قیمت گذاری ریسک کیفیت خدمات الکترونیکی در خدمات مالی

E-service quality is crucial for differentiating e-commerce offers and gaining competitive advantage. E-service quality risk is the risk that a firm’s e-service quality will drop, or improve, relative to competitors. There is evidence that benchmark ratings of e-service quality that are published regularly by third-parties can impact the market value of rated firms. Firms therefore continue investing in IT-related determinants of e-service quality. However, they do so without knowing: (1) the cost or return associated with a unit relative deterioration, or improvement in e-service quality ratings, and (2) how this cost or return may vary across firms. To answer these questions, we adapt a well-established financial risk pricing approach for the case of pricing a single idiosyncratic IT investment risk, where an event study is used to generate the market data needed to price risk (Thompson 1985). We then apply the approach with Keynote’s bi-annual e-service quality ratings for firms in six financial services sectors. We find that firms’ sensitivity to e-service quality risk depends primarily on the sector to which they belong, and also on their size and growth potential. Our results suggest a cap on the amount that different firms ought to spend to achieve a unit improvement in relative e-service quality ratings. The risk pricing approach presented can be applied for other important IT investment risks, and the risk pricing information it yields may open up new ways to approach fundamental IT investment problems.

Announcements of e-service launch initiatives have been seen to benefit the market value of the launching firms (Subramani and Walden, 2001, Geyskens et al., 2002, Cheng et al., 2007 and Lin et al., 2007). E-services involve the use of information technologies (IT) via the Internet to enable, improve, enhance, transform or invent a business process or system to complete tasks, solve problems, conduct transactions or create value for current or potential customers ( Sawhney and Zabin, 2001 and Wu et al., 2003). Using e-services, firms can provide rapid customer response, improve service quality, enhance operational efficiency, and reduce costs. E-service quality is a crucial determinant in differentiating e-service offers and building a competitive advantage (Santos, 2003 and Rust and Miu, 2006). E-service quality is determined by IT-related factors, such as website security and functionality, and by product and process factors, such as product variety and order delivery timeliness ( Collier and Bienstock, 2006 and Rowley, 2006). Superior e-service quality can improve customer satisfaction, customer acquisition, and customer retention ( Boulding et al., 1993, Ranaweera and Neely, 2003 and Lee and Lin, 2005). With the payoff from e-service quality, however, also comes risk. E-service quality risk is the risk that the e-service quality of a firm will change – deteriorate or improve – relative to that of competitors. This definition recognizes that risk can be negative or positive, consistent with the way much finance research defines risk as the possibility that things will deviate from expectations (Elton and Gruber 1995).1 Companies can develop their own measures of e-service quality, but many rely on third-party benchmark measures such as those from Keynote, Bizrate, and ePublicEye. Keynote (www.keynote.com), for example, uses its GomezPro Scorecard (GPSC) to rate companies’ e-service quality based on how customers assess those companies’ websites along IT-related determinants, including: functionality, content availability, accuracy of online transactions, ease of use, and security (Al-Hawari and Ward 2006). Keynote’s benchmark ratings are published regularly for the top 20–30 companies in each of numerous business sectors (e.g., banking, insurance, brokerage). Firms whose benchmark e-service quality ratings showed superiority over competitors have used them to boost their strategic position (e.g., Citigroup Inc 2004), and as a result, firms whose ratings show inferiority feel pressured to invest in improving their e-service quality ( Carpenter, 2005 and Wright and Dawson, 2004). Unless specified otherwise, for brevity we will hereafter use the term benchmark e-service quality ratings to refer to competitive benchmark e-service quality ratings by a third-party. The reality, however, is that most firms invest in controlling e-service quality risk without knowing what level of investment is “right” for each of them. That is, firms do not know the answers to fundamental questions: What is a suitable approach for pricing e-service quality risk? Can the approach inform the firms about the economic cost, or return, associated with a relative deterioration, or improvement, in the third-party benchmark ratings of their e-service quality? And, is the answer to the last question different for different companies, depending on their characteristics (e.g., size, growth potential, industry)? This research seeks to answer these questions by presenting