Managerial Accounting for Decision-Making and Planning

Abstract

Accounting information, as characterized by Simon, Kozmetsky, Guetzkow, and Tyndall (1954), has three main uses: decision-making, planning, and control. Academic research has mostly focused on the third role of accounting information. There is a vast literature examining the role of incentives in aligning the interests of managers and employees with organizational goals, that is, the control function of accounting. Much less effort has been devoted to the problem of the design of information systems to improve decision-making and planning. Yet much of what we teach in the classroom and what practitioners highlight as pressing issues for their organizations has to do precisely with the first two roles of accounting information (Labro 2015).The objective of this paper is to provide a brief, bird's eye overview of decision-making research in accounting. We concentrate here on barriers to decision-making and the tools to overcome them, and largely abstract from agency problems. We also highlight potential avenues for future research that we believe are ripe for more thorough study. The papers published in this special interest forum are a good example of the work that needs to be done to advance this agenda. We hope they spur further research in the area. Rigorous work on management accounting to support the decision-making and planning functions would be very relevant for practitioners and is long overdue.Organizations rely on managers and employees to make decisions that are in the best interest of the entire organization. It is no surprise that there is a great need to understand the tools that are available to aid in those decisions. Several approaches to decision-making are useful for management accounting research. The economic approach assumes that the decision-maker is rational and makes the best choice possible to maximize her expected utility. The behavioral approach recognizes the cognitive limitations of the decision-maker that may lead to biased or suboptimal choices. Finally, the sociological approach explicitly acknowledges that decision-making in organizations is a political process that is affected by the socio-economic interactions among decision-makers whose interests and preferences are not naturally aligned. We review each of these approaches in turn, with suggestions for advancing their contributions.The economic approach has proven very useful for understanding decision-making in many contexts, including those of a more fundamental economic nature, such as consumption decisions, investment choices, or educational and career choices, but also beyond economics, such as family organization.An important feature of many decisions is that they take place under a great deal of uncertainty. One of the central tenets of rational decision-making is that more information always improves the expected outcome of the decision. Because the decision-maker can always ignore any piece of information she is provided with, she cannot be made worse off when offered additional information. Naively, one might interpret this result as suggesting that the job of managerial accounting is to provide all information available to those making decisions. Unfortunately, information is not free: there are costs to acquire it, process it, and communicate it. To economize on those costs, it is important to avoid irrelevant information and to find ways of presenting all the relevant information in a succinct, useful way.In this regard, it would be useful to expand on the study of the informativeness of different accounting measures. There has been a large amount of work studying the properties of different measures in relation to how they are used in contracting (see, for instance, Ittner, Larcker, and Rajan 1997; Ittner, Larcker, and Meyer 2003). But very little is known about their use for decision-making. There is scope for studying the value of financial versus nonfinancial, forward looking versus backward looking, objective versus subjective, and absolute versus relative information in a decision-making context. That value may differ depending on the decision at hand. It would be interesting to see how various measures allow decision-makers to balance tradeoffs between short-term and long-term value creation, or between risk and reward (see, for example, Casas-Arce, Martínez-Jerez, and Narayanan 2017b). This would help us understand if or how the use of such measures for decision-making differs from their use in contracting. It would also be useful to depart from Blackwell's framework, recognizing that incentive systems are unlikely to completely solve agency problems and, hence, the preferences of the decision-maker may not be aligned with those of the principal. Extending the literature on Bayesian persuasion to accommodate the various features of accounting information can provide a fruitful avenue for future research, offering new insights for the design of management accounting information systems.The technologies and tools that produce the information are other important levers to study as well, specifically since a number of new information technologies are designed to aid with decision-making and planning. Predictive analytics that make recommendations for alternative action courses, or customer relationship management systems that provide reminders with actionable suggestions to improve customer satisfaction are just two examples. We can see these tools as decision aids that lower the cost of accessing and processing information, freeing up time for other aspects of the decision-making process. Given the pace of technological change, this is potentially a very fruitful avenue for future research, a topic to which we will return repeatedly.Although assuming that decision-makers are rational is a useful approach in some contexts, decision-makers are often imperfect. Behavioral research (both in psychology and in economics) has shown that people are fallible, and often in predictable ways. Although decision-makers in organizations tend to be experienced professionals, they also fall prey of such