Asia-Pacific Journal of Accounting & Economics ISSN: 1608-1625 (Print) 2164-2257 (Online) Journal homepage: http://www.tandfonline.com/loi/raae20 Value relevance of customer equity beyond financial statements: evidence from mobile telecom industry Yong Bum Choi, Janghyuk Lee, Shijin Yoo & Yong Keun Yoo To cite this article: Yong Bum Choi, Janghyuk Lee, Shijin Yoo & Yong Keun Yoo (2017): Value relevance of customer equity beyond financial statements: evidence from mobile telecom industry, Asia-Pacific Journal of Accounting & Economics, DOI: 10.1080/16081625.2017.1386575 To link to this article: http://dx.doi.org/10.1080/16081625.2017.1386575 Published online: 08 Oct 2017. Submit your article to this journal Article views: 60 View related articles View Crossmark data Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=raae20 Download by: [Tufts University] Date: 28 October 2017, At: 05:59 Asia-Pacific Journal of Accounting & Economics, 2017 https://doi.org/10.1080/16081625.2017.1386575 Value relevance of customer equity beyond financial statements: evidence from mobile telecom industry* Yong Bum Choia, Janghyuk Leeb, Shijin Yoob and Yong Keun Yoob a Martin J. Whitman School of Management, Syracuse University, Syracuse, NY, USA; bBusiness School, Korea University, Seoul, South Korea Downloaded by [Tufts University] at 05:59 28 October 2017 ABSTRACT We investigate the extent of usefulness of customer equity (CE) estimate in explaining market value of equity (MVE) and the determinants of the discrepancy between CE estimate and MVE. Using the data of 17 companies in the mobile telecom companies in 7 countries from 2004 to 2008, we find (1) CE estimate provides incremental value relevant information to explain contemporaneous MVE beyond financial statements, and (2) CE estimate can predict future stock returns by indicating a part of temporal divergence between MVE and a firm’s intrinsic value of equity. ARTICLE HISTORY Received 7 December 2015 Accepted 14 September 2017 KEYWORDS Customer equity; market value of equity; value relevance 1. Introduction Accounting and finance researchers have spent considerable effort investigating how to accurately estimate firm value (e.g. Ohlson 1995; Dechow, Hutton, and Sloan 1999; Lee, Myers, and Swaminathan 1999; Easton et al. 2002). While they have focused on financial statements as the main information source, it will be true that ‘other’ information outside financial statements is also useful for firm valuation. This study seeks to find one important ‘other’ information by examining the information contents of customer equity. Given that the incorporation of all available value relevant information into firm valuation may improve valuation accuracy, whether customer equity provides additional value relevant information is of interest for accounting and finance researchers. The firm needs to understand its customers to develop and implement successful business strategies. Many researchers have been investigating the linkage between the firm’s marketing actions and the value for customers. However, measuring economic value of customers is also an equally important task to fully understand the firm’s customers though its importance has been recently appreciated by marketing researchers. As a response to this call, customer equity (CE) has been suggested as an effective metric of the total value of customers and thus links marketing with other business functions such as accounting and finance within an organization (Gupta and Lehmann 2005). CE is a concept that understands customers as an asset for a firm to measure, manage, and maximize like a financial asset (Blattberg and Deighton 1996; Gupta, Lehmann, and Stuart 2004). According to this concept, marketing resources allocated to customers should be considered as an investment to induce positive behavioral changes of the consumers in the future, not an expenditure to be consumed CONTACT Janghyuk Lee [email protected] * Accepted by Jeong-Bon Kim upon recommendation by Oliver Rui © 2017 City University of Hong Kong and National Taiwan University Downloaded by [Tufts University] at 05:59 28 October 2017 2 Y. B. CHOI ET AL. and disappear. Under this perspective, CE can represent the firm’s intangible asset that reflects various performance indicators such as brand equity, market power, or technological know-how in combination. Therefore, CE is believed to be an effective measure that can enhance the reliability of marketing program’s accountability and long-term profitability (Reinartz and Kumar 2000; Gupta, Lehmann, and Stuart 2004; Rust, Lemon, and Zeithaml 2004; Wiesel, Skiera, and Villanueva 2008). Beyond its role as a measure for the value of customers, CE can also provide the equity value related information to investors in stock market. This is because CE is based on the direct estimation of a firm’s future expected cash flows from its customers, which in turn firm’s equity value consists of. However, there is a paucity of evidence on the value relevance of CE, which is another important metric to evaluate its usefulness. While Gupta, Lehmann, and Stuart (2004) and Kumar and Shah (2009) document that CE estimate is closely associated with a corresponding firm’s market value of equity (MVE), these studies do not show – and do not seek to show – the extent of usefulness of CE to explain MVE. Furthermore, although they found the existence of a significant gap between CE estimate and MVE, they did not further analyze any systematic mechanism behind the discrepancy. For example, while previous studies suggest that the ignorance of risk factors (Kumar and Shah 2009) or debt and non-operating assets (Schulze, Skiera, and Wiesel 2012) is a part of reasons for the discrepancy between CE estimate and MVE, those studies lack of a more comprehensive investigation of the other determinants. Our study attempts to fill this void in the literature. Specifically, we examine whether CE provides incremental value relevant information to explain contemporaneous MVE beyond the information contents financial statements present. If CE does not provide any incremental value relevant information beyond financial statements, CE would be considered to simply repackage the information contents of financial statements and so to be useless at least in explaining a firm’s equity value. Given that financial statements are the most important public information which is necessary for investors to derive the equity value of the firm, the information contents of financial statements will be a good benchmark to evaluate the extent of usefulness of CE in explaining the firm’s equity value. By employing this specific benchmark for the extent of value relevance, our study makes a more conclusive argument about the extent of value relevance of CE than prior literature (e.g. Gupta, Lehmann, and Stuart 2004) which documents a positive association between CE and MVE simply without any benchmark. Then, we explore also whether CE estimate can predict future stock returns. Given that MVE is a noisy proxy for the firm’s intrinsic value of equity, the discrepancy between CE and MVE may capture investors’ temporary mispricing of a firm’s equity value rather than measurement errors in CE estimate. If CE estimate is less biased than MVE in incorporating value related factors in general, the discrepancy between CE and MVE will indicate at least a part of temporal divergence between MVE and the firm’s intrinsic value of equity, and so will predict a future movement of stock prices which will converge into the firm’s intrinsic value of equity as time goes. Given that it is unclear whether CE estimate incorporates value related factors more (or less) properly than MVE in general, it is an empirical question whether CE estimate has the prediction power for future stock returns, which is another metric to evaluate the usefulness of CE. In addition, we investigate which factors explain the discrepancy between CE estimate and MVE. Given that both metrics are the noisy proxies for the unobservable intrinsic value of equity of the firm, it is possible that CE estimate may not properly incorporate the relevant information for a firm’s intrinsic value of equity by a greater extent than MVE, and vice versa. We posit that some firm-level characteristics, which CE estimate and/or MVE incorporate into their valuation by a differing extent, will lead to the discrepancy between CE estimate and MVE. If CE estimate is more biased than MVE in incorporating some value related factors, we can improve the accuracy of CE estimate by reflecting such factors in its estimation. To the contrary, if CE estimate incorporates some value related factors more properly than MVE does, investors can refer to CE estimate to conduct a more efficient and less biased equity valuation. The rest of the article is organized as follows: In the next section, we review the relevant literature on the relationships between CE and a firm’s equity value and develop our research hypotheses. Section ASIA-PACIFIC JOURNAL OF ACCOUNTING & ECONOMICS 3 3 describes our methodology, data-set, and the measurement of CE. Section 4 presents our main empirical results. Finally, we draw implications for researchers and practitioners and discuss further research issues as concluding remarks in Section 5. 