Knowledge that Transforms

We study the nature and effects of cultural biases in choice under risk and uncertainty by comparing peer-to-peer loans the same individuals ( lenders ) make alone and after observing robo-advised suggestions. When unassisted, lenders are more likely to choose co-ethnic borrowers, facing 8% higher defaults and 7.3pp lower returns. Robo-advising does not affect diversification but reduces lending to high-risk co-ethnic borrowers. Lenders in locations with high inter-ethnic animus drive the results, even when borrowers reside elsewhere. Biased beliefs explain these results better than a conscious taste for discrimination: lenders rarely override robo-advised matches to ethnicities they discriminated against when unassisted.

This paper studies the transmission of monetary policy to the stock market through investors’ discount factors. To isolate this channel, we investigate the effect of US monetary policy surprises on the ratio of prices of the same stock listed simultaneously in Hong Kong and Mainland China. We identify a strong discount rate channel driven exclusively by cycle-amplifying surprises, defined as rate cuts during easing cycles and surprise hikes during tightening cycles. A 100 basis point of such cycle-amplifying surprise induces a 30 basis point change in the price ratio within five days.

We document implicit extrapolation in investment decision-making that exceeds the extrapolation inferable from stated expectations. Locally experienced returns predict individual real-estate investment decisions even conditional on an investor’s forecasted home-price growth and risk aversion. Moreover, estimates of this experience effect on investment are larger than implied by the combined effect of past returns on stated expectations and stated expectations on investment. We demonstrate that heterogeneous forecast confidence helps explain why many investors rely on past returns over their survey-elicited forecasts. As their rationale, such survey respondents frequently cite intentional extrapolation or a lack of confidence in other belief factors.

Disagreement about macroeconomic fundamentals accounts for only part of the disagreement about future interest rates, creating a “disagreement correlation” puzzle. This puzzle arises because standard equilibrium models with belief differences predict a strong link between asset return disagreement and fundamental disagreement, a link not supported by the data. We address this puzzle by introducing a model where disagreement about future demand for savings—driven by disagreement over the prevalence of patient versus impatient investors in the economy—generates asset return disagreement. Our mechanism produces stochastic yield volatility, time-varying bond risk premia, and an upward-sloping yield curve. Empirically, we construct a proxy for demand disagreement by isolating the component of yield disagreement unrelated to disagreement about macro-fundamentals. This proxy is positively related to yields and their volatilities, and predicts future bond risk premia, consistent with the predictions of our demand disagreement model.

Research on the impact of inward foreign direct investment (IFDI) on new venture creation has primarily focused on the country level, often yielding inconsistent findings. To reconcile these inconsistencies, we develop a theory on the influence of IFDI stock on new venture creation at the industry-regional level, offering three possible complementary explanations. Drawing on learning and competition theories, we propose that IFDI in a specific industry and region has an inverted U-shaped effect on new venture creation, driven by the dual forces of learning opportunities and competitive threats anticipated by prospective entrepreneurs. Additionally, we found that IFDI in related industries in the focal region and IFDI in the focal industry in adjacent regions positively influence new venture creation. Furthermore, we argue that these effects are amplified in regions with more developed non-state economy. Analyzing data on all registered firms (2013−2023) collected from China National Enterprise Credit Information Publicity System provides robust support for our hypotheses. The findings make important theoretical contributions to research on foreign direct investment and entrepreneurship, as well as to the literature on learning and competition.

