama and French three-factor model is used to gain the measures of systematic and idiosyncratic risk, which is presented in equation (2) (Fama and French (1993)).
!!!!!!!" ! !!"! ! !!"! !!" !!" ! !!" ! !!"!!"#!!! !!!!!"#!!! !!" (2)
Where !!" is the daily return on stock of firm i on day t,!!!" the daily risk-free return
! *&!
the Fama-French size portfolio on day t, and !!"#!! the Fama-French market-to-book ratio portfolio on day t. !!" and the !!" are calculated by the equation (3) and (4):
!!!" ! !"!! !!!!!
!!! !!!!! (3)
!!!" !!"!! !!!!!
!!! !!!!! (4)
Where !!!! is the RI amount of company i on day t, !!!!!!!! is the RI amount of company i on the previous day, !!!! is the RI amount of the MSCI World index on day
t, and !!!!!!!! is the RI amount of the MSCI World index on day t. The SMB variable
is the difference in the return between a portfolio with small capitalization stocks and with high large capitalization stocks. The HML variable is the difference in the return between a portfolio with high book-to-market ratios (value stocks) and a portfolio with low book-to-market ratios (growth stocks). Hence, a “value” stock should have a positive exposure to the book-to-market variable, which should be the opposite for a “growth” stock. The intercepts in the equation are the abnormal portfolio returns relative to the three-factor model. Additionally, the coefficients in the equation except
for !!" are the estimations for the systematic risk of the portfolios. !!!" ! !!"! of
! *'!
To test hypotheses 3a and 3b almost the same methodology as Matzler et al. (2005) is used. In their research they use Tobin’s q since Anderson, Fornell, and Mazvancheryl (2004) provide an overview of advantages and disadvantages of numerous measures of shareholder value. They state that it is the best option since it is forward-looking, comparable across firms, and grounded in economic theory. Tobin’s q is defined as the ratio of market value of the firm to the replacement cost of its tangible assets. Though, it should be noticed that the ratio is difficult to compute since the replacement costs of a company’s tangible assets is hard to measure (Varajya, Kerin, and Weeks (1987)). Therefore, Matzler et al. (2005) use the approximation for Tobin’s q of Chung and Pruitt (1995). The approximations of q is calculated by adding the market value of equity to the book value of debt and divide that total by the total assets of the specific company. In contrary to the study of Matzler et al. (2005) I calculate monthly approximations of q instead of quarterly, since it is more specific than quarterly data. A linear regression analysis on the following two equations is used to test the degree of the relationship between positive and negative changes in ACSI ratings and the approximation of q in different months, and I test the time lag for positive and negative changes:
!"#$%&!!!!!!!!!!!!!!"!!!"#$!! ! !!"##$!"#$%$&' ! !!!!! (5)
!"#$%&!!!!!!!!!!!!!!"!!!"#$!! ! !!"##$!"#!"#$%! !!!! (6)
where !"!!!"#$!!! is the change in the CS rating for the specific companies at each
release moment, !!"##$!"#$%$&' and !!"##$!"#$%&'" are respectively the dummy
variables for a positive and a negative change in customer satisfaction ratings, !! is a constant, !! is the degree of the relationship between the approximation of Tobin’s q
! *(!
represents the month of the release (MofR) of the customer satisfaction rating and t = 1, …, t = 12 the following twelve months, and !!! is the disturbance term.
For the portfolio study I estimate the rate of return for each company at each day by using equation (3). To estimate the beta on the stock of firm j on day t the following market model is used:
!!" ! ! !! ! !!!!"! !!" (7)
where !!" is the rate of return on the stock of firm j on day t, !!" is the market rate of
return using the MSCI world index on day t, !! is an intercept, !! is the slope
parameter that measures the sensitivity of the return of the stock to the market index,
and !!" is the disturbance term with the normal least squares properties.
V. Results
This section discusses the results for the four hypotheses in the same order as they are discussed in section II, IV.
A. Relationship between CS and MVE
! *)!
Therefore, table II also presents the results for the same regression without the CS coefficient. As the results show, omitting the CS variable has almost no effect on the R-square and the other coefficients, and omitting the variable has the most effect on the constant.
Table I – General statistics of the data for equation (1).