2. Background and hypotheses development Downloaded by [Tufts University] at 05:59 28 October 2017 2.1. Literature review Researchers have attempted to reveal the relation of marketing actions with the firm value to enhance marketing’s accountability. In particular, as the global financial investment market has evolved, the increased recognition that the outcome of all strategic decisions of a firm will be eventually incorporated in market value of equity has accelerated this trend of research. For instance, Srivastava, Shervani, and Fahey (1998) showed the process how the marketing related assets, such as customer relationship, channel relationship, and partner relationship, affect a firm’s market value of equity. Ittner and Larcker (1998) showed the impact of customer satisfaction level on the firm performance indicators such as retention, revenue, and revenue changes. Anderson, Fornell, and Mazvancheryl (2004) and Fornell et al. (2006) revealed the positive relationship between customer satisfaction and a firm’s market value of equity. There are diverse studies which have examined the links between marketing actions and firm value, such as brand relationship and firm value (Rao, Agarwal, and Dahlhoff 2004), new product announcement and firm value (Sorescu, Shankar, and Kushwaha 2007), negative consumer voice and stock return (Luo 2007), and advertising and firm value (Joshi and Hanssens 2010).1 Several marketing researchers also have focused on the value of customers. They show interests about studies of marketing strategies which can evaluate, manage, and enhance the value of customers (Blattberg and Deighton 1996; Gupta, Lehmann, and Stuart 2004; Rust, Lemon, and Zeithaml 2004; Venkatesan and Kumar 2004). At the core of this line of research, there is a concept of customer lifetime value (CLV), which is defined by the discounted stream of cash flows generated over the lifetime of a customer. Although the measurement of CLV is slightly different across industries with respect to the relationship between the firm and the customer (Dwyer 1997), the metric provides a basis to understand the value of marketing asset in terms of longer-term profitability, which can be used in measuring return on marketing investment. Based on this concept of CLV, researchers define customer equity as the sum of CLVs of all customers including existing and expected ones in the future (Blattberg, Getz, and Thomas 2001; Gupta, Lehmann, and Stuart 2004). The direct association between CE and MVE has been explored by numerous researchers as well (Gupta, Lehmann, and Stuart 2004; Kumar and Shah 2009; Libai, Muller, and Peres 2009). In particular, Gupta, Lehmann, and Stuart (2004) state that traditional financial index, such as price-to-earnings ratio (PER), is not applicable to explain a certain kind of business (e.g. dot-com companies on the Internet). They suggest that CE can be an alternative metric to measure such firm’s value by documenting highly positive association between CE and MVE. Kumar and Shah (2009) suggest an MVE prediction model using CE and other relevant variables, such as volatility and vulnerability that are reflected in MVE but not in CE. They show that the actual MVE is very close to the predicted MVE from their model, suggesting that some risk factors are the variables that significantly affect the CE–MVE relationship. Recently, Libai, Muller, and Peres (2009) estimate and compare the CE and MVE of seven companies in four different service industries. They divide attrition rate into dis-adoption, the case where a customer stops using the service category itself, and a churn, the case where a customer leaves for rival companies. They suggest a model which considers the influence of these factors to CLV and customer diffusion curve, and find that six of seven companies show similar monetary values between their CE and MVE. Based on the assumption of stock market efficiency, which allows us to derive reversely CE from market value of equity, debt, non-operating assets and tax, Schulze, Skiera, and Wiesel (2012) find the leverage effect of CE on MVE is 1.55 on average from over 2000 listed companies in US. They also report CE with infinite time horizon fits the bracket of min and max MVE from the case of two internet-based companies over multiple periods (Table 1). Current paper Schulze, Skiera, and Wiesel (2012) Computer and retail; 2 firms in US Estimate of CEs are very close to MVE; Retail, financial service, radio, mobile telecom; 7 Considering attrition improves the estimation of CE firms in 3 countries in service industry Study 1: Leverage effect of CE on MVE Over 2000 listed companies in US; Study 2: Multi-year comparison of CE and MVE Netflix and Verizon CE–MVE relationship Mobile telecom; 17 firms in 7 countries • Incremental value relevance of CE estimate beyond financial statement in explaining MVE • Usefulness of CE estimate to find mispriced stock • Determinants of CE–MVE discrepancy Kumar and Shah (2009) Libai, Muller, and Peres (2009) Industry/sample Internet retail and financial service; 5 firms in US Findings High positive correlation between CE and MVE; High R2 when MVE is regressed on CE and the coefficient of the CE is approximately 1 Estimate of CE, with risk factors, improves MVE prediction • Volatility and vulnerability of the firm’s cash flow from customers Paper Gupta, Lehmann, and Stuart (2004) Table 1. Literature review on customer equity and market value of equity. Downloaded by [Tufts University] at 05:59 28 October 2017 Aggregate level; Individual level Aggregate level Aggregate level Individual level Data Aggregate level 4 Y. B. CHOI ET AL. ASIA-PACIFIC JOURNAL OF ACCOUNTING & ECONOMICS 5 Although these studies examined the association between CE and MVE as a metric that proves the appropriateness of CE from the perspective of marketers, however, they did not investigate further the extent of usefulness of CE from the perspective of investors in stock market. Furthermore, they did not analyze a comprehensive set of factors which may cause the discrepancy between CE and MVE. Thus, we will examine how CE is incrementally useful for investors beyond the most representative and easily accessible set of value relevant information, i.e. financial statements, in terms of explaining contemporaneous MVE and predicting future stock returns. In addition, to shed some lights on how to improve the quality of CE estimate itself, we will investigate a comprehensive set of potential determinants of the discrepancy between CE and MVE. Downloaded by [Tufts University] at 05:59 28 October 2017 2.2. Measurement of customer equity This section describes how to measure CE in this study as well as in prior literature. There exist two types of measurement for CE according to data availability and managerial objectives (Kumar and George 2007). First, CE estimation based on internal data (e.g. transactions or managerial judgments) can be viewed as an inductive or disaggregate approach that can assess the individual-level CLVs of all customers and sum them up to get the firm’s CE (e.g. Venkatesan and Kumar 2004). This approach is particularly useful to segment customers in terms of their long term contribution to a firm. By combining the firm’s marketing data (e.g. a time series of price promotion), managers can also examine the implication of a firm’s marketing actions on its CE. However, understanding and reflecting the behaviors of competing companies’ customers are unavailable in this approach. On the other hand, the firm’s CE can be also measured using external data such as panel data, consumer surveys, and company reports. This aggregate method enables outsiders who have limited access to the firm’s proprietary data to estimate the CE values of various firms. The weakness of this approach is that researchers have to estimate CE from a representative customer inferred from aggregated data, which cannot show individual differences among customers. For instance, Gupta, Lehmann, and Stuart (2004) estimated the CLV of an average customer from publicly available financial data and multiplied the average CLV by the number of customers to derive a firm’s CE. Since the objective of this article is to empirically investigate the relationship between CE and a firm’s value of equity based on publicly available data in a multi-firm setting, we will follow the aggregate method, suggested by Gupta, Lehmann, and Stuart (2004), which does not require individual-level customer transactions data. More specifically, an average CLV from a representative customer acquired at time k can be expressed as; CLVk = ∞ ∑ ( ) ARPUi ⋅ AMi − RCi i=k r i−k − ACk (1 + d)i−k (1) where ARPU: average revenue per user; AM: average contribution margin; RC: retention cost; AC: acquisition cost; r: retention rate; d: firm-specific time discount factor; i: time period; k: time period of customer acquisition. Unlike Gupta, Lehmann, and Stuart (2004), we assume that the customer’s revenue (ARPU), contribution margin (AM), retention cost (RC), and acquisition cost (AC) will vary across calendar time rather than each customer’s length of tenure with the firm, since the metrics tend to be highly influenced by the firm’s marketing actions (Yoo and Hanssens 2005).2 Due to the lack of available data with respect to the allocation of marketing costs, we assume that the ratio of retention cost to the total marketing spending is proportional to the firm’s market share in the mobile telecommunications industry. We incorporate both current and future customers to calculate CE, since we evaluate the value of customers from time t to infinity in Equation (2). If there are nk number of customers acquired at time k, then CE evaluated at time t (t ≤ k) will be; 6 Y. B. CHOI ET AL. CEt = ∞ ∑ nk k=t (1 + d)k−t [ ∞ ∑ ) ( ⋅ ARPUi ⋅ AMi − RCi i=k r i−k − ACk (1 + d)i−k ] (2) We calculate the magnitude of CE for 17 mobile telecommunications service providers in 7 countries for 5 years (i.e. from year 2004 to year 2008) based on non-proprietary data. More details regarding the measurement and operationalization of each variable are presented in Section 3. Downloaded by [Tufts University] at 05:59 28 October 2017 2.3. Hypotheses development This section develops our main hypotheses to be tested. First, while several studies report that CE is closely associated with contemporaneous MVE (e.g. Kumar and Shah 2009), these studies