Hart E. Posen

On the Strategic Accumulation of Intangible Assets

The resource-based view holds that firms can earn supranormal returns if and only if they have superior resources and those resources are protected by some form of isolating mechanism preventing their diffusion throughout industry. One isolating mechanism that has been proposed for intangible assets is their accumulation process. The hypothesis is that intangible assets are inherently inimitable because would-be imitators need to replicate the entire accumulation path to achieve the same resource position. Thus, entrants can never catch up to incumbents.An interesting challenge to this hypothesis is counterfactual evidence that entrants sometimes outperform incumbents. Such counterfactual evidence should not exist if the theory is strictly correct. This paper attempts to reconcile resource accumulation theory with the counterfactual evidence. We do so by building an intermediate good-production function for a firms intangible asset stocks. We test the contribution of the intangible asset stock to the firms final good-production function and examine the extent to which that asset stock deters rival mobility in the pharmaceutical industry.We find that the asset accumulation process itself cannot deter rivals, because asset stocks reach steady state rather quickly. Entrants can achieve an incumbents intangible asset stock merely by matching its investment until steady state. Thus, we conclude that the accumulation process per se is not an isolating mechanism. While this is perhaps the most important contribution, another contribution is an empirical methodology for characterizing the accumulation function.

Chasing a Moving Target: Exploitation and Exploration in Dynamic Environments

A common justification for organizational change is that the circumstances in which the organization finds itself have changed, thereby eroding the value of utilizing existing knowledge. On the surface, the claim that organizations should adapt by generating new knowledge seems obvious and compelling. However, this standard wisdom overlooks the possibility that the reward to generating new knowledge may itself be eroded if change is an ongoing property of the environment. This observation in turn suggests that environmental change is not a self-evident call for strategies of greater exploration. Indeed, under some conditions the appropriate response to environmental change is a renewed focus on exploiting existing knowledge and opportunities. We develop a computational model based on the canonical multiarmed bandit formulation of exploration and exploitation. We endeavor to understand the mechanisms by which environmental change acts to make purposeful efforts at organizational adaptation less (or more) valuable. This paper was accepted by Jesper Sorensen, organizations.

Do Product Architectures Affect Innovation Productivity in Complex Product Ecosystems

Abstract In this paper, we seek to understand how changes in product architecture affect the innovation performance of firms in a complex product ecosystem. The canonical view in the literature is that changes in the technological dependencies between components, which define a product’s architecture, undermine the innovation efforts of incumbent firms because their product development efforts are built around existing architectures. We extend this prevailing view in arguing that component dependencies and changes in them affect firm innovation efforts via two principal mechanisms. First, component dependencies expand or constrain the choice set of firm component innovation efforts. From the perspective of any one component in a complex product (which we label the focal component), an increase in the flow of design information to the focal component from other (non-focal) components simultaneously increases the constraint on focal component firms in their choice of profitable R&D projects while decreasing the constraint on non-focal component firms. Second, asymmetries in component dependencies can confer disproportionate influence on some component firms in setting and dictating the trajectory of progress in the overall system. Increases in such asymmetric influence allow component firms to expand their innovation output. Using historical patenting data in the personal computer ecosystem, we develop fine-grained measures of interdependence between component technologies and changes in them over time. We find strong support for the empirical implications of our theory.

Resource Allocation in Strategic Factor Markets: A Realistic Real Options Approach to Generating Competitive Advantage

This paper develops a realistic real option theory of resource allocation decisions in strategic factor markets. Competitive advantage in factor markets is underpinned by market failures that allow firms to acquire assets at less than their value in use. We recognize that market failure may result from uncertainty regarding the current and/or future value of an asset, which map, respectively, to uncertainty as modeled in the feedback learning and real options literatures. The realistic real option framework we develop grafts insights from the strategic factor market, feedback learning, and real option valuation literatures. We argue that competitive advantage may emerge not only from luck, or ex ante differences in information or complementary assets, but also because firms differ in a specific type of learning ability — the ability to integrate new information to exercise a contingent claim on an asset in a factor market. We dimensionalize these differences in terms of information processing and belief updating, argue that these differences lead to different resource allocation decisions, and suggest how these decisions may generate competitive advantage.

Adaptive Capacity to Technological Change: A Microfoundational Theory of the Dynamics of Routines

We take a microfoundational approach to understanding the origin of heterogeneity in firms’ capacity to adapt to technological change. We develop a computational model of individual-level learning in an organizational setting characterized by interdependence and ambiguity. The model leads to organizational outcomes with the canonical properties of routines: constancy, efficacy, and organizational memory. At the same time, the process generating these outcomes also produces heterogeneity in firms’ adaptive capacity to different types of technological change. An implication is that exploration policy in the formative period of routine development can influence a firm’s capacity to adapt to change in maturity. This points to a host of strategic trade-offs, not only between performance and adaptive capacity, but also between adaptive capacities to different forms of change.