a finance-based approach for pricing risk and applying it to the case of e-service quality risk. Finance research prices risk in terms of two parameters: the sensitivity of an asset to a particular risk, and the risk premium measuring the extra return that the stock investor community expects to earn on the asset per unit exposure to that risk (Elton and Gruber 1995). By telling us the value that stock investors associate with a unit change in exposure to a particular risk, these risk pricing parameters could suggest a limit on the amount that a firm ought to spend to achieve that level of improvement in exposure to that risk. The financial risk pricing approach used to estimate these parameters works as follows. It starts with a linear multi-factor model linking the expected excess return on assets to the behavior of multiple systematic (firm-independent) risk factors,2 and then uses arbitrage pricing theory to estimate the risk pricing parameters based on market data (Elton and Gruber 1995). We will adapt this risk pricing approach for our needs because it makes some assumptions that may not apply in our context. The adapted approach starts with a single-factor model that is conditional on events reflecting the effect of a single idiosyncratic (firm-specific) IT investment risk factor. Thompson (1985) shows that such a conditional single-factor model captures the essence of the event-study methodology, which isolates abnormal stock returns reflecting the impact of unanticipated idiosyncratic economic events on the market value of firms experiencing those events. Here, the events of concern are the periodic publication of third-party benchmark ratings of e-service quality that show changes in firms’ relative standing. We use five years worth of data of Keynote’s bi-annually published GomezPro e-service quality ratings for firms in six financial services sectors (banking, mortgages, insurance, etc.). There is evidence that Keynote’s benchmark ratings change the perception of stock market investors about firms’ relative e-service quality and, in turn, about the firms’ market value (Chen and Hitt 2002, Kotha et al. 2004). The adapted approach then uses arbitrage pricing theory to estimate the risk pricing parameters based on abnormal returns that firms experience as a result of changes in their relative e-service quality. This article makes a contribution to IT and marketing research on e-service quality and firm value. It is the first to present an approach for measuring and pricing the risk associated with e-service quality. This approach goes well beyond extant research that only links individual aspects of e-service quality to firms’ financial performance (Barua et al., 2004, Anderson et al., 2004, Kotha et al., 2004 and Chen and Hitt, 2002). Risk pricing information opens new ways to think about the economics of a firm’s e-service quality falling behind, or moving ahead, of the competition. In particular, it could help firms determine how much they should be willing to invest in improving their relative benchmark e-service quality ratings. Our results indicate that firms in only certain financial services sectors have a significant level of exposure to e-service quality risk as measured by third-party benchmark ratings, and that level appears to vary across sectors. Further, our results suggest that firm size and growth potential also influence how investors react to relative changes in firms’ e-service quality risk, albeit their influence is notably lesser in magnitude. The latter means that firms within a particular sector have only slightly different sensitivities to e-service quality risk as measured by benchmark ratings, due to their firm-specific characteristics. This article makes a broader contribution to the literature on IT investment and risk management. We believe that it is the first to adapt and apply a well-established financial risk pricing approach to idiosyncratic IT investment risk. Another adaptation of the approach has been presented and applied elsewhere for the case of software development risks, a somewhat narrower application (Benaroch and Appari 2010). The approach presented here permits the pricing of a range of IT investment risks that are of prime concern to organizations, including security risks, customer adoption risks, and technology maturity risks. The significance of this contribution is also in supporting the solution of fundamental IT investment problems, including the management of IT investment risk and of IT investment portfolios. The remainder of the article proceeds as follows. Section 2 reviews literature on e-service quality and its relation to IT investment and financial performance. Section 3 presents our adapted risk pricing approach. Section 4 empirically applies the approach in the financial services context. Section 5 concludes with a discussion of the empirical results and the reasons behind them. It also discusses the implications of our results for research and practice, and the limitations of our work along with directions for future research.