do not further investigate whether investors can get incremental value relevant information from the value of customers, measured by CE, beyond easily accessible public information. Since accounting and finance literature concludes that financial statements provide much information which is relevant for investors to evaluate a firm’s equity value, we consider the information contents financial statements present as the benchmark to evaluate the incremental value relevance of our CE estimate. Without such a benchmark, it is not possible to evaluate the incremental usefulness of CE to investors in stock market. Thus, our first empirical hypothesis (in an alternative form) is as follows. Hypothesis 1: Customer equity provides incremental value relevant information beyond financial statements to explain contemporaneous market value of equity. Second, to address a potential concern that MVE, our benchmark for the firms’ intrinsic value of equity when testing Hypothesis 1, is also a noisy measure of true intrinsic value, we also investigate whether our CE estimate can predict future stock returns. If the discrepancy between CE and MVE captures investors’ temporary mispricing of a firm’s equity value rather than measurement errors in CE estimate, our CE estimate will help predict future stock returns over the period during which market value of equity converts to its intrinsic value. As a measure of the divergence between CE and MVE, we propose the ratio of CE to MVE (CM ratio, hereafter). If our CE estimate successfully captures at least a part of temporal divergence between MVE and the firm’s intrinsic value of equity, higher (lower) CM ratio will indicate the undervaluation (overvaluation) of the firm’s value of equity, which leads to positive (negative) abnormal stock returns in the future. Thus, we construct our second alternative hypothesis as follows. Hypothesis 2: The firm’s future stock return is positively associated with CM ratio. Even though our CE estimate can provide value relevant information to explain the firm’s contemporaneous MVE, there still exists a significant discrepancy between CE estimate and MVE across firms. For example, while Gupta, Lehmann, and Stuart (2004) found reasonably similar CE estimate as MVE for three out of five firms, two firms – Amazon and eBay – showed substantial differences between the two values. Thus, we will investigate the potential determinants of the discrepancies between CE and MVE, which will be helpful to improve the quality and so the usefulness of our CE estimate. While using CM ratio as a focal variable to understand the discrepancy between CE and MVE across firms, we categorize various explanatory factors into three groups on the assumption that both CE and MVE are noisy proxies for the firm’s intrinsic value of equity. The first group includes the value relevant factors that investors may incorporate less properly into MVE than our CE estimate. The second group includes the value relevant factors that our CE estimate does not fully incorporate while investors do. We classify the other factors to which CE and MVE would react by a differing extent in unexpected directions as the third group. For each group, we bring related theoretical backgrounds and set up hypotheses that will be tested by our empirical analysis as follows. ASIA-PACIFIC JOURNAL OF ACCOUNTING & ECONOMICS 7 Downloaded by [Tufts University] at 05:59 28 October 2017 2.3.1. Complementary factors to market value of equity Contrary to the efficient market hypothesis, previous research pointed out some factors that make the firm’s market value of equity deviate from its intrinsic value of equity. We list up two major ones with their perspectives. 18.104.22.168. Sales growth rate. Lakonishok, Shleifer, and Vishny (1994) find that firms with higher past sales growth earn lower subsequent stock returns suggesting that investors overvalue the firms’ higher past sales growth. That is, when a firm reports higher sales growth rate in previous year, investors tend to over-estimate the future revenue, and thus the firm’s value of equity. Therefore, the stock price of firm with higher past sales growth rate will deviate upward from the intrinsic value at a point of time. To the contrary, the implication of past sales growth to future revenue will be adequately incorporated into our CE estimate through the reasonable consideration of customer retention rate and average revenue per user. In this case, our CE estimate will not overreact to past sales growth rate in terms of forecasts of future revenue. That is, our CE estimate will be relatively lower than market value of equity for a firm with higher past sales growth rate. Thus, CM ratio may capture investors’ overreaction to past sales growth by presenting a lower value for firms with higher past sales growth rate. Consequently, if CM ratio has an ability to capture investors’ over-evaluation of past sales growth, the firms’ sales growth rate will be negatively associated with CM ratio as in the following alternative hypothesis. Hypothesis 3-1: Sales growth rate is negatively associated with CM ratio. 22.214.171.124. Market-to-book ratio. Lakonishok, Shleifer, and Vishny (1994) also suggest that higher (lower) market-to-book ratio (MTBR) indicates investors’ overvaluation (undervaluation) of equity value relative to the used net assets which are the proxy for the firm’s fundamental on the balance sheet. Thus, they report that firms with relatively high (low) MTBR tend to show lower (higher) subsequent stock returns. However, MTBR is less likely to be associated with high or low CE estimate, since our CE estimate is mainly calculated by using the income statement information (e.g. revenue per user and contribution margin) and other operational reports (e.g. retention rate). As a result, we expect that higher (lower) MTBR leads to smaller (greater) CM ratio. Thus, we set up the following hypothesis. Hypothesis 3-2: Market-to-book-ratio is negatively associated with CM ratio. 2.3.2. Complementary factors to customer equity Although there exist some advantages as described in Section 2.2, our estimation of CE from secondary data is also embedded with some systematic limits to approximate the firm’s intrinsic value of equity, compared with investors’ equity valuation. We propose three factors that our CE estimate may not fully incorporate, creating a systematic discrepancy between CE and MVE as follows: firm size, non-operating income, and the revenue share of focal business within the firm. 126.96.36.199. Firm size. We use the firm-year specific discount rate to estimate CE of each firm by applying Weighted Average Cost of Capital (WACC) based on the Capital Asset Pricing Model (CAPM). However, it is necessary to customize further the discount rate in line with the firm’s other risk characteristics, notably the size of the firm. Penman (2004) discusses the importance of liquidity in explaining the cost of capital, and Amihud and Mendelson (1986) argue that firm size is a good proxy for liquidity. Firm size is further identified as a risk proxy by Fama and French (1995). They suggest that the cost of capital (in turn, discount rate) will be lower (higher) for bigger (smaller) firms. Even though there does not exist any systematic conversion rule from the firm size to the appropriate discount rate, we can hypothesize that CE will be underestimated (overestimated) for bigger (smaller) firms since their true discount rate will be relatively lower (higher). Since the denominator, MVE, may adequately incorporate the firm size factor in investors’ application of discount rate, the CM ratio is expected to be smaller (greater) for bigger (smaller) firms due to the underestimated (overestimated) numerator, CE. Therefore we construct the following hypothesis. 8 Y. B. CHOI ET AL. Downloaded by [Tufts University] at 05:59 28 October 2017 Hypothesis 3-3: Firm size is negatively associated with CM ratio. 188.8.131.52. Non-operating income. Although operating income is the main component of firms’ profitability, non-operating income also consists of firms’ overall profitability. While investors may incorporate the comprehensive income (sum of operating and non-operating income) into MVE, our CE estimate incorporates only the operating income. This is because our CE measure captures only the profits from customers while ignoring the other non-operating income, such as gains from asset disposals and foreign currency translation gains etc. Thus, our CE estimate will be relatively underestimated (overestimated) for the firms with higher (lower) proportion of non-operating income. Schulze, Skiera, and Wiesel (2012) also propose that MVE is sum of customer equity and non-operating assets (minus debt and tax paid) in balance sheet. If there exists a positive relationship between non-operating assets and non-operating income, the theoretical relationship suggested by Schulze, Skiera, and Wiesel (2012) also supports the above hypothesis. Thus, we expect that CM ratio is smaller (greater) for firms with higher (lower) proportion of non-operating income due to the underestimated (overestimated) numerator, CE. Therefore, we construct the following hypothesis. Hypothesis 3-4: The proportion of non-operating income is negatively associated with CM ratio. 184.108.40.206. Share of focal business. Compared to MVE that encompasses the firm’s value of equity in total, our CE estimate is based on the specific performance indicators, such as customer base, average revenue, cost of customer acquisition and retention, and retention rate, from a focal business unit (i.e. mobile telecommunications service in our empirical setting). If a company owns multiple business units, the separate estimation of CE for a focal business unit may neglect to incorporate the synergy effect among multiple business units within a firm. For example, customers exposed to multiple business units of a firm may show higher probability of cross-purchases, resulting in higher customer values. While investors incorporate such a synergy effect properly into MVE, our CE estimate ignores such indirect values of customers. As a result, high share of a focal business unit, the inverse of synergy effect, positively influences the CM ratio as in the following hypothesis. Hypothesis 3-5: The share of a focal business unit is positively associated with CM ratio. 