Fortune Favors Fools: How Confidence Can Compensate for Competence in Learning

An important strategic challenge facing entrepreneurs and managers is the need to make choices across a set of policy alternatives (e.g., strategies, technologies, product designs), the merits of which are not initially well known. The efficacy of these choices is an increasing function of knowledge — competence that takes the form of accurate beliefs about the relative merits of the alternatives. Yet knowledge has a second, equally important but underexplored, dimension related to confidence in those beliefs. We examine confidence from a behavioral theory perspective that points to its role in learning processes, rather than a psychological perspective. We employ a formal computational model of learning where uncertainty may make any particular experience with an alternative misleading or unrepresentative. While our model is general in its applicability, we apply it to the case of an entrepreneur pursuing a particular market opportunity. Confidence acts to moderate an entrepreneur’s willingness to disregard feedback that conflicts with her beliefs. We find that there are conditions (e.g., high uncertainty) under which confidence in beliefs, rather than the accuracy of those beliefs, is the primary driver of the efficacy of entrepreneurial choice. One implication of this observation is that a less competent (less accurate) but more confident entrepreneur may outperform a highly competent entrepreneur lacking confidence in her accurate beliefs.

The Impact of Learning and Overconfidence on Entrepreneurial Entry and Exit

Empirical evidence suggests that entrepreneurs make mistakes: too many enter markets and, once there, persist too long. While scholars have largely settled on behavioral bias as the cause, we suggest that this consensus is premature. These mistakes may also arise from a process in which entrepreneurs continually learn about their prospects, and make entry and exit decisions from what they have learned. We develop a computational model of this process that connects pre- and post-entry learning and can be directed to analyze Bayesian rational or biased entrepreneurs. The model suggests that, to outside observers, rational entrepreneurs may appear overconfident, seem to take too long to exit, and exhibit a positive correlation between entry cost and persistence in the market. When examining confidence biases, the model suggests that entrepreneurs whose biases cause them to perform the worst post-entry will be most likely to enter, that pre-entry learning induces a positive correlation between distinct confidence biases among entrants, and that exit changes the prevalence of certain biases in the surviving population of entrants over time. Our study also speaks to recent work on pre-entry experience that documents the transfer of knowledge from parent to progeny firms, suggesting that, in addition to inheritance, differential performance may also be the result of heterogeneity in the length and quality of pre-entry learning during which an opportunity is assessed.

Learning-By-Participating in Decision-Making: Broadening Participation, Narrowing Feedback

A central tenet of work in the Carnegie School tradition is the notion of “learning-by-doing”— organizations learn over time through feedback. In this paper we argue that the learning-by-doing account overlooks the fact that an organization’s decision-making structure is often participatory—i.e., organizational decisions often involve multiple individuals aggregating opinions through a process such as voting. In such contexts, individuals in the organization do not themselves learn-by-doing. Rather, when participating in the decision, they may vote for an alternative that is different from the one eventually selected by the organization. A key consequence of this is that these individual participants do not always receive feedback on their own choices; rather, they receive feedback on the choice made by the organization. We call this “learning-by-participating,” and we seek to understand the implications of this form of learning by comparing it to learning-by-doing, where an individual in the organization (such as the CEO) makes decisions on her own. Using a computational model of decision-making under uncertainty, we find that learning-by-participating leads to distinct patterns of individual learning that create trade-offs at the individual and organizational-levels. For example, while learning-by-participating is beneficial with respect to organization-level performance, it causes a minority of individuals within the organization to hold overly-optimistic views of low-payoff alternatives. We discuss the implications of our findings for research on learning and information aggregation.

E Pluribus Unum: Organizational Size and the Efficacy of Learning

Learning from experience is a central theme in the management literature. While in general experiential learning is viewed as efficacious, the literature increasingly points to the difficulties inherent in the learning process — many of which stem from a deficit of information about the merits of alternative solutions. It seems plausible that larger organizations, with their capacity to simultaneously pursue a variety of potential solutions to a given challenge, may overcome this deficit. Such a perspective suggests that the efficacy of an organizations learning process should be an increasing function of organizational size. While this logic is intuitively appealing, we find that it does not fully capture the nuances of the organizational learning process. We employ a computational model and find that larger organizations, as characterized by their scale in pursuing parallel initiatives: (a) explore less than smaller organizations, (b) are less likely to discover the very best alternative, and yet (c) on average identify better alternatives. Increasing the number of parallel initiatives guides the search process towards viable alternatives, but it does so at the cost of inhibiting search breadth. Thus, in our model, the characteristics of learning by larger organizations do not result from differences in inertia or incentives that may impede learning and innovation, but rather from the properties of the organizational learning process itself.

How Can Imitation Increase Inter-Firm Heterogeneity?

Imitation is thought to lead to a decrease in inter-firm heterogeneity. This outcome rests on the assumption that imitation engenders only one type of implication for the imitating firm: an endowment of knowledge (technologies, product designs, strategies, etc.) about successful practices that makes the imitator more similar to its target. Yet research in the Carnegie tradition points to a second mechanism — imitation moderates a firms post-imitation adaptation, which we term the generative effect of imitation. In this paper, we use a computational model to examine the implications of this dual role of imitation for inter-firm heterogeneity. As intuition suggests, we find the endowment effect of imitation reduces inter-firm heterogeneity. However the generative effect of imitation is double-edged: it tends to make firms more similar in practices but more different in performance. Our results suggest that the generative effect of imitation produces a bimodal performance distribution; some firms achieve significant benefits from imitation, while many others find themselves significantly worse-off than they would have been had they foregone imitation. Because of its generative effect, imitation can lead to an increase in performance heterogeneity.