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Average CS rating 68.83 71.34 71.94 70.18 70.67 73.12 71.74 71.7 68.23 69.08 69.01
Average LN MVE 8.73 8.76 8.78 8.92 8.69 9.18 9.07 9.02 8.47 9.96 9.14
Average LN BVA 18.3 18.42 18.7 17.75 17.52 17.19 16.78 16.98 16.98 17.1 17.27 Average LN BVL 18.35 18.41 18.65 17.23 16.71 16.79 16.37 16.54 16.54 16.69 16.88
Table II - The effect of the customer satisfaction rating on the market value of equity.
Ln(Market value) Ln(Market value)
CS (lnCS) !! 0.2179
Significance 0.4363
Total assets (lnBVA) !! 1.1329 1.1322
Significance 0.0*** 0.0*** Total liabilities (lnBVL) !! -0.4619 -0.4634 Significance 0.0**** 0.0*** Constant !! -3.647 -2.6941 Significance 0.0038*** 0.0*** N 952 952 Adjusted R-square 0.648 0.6472 F-statistic 5.8469 8.7709
Notes: The third column is the result for !"#$%!!!!!! !!!!"#$%!! !!!"#$!!! !!!"#$!! !!, where the last column is the result for the same equation as column three only without the CS variable.
P-value < 0.1 *, P-value < 0.05 **, P-value < 0.01 ***
! "+!
include companies for which I have both datasets and if I only include companies with indirect data. As an example for indirect data, I have customer satisfaction ratings for “Lattlecom”, but this is a daughter company of “Telia Sonera”. Therefore, the market value of equity of “Telia Sonera” is used, and if this is the case it is labeled as an indirect relationship.
Table III – The effect of CS ratings on the market value of equity (direct vs. indirect).
Ln(Market value direct) Ln(Market value indirect)
CS(lnCS) !! 0.1238 0.2809
Significance 0.7406 0.0774*
Total assets (lnBVA) !! 1.6085 0.768
Significance 0.0*** 0.0*** Total liabilities (lnBVL) !! -0.8541 -0.3269 Significance 0.0*** 0.0*** Constant !! -4.8668 0.3848 Significance 0.004*** 0.0315** N 622 330 Adjusted R-square 0.719 0.9608 F-statistic 5.3078 2.6923
Notes: The third and last column are the result for !"#$%!!!!!! !!!!"#$%!! !!!"#$!!!
!!!"#$!! !!, where the third column is the result for the direct relationships, where the last column is the result for the indirect relationships.
P-value < 0.1 *, P-value < 0.05 **, P-value < 0.01 ***
! "*!
and could have been rated differently if people knew the fact that they are part of each other.
B. The relationship between CS and risk
Table IV presents the results for equation (2) of the Fama and French three-factor model. The results show that the three-three-factor model explains a significant proportion of the portfolios’ risk that is shown by the adjusted R-square of the five portfolios. For example for the portfolio of the highest 20% customer satisfaction ratings the model explains 88% of the volatility. The results show that portfolios two and four are almost equal to the risk in the market (!!"!of 1.08 and 0.99). Portfolios high, three, and low are more risky than the market (!!"!of 1.23, 1.13, and 1.13). These five results for the !!"!are highly significant. Furthermore, the results for the
!!" (size factor) show that portfolios high until three have positive signs and are
significant, which suggest that these portfolios consist of small capitalization stocks. This is in contradiction with results of Aksoy et al. (2008) who find negative signs for their portfolios based on ASCI ratings. For the !!! (book-to-market) variable only
! ""!
Table IV - Customer satisfaction portfolio risk estimates using the Fama and French three-factor model. Portfolio !!" !!" !!! Constant Adjusted R-square High Estimate 1.2364 0.4606 0.0379 0.0076 0.8817 Significance 0.0*** 0.0*** 0.542 0.0399** 2 Estimate 1.0821 0.2985 0.2701 0.0018 0.8865 Significance 0.0*** 0.001*** 0.0*** 0.6207 3 Estimate 1.1289 0.1224 0.2695 0.0056 0.8833 Significance 0.0*** 0.099* 0.0*** 0.1156 4 Estimate 0.9946 0.077 0.1139 0.0038 0.8339 significance 0.0*** 0.2796 0.05* 0.2591 Low Estimate 1.129 -0.018 0.2832 0.0052 0.8665 Significance 0.0*** 0.8149 0.0*** 0.1562 High-Low Estimate 0.1074 0.4786 -0.2453 0.0023 0.2207 Significance 0.2345 0.0*** 0.0016*** 0.5977
Notes: The five portfolios high until low are portfolios based on CS ratings, where portfolio high
consist of firms that are the top 20% of CS ratings and portfolio 2 the next 20% etcetera. The results are from release moment 12 until 102.