2.3.3. Ambivalent factors We list up the other factors to which CE and MVE would react by a differing extent in unexpected directions as the third group. Thus, each factor’s relationship with the CM ratio remains as an empirical question without directional hypothesis developed a priori. 220.127.116.11. Loyalty program. Customer loyalty increases revenue, enhances retention rate, and reduces cost to acquire and retain customers (Reichheld 1996). Firms in diverse industries, including the mobile telecommunications industry, are implementing loyalty programs to increase customer loyalty and thus the value of their customers. However, loyalty programs under certain circumstances may not be effective (Liu and Yang 2009). Depending on the program’s cost effectiveness (Reinartz and Kumar 2002), the existence of loyalty programs may lead to increase both customer equity and market value of equity by a differing extent. We cannot develop a directional hypothesis regarding this factor since the direction depends on the relative appreciation of a loyalty program by customers and by investors. If customers more (less) appreciate the value of a loyalty program than investors do, the firm’s CM ratio may be positively (negatively) associated with the existence of loyalty programs. Hypothesis 3-6: The existence of loyalty programs is associated with CM ratio without a priori direction. 18.104.22.168. Market share. Market share is positively related to business profitability (Szymanski, Bharadwaj, and Varadarajan 1993), and amplifies the positive impact of business actions (e.g. innovation) on the firm’s market value (Blundell, Griffiths, and Van Reenen 1999). Though it is not ASIA-PACIFIC JOURNAL OF ACCOUNTING & ECONOMICS 9 required to report market share in financial statements, it is one of the important factors that determine the firm’s equity value. Likewise, high share firms tend to have a high level of customer equity due to a bigger customer base, a higher level of marketing investment for customer retention, better reputation, etc. However, the relative magnitude of the reaction of CE estimate and MVE to market share is an empirical question. For example, high share firms may show lower customer satisfaction due to the customer’s higher expectation (Anderson, Fornell, and Lehmann 1994), which may negatively affect the total value of customers. Therefore, we develop a non-directional hypothesis as follows. Downloaded by [Tufts University] at 05:59 28 October 2017 Hypothesis 3-7: Market share is associated with CM ratio without a priori direction. 22.214.171.124. Cross-selling efforts. Like customer loyalty programs, the firm’s cross-selling efforts may positively affect both CE and MVE. Since the firm’s top- and bottom-line performance can be improved by successful cross-selling efforts (e.g. product bundle packages), more cross-selling efforts will increase a firm’s MVE. The firm’s CE also will be improved by successful cross-selling efforts due to the direct increase through higher retention rate (i.e. customers may repeatedly purchase the focal product from the firm due to attractive cross-purchase opportunities) and indirect increase through higher crosspurchases (i.e. customers can contribute more to the firm by purchasing other products provided by the firm). Firms may also be able to acquire more new customers by allocating more marketing resources to customer acquisition given effective cross-selling results from existing customers. In fact, the fundamental equation of customer equity proposed by Blattberg, Getz, and Thomas (2001) includes cross-selling efforts as an important component of CE. However, it is not clear whether our CE estimate appreciates the cross-selling efforts more or less than MVE does. We therefore hypothesize that the CM ratio will be associated with the firm’s cross-selling efforts without any a priori direction. Hypothesis 3-8: Cross-selling efforts are associated with CM ratio without a priori direction. 3. Methodology and data 3.1. Methodology This section describes how to test our hypotheses. First, to empirically examine our first hypothesis, we need to measure the overall value relevance of financial statements as a benchmark and also the incremental value relevance of CE as the other information beyond financial statements. To achieve this purpose, we employ a specific equity valuation model developed by Ohlson (1995; hereafter Ohlson model). This is because Ohlson model expresses the equity value as a linear function of summary information from financial statements and the other information beyond financial statements as described below.3 First, Ohlson (1995) assumes that firms’ equity value is equal to the present value of expected future dividends: ] [∞ ∑ DIVt+n Assumption 1: dividend discount model; Vt = Et ( )n (3) n=1 1 + COEt where Vt is the firm’s equity value at time t, DIVt is the dividends net of capital contributions during period t, and COEt is the firms’ cost of equity capital at time t. Then, Ohlson (1995) imposes the basic structure of accounting system, called as clean surplus relation, into the dividend discount model. This assumption allows future dividends to be expressed in terms of future earnings and book values of equity, which leads to express the firm’s equity value as a function of those accounting information: Assumption 2: clean surplus relation; BVt+1 = BVt + Et − DIVt where BVt is the firm’s book value of equity at time t, Et is the earnings during period t. (4) 10 Y. B. CHOI ET AL. Downloaded by [Tufts University] at 05:59 28 October 2017 Lastly, Ohlson (1995) applies the following assumption about the autoregressive process of abnormal earnings (see the definition below) and the other information beyond abnormal earnings. This assumption specifies the nature of the relation between future earnings and current information within or outside of financial statements: Assumption 3: linear information dynamics; AEt+1 = AEt + t + 1(t+1) (5) t+1 = t + 2(t+1) (6) where AEt is the abnormal earnings which is defined by (Et − COEt * BVt−1), υt is the information other than abnormal earnings, and ε1t and ε2t are the disturbance terms which are unpredictable zero mean variables. By combining these three assumptions, Ohlson (1995, 670) expresses a firm’s equity value as a linear function of reported book value of equity (as summary information from balance sheet) and reported earnings (as summary information from income statement) with the other information, which indicates the value relevant information contents beyond financial statements as follows: Vt = 1 BVt + 2 Et + 3 t (7) Since Ohlson (1995) introduced this simplified but theoretically driven equity valuation model, accounting researchers have thoroughly investigated and widely accepted the Ohlson model due to its systematic link from accounting information to equity value. Thus, many accounting researchers have used the Ohlson model for various research purposes. For example, Dechow, Hutton, and Sloan (1999) investigate the ability of the Ohlson model to explain the cross-sectional distribution of a firm’s market value of equity. Then, Collins, Maydew, and Weiss (1997) and Francis and Schipper (1999) use the Ohlson model to examine the change of value relevance of financial statements for last 40 years in US stock market. Furthermore, Ali and Hwang (2000) also use the Ohlson model to examine the inter-country differences of value relevance of financial statements across various countries. Following previous literature in this area, we evaluate the overall value relevance of financial statements by its ability to explain the cross-sectional distribution of firms’ market value of equity, which is measured by the adjusted R2 of the following regression equation4: MVEt = 0 + 1 BVt + 2 Et + u1t (8) where MVEt is the firm’s market value of equity at time t, and u1t is the error term. While Equation (8) ignores the other information which is not available on financial statements, we explicitly consider our measure for customer equity as the other information beyond financial statements as in following Equation (9)5: MVEt = 0 + 1 BVt + 2 Et + 3 CEt + u2t (9) where CEt is the measure of customer equity at time t. Then, we measure the incremental value relevance of customer equity beyond financial statements by the differences of adjusted R2 between the regression Equations (8) and (9). If our CE estimate provides incremental value relevant information beyond financial statements, we will observe the significant increase of adjusted R2 from the regression Equations (8) to (9), which will support our first hypothesis. To test our second hypothesis of whether our CE estimate can predict future stock returns, we form a quintile portfolio of firms based on the magnitude of CM ratio by each year then examine the mean differences of one-year-ahead stock returns across quintiles. If our CE estimate can indicate the temporal divergence between MVE and intrinsic value of equity, the firms within the quintile with lower (higher) value of CM ratio will show lower (higher) one-year-ahead stock returns. In this analysis, we assume that the market value of equity will convert to its intrinsic value within a year, which is a standard in accounting and finance literature. In addition, stock returns we use in our analysis refer