PurposeThe purpose of this paper is to examine the dynamics between banking penetration, infrastructure development and regional growth within a multivariate framework in 23 Indian states over the period 2000-2012.Design/methodology/approachThe study employs the multivariate panel data framework to analyze the dynamics between banking penetration, infrastructure development and regional growth within the vector error correction model (VECM) framework.FindingsThe findings confirm the long-run equilibrium relationship between banking penetration, infrastructure and income for the panel. Long-run income elasticity of infrastructure, estimated using Panel dynamic ordinary least square, is positive, statistically significant and has a value of 0.1531. Further, results show bidirectional causality between income and aggregate infrastructure and unidirectional causality running from banking penetration to income and aggregate infrastructure in the long run. However, there is unidirectional causality running from income to banking penetration and aggregate infrastructure and from banking penetration to aggregate infrastructure in the short run.Research limitations/implicationsThe study mainly concentrates on the 2000-2012 period and includes transportation (roadways and railways), energy (including electricity) and telecommunication as indicators for infrastructure, as the data for these sectors are easily available at the state level. Second, this study employs the panel data technique as it has a shorter data count.Practical implicationsIn order to minimize the existing regional disparity in a developing India, national infrastructure policies should be aimed toward improving the overall access to as well as the quality of infrastructure (existing as well as newly planned). Further, widening the banking outreach at the bottom level may further help the economy as well as the infrastructure sector in mobilizing long-term finances for productive investments, in order to have a balanced, more inclusive and faster growth in the long run.Originality/valueThe study employs panel unit root, cointegration and Granger causality tests within the panel VECM framework to explore the dynamics among the system variables. Further, the study creates a composite index of infrastructure with principle component analysis.