Persistent social and environmental challenges require social entrepreneurs to work towards systems change, which is defined as fundamental shifts in social processes, practices, and structures. However, we do not understand well why some social entrepreneurs achieve systemic changes, while others fail to do so. In particular, we do not know which mechanisms help social entrepreneurs produce systemic transformations in society. This article advocates for harnessing Giddens' structuration theory to improve our understanding. I develop a structuration model of systems change through social entrepreneurship, which shows that connected modalities, that is, intricate combinations of (1) interpretative schemes, (2) resources, and (3) norms, act as the key mechanisms that enable systemic effects. I argue when social entrepreneurs mobilize connected modalities, they can realize high integrativeness, i.e., cover a wide variety of interconnected issues and trigger shifts at different societal levels. Connected modalities also enable high inclusiveness, i.e., unify a wide diversity of stakeholders. Both elements are needed to achieve systems change. I support the model's development with illustrations from research on social entrepreneurship with and without a systemic character. Thereby, I alter how we think about social entrepreneurship and effective ways of supporting it when the aim is to generate systems change. Executive summary Persistent social and environmental challenges are affecting the world. We tend to assume social entrepreneurs help address them and promote so-called systems change. Systems change means creating fundamental shifts in social processes, practices and structures so that problems are prevented or new kinds of lasting and major solutions surrounding those problems can emerge. Unfortunately, we do not understand well why some social entrepreneurs achieve systemic changes, while others fail to do so. Especially, we lack knowledge about which mechanisms help social entrepreneurs produce systemic transformations in society. This article uses Anthony Giddens' structuration theory to develop an explanation. It introduces the theoretical idea of connected modalities. These are close and strong combinations of new (1) ideas, mindsets or imaginations, (2) resources, and (3) norms and values. When combining these modalities, that is, different ways of acting within society, social entrepreneurs can realize high integrativeness. This means they can affect several societal fields, industries, or issues at the same time. Combining modalities also helps social entrepreneurs achieve high inclusiveness. This means they can activate and work with many different stakeholders. Both elements are needed to achieve systems change. Research about microfinance underpins my arguments. Microfinance achieved systems change not primarily because it was a commercially successful product but because it triggered changes in various social sub-systems—ranging from banking to national government regulations to development assistance. It did so by establishing the new idea that poor people were bankable, by mobilizing resources across the philanthropic and business sectors, and by establishing a new social norm of empowerment in poverty alleviation. I discuss other cases that were not able to connect modalities in the same way and therefore faced restrictions in promoting systems change. The approach proposed in this article thus changes how we think about social entrepreneurship and effective ways of supporting it, when the aim is to generate systems change. In particular, it encourages more support for entrepreneurs in creating and applying connected modalities to cover various issue areas and mobilize diverse stakeholders, instead of mimicking the incubation practices of its commercial counterpart. While these are important for creating viable organizations, established practices will not be sufficient for changing systems fundamentally.