behavioral biases. This is more so these days of rapid technological change, in which managers have to adapt to fast-evolving market conditions and where knowledge becomes rapidly obsolete. In those circumstances, there is less scope for learning as a way to mitigate those biases. This opens the scope for information systems to provide information in such a way that mitigates the effect of those biases and facilitates optimal decision-making (see, for instance, Casas-Arce, Lourenço, and Martínez-Jerez 2017a).Technology provides us with the ability to transform unstructured and semi-structured data (e.g., customer comments and email content) into quantifiable measures. However, the measurement properties of these new types of information, and the extent to which they capture the performance dimensions of interest, may be hidden to those who are impacted by them. A fruitful area of research is the study of individuals' willingness to rely on performance measures that they do not truly understand, and the extent to which individuals consider measurement errors associated with these less-familiar types of information when making decisions.In this special interest forum, Black, Meservy, Tayler, and Williams (2022) show that the mere awareness of measurement can lead to strategy surrogation—a tendency for individuals to make decisions based on imperfect performance measures as if these measures are the actual strategy (Choi, Hecht, and Tayler 2012). Their study raises an interesting future research question: as technology continues to transform previously unquantifiable performance into analytics, what is technology's impact on strategy surrogation when everything can be measured but, at the same time, the measures themselves may be difficult to comprehend? Technology-enabled dissemination of information is also becoming increasingly cost effective. Newman, Stikeleather, and Waddoups (2022) find that the (mis)alignment between relative performance information's (RPI) social motivational role and its informational role affect employees' pursuit of exploratory activities. Behavioral implications of employees receiving real-time and transparent feedback, and the possibility of sharing more varieties of RPI with increased measurement sophistication, are fertile grounds for research.In addition to changing the types and availability of information we use to make decisions, technological advances have also given rise to decision algorithms that may partially, or even fully, replace human decisions. Prior research shows that individuals tend to underuse predictive algorithms that are seen as imperfect (e.g., Dietvorst, Simmons, and Massey 2015), especially in settings with a high level of inherent uncertainty (Dietvorst and Bharti 2020). However, in other circumstances individuals have a preference for algorithmic advice over human advice (Logg, Minson, and Moore 2019). As decision algorithms become more commonplace, from sales forecasting to supply chain management, future research can examine factors that influence the over- versus under-reliance on decision algorithms, and interventions that ensure their optimal use (e.g., Kesavan and Kushwaha 2020).Technology has also changed the way we interact with information, with personalization a key new feature. This presents researchers with opportunities to bring “back in vogue” research into personality traits, cognitive styles, and other individual differences. In the past, such research has been somewhat constrained by its practical implications, because it is often unclear as to what realistic actions organizations can undertake to utilize knowledge about employee heterogeneity. Technology now enables users to customize visualization, report format, and more. Further research into the implications of individual characteristics on decision-making can inform managers on how to personalize their interactions with management accounting systems in a way that can reduce decision biases and errors.In summary, fast evolving technologies have transformative impacts on information availability, decision processes, and how decision-makers interact with management accounting information.In contrast to economic and behavioral approaches, sociologically informed research on managerial accounting draws attention to the inherently political nature of organizational decision-making and often emphasizes the serendipitous and seemingly unstructured nature of such decision-making. Although this research is not strongly represented in the papers included in this special interest forum, it can extend our understanding of managerial accounting and decision-making. One of the key theoretical perspectives informing such research is that evolving from the garbage can model originally advanced by Cohen, March, and Olsen (1972). The original garbage can model broke with the view of organizational decision-making as a relatively orderly process, where clearly defined problems trigger a rational search for appropriate solutions and, rather, portrayed decision-making as a matter of fitting largely predefined solutions to diverse problems in a seemingly random manner. It also saw the complexity of organizational decision-making as being compounded by the fluidity of participation that is due to the changing constellations of actors influencing particular decisions, and the existence of multiple choice opportunities, or arenas for decision-making, where problems and solutions are being matched. As actors with diverse and often conflicting priorities enter and exit different choice opportunities, the political nature of decision-making is amplified and outcomes become increasingly difficult to predict. A key insight from the garbage can literature is that, due to the resultant complexity, many organizational problems remain unresolved for extended periods of time.Since its inception, the garbage can model has undergone considerable refinement (see Lomi and Harrison 2012) and has been recognized as a potentially useful basis for examining how accounting is implicated in the politics of decision-making (Cooper, Hayes, and Wolf 1981; March 1987). However, its