to ASIA-PACIFIC JOURNAL OF ACCOUNTING & ECONOMICS 11 the firm-idiosyncratic cum-dividend stock returns, which are defined as firm-specific cum-dividend stock returns over the market portfolio returns (Fama and French 1993, 2005; Campbell et al. 2001; Luo 2007). Since investors’ mispricing of equity value will be restricted to a specific group of firms rather than the whole stock market, it is desirable to focus on abnormal firm-specific stock returns over the influence of market factors (e.g. global financial risk) to examine the conversion of market value of equity into its intrinsic value. Thus, we measure the abnormal cum-dividend stock returns of individual firms by deducting the market portfolio returns from individual cum-dividend stock returns. Lastly, we investigate the determinants of the discrepancy between the firm’s CE estimate and MVE by regressing the firm-level CM ratio on several firm-specific characteristics as follows. Downloaded by [Tufts University] at 05:59 28 October 2017 CMit = 0 + 1 SGRWit + 2 MTBRit + 3 SIZEit + 4 NOINit +5 STBUit + 6 LOTYit + 7 MSHRit + 8 CRSLit +1 DEBTit + 2 PENEit + 3 IGRWit + 4 YEARt + it (10) According to our Hypotheses 3-1 and 3-2, we include two complementary factors to market value of equity: past sales growth rate (SGRW) and market-to-book ratio (MTBR). We expect that the coefficients of these two variables, ψ1 and ψ2, will be negative. Next, following our Hypotheses 3-3, 3-4, and 3-5, we add three complementary factors to CE estimate, firm size (SIZE), non-operating income (NOIN), and share of the mobile telecommunications business unit (STBU). We posit that the coefficients of SIZE and NOIN (ψ3 and ψ4) are negative while that of STBU (ψ5) is positive. In addition, to test our Hypotheses 3-6, 3-7, and 3-8, we include the other ambivalent factors, such as customer loyalty programs (LOTY), market share (MSHR), and cross-selling efforts (CRSL), into the regression equation. For these variables, we do not have any ex-ante expectation of the directional association with the CM ratio. Lastly, to control for the well-known leverage effect on the discrepancy between CE and MVE (Schulze, Skiera, and Wiesel 2012), we include debt-to-asset ratio as a control variable. In addition, to control for the potential country- and time-specific factors, we include two country-level variables, such as the penetration rate of mobile telecommunications (PENE) and market index growth rate (IGRW), as well as a year dummy variable (YEAR).6 3.2. Sample selection We empirically test research hypotheses proposed in the previous section with a data-set collected from the mobile telecommunications industry. We selected this industry for the current study due to several reasons. First, since many firms in the mobile telecommunications industry regularly unveil critical information (e.g. the number of customers, churn rate, and average revenue per user) necessary to calculate CE, this industry fits well to our current investigation even though we cannot access to any internal proprietary data (e.g. individual transactions data). Second, most firms in the mobile telecommunications industry are listed on the stock market so that we can easily obtain financial statement information required to conduct our empirical analyses. Moreover, since the mobile telecommunications business unit is a major strategic business unit for most of sample firms,7 we can more directly compare CE with MVE. Third, since the nature of business in the mobile telecommunications industry is relatively similar in most of countries, we can increase the degree of freedom by pooling the data-set from different countries. Lastly, by focusing on one industry, we can control for any systematic difference across various industries in testing our research hypotheses. We selected 17 firms in the mobile telecommunications industry of 7 countries mainly from Asia Pacific region (Australia, Canada, China, Japan, South Korea, United States) and Belgium as our sample.8 We first obtained the number of customers of these firms up to 12 years before 2008 given data availability. This information will be used to estimate the future number of customers, a critical component to calculate our CE estimate for each firm. We then obtained annual financial information data for five years from 2004 to 2008 to conduct our empirical analyses.9 All the data were obtained Firm Hutchison Australia Telstra Belgacom Mobistar Bell Rogers Telus China Mobile China Unicom KDDI NTT Docomo KTF LGT SKT AT&T Sprint Nextel Verizon 30,000 49,073 6626 3809 9269 49,335 30,427 49,322 30.8 54.6 14.4 8.2 23.0 77.0 40.3 72.1 9.3 4.8 3.7 6.5 7.9 6.1 457.3 133.4 9075 2802 2065 3677 5198 3823 60,439 12,261 41.09 54.33 21.65 11.65 33.49 102.01 51.81 93.58 11.29 9.37 5.49 7.11 10.43 7.38 649.72 147.58 α 2.75 −1.98 −1.64 −2.04 −1.90 −1.66 −3.72 −3.99 −2.59 −1.78 −0.63 −2.04 −2.31 −2.46 −2.68 −4.08 −4.06 β −5.56 0.25 0.45 0.25 0.24 0.20 0.38 0.51 0.29 0.27 0.05 0.22 0.35 0.28 0.32 0.40 0.89 γ 0.50 0.100 0.025 0.756 0.427 0.148 0.128 0.297 0.072 0.040 0.775 0.193 0.050 0.099 0.082 0.366 0.089 MAPE 0.417 62.2 62.7 37.0 33.4 41.6 50.0 59.0 50.0 55.1 53.3 48.0 44.3 47.8 53.9 11.7 7.2 ARPU (USD) 85.9 23.2 27.7 26.9 25.6 44.5 24.4 18.2 36.2 35.9 39.2 39.0 40.8 38.8 34.4 47.4 31.5 Margin (%) 0.5 2.6 4.1 3.4 1.2 3.1 2.0 1.0 2.0 2.6 0.0 1.3 2.6 2.6 1.7 1.0 0.5 Retention cost (USD) 2.6 380.4 309.2 119.7 177.3 239.1 146.0 191.0 213.0 150.0 41.3 145.3 236.5 155.7 201.7 19.2 33.4 Acquisition cost (USD) 1.207.7 1.15 0.84 2.97 3.16 1.99 2.12 2.59 1.34 1.73 1.40 2.03 1.52 1.89 1.42 1.42 2.33 Churn rate (%) 1.57 33.8 109.1 2.1 0.9 20.8 123.4 −13.0 161.8 15.7 12.2 13.2 8.1 12.8 10.7 692.1 16.3 CE (USD billion) 1.7 Notes: This table contains the sales and the number of customers of mobile telecommunications business unit of the firm in 2008. α means the firm’s expected number of customers (in millions) as time approaches infinity. MAPE is an average mean absolute percentage error between March 2004 and March 2008. ARPU (average revenue per user, monthly) = sales/the number of customers/12, margin (monthly) = (total revenue − (operating expenses − marketing expenses))/total revenue, retention cost (monthly) = marketing expenses * market share/the number of customer/12, acquisition cost (annual) = marketing expenses * (1 − market share)/the number of new customers, and churn rate (monthly) are computed based on each firm’s annual report. United States South Korea Japan China Canada Belgium Country Australia Number of Customers (million) 2.0 Sales (USD million) 2310 Table 2. Descriptive data and estimated parameter of sample firms. Downloaded by [Tufts University] at 05:59 28 October 2017 12 Y. B. CHOI ET AL. ASIA-PACIFIC JOURNAL OF ACCOUNTING & ECONOMICS 13 from secondary data sources such as annual reports, company websites, and databases such as Data Stream and fnguide.com. We also obtained data from Journal of Mobile Communications, marketresearch.com, and celluar-news.com to complement our data. Finally, our sample consists of 85 firm/ year observations from distinct 17 firms. Table 2 presents the description of our sample companies as well as their CE parameter estimates. 3.3. Measurement of variables Downloaded by [Tufts University] at 05:59 28 October 2017 This section describes how to measure some key components of our CE estimate and other variables for our empirical analyses with their distributions. 3.3.1. The number of customers Since our measurement of CE estimate includes the value of future customers, we need to estimate the number of customers in the future. As the customer growth of each firm in our sample shows a very similar pattern as a conventional diffusion theory, we follow the diffusion model applied in Gupta, Lehmann, and Stuart (2004). The number of customers (N) at time t is modeled as: Nt = 1 + exp (− − t) (11) where α, β, and γ are parameters to be estimated from the past data. Therefore, considering the retention rate r, the number of newly acquired customers (n) at time t can be calculated as follows10: nt = Nt − rNt−1 (12) The parameter estimation results for customer growth are provided in Table 2. Mean absolute percentage error (MAPE) shows that our model reflects the actual growth pattern quite closely. Although the sum of ultimate market potential (α) within some countries is greater than the current population of those countries (e.g. 66.8 million vs. 48.6 million for Korea), it seems reasonable considering the high mobile penetration rate in several countries. For instance, the penetration rates of most European countries are higher than 100%, and those of Korea, Australia, Hungary, Italy, and Hong Kong are 94, 110, 120, 156, and 163% respectively (Wireless federation and China wireless news, April, 2009). 