P-value < 0.1 *, P-value < 0.05 **, P-value < 0.01 ***.
Although the three-factor model explains most of the volatility for the portfolios, it only explains 22% of the volatility in the high-low portfolio. This portfolio has a risk-adjusted abnormal return of 0.23% for each period between the choice moments even though I take the three risk factors of Fama and French into account. The risk-adjusted abnormal return for the whole period of 91 choice moments is 20.9%, however this return is not significant. The !!" and the !!! variable
! "#!
Overall, these results show that a portfolio based on high customer satisfaction rating stocks has an abnormal return of 0.76% for each period between the choice moments (69% abnormal return for the whole period). However, this return can be explained by the fact that the portfolio is more risky than the market and the portfolio consists of small capitalization stocks, which usually generate higher returns (Bernard and Thomas (1989); Chan, Chen, and Hsieh (1985); Chan and Chen (1991)).
C. The effect of change in CS and time lag
The results for hypothesis 3a and 3b are presented in table V and VI. These results show that the relationship between a positive as well as a negative change in the customer satisfaction rating and shareholder value is not significant.
Table V – The relationship between a positive change of the customer satisfaction rating and shareholder value (Tobin’s q).
!! Significance Constant Significance
MofR -0.0340 0.8844 0.2780 0.0*** M2 0.0315 0.8941 0.2772 0.0*** M3 -0.0546 0.8147 0.2790 0.0*** M4 0.0074 0.9762 0.2829 0.0*** M5 -0.0030 0.9905 0.2829 0.0*** M6 0.0671 0.7871 0.2815 0.0*** M7 0.0512 0.8338 0.2795 0.0*** M8 0.0568 0.8166 0.2790 0.0*** M9 0.0859 0.7286 0.2812 0.0*** M10 0.0938 0.7045 0.2804 0.0*** M11 0.0699 0.7757 0.2815 0.0*** M12 0.2607 0.3228 0.2771 0.0***
Notes: MofR is the month of the release of a new CS rating, and M2 until M12 the following months.
! "$!
Table VI – The relationship between a negative change of the customer satisfaction rating and shareholder value (Tobin’s q).
!! Significance Constant Significance
MofR 0.0929 0.7713 0.2784 0.0*** M2 0.1331 0.6798 0.2792 0.0*** M3 0.0577 0.8560 0.2787 0.0*** M4 0.2989 0.3675 0.2865 0.0*** M5 0.2892 0.3836 0.2862 0.0*** M6 0.3201 0.3340 0.2863 0.0*** M7 0.2498 0.4435 0.2832 0.0*** M8 0.2575 0.4312 0.2830 0.0*** M9 0.2802 0.3975 0.2859 0.0*** M10 0.2681 0.4174 0.2852 0.0*** M11 0.3842 0.2429 0.2875 0.0*** M12 0.3522 0.3052 0.2857 0.0***
Notes: MofR is the month of the release of a new CS rating, and M2 until M12 the following months.
Where!!!! is the degree of the relationship With 317 negative changes in CS from 2001 to 2011. P-value < 0.1 *, P-P-value < 0.05 **, P-P-value < 0.01 ***.
However, it can be argued that when you relate a change in customer satisfaction ratings to Tobin’s q instead of the change in Tobin’s q, you relate variables at different levels. Hence, I also tested the relationship between a positive / negative change of customer satisfaction ratings on the change in Tobin’s q. The outcomes for these tests are presented in appendix B, since only one significant outcome is important to be discussed. The results show that the relationship between a positive / negative change in the customer satisfaction rating and the change in Tobin’s q is generally not significant. However, the relationship in month seven after the announcement of a negative change in the customer satisfaction is highly significant related to a change in Tobin’s q. In this month the !!!has a value of -0.153 and the
significance level is 0.0286.
! "%!
customer satisfaction ratings and Tobin’s q. Consequently; I also test the relationship between the customer satisfaction ratings, so not the change, and Tobin’s q. The results, presented in appendix C, show that the relationship in months three, four, seven until ten, and twelve are significant. However, the degree of !! is much lower
compared to the results of Matzler et al. (2005). Additionally, I test the relationship of the customer satisfaction ratings times the dummy for a negative or a positive change to Tobin’s q. These results are also presented in appendix C. The outcomes of the tests show that solely month eight after the release date is highly significant for the positive change dummy times the CS rating. For the negative change dummy times the CS ratings months four, six, and eight until twelve after the release date are significant.