use in empirical accounting research has been relatively limited (Mezias and Scarselletta 1994; Wiesel, Modell, and Moll 2011). One reason for this may be the methodological challenges of studying complex decision-making processes in vivo. For instance, Wiesel et al. (2011) relied on close observations of the meetings involved in an intricate decision-making process, which unfolded over a period of eight months, and combined these observations with interviews and document analyses to situate this process in its broader organizational context. Access to such processes may be difficult to negotiate and this may discourage researchers from delving into the complexities of decision-making. However, another reason for the limited use of the original garbage can model in empirical accounting research may be that many decision-making processes, in which accounting is implicated, are not as unstructured as the model initially envisaged (Perrow 1977; Levitt and Nass 1989). As noted by Mezias and Scarselletta (1994, 676), the institutional environment in which accounting practices evolve “can produce both order and disorder” and thus needs to be considered as a potential constraint that structures decision-making and imbues it with a degree of predictability.One promising avenue for enhancing our understanding of how decision-making is structured and how garbage can situations can be imbued with a degree of order is that advanced by Nils Brunsson (1982, 1985) and others working within the broader research tradition known as “Scandinavian institutionalism” (see Boxenbaum and Pedersen 2009). Combining insights from the garbage can literature with institutional sociology, Brunsson (1982, 1985) conceived of decision-making as an inherently political phenomenon, which gives rise to seemingly “irrational” choices, but that can be guided by powerful, organizational ideologies. Such ideologies take the form of widely shared, socio-cognitive schema that prescribe what constitutes appropriate, or legitimate, solutions to particular organizational problems. Insofar as such ideologies are mobilized and gain traction, the range of potential solutions will be narrowed down quite considerably, and this is likely to reduce the complexity of decision-making and enhance the likelihood that particular organizational problems are actually resolved, although the solutions may not be entirely optimal, or “rational,” from an objectivist point of view. A similar and more recent approach to decision-making, inspired by “Scandinavian institutionalism,” is that of considering decisions as promises (Mouritsen and Kreiner 2016). If decisions are viewed as promises, to which decision-makers have committed themselves, this provides a basis for holding them accountable for the consequences of their decisions. This will, in turn, increase the likelihood of decision-makers working hard to implement decisions despite the many obstacles that may hamper implementation. According to Mouritsen and Kreiner (2016), accounting can play an important role by solidifying promises and documenting the outcomes for which decision-makers are held accountable.While the discussion above provides some clues as to how accounting is implicated in the politics of decision-making, more research is required into how this phenomenon is affected by the introduction of artificial intelligence and related technologies. We currently know little about whether such technologies will amplify or reduce the political nature of decision-making. Insofar as they automate decision-making and reduce the need for human judgement, it is possible that such technologies will reduce the likelihood of decisions being politicized. However, it is also possible that the abundance of information that is being made available will present decision-makers with a nearly endless range of problems and solutions, which will, in turn, find their way into diverse choice opportunities where problems are less likely to be resolved due to the politics of decision-making.Gartner (2021) defines advanced analytics as “the autonomous or semi-autonomous examination of data or content using sophisticated techniques and tools, typically beyond those of traditional business intelligence (BI), to discover deeper insights, make predictions, or generate recommendations. Advanced analytic techniques include those such as data/text mining, machine learning, pattern matching, forecasting, visualization, semantic analysis, sentiment analysis, network, and cluster analysis, multivariate statistics, graph analysis, simulation, complex event processing, neural networks.”The proliferation of advanced analytics (AA) in organizational decision-making, accelerated by the availability of Big Data, is having a tremendous impact within our society and within many industries (Bean 2018; Wamba et al. 2017). The retail industry pioneered with the application of analytical tools to make just-in-time recommendations or increase the customer experience (Wamba et al. 2017). The healthcare industry followed by implementing analytics to cut costs and improve treatment quality. Predictive maintenance, supply chain optimization, and industrial automation are illustrative examples where analytics plays a significant role in manufacturing. While one can observe first attempts of analytics also in the finance function (Mehanna, Tatzel, and Vogel 2016; Schäffer 2017; Rikhardsson and Yigitbasioglu 2018), interestingly the adoption of AA in this area is lagging behind. Consequently, studies on the determinants and consequences of applying or choosing not to apply AA in management accounting domains such as forecasting, budgeting, and performance evaluation are rare.There is huge potential for theory-based evidence on the adoption patterns and implementation hurdles of AA in management accounting. First, future research could address the reasons for the lower adoption of AA in the finance function as compared to other business functions. We need to better understand the characteristics of corporate planning and how they differ from areas with high adoption (e.g., retail). For example, how is