We proceed to interpret the main risk pricing results of e-service risk for financial services firms, present implications for research and practice of these results and the risk pricing approach presented, and highlight limitations of our work and directions for future research. 5.1. Main results and interpretation One of our main empirical results is that the exposure of firms’ B2C e-channels to e-service quality risk is significant. The abnormal returns recorded after the publication of competitive benchmark ratings of firms’ e-service quality tell us that some firms benefit while others suffer as a result of these ratings. On the whole the mean abnormal returns are somewhat positive, rather than being zero as if the market reaction to competitive benchmark ratings were following a zero-sum game. One explanation, which is supported by our data, is that the market also rewards a consistent overall improvement in firms’ absolute e-service quality ratings resulting from a continued ongoing investment in IT-related determinants of e-service quality. Another main result is that the sensitivity of financial services firms to e-service quality risk varies across sectors. Firms in three sectors have statistically significant but different sensitivities to relative changes in e-service quality. Credit card firms have the largest sensitivity, followed by discount brokerage and by mortgage firms. This pattern could be due to the rate at which e-services have been embraced in these sectors. Among the most successful early e-commerce businesses were discount brokerage (Bakos et al. 2005) and online mortgage services (Hunt and Menon 2006). E-services in these sectors grew fast, in part, because of the fierce competition and aggressive customer acquisition tactics (Chen and Hitt 2002) from new entrants in the discount brokerage sector (e.g., eTrade, Datek, AmeriTrade) and the mortgage sector (e.g., E-Loan, LendingTree, Floan). The credit card and mortgage sectors faced other growth drivers (Osho 2008). Beginning in 1996, traditional banks pushed to expand their product offerings and customer base on loan products like credit cards and mortgages (Osho 2008). One reason is the erosion of margins in their liability businesses (savings, money market, etc.) due to the emergence of loan aggregators and Web banks. Another reason is that Internet companies with strong brands and highly trafficked websites (e.g., Microsoft, Intuit, Yahoo!) began earning fees for generating mortgage and credit card leads (Hunt and Menon 2006), and the concern was that they would move into fulfillment and funding (HFI June 1997). Lastly, there was also a concern that the Internet will increase the transparency and accelerate concentration among originators (Morgan Stanley Dean Witter 2000). The end result consequence was that e-service quality became a crucial determinant of differentiation (Bakos et al., 2005 and Chen and Hitt, 2002). By contrast, the other three sectors were not found to have statistically significant sensitivities to e-service risk. While we cannot make reliable inferences about these sectors, we can speculate about the reasons. Banking, insurance and full-service brokerage services are the largest sectors in the financial services economy (Bakos et al. 2005). The sheer size of these sectors may be at the root of the slow growth in their e-services. These sectors encompass many large traditional firms that are better characterized by their brick-and-mortar side of the business. Their business is inherently about relationships with customers, but at the time, the e-service channels probably did not represent an acceptable substitute for the conventional sales channels. Moreover, the potential for channel conflicts was significant. This factor was at the core of the full-service brokerages’ slow responses to competition from the discount brokerages (Bakos et al. 2005). For similar reasons, major commercial houses including JP Morgan and Chase (when they were separate firms) and Bankers Trust did not consider moving into e-services until 1998 (Osho 2008). The insurance sector faced an additional challenge. While considerable growth occurred with commodity-type (motor, travel, and home) insurance products, growth in sales of more complex insurance products could not occur before their lengthy online application process and the need for an offline verification and approval were reengineered (Hunt and Menon 2006).13 These factors limited the value of e-service channels, and made firms in these sectors insensitive to e-service quality risk. This explanation is consistent with the extant research and calls for additional research into the influence of market growth. In the context of new service and e-service entry, the market growth rate has been observed to have a positive relationship with firm performance (Geyskens et al., 2002 and Bowman and Gatignon, 1995), as it can characterize the ease of gaining access to a market (Geyskens et al., 2002 and Ramaswamy et al., 1994). Because rapid market growth also offers opportunities to broaden e-service operations, it may positively enhance the overall performance improvement that is anticipated due to firms’ improvement in their e-service quality ratings. Future research should verify this