3.3.2. Average contribution margin It is necessary to separate marketing costs from the other operating expenses to estimate customer equity as proposed in Equation (1). However this task is challenging for some firms since detailed marketing costs are not required items to be posted in the financial reports for most of countries. For instance, US firms disclose only selling, general and administrative (SG&A) costs and provide limited information about advertising costs. Due to this lack of disclosure of detailed marketing costs in many countries, we define marketing costs as the sum of advertising, promotion, and other selling costs. In particular, we apply the industry average ratio of marketing costs to SG&A among Canadian firms to the firms in the US and Japan, where firms are not required to separately report the marketing costs. In the case of firms that only report the consolidated expenses with other business units (e.g. cable TV business), we apply the revenue share of mobile telecommunications business to total revenue in order to obtain the cost information only for the focal business unit. We do not include acquisition costs (AC) in Equation (13) since AC is separately incorporated in CLV and CE calculation in Equations (1) and (2), respectively. Therefore, the average contribution margin (AM) in Equation (13) is before acquisition costs. As a result, the average contribution margin (AM) at time t is obtained from the following equation. AMt = Total Salest − Total Operating Costst + Marketing Costst Total Salest (13) 14 Y. B. CHOI ET AL. Table 3. Variable operationalization, data sources and descriptive statistics. Variables CE MVE CM BV E SGRW MTBR SIZE NOIN STBU LOTY MSHR Downloaded by [Tufts University] at 05:59 28 October 2017 CRSL DEBT PENE IGRW Measurement See Section 3.1 Stock price × number of shares CE/MVE Book value of common equity Net income Mobile sales growth rate MVE/book value of common equity Log of total assets (USD) Operating margin – net margin Mobile business sales/total sales Yes = 1/No = 0 Firm’s share of mobile market based on sales The number of bundled services (unit) Debt/asset ratio The penetration rate of mobile telecommunication Local stock market index growth rate Sources Annual Reports Data Stream Data Stream Data Stream Annual Reports Data Stream Annual Reports Annual Reports Annual Reports Annual Reports Annual Reports Annual Reports Annual Reports Annual Reports, KISDI, Journal of Mobile Communication Data stream Mean 1.746 1.013 0.992 0.408 0.064 0.106 2.415 10.315 0.098 0.708 0.412 0.344 Std. dev. 1.745 0.764 4.050 0.237 0.075 0.119 1.577 0.578 0.102 0.303 0.495 0.144 3.235 0.548 0.738 1.212 0.175 0.202 0.127 0.152 Notes: This table reports the variable operationalization, data sources and descriptive statistics of main variables used in our tests. We used pooled data of 17 firms from 2004 to 2008 (N = 85). ‘CE’ is customer equity measured annually, ‘MVE’ is firm’s market value of equity at the end of March of each year, ‘BV’ is book value of common equity, and ‘E’ is net income. ‘CE’, ‘MVE’, ‘BV’ and ‘E’ are converted into US dollars and scaled by total assets. The values of ‘MVE’, ‘BV’ and ‘E’ are adjusted by multiplying the firm’s ratio of mobile sales to total sales. ‘CM’ is CM ratio which is equal to CE divided by MVE. ‘SGRW’ is past sales growth rate, ‘MTBR’ is marketto-book ratio, ‘SIZE’ is firm size in terms of total assets, ‘STBU’ is share of the mobile telecommunications business unit, ‘LOTY’ is a dummy variable for the existence of loyalty programs, ‘MSHR’ is market share of the firm in local mobile telecommunications market in terms of sales, and ‘CRSL’ is cross-selling efforts of the firm, measured by the number of bundled services. For the control variables, ‘DEBT’ is the total debt divided by total asset, ‘PENE’ is penetration rate of mobile telecommunication, and ‘IGRW’ is stock market index growth rate in each local market. 3.3.3. Retention rate The customer churn rate is regularly reported in the mobile telecommunications industry. Most of firms include annual or quarterly churn rate in their annual reports because the number of subscribers is one of the most important metrics directly related to the firm’s profitability in this industry. Moreover, as customers are allowed to switch service providers without changing their existing phone numbers in most countries, effective management of the churn rate has become a major concern of the industry. We use one minus annual average churn rate as our estimate for retention rate. 3.3.4. Distribution of variables By using the estimated components consisting of CE based on Equation (1), we estimated annual values of CE for our 17 sample firms from 2004 to 2008. Since we were not able to observe any pattern in margins, acquisition costs, retention costs, and retention rates, we decided to use an average of previous 2 years for each metric to estimate the current CE. To estimate CE, we use firm-year specific discount rate by applying Weighted Average Cost of Capital (WACC) based on the Capital Asset Pricing Model (CAPM). More specifically, to calculate the cost of equity capital, we have regressed at least 12 prior monthly stock returns against corresponding country’s stock market index up to 60 months preceding the current period to estimate market beta. The resulting market beta estimate is used in conjunction with realized treasury-bill rates in each country as risk-free rates and 7% as the market risk premium. Then, we estimate the cost of debt capital by the ratio of interest expenses relative to total debt on the financial statements. Lastly, we calculate WACC by the weighted average of cost of equity and cost of debt capital, where the respective weight is market value of equity or book value of debt relative to the sum of those two values. Each firm’s CE is scale-adjusted by the size of total assets resulting in a mean value of 1.746. Table 3 summarizes how to measure main variables for our empirical analyses with corresponding data sources, and presents the basic statistics such as mean and standard deviation of each variable. Downloaded by [Tufts University] at 05:59 28 October 2017 ASIA-PACIFIC JOURNAL OF ACCOUNTING & ECONOMICS 15 Since annual reports and other financial statements are posted around March in most countries, we use data from the end of March in one year to the end of March in next year to measure stock market related variables. We obtained monthly market value of equity (MVE) by multiplying monthly stock prices by monthly total number of outstanding shares of a firm (i.e. market capitalization). We then picked a firm’s MVE at the end of March of each year. We defined book value of equity (BV) as the book value of common equity, and annual earnings (E) as net income. Note that market value of equity, book value of equity, and earnings are adjusted by the revenue share of mobile telecommunications business within a firm to fairly evaluate multi-business firms. As in the CE variable, the three variables are also scale-adjusted by total assets. As explained earlier, CM ratio (CM) was obtained by dividing each firm’s CE by MVE. The average CM across all observations is 0.992 with a standard deviation of 4.050, which shows diverse distribution. We obtained past sales growth rate (SGRW) of the mobile telecommunications business from sample firms’ annual reports (mean = 10.6%). Market-to-book-ratio (MTBR) was calculated by MVE/ BV (mean = 2.415). The firm size (SIZE) variable was obtained from natural logarithm of total assets to minimize a scale issue (mean = 10.315), and non-operating income (NOIN) from deducting net margin from operating margin (mean = 0.098). The revenue share of mobile telecommunications business (STBU) was calculated by dividing revenue of mobile telecommunications business by the firm’s total revenue (mean = 0.708). Three ambivalent factors were also obtained from annual reports of each firm. The existence of loyalty programs (LOTY) is dummy coded with one if a firm provides any program, zero otherwise (mean = 0.412). Market share (MSHR) stands for the firm’s revenue share in a corresponding country’s mobile telecommunications industry (mean = 0.344). For measuring cross-selling efforts (CRSL), we counted the number of bundled services (e.g. cable TV) with mobile telecommunications (mean = 3.235). As a firm-level control variable, Debt-to-Asset ratio (DEBT) has the mean of 0.548. Two country-specific control variables were collected from various sources (e.g. Journal of Mobile Communication). We measured the penetration rate of mobile telecommunications (PENE) and the growth rate of stock market index (IGRW) for each country. 4. Empirical results This section documents our main empirical results.11 First, we test our first hypothesis of incremental value relevance of CE estimate by comparing the adjusted R2 between the regression Equations (8) and (9). As reported in Table 4, the explanatory power (measured by adjusted R2) of two summary information from financial statements on contemporaneous MVE is 0.638. However, when adding our CE estimate into the regression, it increases into 0.729. This result indicates that our CE estimate provides incremental value relevant information beyond financial statements and so investors may be able to Table 4. Incremental value relevance of customer equity. MVEt = α0 + α1 BVt + α2 Et + u1t Model 1 Model 2 Variables Coefficients BV −0.075 E 8.305*** MVEt = β0 + β1 BVt + β2 Et + β3 CEt + u2t Std. Error 0.260 0.819 Variables BV E CE Std. Error 0.232 1.184 0.044 Coefficients 0.248 3.219*** 0.238*** Adj. R2 0.638*** Adj. R2 Changes 0.729*** 0.091*** Notes: N = 85. This table reports the results of regression analyses for Model 1 and 2. See the note of Table 3 for the definition of variables. ***p < .01. 