! "&!
VI. Portfolio study
Fornell et al. (2006) show that new customer satisfaction ratings do not affect stock prices on the announcement day. This is in line with the efficient market theory that stock prices include all information (Fama (1970)). Under the conditions of the efficient market theory it would not be possible to create excess returns with a trading strategy that is based on announcements of customer satisfaction ratings. However, the previously discussed theories suggest that companies with satisfied customers will make higher profits. Therefore, I test whether it is possible to generate excess returns with a trading strategy based on announcements of customer satisfaction ratings. Fornell et al. (2006) performed a portfolio study; they create two kinds of stock portfolios a hypothetical paper portfolio with simple trading rules and a real-world portfolio. Their study shows that the paper portfolio outperforms the Dow Jones Industrial Average (DIJA), the S&P500, and the NASDAQ. The real-world portfolio also outperformed the DIJA and the S&P500. I perform the same two portfolio studies for my dataset.
! "'!
years 2001 until 2011, the dates for the 102 moments are presented in appendix D. For each announcement moment the portfolio is reexamined and companies will be included or excluded when they do / do not meet the criteria of the portfolio. For example, “Handelsbanken” is included in the portfolio from the first announcement moment until the last, but “Omnitel” is only included in the announcement moment 6 until 65. The amount of companies included in the portfolio during the 102 moments ranges from 2 in the first four years, until a maximum of 42 companies at announcement moment 68 (8-11-2009 until 20-12-2009).
! "(!
Figure 3 - The cumulative returns for the top 20% portfolio and the MSCI world (2006-2012).
Notes: The top 20% portfolio consisting of stocks based on the customer satisfaction rating. The period
is from 1-2-2006 (first announcement date in 2006) until 31-1-2012.
The graph shows that high customer satisfaction ratings pay off in growing markets and not in downward moving markets. This is in contradiction with the theory, since that suggests that high customer satisfaction should create stability in the turnover of a company, which is already explained in the theory section. Additionally, the portfolio study of Fornell et al. (2006) shows that customer satisfaction pays off in growing markets and downward moving markets. An explanation for the contradiction in the results can be that during the downward moving markets, customer’s focus more on the price of products/services instead of their loyalty toward a company. This can be strengthened by the fact that during the downward period the financial crisis was at a peak and the customers focus even more on prices.
Similar to the study of Fornell et al. (2006) I discuss possible other explanations besides the customer satisfaction ratings for the results that the portfolio
! ")!
outperforms the benchmark portfolio. First, the chance that professional stock traders constantly beat the market is very small, regardless of the fact whether the efficient market theory holds or not. Secondly, the results are supported by empirical findings in the literature (Aksoy et al. (2008); Fornell et al. (2006); Fornell, Mithas, and Morgeson (2009)). Thirdly, one could question if the outperformance is caused by an increase in risk, beta, of the portfolio. However, the average beta relative to the benchmark portfolio for the portfolio during 2006 until 20125 is 0,95 that is only a little bit less than the market. The separate betas for all years and for the two different periods are presented in appendix F. The firms included in the customer satisfaction ratings are selected based on size, and the total of the customer satisfaction firms should mirror large cap indexes. Therefore, there is no designation to the fact that the bottom 80% should perform the same as the top 20% of companies. To see whether the bottom 80% also outperforms the benchmark portfolio, I calculate the cumulative return for the bottom 80% as well. The cumulative return for the bottom 80% for the period 2006 until 2012 is 14% and for the MSCI world it is also 14%. The cumulative return for the bottom 80% of companies is almost the same as the cumulative return for the MSCI world, which rule out the selection criteria as a covariate of the stock performance. Another explanation for the long-term abnormal stock returns is mentioned by Fama and French (1992) they say that book-to-market ratios are possible explanations for long-term abnormal stock returns. For this reason, the market ratios for the top 20% and the bottom 80% are compared. The book-to-market ratios are 0.72 for the bottom 80% and 0.79 for the top 20% during the 2006 – 2012 period. The ratios are higher than in the study of Fornell et al. (2006) but are just like their study fairly close together, which rules out the possibility that the book-to-market ratio explains the difference in long-term stock returns. The results are !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
! #+!
