|
Article Excerpt
This study investigates the extent to which investors believe that enterprise resource planning (ERP) systems enhance firm value by examining changes in financial analysts' earnings predictions before and after they receive an announcement that a firm plans to implement an ERP system. A total of 63 analysts participated in a two (firm size: small and large) by two (firm health: unhealthy and healthy) randomized between-subjects design. The ERP announcement represented a within-subjects manipulation. The analysts' overall reaction to ERP implementation plans was positive, as mean post-announcement earnings forecasts were significantly higher than mean pre-announcement forecasts. Additionally, as expected, mean earnings forecast revisions in the small/healthy and large/unhealthy firm conditions were significantly greater than mean forecast revisions in the small! unhealthy firm condition.
Experimental results from the current study support archival findings reported by Hayes et al. (2001), who explored the same research questions, among others, by examining cumulative abnormal returns surrounding ERP announcements. Triangulation studies of this nature using multimethods (e.g., behavioral vs. archival) and complementary criterion variables (e.g., earnings forecasts vs. cumulative abnormal returns) are important to social scientists, as they provide insight into the reliability, consistency, and validity (both internal and ecological) of proposed theoretical relationships (Boyd et al. 1993; Flick 1992; Libby et al. 2002).
Keywords: enterprise resource planning; ERP; investors; financial analysts; triangulation.
Data Availability: Data will be made available upon written request and justification.
I. INTRODUCTION
Enterprise resource planning (ERP) systems promise to integrate business processes within and across functional areas in organizations. Early ERP systems primarily included inventory control software, material requirements applications, and manufacturing planning modules. The continual evolution of ERP systems has subsequently encapsulated the full spectrum of business processes such as selling, marketing, purchasing, warehousing, accounting, and human resource planning into tightly integrated enterprise-wide information databases. The latest generation of ERP systems extends beyond the organization by capturing inter-organizational processes such as customer and vendor relationship management (Kumar and Van Hillegersberg 2000).
The pervasive organizational effects of ER? systems have been widely addressed in the popular literature. However, there is a paucity of empirical research examining the impact of ERP system implementations on firm performance, which is surprising considering that ERP systems inextricably intertwine a vast array of intra- and inter-organizational business processes with the accounting information system. Recently, calls for empirical studies into the effects of ERP system implementations on firm value have appeared in the literature (Lee 2000).
In response to such calls, a recent accounting archival study investigated the market reaction to ERP implementation announcements (Hayes et al. 2001). Overall, they found that the market reacted favorably to ERP announcements, as cumulative abnormal returns surrounding the announcement date were significantly positive. Additionally, they hypothesized that market reactions to small/healthy and large/ unhealthy firms would be more positive than the reaction to small/unhealthy firms. However, this effect was realized only for the small/healthy firms. They suggested that the nonsignificant effect for large/ unhealthy firms could have been due to low power, as the mean reaction was in the anticipated direction, but sample sizes were somewhat small.
In the current study, we examine the same questions set forth by Hayes et al. (2001): do investors (1) believe that ERP systems add value to organizations and (2) react differently to ERP system announcements contingent on the firms' relative size and financial health? However, we use a different research approach (experimental) and examine a different, yet complementary, criterion variable (earnings forecasts). As compared to Hayes et al. (2001), our experimental design allows us to draw causal inferences, which could only be implied through correlation in the archival study, and yield greater statistical power to determine if the lack of significance between large/unhealthy and small/unhealthy firms in the archival study was a power problem or a theoretical issue.
As suggested by Boyd et al. (1993), Flick (1992), and Libby et al. (2001), the use of multimethods (e.g., behavioral vs. archival) and complementary dependent variables (e.g., earnings forecasts vs. cumulative abnormal returns) to investigate underlying phenomena are very important to social science research, as a convergence of perceptual (e.g., financial analysts' beliefs) and objective (e.g., stock investors' reactions) measures of the environment can increase the internal and ecological validity of theoretical relationships pertaining to phenomena of interest. In the next section, we present relevant theory and offer study hypotheses. In the following sections, we describe the research method, analyze the experimental data, and discuss the study findings.
II. THEORY AND HYPOTHESES
The purpose of this study is to experimentally test theoretical relationships developed by Hayes et al. (2001), who examined the market reaction to ERP implementation announcements via cumulative abnormal returns surrounding announcement dates. While Hayes et al. (2001) attributed positive, significant, cumulative abnormal returns to ERP announcements, they warned readers to be cautious about drawing causal inferences from event studies of this nature. Another limitation of the Hayes et al. (2001) study is that one of their hypotheses did not obtain statistical significance (large/unhealthy firms> small/ unhealthy firms). While they attributed the lack of significance to low power, it is also possible that the underlying theory leading to their hypothesis needs refinement.
We examine the same issues as Hayes...
|