the forecasting task itself different? One characteristic of the forecasting task is the importance of the decisions made. Applications in finance can be expected to have high impact, in the sense that decisions made in the finance function have spillover effects and interdependencies for other areas, which might hamper adoption. Relatedly, typically there are personal consequences of forecast accuracy and budget target achievement for the people involved (e.g., Feichter, Grabner, and Moers 2018). Thus, there might be resistance of finance people to AA interventions in the budgeting process due to a lower possibility to influence targets ex ante or game numbers ex post. Maybe it is not about not trusting the algorithm, but more about being able to control the result.Second, another fruitful area is to study the implications of AA use in the finance function for firms' external disclosure decisions. For example, can AA in internal forecasting help to improve the likelihood of beating earnings forecasts? The study by Kroos, Schabus, and Verbeeten (2022) in this special interest forum provides first evidence of such a link (see, also, Ittner and Michels [2017] for related evidence). In particular, they show that firms with more sophisticated internal forecasting systems are less likely to engage in misreporting, given higher ex ante forecasting capability, and consequently more accurate forecasts in the first place. The benefits of being able to avoid manipulations (potentially through better internal planning procedures) are demonstrated by another study in this special interest forum. Cook, Huston, Kinney, and Smith (2022) show that the market premium from meeting or beating analyst earnings forecasts by manipulating inventory is smaller than the premium for achieving this threshold absent inventory manipulation. They further show that the market consistently discounts earnings beats for “serial manipulators,” suggesting that firms' manipulation efforts erode investor confidence in earnings.Finally, we need to understand the implications of AA use for managerial decision-making and organizational design. Given the interdependence of organizational design choices (Grabner and Moers 2013), the progressing implementation of AA in various functions of the firm will not happen in isolation, but will also trigger changes to other important organizational (control) practices. Labro, Lang, and Omartian (2020) are the first to document a relation between predictive analytics use and organizational architecture choices and conclude, “the information generated through predictive analytics and related Big Data techniques has the potential to fundamentally alter the optimal organizational design and control structure within the firm.”Another reason for the adoption hesitation might be the fear that machines will make humans obsolete. As the computational and processing capabilities of AA grow with immunity to cognitive biases and organizational politics that can often undermine human judgment, organizations face a critical question: is human involvement necessary when making decisions with AA, or are analytics best left alone? This question offers huge opportunities for in-depth field studies (see, e.g., Choi, Forker, Grabner, and Sedatole 2021), as well as in experimental environments (where psychology research has already made a head start). What are the conditions where augmented decision-making—that is, a man-machine symbiosis—is superior to predictive analytics models alone? Does it depend on task types or people types? How do a firm's management control choices affect the quality of human judgment over AA? With the ever-increasing use of analytics in practice, understanding the conditions under which advanced analytics and human judgment can best work together is critical for the design of effective management control systems.As we have been emphasizing throughout the paper, information technology is a potentially fruitful avenue for research. It has drastically increased the amount of available information in organizations, while at the same time increasing the power to analyze these data. Although this technological revolution has the potential to drastically improve decision-making and planning, organizations are still learning how to unleash all its potential. Part of the difficulty is not just purely technological. Indeed, information technology creates its own challenges by vastly increasing the amount of information available, both relevant, but also irrelevant. Management accountants are best placed to help organize that information and present it in its most useful form. Some of it is simply driven by the need to lower the cost of accessing and processing that information for decision-makers. Some of it requires more careful thinking to address potential behavioral biases, such as focusing attention on salient, but irrelevant or unrepresentative information. Yet others respond to social dynamics and internal politics that hinder the use of that information in a rational and predictable manner.Addressing all these research questions would be best served by a wide range of methods to leverage the unique benefits that each of them offer. Archival data and surveys provide a lens into what firms do that may be illuminating about the decision-making and planning processes, as well as how firms adapt with the use of technology. Observational field research can provide more texture and details about what is really going on inside firms, providing a more detailed and nuanced view than what can be readily measured archivally. Laboratory experiments can provide a more reliable inference about causality, as well as offer a window into counterfactuals that might otherwise be unobservable in the real world (for instance, the outcome from a decision that has not been taken). In addition, if those experiments are conducted in the field, we have the added benefit of using professionals in a real decision-making setting that is relevant for practice. Leveraging all these tools is likely to produce vibrant and insightful new evidence that takes management accounting forward.