consistent with the study of Fornell et al. (2006) that the betas are close together and lower than one. The last possible explanation for the difference in the returns for the top 20% and the bottom 80% can be the “size effect”. Previous studies show that the stocks of smaller firms tent to generate higher returns (Bernard and Thomas (1989); Chan, Chen, and Hsieh (1985); Chan and Chen (1991)). Therefore, I calculate the average size for the top 20% (187 million) and the bottom 80% (577 million). The average size is calculated by using yearly market capitalization data from DataStream. The results show that the top 20% portfolio consists of smaller firms than the bottom 80%, which therefore could be an explanation for the long-term abnormal returns according to the previous studies. The fact that the average size in the top 20% portfolio is lower is due to the inclusion of Internet companies that have high satisfaction ratings but low market capitalization (for example companies as Apple, Ebay.co.uk, Play.com, and Ticketmaster). This result is in contradiction with the results of Fornell et al. (2006), since they find that the top 20% portfolio consist of larger sized companies than the bottom 80%. Overall, it can be concluded that the abnormal returns can be explained by the customer satisfaction ratings and due to the fact that the firms are smaller in the top 20% portfolio. The fact that these firms are smaller indicates a higher risk profile for the top 20% portfolio, although the beta and the book-to-market ratio indicate otherwise.
! #*!
it is always easier to find a successful trading strategy, however the trading strategy is extremely easy and straightforward.
Nevertheless, there were also some bad trades made with this trading strategy. A good example is the purchase of “Aksigorta”, which is purchased at 28-3-2007 and is sold at 27-3-2009. “Aksigorta” had a customer satisfaction rating of 80 and during the period the cumulative return was -98% where the customer satisfaction rating only dropped to 79 during that period. At 29-3-2010 “Aksigorta” is repurchased again and is hold on to until 24-3-2011, when including this period the cumulative return is -108%.
! #"!
! ##!
invest in 39 stocks of “Sparebank1 Gruppen”. Another extreme example is moment 78, where I invest in one stock of “Borneo motors”, but to equal the amount invested in the other companies I have to invest in 491 stocks of “Comfortdelgro”, 491 stocks of “ SingPost”, and 246 stocks of “Cathay Pacific”. These problems will influence the results of this portfolio study drastically. Consequently, I choose to not perform the real-world portfolio study of Fornell et al. (2006).
VII. Conclusion
! #$!
portfolio strategy can be used to generate excess returns although the additional risk should be taken into account. Even though the positive relationship between customer satisfaction ratings and the market value of equity of the firm is not significant for the whole dataset, this positive relationship can still have implications for managers of firms to focus more on customer satisfaction to generate more shareholder value.
Whether there is a difference in the degree of the relationship between a reward/punishment of shareholder value after an increase/decrease in the customer satisfaction rating cannot be concluded. However, the results for the relationship between the change in customer satisfaction rating and the change in shareholder value indicate that there is a relation between a negative change in the rating and the change in shareholder value since seven months after a decrease in the satisfaction rating the relationship is significant. Furthermore, it cannot be concluded if there is a difference in the time lag between a reward/punishment of shareholder value after an increase/decrease in the customer satisfaction rating since the results are not significant.
! #%!
! #&! Reference list
Aksoy, Lerzan, Bruce Cooil, Christopher Groening, Timothy L. Keiningham, and Atakan Yalcin, 2008, The long-term stock market valuation of customer satisfaction,
Journal of Marketing, 72, 105-122.
Ambler, Tim, 2000, Marketing and the bottom line, London: Financial Times, Prentice Hall.
Anderson, Eugene W., and Mary W. Sullivan, 1993, The antecedents and
consequences of customer satisfaction for firms, Marketing Science, 12, 125–43.
Anderson, Eugene W., Claes Fornell, and Donald R. Lehmann, 1994, Customer satisfaction, market share, and profitability: Findings from Sweden, Journal of
Marketing, 58, 53–66.
Anderson, Eugene W., 1996, Customer satisfaction and price tolerance, Marketing
letters, 7, 19-30.
Anderson, Eugene W., Claes Fornell, and Roland T. Rust, 1997, Customer
satisfaction, productivity, and profitability: Differences between goods and services,
Marketing Science, 16, 129–45.
! #'!
Barth, Mary E., and Maureen F. McNichols, 1994, Estimation and market valuation of environmental liabilities relating to superfund sites, Journal of Accounting
Research, 32, 177–209.
