The Multi-Level Perspective (MLP) has been the dominant framework for understanding sustainability transitions for two decades. It describes how niche innovations interact with incumbent socio-technical regimes and landscape pressures to produce systemic change — the transition from horses to automobiles, from coal to renewable energy, from linear to circular economies. But the MLP has always been better at explaining transitions after they happen than predicting when they will happen. Three recent lines of research are pushing the framework from retrospective narrative toward prospective analysis.
Speciation and Aggregation
The concept of projects as speciation and aggregation in transitions offers a micro-level mechanism that the MLP's macro-level framework has lacked. Individual projects — pilot programs, demonstration sites, experimental initiatives — function as the "species" of innovation ecosystems. They speciate (diversify into variants) during the niche phase and aggregate (consolidate into dominant configurations) during the regime shift. The rate of speciation relative to aggregation may be a leading indicator of transition timing: when niche innovations stop diversifying and start consolidating, the transition may be approaching a tipping point.
Socio-Technical Experiments
Research on autonomous mobility transitions examines the role of socio-technical experiments — real-world trials of autonomous vehicles in specific cities and contexts — in shaping transition dynamics. The systematic review reveals that these experiments serve multiple functions beyond technical testing: they build social acceptance, reveal regulatory gaps, create stakeholder coalitions, and generate the practical knowledge that transitions require.
The key finding is that the transition toward autonomous mobility is not progressing uniformly but through a patchwork of local experiments with highly variable outcomes. Some experiments accelerate transition by demonstrating feasibility and building political support. Others stall it by revealing safety risks or public opposition. The aggregate pattern is one of distributed exploration rather than coordinated transformation — a dynamic that the MLP framework captures at a high level but struggles to model with predictive precision.
Circular Economy Mainstreaming
The challenge of mainstreaming circular economy innovations illustrates the regime resistance that MLP predicts but rarely quantifies. Circular economy solutions — product-as-service models, remanufacturing, industrial symbiosis — exist as proven niche innovations. The technical feasibility is established. The economic viability is demonstrated in specific contexts. Yet mainstreaming remains elusive because the incumbent regime — linear production, disposable consumption, externalized waste costs — is stabilized by complementary institutions, infrastructure, and consumer habits.
The transition research community is increasingly recognizing that predicting tipping points requires not just understanding niche innovation dynamics but modeling regime vulnerability — the conditions under which incumbent systems lose their institutional stabilization. Landscape pressures (climate change, resource scarcity, regulatory tightening) weaken the regime. Niche innovations accumulate. At some point, the combination crosses a threshold and the system tips. The theoretical question that remains unanswered is whether this threshold can be identified in advance or only recognized in retrospect.
The Prediction Challenge
The fundamental question remains open but is becoming more tractable. Three approaches show promise. Agent-based models simulate the interactions between niche innovations, regime incumbents, and landscape pressures, generating distributions of possible tipping points under different scenarios. Network analysis examines the structure of innovation ecosystems for signatures of approaching transitions, including increasing connectivity between niche actors and weakening of regime-supporting coalitions.
The most promising but least developed approach draws on complexity science: treating socio-technical systems as complex adaptive systems and looking for early warning signals that precede critical transitions, such as increased variance, critical slowing down, and flickering between alternative states. These signals have been observed in ecological, financial, and climate systems. Whether they generalize to socio-technical transitions remains an empirical question, but preliminary evidence from energy system transitions is encouraging.
The practical implication for policymakers is that transition management requires monitoring not just the performance of niche innovations but the structural dynamics of the entire socio-technical system, including the stability of the incumbent regime, the diversity of niche alternatives, and the intensity of landscape pressures.
Beyond the MLP
The MLP framework itself is evolving under the pressure of these empirical findings. The original three-level model (niche, regime, landscape) is being supplemented with more granular concepts: transition pathways (how different configurations of niche-regime-landscape interaction produce different transition dynamics), transition intermediaries (organizations that facilitate connections between levels), and transition politics (the power struggles that determine which niche innovations are supported and which are suppressed).
These refinements do not replace the MLP but add the analytical precision needed for policy-relevant prediction. A policymaker who can identify which transition pathway a system is following, which intermediaries are active, and which political dynamics are accelerating or blocking change has substantially better foresight than one working with the broad strokes of the original framework. The question is no longer whether the MLP is useful but whether it can be made precise enough to guide real-time policy decisions in systems undergoing active transformation.