explanation by controlling for important firm-specific characteristics, for example, the e-service share of a company’s entire business. This characteristic probably tends to be low for those sectors for which we did not find a significant level of sensitivity to e-service quality risk. A third main result comes out of our robustness test. It shows the market reaction observed in connection with changes in e-service quality ratings to be stronger for high growth firms and for large firms. Two factors may explain the relationship with firm growth. On the one hand, high growth companies may have greater opportunities to capitalize on relative improvements in e-service quality ratings due to their earlier IT investments; IT investments are known to have a strong positive complementarities effect with e-commerce capability that positively contributes to firm performance (Zhu 2004). On the other hand, high growth firms experiencing a relative drop in their e-service quality ratings may be constrained in their ability to make the IT investments necessary to recover their relative e-service quality ratings, because growth imposes greater cash flow demands (due to the existence of many high NPV projects) and leaves less funds for IT budgets (Dewan et al. 1998). Both possibilities suggest that high growth firms could be more sensitive to e-service quality risk. It is puzzling, however, that the market reaction to changes e-service quality risk may be stronger for large firms. Some research has shown that firm size is negatively associated with the market reaction to e-service launch initiatives (Lin et al. 2007). Large firms are less exposed to small changes in any single risk factor, including e-service quality, because they are generally better diversified in relation to their products, services and sales channels. Moreover, large firms often have greater established industry experience, which means a better-established customer base and brand name as well as lower risk (Gatignon et al. 1990, Geyskens et al. 2002). By contrast, small firms stand to lose more from exposure to any single risk, including e-service quality risk, although they may also stand to make inroads into opportunities that larger firms have already exploited, such as extending their geographic reach (Alba et al., 1997 and Geyskens et al., 2002). Our result goes against both assertions concerning the impact of firm size, and so more research is needed to verify and explain this result. 5.2. Implications for practice and research The implications of our work for practice and research are numerous. On a practical level, risk pricing information of the kind we derived can aid managers making decisions on investments in IT-related determinants of e-service quality. Qualitatively, our results suggest that small firms, especially in fast-growing industries, stand to gain more from relative improvements in their e-service quality ratings. Quantitatively, for credit card companies, return on IT investment in e-service quality may be twice as high as that of discount mortgage firms and almost five times that of mortgage firms, at least for the timeframe our data covers. Repeating our analysis with recent data may yield relative return on investment figures that apply today. This information could be of greater value to large financial services firms with business units operating in multiple different sectors (credit card, mortgages, brokerage, etc.), as their investments in IT determinants of e-service quality could be allocated more optimally across units. Going a step further, risk pricing information could also benefit firms seeking to manage their relative e-service quality standing according to benchmark ratings. The premium return on a one-unit change in relative e-service quality ratings is computed as risk premium × sensitivity to risk. For example, with the risk premium of 0.112% we obtained, the premium return for mortgage firms is 0.112 × 4.894 = 0.548%, for discount brokerage firms is 1.358%, and for credit card firms is 2.344%. Multiplying the premium return by the number of outstanding shares of a firm, premium return × number of shares, can approximate the gain (or loss) in market value that a firm can expect to see from a one-unit improvement (or drop) in its relative benchmark e-service quality ratings. It remains to be seen, however, whether such an approach also reflects the long-term return on investments in IT-related determinants of e-service quality. For research, the implications of our work are twofold. Compared to prior research linking individual IT-related determinants of e-service quality to firm financial performance (Anderson et al., 2004, Barua et al., 2004, Chen and Hitt, 2002 and Kotha et al., 2004), our work goes further by pricing the risk associated with those determinants. It prices e-service quality risk in terms of the economic consequences for a firm if the IT determinants of e-service quality fall behind or move ahead relative to the competition. This can open up new ways for IT and marketing research to think about the economics of investments in e-service quality. Second, the well-established risk pricing approach we presented could work equally well in the context of other idiosyncratic IT investment risks. Security risk, customer adoption risk, and technology maturity risk are just a few of the IT risks that should be of great interest to academics and practitioners. As we mentioned earlier, another adaptation of the approach has been presented and applied elsewhere for the case of pricing software development risks (Benaroch and Appari 2010).14 Going beyond the pricing of single IT investment risks, though, availability of risk pricing information could open new research directions on fundamental IT investment management problems, including the valuation and ranking of IT investments, the economic-based management of IT investment risk, and the management of IT investment portfolios. 