Barwise, Patrick, Paul R. Marsh, and Robin Wensley, 1989, Must finance and strategy clash?, Harvard Business Review, 67, 865-890.
Bearden, William 0., and Jesse E. Teel, 1983, Selected determinants of consumer satisfaction and complaint reports, Journal of Marketing Research, 20, 21-28.
Bernard, Victor L., and Jacob K. Thomas, 1989, Post-earnings-announcement drift: Delayed price response or risk premium?, Journal of Accounting Research, 27, 1-48.
Bolton, Ruth N., and James H. Drew, 1991, A multistage model of customers'
assessments of service quality and value, Journal of Consumer Research, 17, 375-84.
Bolton, Ruth N., 1998, A dynamic model of the duration of the customer’s
relationship with a continuous service provider: The role of satisfaction, Marketing
Science, 17, 45–65.
! #(!
Boulding, William, Ajay Kalra, Richard Staelin, and Valarie Zeithaml, 1993, A dynamic process model of service quality: From expectations to behavioral intentions,
Journal of Marketing Research, 30, 7- 27.
Chan, K.C., and Nai-Fu Chen, 1991, Structural and return characteristics of small and large firms, Journal of Finance, 4, 1467 – 1484.
Chan, K.C., N. Chen, and D. Hsieh, 1985, An exploratory investigation of the firm size effect, Journal of Financial Economics, 14, 451 – 471.
Chung, Kee H., and Stephen W. Pruitt, 1995, A simple approximation of Tobin’s Q,
Financial Management, 23, 70-79.
Day, George, and Liam Fahey, 1988, Valuing market strategies, Journal of
Marketing, 52, 5-57.
Fama, Eugene F., 1970, Efficient capital markets: A review of theory and empirical work, Journal of Finance, 25, 383-417.
Fama, Eugene F., and Kenneth R. French, 1992, The cross-section of expected stock returns, Journal of Finance, 47, 427–65.
! #)!
Fornell, Claes, 1992, A national customer satisfaction barometer: The Swedish experience, Journal of Marketing, 56, 1-21.
Fornell, Claes, Michael D. Johnson, Eugene W. Anderson, Jaesung Cha, and Barbara E. Bryant, 1996, The American customer satisfaction index: Nature, purpose, and findings, Journal of Marketing, 60, 7-18.
Fornell, Claes, Sunil Mithas, Forrest V. Morgeson III, and M.S. Krishan, 2006, Customer satisfaction and stock prices: High returns, low risk, Journal of Marketing, 70, 3-14.
Fornell, Claes, Sunil Mithas, and Forrest V. Morgeson III, 2009, The economic and statistical significance of stock returns of customer satisfaction, Marketing science, 28, 820-825.
Hallowell, Roger, 1996, The relationship of customer satisfaction, customer loyalty and profitability: an empirical study, International Journal of Service Industry
Management, 7, 27–42.
Homburg, Christian, and Heiko Schäfer, 2002, Die erschließung von
! $+!
Ittner, Christopher D., and David.F. Larcker, 1996, Measuring the impact of quality initiatives on firm financial performance, in Advances in the Management of
organizational Quality, Vol. 1, Soumen Ghosh and Donald Fedor, ed. Greenwich,
CT: JAI Press.
Jacobson, Robert, and Natalie Mizik, 2009, The financial markets and customer satisfaction: Reexamining possible financial market mispricing of customer satisfaction, Marketing science, 28, 810-819.
Johnson, Michael D., and Claes Fornell, 1991, A framework for comparing customer satisfaction across individuals and product categories, Journal of Economic
Pshychology, 12, 267-286.
Krishnamurthi, Lakshman, and S. P. Raij, 1991, An empirical analysis of the
relationship between brand loyalty and consumer price elasticity, Marketing Science, 10, 172–183.
Landsman, Wayne, 1986, An empirical investigation of pension fund property rights,
Accounting Review, 61, 662–91.
Matzler, Kurt, and Heinz K. Stahl, 2000, Kundenzufriedenheit and unternehmenswertsteigerung, Die Betriebswirtschaft, 60, 626-640.
Matzler, Kurt, Hans H. Hinterhuber, Christian Daxer, and Maximilian Huber, 2005, The relationship between customer satisfaction and shareholder value, Total quality
! $*!
Mittal, Vikas, and Wagner Kamakura, 2001, Satisfaction, repurchase intent, and repurchase behavior: Investigating the moderating effect of customer characteristics,
Journal of Marketing Research, 38, 13 1-42.