In recent years, the rapid adoption of Generative AI tools by entrepreneurs is transforming entrepreneurial ideation processes. Powered by increasingly sophisticated algorithms and massive computing facilities, Gen AI systems are capable of generating extraordinarily creative ideas that often surpass the abilities of human entrepreneurs. Yet, despite these benefits, Gen AI systems also create a series of important epistemic risks for entrepreneurs, most notably: algorithmic hallucinations and ‘alien minds.’ In the near term, the tendency of Gen AI systems to ‘hallucinate’ new ideas that appear plausible but lack a logical or factual basis amplifies the risks that entrepreneurs will invest valuable time, effort, and resources in pursuit of flawed ideas. As the capabilities and intelligence of these systems continue to grow, however, entrepreneurs also face an emerging risk of falsely rejecting breakthrough ideas recommended by ‘alien minds’ they do not understand. For entrepreneurs, the opaque processes through which Gen AI systems generate new ideas create an ideator's dilemma where entrepreneurs do not know if a Gen AI idea is a true, breakthrough innovation or simply a hallucinated mirage. In this study, we extend emerging theory on entrepreneurial work to examine the complementary roles of two distinct types of judgments – possibility and plausibility judgments – in enabling entrepreneurs to evaluate Gen AI ideas. Towards this end, we integrate these judgments into a comprehensive Popperian approach to entrepreneurial work, enabling entrepreneurs to more effectively address the inherent epistemic risks associated with using Gen AI in entrepreneurial ideation. In doing so, our study contributes important new insights regarding the fundamental role of entrepreneurial work in addressing the ideator's dilemma in entrepreneurial ideation processes. Executive summary Generative AI is fundamentally transforming how entrepreneurs identify and develop new venture opportunities. With 89 % of founders now using at least one AI model and 50 % employing four or more in daily ideation work, these tools have become integral to entrepreneurial innovation. However, this rapid adoption creates unprecedented challenges that demand new frameworks for evaluating AI-generated business concepts. The democratization of AI tools creates fundamental challenge for entrepreneurs: if every entrepreneur can generate breakthrough ideas at scale, where lies competitive advantage? Our research reveals that judgment – not idea generation – becomes the scarce resource. As ventures pursue increasingly speculative concepts to maintain competitive edge (evidenced by Y Combinator's shift towards deep tech ventures), the ability to efficiently evaluate and actualize AI concepts becomes paramount. At the same time, these fundamental challenges are amplified by the growing sophistication of the reasoning capabilities of frontier Gen AI systems. To illustrate these challenges, we posed a single question to the 7 leading Gen AI models, which all provided forecasts estimating the probability within five years that AI systems will generate venture concepts beyond expert entrepreneurs' full comprehension. Current systems already demonstrate this capability in specialized domains. The integration of multimodal models, real-time data, and agentic systems will accelerate this trend. For entrepreneurs, this means reconceptualizing the entrepreneurial process. Rather than competing on idea generation – where AI increasingly dominates – successful ventures will differentiate through superior judgment, interpretation, and actualization capabilities. The entrepreneurs who thrive will be those who can effectively orchestrate human-AI collaboration, maintaining human agency while leveraging AI's superhuman pattern recognition. The ideator's dilemma: Hallucinations and alien minds Building on these arguments, one of the central contributions of our study examines what we refer to as the Ideator's Dilemma. Given the training methods – especially the growing use of deep reinforcement learning methods – leading AI systems are creating two critical, interrelated epistemic risks for entrepreneurs using these tools in ideation processes: First, algorithmic hallucinations occur when AI systems generate plausible-sounding ideas that lack factual or logical basis. These can lead entrepreneurs to waste valuable resources pursuing impossible ideas. Second, entrepreneurs are facing a growing set of challenges in addressing the “alien minds” problem – where AI systems generate genuinely breakthrough concepts through reasoning processes so advanced or unconventional that human entrepreneurs cannot fully comprehend them. For entrepreneurs, these two epistemic risks combine to create what we term the “ideator's dilemma”: entrepreneurs cannot determine whether an AI-generated idea represents a revolutionary opportunity or a convincing mirage. As AI systems grow more sophisticated, particularly through deep reinforcement learning methods that operate without human oversight, this dilemma intensifies. Recent developments like DeepSeek's autonomous discovery of novel problem-solving strategies and AlphaDev's counterintuitive algorithmic breakthroughs demonstrate that AI systems are already generating solutions that initially appear incorrect but prove revolutionary. A novel framework for AI-enabled entrepreneurship To address Ideator's Dilemma, our study extends recent work in the field of entrepreneurship to build a Popperian approach to address these challenges through two complementary types of entrepreneurial judgment: Possibility Judgments focus on falsification – rigorously testing whether necessary conditions exist for an idea to be actualizable. Entrepreneurs should ask “why won't this work?” before investing resources. This approach quickly eliminates algorithmic hallucinations by establishing clear boundaries between possible and impossible new venture concepts. The key is developing bold, testable conjectures that can be decisively falsified if the concept is fundamentally flawed. Plausibility Judgments employ corroboration to identify sufficient conditions for actualization. Once an idea passes falsification tests, entrepreneurs assess how to configure enabling factors – market conditions, technical infrastructure, regulatory environment – to bring the concept to market. This stage addresses the alien minds problem by translating opaque AI reasoning into actionable business strategies. The Popperian approach we outline provides a practical methodology for navigating this transition. To illustrate the benefits of such an approach, we provide an in-depth example of the Ideator's Dilemma through a Gen AI idea building on an extremely speculative, advanced integration of ‘new quantum physics.’ This idea illustrates the unique epistemic risks entrepreneurs already face in evaluating and vetting Gen AI ideas to identify algorithmic hallucinations while addressing the challenges of verifying an advanced concept that is well outside of our current capabilities. Overall, our study illuminates an important set of fundamental shifts in how opportunities are identified, evaluated, and actualized. Startups that develop robust frameworks for human-AI collaboration in ideation will capture disproportionate value in the emerging economy, while those that fail to adapt risk being overwhelmed by the Ideator's Dilemma.

Entrepreneurial pivots can mean the difference between venture failure and survival. Mentor feedback often illuminates the essential need to pivot, but the manner in which feedback is delivered may determine whether entrepreneurs act on it and pivot. Drawing on the feedback literature and construal level theory, we hypothesize that mentor feedback provided in a concrete manner will increase the likelihood that entrepreneurs will pivot. Further, we examine how temporal distance, mentor expertise distance, and entrepreneurial experience, affect entrepreneurs' responses to feedback. We test our hypotheses using a conjoint analysis with a validation study; gain additional insights from a qualitative study of mentors; and test a way to improve entrepreneurial receptiveness to abstract feedback from mentors through a second conjoint analysis design. In doing so, we contribute to the entrepreneurship literature on pivoting and mentoring.