16 Y. B. CHOI ET AL. Table 5. Difference of abnormal stock returns among groups based on CM ratio. Between groups (ANOVA) Multiple comparison (LSD) Group 1 2 3 4 5 Total N 15 20 16 19 15 85 Mean 0.075 −0.049 0.044 −0.137 −0.210 −0.058 Mean difference from group 5 0.285 0.161 0.254 0.073 F 3.401 Sig. 0.013** Std. error 0.094 0.088 0.092 0.089 Sig. 0.003*** 0.071* 0.007*** 0.417 Downloaded by [Tufts University] at 05:59 28 October 2017 Notes: N = 85. This table reports the mean of one-year-ahead abnormal stock returns for each quintile portfolio based on CM ratio. We use pooled data of annual stock returns. Each group is divided in accordance with rank of CM ratio by year. The group with the highest (lowest) CM ratio is named as Group 1 (Group 5). *p < 0.10; **p < 0.05; ***p < 0.01. Table 6. Determinants of CM ratio. Adj. R2 Model (Constant) SGRW MTBR SIZE NOIN STBU LOTY MSHR CRSL DEBT PENE IGRW YEAR Coefficients 27.367 −0.340 0.363 −2.335 −26.541 0.224 −2.998 7.855 1.149 −6.625 −0.231 −4.589 −0.307 Std. error 8.855 2.448 0.205 0.786 3.626 0.372 1.155 2.929 0.288 2.365 2.061 2.305 0.275 Sig. 0.003*** 0.890 0.082* 0.004*** 0.000*** 0.549 0.011** 0.009*** 0.000*** 0.007*** 0.911 0.050** 0.268 0.711*** Hypothesized direction H3-1 (−) H3-2 (−) H3-3 (−) H3-4 (−) H3-5 (+) H3-6 (+/−) H3-7 (+/−) H3-8 (+/−) Notes: N = 85. This table reports the result of pooled regressions of the following equation. See the note of Table 3 for the definition of variables. *p < 0.10; **p < 0.05; ***p < 0.01. CMit = 0 + 1 SGRWit + 2 MTBRit + 3 SIZEit + 4 NOINit +5 STBUit + 6 LOTYit + 7 MSHRit + 8 CRSLit +1 DEBTit + 2 PENEit + 3 IGRWit + 4 YEARt + it get more value relevant information by considering the value of customers. Furthermore, note that the coefficient of earnings decreases from 8.305 to 3.219 before and after adding our CE estimate respectively, while the coefficient of book value of equity remain insignificant for both regressions. This result implies that our CE estimate may subsume a part of information contents of reported earnings to explain contemporaneous market value of equity, which is natural given that our CE estimate is measured based on information from income statement. Second, we test our second hypothesis on whether CE estimate can predict future stock returns by examining the mean differences of one-year-ahead stock returns across quintiles, which are formed on the basis of the magnitude of CM ratio. As reported in the third column of Table 5, the firms within the quintile with the lowest CM ratio (Group 5 in Table 5) show the lowest mean of one-year-ahead abnormal stock returns (−21.0%) while the other groups present stock returns between −13.7 and 7.5%. The results of ANOVA show that the differences of one-year-ahead abnormal stock returns among five groups are significant (sig. = 0.013). The last column of Table 5 reports the result of least-significant difference (LSD) tests showing that the differences of one-year-ahead abnormal stock returns between Group 5 and the other three groups Downloaded by [Tufts University] at 05:59 28 October 2017 ASIA-PACIFIC JOURNAL OF ACCOUNTING & ECONOMICS 17 (except Group 4) are statistically significant (p < 0.10). Although we cannot observe the monotonic pattern of abnormal stock returns proportional to the magnitude of CM ratio as our second hypothesis expects, it is clear that our CE estimate may capture at least the overstated market value of equity of the firms within the extreme quintile with the lowest CM ratio, whose one-year-ahead stock returns are significantly lower. Thus, we conclude that our CE estimate is useful for investors to identify overvalued stocks by capturing the divergence between overstated market value of equity and its intrinsic value. Lastly, Table 6 documents the results of regression of CM ratios on various firm characteristics to identify the determinants of the discrepancy between CE estimate and MVE. First, we find that the coefficients of firm size (SIZE) and non-operating income (NOIN) are significantly negative which is consistent with our Hypotheses 3-3 and 3-4, respectively. This result indicates that our procedure to estimate CE may ignore these two pieces of important value relevant information, and so investors can improve the CE estimation procedure by considering these factors. Second, while past sales growth rate (SGRW) is negatively associated with CM ratio and share of the mobile telecommunications business unit (STBU) is positively associated with CM ratio, as our Hypotheses 3-1 and 3-5 expect, the associations are not statistically significant. One of possible reasons for this result is the weak testing power due to the small number of sample (85 observations). Third, regarding ambivalent factors, the existence of customer loyalty programs (LOTY) shows significantly negative coefficient, while market share (MSHR) and cross-selling efforts (CRSL) are significantly positively associated with CM ratio. This result suggests that while investors may more appreciate the value of a loyalty program than customers do, our CE estimate reacts further to the effect of market share on future profitability and to the higher retention rate and more purchasing by successful cross-selling efforts than investors do. In addition, while the coefficient of market-to-book ratio (MTBR) is positive, which is contradictory to our Hypothesis 3-2, such an association is found to be marginally significant. Taken together, we identify firm size, non-operating income, the existence of customer loyalty programs, market share, and cross-selling efforts as the potential determinants of the discrepancy between our customer equity estimate and contemporaneous market value of equity. We believe that this finding will be helpful for researchers and practitioners to improve the quality of CE estimation procedure and to further understand the appreciation of marketing actions by investors and customers. 5. Discussion and conclusion As Levitt (1960) suggested in his seminal paper, the firm needs to buy customers and sell satisfaction to generate repeating purchases from its customers, which stressed out that the firm’s operational focus should lie in customers not in products or services. Following this perspective, many firms have spent lots of marketing effort to provide superior value for customers and in turn to get the maximum value of customers. Furthermore, marketing investments should be linked with the firm value in theory, but this task is challenging in practice due to a blurry causal link between the firm’s marketing actions and the firm value in stock market. Many firms’ shareholders and CEOs have concerns about the long-term financial contribution of marketing investment beyond its role on conventional marketing metrics such as customer satisfaction and market share. Customer equity has been proposed as a new metric to link between marketing and other functions (e.g. finance and accounting) within an organization by quantifying the value of customers. This paper explores the role of CE as the source of value relevant information for investors in the stock market. We show that CE, a proxy of the firm’s equity value from marketing perspectives, provides investors with incremental value relevant information beyond financial statements to explain contemporaneous market value of equity. In other words, CE can be considered by investors as a significant piece of other information incorporating the value of (arguably most important) intangible asset of firms, the value of customers, which is not captured by conventional financial reports. Thus, marketing can be connected to other managerial fields within an organization through CE. For example, the impact of marketing efforts on firm value can be measured by investigating how CE is moving according to different levels of marketing investments. Downloaded by [Tufts University] at 05:59 28 October 2017 18 Y. B. CHOI ET AL. We also find that CE can indicate the investors’ temporal mispricing of the firm’s equity value. This result implies that investors in the stock market may be biased in incorporating the impact of marketing actions on firm value but that they can reduce such a bias by referring to CE. Therefore, CE will be able to contribute to the enhancement of stock market efficiency, which is a necessary condition for the efficient allocation of limited resources in the economy. We also document that several key determinants account for the discrepancy between CE and MVE, of which the extant literature fails to provide a deeper investigation. CE estimate should be adjusted by firm-level characteristics such as firm size and non-operating income to more accurately approximate the firm’s market value of equity (or intrinsic value of equity). More specifically, since a firm’s non-operating income is not incorporated in the conventional calculation of CE, a metric based on customer values from a fundamental business operation, our measurement of CE tends to underestimate the value of equity for firms with a relatively high level of non-operating income. In addition, the measurement of CE should consider firm size to apply more appropriate firm-specific discount rate. On the other hand, there are other significant factors that affect the relationship between CE and MVE such as loyalty programs, market share, and cross-selling efforts. First, the service providers’ loyalty programs in the mobile telecommunications industry are less effective in customer acquisition and retention than investors expect. Though further research is required, this result proposes that loyalty programs may have signaling and spillover effects as advertising on market value of equity (Joshi and Hanssens 2010). Second, the financial market seems to less appreciate high market share than customers do. Third, the firm’s cross-selling efforts by providing other related services such as the broadband Internet may increase the synergy among related business more than investors expect. These findings also improve our understanding of the CE–MVE relationship. As a closing remark, there are also several limitations of this study that open areas for further research. First, although we aimed to extend the literature on the CE–MVE relationship, our results are based only on one industry under the lost-for-good business circumstance (Dwyer 1997). To generalize our results, similar research on various industries should confirm our findings. For example, investigating industries under non-contractual setting (e.g. personal computers) will provide an interesting comparison to our findings. Second, a temporal change of the CE–MVE relationship within a firm would be worthwhile to examine. Though we could not investigate such issues due to the limited number of data points in the current paper, longer time-series will enable researchers to study a dynamic nature of the CE–MVE relationship. For example, it will be interesting to examine how the changes in a firm’s business portfolio affect the CE–MVE relationship. Third, competition and interaction among rivalry firms would play an important role on the CE–MVE relationship as well. Models that explicitly accommodate the competitive nature of firms will be necessary to understand this phenomenon. Finally, one can obtain a more complete picture of the CE–MVE relationship if a firm’s internal data (e.g. customer-specific transactions history) complement our proposed methodology in calculating customer equity. For example, decomposing acquisition and retention costs based on the firm’s accounting data (if any) will provide more accurate CE estimate. We leave all these issues for future research. Notes 1. See Srinivasan and Hanssens (2009) for a comprehensive review on this issue. 