Economics is producing more research than ever while losing policy influence, study finds

Economics is producing more research than ever before, yet that research appears to be losing its grip on the real world. A new study published in the journal Digital Economy and Sustainable Development argues that the discipline has entered a two-stage crisis it calls the “Great Digital Decoupling”: a widening gap between the sheer volume of academic economic output and the macroeconomic health of the societies that output is meant to serve. Analyzing data from 2013 to 2023, the author reports an approximately 80 percent increase in indexed academic publications, relevant digital patents, and economics graduates over the period, even as productivity growth in advanced economies averaged below 1 percent and global debt-to-GDP ratios climbed to record levels exceeding 300 percent, citing figures from the Institute of International Finance.

The study, an open-access article by Ahmed Shalaby, does not claim that more research caused economic stagnation. Its statistical evidence is explicitly correlational. The correlation coefficient between publication volume and GDP growth was just +0.08 with a p-value of 0.25, which is statistically insignificant. In other words, across the decade examined, the growth of the economics knowledge industry showed essentially no measurable association with aggregate growth outcomes. The author frames this as evidence that the marginal utility of additional academic output, under current paradigms, has diminished in terms of driving aggregate growth, while carefully noting that the 80 percent figure reflects growth within the selected databases rather than a universal constant.

The first stage of the decoupling, termed the Conventional Decoupling, traces back to the aftermath of the 2008 global financial crisis. The paper argues that prevailing macroeconomic models, particularly dynamic stochastic general equilibrium models built on assumptions of equilibrium and representative rational actors, were structurally incapable of foreseeing the systemic financial contagion that precipitated the collapse. Subsequent shocks, including the European sovereign debt crisis and the COVID-19 pandemic, further exposed what the author describes as deep-seated fragilities in global labor markets, public finance, and demographic structures. Meanwhile, the institutional apparatus of economics flourished: the RePEc repository now indexes more than five million publications from over 9,100 institutions worldwide, and the number of economics graduates and specialized research centers continued to climb.

The second stage, the Digital Decoupling, is the study’s central and more provocative claim. The author argues that the rise of artificial intelligence and digital transformation is not merely a continuation of the earlier trend but a qualitative escalation. The digital economy, with its network effects, data-as-capital, and algorithmic governance, operates on principles that traditional equilibrium-based models cannot capture, the paper contends. Data, for instance, is treated in conventional production functions as a secondary input, but in the digital economy it functions as the core asset, driving value creation in ways that GDP cannot measure. The author characterizes this as an epistemological mismatch rather than a simple forecasting failure, and describes the resulting threat to the discipline’s relevance in stark terms.

One of the study’s more concrete empirical contributions is what it calls a policy relevance audit. Using citation analysis of IMF Working Papers and World Bank Policy Research Working Papers, the author measured the share of citations in these reports that reference articles from academic economics journals. Over the study period, that share declined by approximately 38 percent, a figure the paper attributes to a relative drop in academic citation shares compared with internal institutional reports and grey literature. The decline showed a negative correlation with productivity growth, with a correlation coefficient of -0.22. The author suggests that as the economy becomes more digital, policymakers are increasingly bypassing traditional academic economics, likely because its models fail to capture digital dynamics, while acknowledging that citation analysis captures only formal influence and misses informal advisory roles.

The study also examined how far the discipline itself has adapted to digital paradigms. A manual content analysis of the top 50 global economics PhD programs, ranked by RePEc, coded whether their core curricula included dedicated courses on artificial intelligence and machine learning, agent-based modeling, and network science. By 2025, the integration of AI and data science modules had reached only about 25 percent of these programs. The author argues that this sluggish adaptation deepens the decoupling, since the models that failed to predict the last crisis are, in his words, now being used to analyze an economy they are structurally blind to. A correlation of -0.28 between publication volume and policy influence is presented as supporting the view that academic incentives favor technical sophistication over practical relevance.

Methodologically, the study combines descriptive trend analysis, panel regressions with country and year fixed effects, Granger causality tests described as suggestive evidence of temporal precedence, and qualitative thematic coding of American Economic Association survey responses on curriculum relevance and policy influence. The regression framework relates outcomes such as GDP growth and debt-to-GDP to knowledge production metrics, including publications and patents per capita, alongside digital adaptation metrics and controls for political stability, a COVID-19 pandemic dummy, and commodity terms of trade. The author is explicit about the limits: panel regressions with fixed effects cannot establish definitive causation, omitted variable bias such as a global decline in institutional quality could drive the results, and the focus on top-50 PhD programs and English-language journals introduces a Global North bias that may not represent economics globally.

Beyond diagnosis, the paper proposes a way forward: the Digital Sustainable Growth Model, or DSGM, presented not as a finished model but as a research agenda. In a DSGM-aligned framework, the author proposes expanding the standard production function to explicitly include data and network effects as distinct factors of production, alongside capital and labor. Data would be characterized by non-rivalry and increasing returns to scale, while network effects would capture value generated by platform connectivity and algorithmic coordination. Unlike traditional total factor productivity, which treats technological progress as exogenous, the model posits that it becomes endogenous to the quality of digital infrastructure and the ethical governance of algorithms. The author concedes that the paper does not perform a full econometric estimation of this expanded function, due to data limitations on global data stocks, and that critics are correct that the framework currently lacks the mathematical rigor of established growth models.

The study situates itself within a broader critical literature. It draws on arguments that the dominance of GDP as a policy objective is itself a cause of the decoupling, since GDP counts environmental destruction as a positive gain and ignores the value of digital services, unpaid labor, and social well-being. It also reviews alternative paradigms, including complexity economics with its agent-based modeling tools, doughnut economics with its social foundation and ecological ceiling, and digital and AI economics, arguing that these remain fragmented and in need of synthesis. The author describes the DSGM as a blueprint for a “3D Economy” that is dynamically complex, digitally native, and developmentally sustainable, and points to testable hypotheses, such as whether policies restricting data flow would have measurable negative impacts on total factor productivity distinct from traditional capital constraints.

The paper’s conclusions are deliberately framed as a warning rather than a settled verdict. The author acknowledges that the analysis establishes correlation, not incontrovertible causation, and that the DSGM is a nascent framework whose full operationalization presents practical challenges. He also concedes that some may argue the discipline is already self-correcting, while responding that the slow pace of curricular change and the continued dominance of traditional methods in top journals suggest internal reform is insufficient to meet the urgency he perceives. Among the recommendations offered are curricular experimentation that complements rather than replaces traditional theory, diversified evaluation criteria for journals and funders including greater weight on reproducibility and code availability, and for policymakers, diversified advisory teams and multi-dimensional dashboards that track well-being and sustainability alongside GDP. Whether the profession accepts the diagnosis, the study’s empirical core stands as a measurable claim: over a decade in which economic research output grew by roughly 80 percent, its statistical association with growth was indistinguishable from zero, and its formal citation presence in major policy institutions fell by more than a third.

Subject of Research: The decoupling of academic economics research output from macroeconomic outcomes and policy influence in the digital era

Article Title: The great digital decoupling: an existential challenge for economics and economists

Article References: Shalaby, A. (2026). The great digital decoupling: an existential challenge for economics and economists. Digital Economy and Sustainable Development, 4(1), Article 9. https://doi.org/10.1007/s44265-026-00084-x

Image Credits: AI Generated

DOI: 10.1007/s44265-026-00084-x

Keywords: economics, digital economy, Great Digital Decoupling, policy influence, citation analysis, GDP, productivity growth, artificial intelligence, DSGM, complexity economics, debt-to-GDP, economic methodology