Title
Financial Digitalization and Urban Green Innovation: Evidence from Panel Quantile ARDL Model
Authors
Abstract
Purpose – Digital finance is spreading unevenly across cities, and whether it narrows or widens the green-innovation gap is unknown because existing evidence is mean-based. This study asks where in the distribution of green patenting digital finance matters, for 16 large U.S. metropolitan areas observed annually over 2000–2025.
Design/methodology/approach – A panel quantile ARDL model, derived from a directed-innovation model with financing frictions, is estimated in error-correction form. Linearity, cointegration and spurious-regression checks precede estimation, and inference uses metro-level cluster and wild cluster bootstraps.
Findings – The long-run semi-elasticity is positive at every quantile and rises monotonically, roughly doubling across the interquartile range from 0.1370 to 0.2860. Error correction is negative throughout and faster in the upper quantiles, so leading metros gain more and converge sooner.
Originality/value – This is the first metropolitan-level distributional treatment of the digital-finance and green-innovation link, and it shows that the substitution and complementarity accounts are opposite signs of one cross-partial derivative that only a quantile design can identify.
Implications – The estimand is a decision input rather than a descriptive average: it tells an agency allocating scarce innovation funds what an extra unit of digital finance buys in a metro of given rank, which is the sense in which this is a decision-sciences contribution. Because gains accrue where capacity is already deep, promoting digital finance is necessary but not sufficient.
Keywords
digital finance, green innovation, quantile ARDL, panel cointegration, metropolitan areas
Classification-JEL
G20, O31, Q55, R11, C31
Pages
188-219
How to Cite
Laurinavicius, A., Laurinavicius, A., Salman, A., Razzaq, M. G. A., & Ghanem, M. E. A. (2026). Financial Digitalization and Urban Green Innovation: Evidence from Panel Quantile ARDL Model. Advances in Decision Sciences, 30(4), 188-219.