Narayandas, Das, 1998, Measuring and managing the benefits of customer retention: An empirical investigation, Journal of Service Research, 1, 1-10.
Oliver, Richard L., 1980, A cognitive model of the antecedents and consequences of satisfaction decisions, Journal of Marketing Research, 17, 460-469.
Oliver, Richard L., and John E. Swan, 1989, Consumer perceptions of interpersonal equity and satisfaction in transactions: A field survey approach, Journal of Marketing, 53, 21-35.
O’Sullivan, Don, Mark C. Hutchinson, Vincent O’Connell, 2009, Empirical evidence of the stock market’s (mis)princing of customer satisfaction, International Journal of
Research Marketing, 26, 154-161.
Reichheld, Frederick F., and W. Earl Sasser Jr., 1990, Zero defections: Quality comes to services, Harvard Business Review, 68, 105–111.
! $"!
Rust, Roland T., Anthony J. Zoharik, and Timothy L. Keiningham, 1995, Returnon Quality (ROQ): Service quality financially accountable, Journal of marketing, 59, 58-70.
Srivastava, Rajendra K., Tasadduq A. Shervani, and Liam Fahey, 1998, Market-based assets and shareholder value: A framework for analysis, Journal of Marketing, 62, 2-18.
Stock, Ruth, 2003, Der einfluss der kundenzufriedenheit auf die preissensitivität von firmenkunden, Die Betriebswirtschaft, 63, 333–348.
Tuli, Kapil R., Sundar, G. Bharadwaj, 2009, Customer satisfaction and stock returns risk, Journal of Marketing, 73, 184-197.
Varajya, Nikhil, Roger A. Kerin, and David Weeks, 1987, The relationship between growth, profitability and firm value, Strategic Management Journal, 8, 487-497.
Verhoef, Peter C., Philip H. Franses, and Janny C. Hoekstra, 2001, The impact of satisfaction and payment equity on cross-buying: A dynamic model for multi service providers, Journal of Retailing, 77, 359-378.
Yi, Youjae, 1991, A critical review of customer satisfaction, in Review of Marketing, 1989, Valarie A. Zeithaml, ed. Chicago: American Marketing Association, 68-123.
! $#!
Appendix A. The definitions of the Customer satisfaction model
Customer Expectations: Customer expectation is a measure of the customer's
anticipation of the quality of a company's products or services. Expectations represent both prior consumption experience, which includes some non-experiential
information like advertising and word-of-mouth, and a forecast of the company's ability to deliver quality in the future.
Perceived Quality: Perceived quality is a measure of the customer's evaluation via
recent consumption experience of the quality of a company's products or services. Quality is measured in terms of both customization, which is the degree to which a product or service meets the customer's individual needs, and reliability, which is the frequency with which things go wrong with the product or service.
Perceived Value: Perceived value is a measure of quality relative to price paid.
Although price (value for money) is often very important to the customer's first purchase, it usually has a somewhat smaller impact on satisfaction for repeat purchases.
Customer Complaints: Customer complaints are measured as a percentage of
respondents who indicate they have complained to a company directly about a product or service within a specified time frame. Satisfaction has a negative
relationship with customer complaints, as the more satisfied the customers, the less likely they are to complain.
Customer Loyalty: Customer loyalty is a combination of the customer's professed
! $$!
Appendix B. The results for the relationship between the change of CS and the change of shareholder value (Tobin’s q).
!! Significance C Significance MofR -0.0004 0.9954 -0.004 0.0312 M2 -0.0159 0.9968 0.1244 0.3506 M3 -0.4326 0.7158 0.0493 0.211 M4 0.1091 0.6914 0.026 0.0048 M5 -0.0853 0.3968 -0.0005 0.8853 M6 -5.4564 0.5793 0.3767 0.253 M7 -0.0823 0.1162 -0.0039 0.0249** M8 0.0368 0.7052 -0.004 0.2184 M9 0.3462 0.3106 0.0046 0.6933 M10 0.0306 0.5763 -0.0009 0.6173 M11 0.4287 0.5941 0.0415 0.1318 M12 -1.7771 0.5889 0.1327 0.2188
Notes: Relationship between positive change of customer satisfaction and the change in Tobin’s q, with
391 positive changes. MofR is the month of the release of a new CS rating, and M2 until M12 the following months. Where!!!! is the degree of the relationship. N = 391. P-value < 0.1 *, P-value < 0.05 **, P-value < 0.01 ***. !! Significance C Significance MofR 0.0034 0.9627 -0.0039 0.0243 M2 3.2862 0.5466 0.1621 0.2173 M3 1.1106 0.4931 0.0549 0.158 M4 0.4015 0.2788 0.0325 0.0004*** M5 -0.1773 0.1884 -0.004 0.2265 M6 -8.628 0.5111 0.1816 0.5797 M7 -0.153 0.0286** -0.0071 0.0*** M8 0.0632 0.6266 -0.0026 0.4195 M9 0.2648 0.5619 0.0135 0.2426 M10 -0.0029 0.968 -0.0005 0.8046 M11 -0.1588 0.8831 0.0467 0.0914* M12 -4.458 0.298 0.0496 0.6436
Notes: Relationship between negative change of customer satisfaction and the change in Tobin’s q.