This study investigates how support organisations for marginalised entrepreneurs (SOMEs), typically peripheral members within entrepreneurial ecosystems (EEs), facilitate the inclusive evolution of EEs. Employing boundary theory and ethnographic research conducted over three years within a refugee entrepreneurship support organisation in Birmingham, UK, this study identifies a four-stage boundary work process: Knowledge brokering, Boundary buffer spacing, Boundary object developing, and Boundary practice institutionalising. These interconnected strategic stages enable SOMEs to reconfigure the knowledge-cognitive, resource-opportunity, and social network exclusionary boundaries of EEs progressively, facilitating EEs' adaptation to marginalised entrepreneurs' diverse needs and pursuits within the overarching growth-orientation of EEs. Theoretically, this study introduces a “periphery-to-centre” model of inclusive evolution, expanding the prevalent centre-driven perspective of EE inclusive evolution, and demonstrates how inclusion could coexist with EE's growth-orientation because of SOMEs' boundary work. The study also unfolds enablers for such effective boundary work, emphasising the effects of SOMEs' dual knowledge capabilities, dual network embeddedness, institutional rhetoric, and the path dependency of evolution. Executive summary In entrepreneurial ecosystem (EE) research, an important but overlooked issue is how ecosystems can become more inclusive, allowing entrepreneurs from different backgrounds pursuing varied objectives to obtain the necessary support and resources. However, given that EEs often prioritise innovation and high-growth ventures, their ideologies and structures are hardly responsive to diverse entrepreneurs' distinctive pursuits and needs for support, causing their marginalisation. Despite community-based and non-profit support organisations for marginalised entrepreneurs (SOMEs) emerging to address this situation, their strategies and role in EEs' inclusion evolution remain understudied. Addressing this gap is critical as enhanced inclusion not only advances social equity but also strengthens EEs' resilience and innovation through the integration of diverse entrepreneurs and entrepreneurial activities. Leveraging boundary theory to conceptualise SOME's boundary work and how it affects the exclusionary boundaries (knowledge-cognitive, resource-opportunity, and social network boundaries) of EE, this paper develops a four-stage process model based on a three-year ethnographic study of a refugee entrepreneurship support organisation in Birmingham, UK. The study conceptualises a “periphery-to-centre” inclusive evolution pathway driven by SOMEs, expanding beyond the dominant “centre-to-periphery” evolutionary perspective prevalent in existing EE literature. SOMEs leverage their distinctive position on the EE's internal periphery to simultaneously understand both marginalised entrepreneurs' distinctive needs and aspirations, and the operational dynamics and value orientations of EEs. Based on this dual understanding, rather than directly challenging the growth-oriented values of EEs, SOMEs strategically mobilise these established priorities to legitimise their innovative boundary work, gradually restructuring exclusionary boundaries while respecting existing EEs' priorities. This study reveals four interconnected stages through which this boundary work unfolds, with corresponding evolution of EEs' exclusionary boundaries: (1) Knowledge brokering - translating MEs' values and needs into EE-compatible discourse, creating initial shifts in EE members' knowledge-cognitive boundary and establishing sole resource conduits into EE through SOMEs; (2) Boundary buffer spacing - establishing specialised subsystems with complementary actors (social service providers outside EE) to deal with marginalised entrepreneurs' distinctive needs, forming externally-driven temporary expansion of resource and network boundaries while reducing EE members' perceived risks in supporting MEs; (3) Boundary object developing - creating collaborative initiatives that satisfy diverse stakeholders' interests, enabling internally-driven structural boundary expansion with bidirectional knowledge exchange between EE members and marginalised entrepreneurs, and facilitating direct resource and network access in EE; and (4) Boundary practice institutionalising - embedding accumulated boundary practices across multiple ecosystem levels, transforming temporary interventions into self-sustaining and self-expanding boundary changes across all dimensions. The study also identifies critical enabling conditions for effective boundary work: SOMEs' dual knowledge capabilities; their dual networks with both social service providers and EE members; supportive rhetoric for economic diversity and inclusion in the institutional environment; and path-dependent factors where early boundary changes enable subsequent evolution. Moreover, this study challenges the binary opposition between growth orientation and inclusion in EEs, showing how boundary changes create structures facilitating bidirectional adaptation without sacrificing the growth objectives of EEs or forcing marginalised entrepreneurs to conform to mainstream definitions of success. These insights offer important implications for policymakers and practitioners. The study suggests that supporting “periphery-to-centre” evolution through stable funding and supportive rhetoric, while providing SOMEs with a blueprint for conducting progressive boundary work.

Amid accelerating global change, entrepreneurship is increasingly shaped by political, social, and technological shifts that cut across borders. Yet, despite these pressing realities, international entrepreneurship (IE) scholarship has paid limited attention to how such global phenomena impact entrepreneurial actors, behaviors, and circumstances. This suggests a disconnect between the evolving global context and current academic attention to the field. Drawing on our perspective as area editors for IE at the Journal of Business Venturing , we reflect on this gap and the concerns voiced across the scientific community about the field's scope and future. In response, we propose an agenda for future research that treats entrepreneurship as inherently international and, in turn, invites future contributions not only by IE researchers but also from the wider entrepreneurship community.