2. Gupta, Lehmann, and Stuart (2004) also used calendar-time-based metrics in their empirical investigation due to the lack of individual-level data. However, this assumption may not hold for business-to-business industries where the tenure of each customer significantly affects the metrics (i.e. ARPU, AM, RC, and AC). 3. See Ohlson (1995) for the details of how to theoretically derive Ohlson model. 4. All regression variables are scaled by the book value of total assets to address scale issue. 5. The market value of equity, book value of equity, and earnings in our studies are adjusted values which are multiplied by the ratio of mobile telecommunications business sales to total sales of the firm. This is because our customer equity estimate is based on the sales of mobile telecommunications business even though the firm may have multiple business units. This adjustment will enhance the comparability across variables. ASIA-PACIFIC JOURNAL OF ACCOUNTING & ECONOMICS 19 6. See Section 3.3 for the details of how to measure the variables. 7. An average revenue share of mobile telecommunications business is 70% across all sample firms. 8. We rule out the firms operating in multiple countries since it is hard to figure out the contribution of each country’s market to the items in consolidated financial statements. 9. We measure the market value of equity at the end of March of each year to allow investors to incorporate the information contents of annual reports which are required to be announced by the three months after the December fiscal-year end in most countries. 10. We use the same approach of Schulze, Skiera, and Wiesel (2012) to calculate the future number of customers in this study. Equations (11) and (12) in our paper are identical with Equations (4) and (5) in Schulze, Skiera, and Wiesel (2012), respectively. 11. We winsorize all variables at the top and bottom 3% of pooled distributions of each variable to minimize the bias from extreme values in all of our empirical analyses. Disclosure statement Downloaded by [Tufts University] at 05:59 28 October 2017 No potential conflict of interest was reported by the authors. Funding This study was supported (partially) by Korea University Business School Research Grant. References Ali, A., and L. Hwang. 2000. “Country-specific Factors Related to Financial Reporting and the Value Relevance of Accounting Data.” Journal of Accounting Research 38 (1): 1–21. Amihud, Y., and H. Mendelson. 1986. “Asset Pricing and the Bid-ask Spread.” Journal of Financial Economics 17 (2): 223–249. Anderson, E., C. Fornell, and D. R. Lehmann. 1994. “Customer Satisfaction, Market Share, and Profitability: Findings from Sweden.” Journal of Marketing 58 (3): 53–66. Anderson, E., C. Fornell, and S. K. Mazvancheryl. 2004. “Customer Satisfaction and Shareholder Value.” Journal of Marketing 68 (4): 172–185. Blattberg, R., and J. Deighton. 1996. “Managing Marketing by the Customer Equity Test.” Harvard Business Review 75 (July–August): 136–144. Blattberg, R., G. Getz, and J. S. Thomas. 2001. Customer Equity: Building and Managing Relationships as Valuable Assets. Boston, MA: Harvard Business School Press. Blundell, R., R. Griffiths, and J. Van Reenen. 1999. “Market Share, Market Value and Innovation in a Panel of British Manufacturing Firms.” Review of Economic Studies 66 (3): 529–554. Campbell, J., M. Lettau, B. Malkiel, and Y. Xu. 2001. “Have Individual Stocks Become More Volatile? An Empirical Exploration of Idiosyncratic Risk.” Journal of Finance 56 (1): 1–43. Collins, D. W., E. L. Maydew, and I. S. Weiss. 1997. “Changes in the Value Relevance of Earnings and Book Values over the past Forty Years.” Journal of Accounting and Economics 24 (1): 39–67. Dechow, P., A. Hutton, and R. Sloan. 1999. “An Empirical Assessment of the Residual Income Valuation Model.” Journal of Accounting and Economics 26 (1–3): 1–34. Dwyer, F. R. 1997. “Customer Lifetime Valuation to Support Marketing Decision Making.” Journal of Interactive Marketing 11 (4): 6–13. Easton, P., G. Taylor, P. Shroff, and T. Sougiannis. 2002. “Using Forecasts of Earnings to Simultaneously Estimate Growth and the Rate of Return on Equity Investments.” Journal of Accounting Research 40 (June): 657–676. Fama, E. F., and K. R. French. 1993. “Common Risk Factors in the Returns on Stocks and Bonds.” Journal of Financial Economics 33 (1): 3–56. Fama, E. F., and K. R. French. 1995. “Size and Book-to-Market Factors in Earnings and Returns.” The Journal of Finance 50 (1): 131–155. Fama, E. F., and K. R. French. 2005. “Financing Decisions: Who Issues Stock?” Journal of Financial Economics 76 (3): 549–582. Fornell, C., S. Mithas, F. Morgeson, and M. S. Krishnan. 2006. “Customer Satisfaction and Stock Prices: High Returns, Low Risk.” Journal of Marketing 70 (1): 3–14. Francis, J., and K. Schipper. 1999. “Have Financial Statements Lost Their Relevance?” Journal of Accounting Research 37 (2): 319–352. Gupta, S., and D. R. Lehmann. 2005. Managing Customers as Investments: The Strategic Value of Customers in the Long Run. Upper Saddle River, NJ: Wharton School Publishing. Downloaded by [Tufts University] at 05:59 28 October 2017 20 Y. B. CHOI ET AL. Gupta, S., D. R. Lehmann, and J. A. Stuart. 2004. “Valuing Customers.” Journal of Marketing Research 41 (1): 7–18. Ittner, C. D., and D. F. Larcker. 1998. “Are Nonfinancial Measures Leading Indicators of Financial Performance? An Analysis of Customer Satisfaction.” Journal of Accounting Research 36 (Supp): 1–35. Joshi, A. M., and D. M. Hanssens. 2010. “The Direct and Indirect Effects of Advertising Spending on Firm Value.” Journal of Marketing 74 (1): 20–33. Kumar, V., and M. George. 2007. “Measuring and Maximizing Customer Equity: A Critical Analysis.” Journal of the Academy of Marketing Science 35 (2): 157–171. Kumar, V., and D. Shah. 2009. “Expanding the Role of Marketing: From Customer Equity to Market Capitalization.” Journal of Marketing 73 (6): 119–136. Lakonishok, J., A. Shleifer, and R. Vishny. 1994. “Contrarian Investment, Extrapolation, and Risk.” Journal of Finance 49 (5): 1541–1578. Lee, C., J. Myers, and B. Swaminathan. 1999. “What is the Intrinsic Value of the Dow?” Journal of Finance 54 (September): 1693–1741. Levitt, T. 1960. “Marketing Myopia.” Harvard Business Review 38 (July–August): 45–56. Libai, B., E. Muller, and R. Peres. 2009. “The Diffusion of Services.” Journal of Marketing Research 46 (2): 163–175. Liu, Y., and R. Yang. 2009. “Competing Loyalty Programs: Impact of Market Saturation, Market Share, and Category Expandability.” Journal of Marketing 73 (1): 93–108. Luo, X. 2007. “Consumer Negative Voice and Firm-Idiosyncratic Stock Returns.” Journal of Marketing 71 (3): 75–88. Ohlson, J. 1995. “Earnings, Book Values, and Dividends in Equity Valuation.” Contemporary Accounting Research 11 (2): 661–687. Penman, S. H. 2004. Financial Statement Analysis and Security Valuation. 2nd ed. New York: McGraw-Hill. Rao, V. R., M. K. Agarwal, and D. Dahlhoff. 2004. “How is Manifest Branding Strategy Related to the Intangible Value of a Corporation?” Journal of Marketing 68 (4): 126–141. Reichheld, F. F. 1996. The Loyalty Effect: The Hidden Force behind Growth, Profits, and Lasting Value. Boston, MA: Harvard Business School Press. Reinartz, W. J., and V. Kumar. 2000. “On the Profitability of Long-life Customers in a Noncontractual Setting: An Empirical Investigation and Implications for Marketing.” Journal of Marketing 64 (4): 17–35. Reinartz, W. J., and V. Kumar. 2002. “The Mismanagement of Customer Loyalty.” Harvard Business Review 80 (July): 86–94. Rust, R. T., K. N. Lemon, and V. A. Zeithaml. 2004. “Return on Marketing: Using Customer Equity to Focus Marketing Strategy.” Journal of Marketing 68 (1): 109–127. Schulze, C., B. Skiera, and T. Wiesel. 2012. “Linking Customer and Financial Metrics to Shareholder Value: The Leverage Effect in Customer-Based Valuation.” Journal of Marketing 76 (2): 17–32. Sorescu, A., V. Shankar, and T. Kushwaha. 2007. “New Product Preannouncements and Shareholder Value: Don’t Make Promises You Can’t Keep.” Journal of Marketing Research 44 (3): 468–489. Srinivasan, S., and D. M. Hanssens. 2009. “Marketing and Firm Value: Metrics, Methods, Findings, and Future Directions.” Journal of Marketing Research 46 (3): 293–312. Srivastava, R. K., T. A. Shervani, and L. Fahey. 1998. “Market-based Assets and Shareholder Value: A Framework for Analysis.” Journal of Marketing 62 (1): 2–18. Szymanski, D. M., S. G. Bharadwaj, and P. R. Varadarajan. 1993. “An Analysis of the Market Share-profitability Relationship.” Journal of Marketing 57 (3): 1–18. Venkatesan, R., and V. Kumar. 2004. “A Customer Lifetime Value Framework for Customer Selection and Optimal Resource Allocation Strategy.” Journal of Marketing 68 (4): 106–125. Wiesel, T., B. Skiera, and J. Villanueva. 2008. “Customer Equity: An Integral Part of Financial Reporting.” Journal of Marketing 72 (2): 1–14. Yoo, S., and D. M. Hanssens. 2005. “Modeling the Sales and Customer Equity Effects of the Marketing Mix.” Working Paper. Los Angeles, CA: UCLA Anderson School of Management.