! $%!
Appendix C. The results for the relationship between the Customer satisfaction ratings and shareholder value (Tobin’s q).
!! Significance Constant Significance
MofR 0.0007 0.3923 0.2303 0.0001*** M2 0.0012 0.1696 0.1989 0.0009*** M3 0.0014 0.0868* 0.1813 0.002*** M4 0.0017 0.0493** 0.1658 0.0069*** M5 0.0019 0.0332** 0.1557 0.0111** M6 0.0018 0.041** 0.1607 0.0087*** M7 0.0017 0.0468** 0.164 0.0064*** M8 0.0017 0.0548* 0.1668 0.006*** M9 0.0016 0.0737* 0.1757 0.0044*** M10 0.0015 0.0898* 0.1807 0.0032*** M11 0.0013 0.1331 0.194 0.0015*** M12 0.0021 0.0278** 0.1314 0.0583*
Notes: Relationship between customer satisfaction ratings and Tobin’s q. MofR is the month of the
release of a new CS rating, and M2 until M12 the following months. Where!!!! is the degree of the relationship. N=737. P-value < 0.1 *, P-value < 0.05 **, P-value < 0.01 ***.
!! Significance Constant Significance
MofR 0.0001 0.7382 0.2752 0.0*** M2 0.0001 0.541 0.2737 0.0*** M3 0.0001 0.6301 0.275 0.0*** M4 0.0003 0.1693 0.2735 0.0*** M5 0.0003 0.1515 0.2728 0.0*** M6 0.0003 0.1381 0.2723 0.0*** M7 0.0003 0.1976 0.2715 0.0*** M8 0.0003 0.0199** 0.2711 0.0*** M9 0.0003 0.1761 0.2731 0.0*** M10 0.0003 0.1672 0.2723 0.0*** M11 0.0003 0.1542 0.2827 0.0*** M12 0.0003 0.1118 0.2696 0.0***
Notes: Relation between customer satisfaction ratings times the dummy for a positive change and
! $&!
!! Significance Constant Significance
MofR -0.0001 0.5011 0.281 0.0*** M2 -0.0002 0.3649 0.2827 0.0*** M3 -0.0001 0.4586 0.2821 0.0*** M4 -0.0004 0.0819* 0.2931 0.0*** M5 -0.0003 0.1044 0.2923 0.0*** M6 -0.0004 0.088* 0.2926 0.0*** M7 -0.0003 0.1064 0.2897 0.0*** M8 -0.0004 0.0907* 0.29 0.0*** M9 -0.0004 0.0853* 0.2929 0.0*** M10 -0.0004 0.0961* 0.2919 0.0*** M11 -0.0005 0.0263** 0.2963 0.0*** M12 -0.0005 0.0202** 0.2958 0.0***
Notes: Relation between customer satisfaction ratings times the dummy for a negative change and
! $'!
! $(!
Appendix E. The cumulative returns for the top 20% portfolio and the MSCI world (2001 – 2012).
Notes: The top 20% portfolio consists of stocks based on the customer satisfaction rating. The period
! $)!
Appendix F. The Betas for the top 20%portfolio for 2001-2012 and 2006 until 2012. Beta (2001 – 2012) Beta (2006 – 2012) 2001 0.48 2002 0.43 2003 0.22 2004 0.65 2005 0.33 2006 1.26 1.32 2007 1.22 1.22 2008 0.89 0.89 2009 0.88 0.88 2010 1.02 1.02 2011 0.96 0.96 2012 1.15 1.15