5.3. Limitations and future research Our study has several limitations which future research ought to try to resolve. First, as we explained earlier, our data sample stops at the fourth quarter of 2004 primarily because of a concern over the objectivity of the third-party ratings posted at that point in time. As a consequence, to obtain a sufficient size data set we have included data from two different periods, before and after the dotcom bust, but we did not control for the possibility of a shift in stock market returns due to the dotcom bust (Dehning et al. 2004). Considering that our goal is one of estimating risk pricing parameters, rather than testing hypotheses, adding time dummies for the before and after dotcom bust periods is not an adequate solution. Rather, it would be necessary to split the data into two subsamples, before and after the dotcom bust, and replicate the analysis for each subsample separately. Our overall data sample was insufficiently large to create both subsamples. A second limitation is the relatively low R2 value that we obtained for the estimation of the risk premium parameter (Table 4). While this R2 is well within the range of results reported in financial research dealing with risk pricing ( Stickel, 1992 and Schadewitz and Kanto, 2002), the skeptics may argue that the observed significance of the coefficients estimated in this regression could be due to sample size as much as it could be due to the presence of a meaningful relationship in the data. Last, our data set does not include the granular components of Keynote’s GomezPro ratings of e-service quality: the ratings for website privacy and security, functionality, ease of use, and information content and quality. Our empirical analysis cannot tell us how much each component alone contributes to the sensitivity of firms to e-service quality risk. Future research ought to replicate our empirical effort with data covering a longer period and containing the granular components of competitive benchmark ratings of e-service quality. Some of the granular components of e-service quality ratings are in themselves recognized as separate IT investment risks (e.g., website security and availability). Therefore, replicating our empirical effort for the granular components could lead to more fine-grained and valuable risk pricing information. Of course, it is possible that none of these components on its own generates measurable abnormal market reactions to relative changes in e-service quality ratings. Additional directions for future research follow from our earlier discussion of the empirical results. One direction pertains to the mean abnormal returns computed for announcements of competitive benchmark ratings. These are positive on the whole, whereas one would expect them to be zero (or close to zero) if the impact of changes in relative e-service quality ratings followed a zero-sum game (Table 2). One possibility that we did not mention earlier is that the market reacts more positively to relative improvements in e-service quality than it does negatively to relative drops in e-service quality. This possibility deserves further exploration as it may have implications for the validity of the linear risk pricing model that we presented. However, it is worth keeping in mind that financial research may have considered this possibility in other contexts and yet it continues using the same linear model. Another direction worth exploring is how firm-specific characteristics may influence firms’ sensitivity to e-service quality risk. Neither the sensitivities we have derived empirically nor the adapted risk pricing model we presented directly consider this issue. Yet, the results of our robustness test indicate the possibility that, at least, firm growth and firm size may moderate firms’ sensitivity to e-service quality risk. Future research should consider using a multi-factor version of the risk pricing model we have presented as one direct way to account for firm-specific characteristics. In summary, this article is the first to present a risk pricing approach in the IT investment risk context and its application to e-service quality risk. Still, it ought to been seen as only a first step in a journey towards a better understanding of the challenges involved in pricing IT investment risk, as well as a fuller appreciation of the theoretical and practical value of risk pricing information. We are confident that the ideas and results that we have presented will inspire and encourage researchers to join this journey and pursue some of the research directions we have touched upon.