Structural Breaks and Growth Regimes in India's UPI Ecosystem: A Piecewise Regression Analysis of Policy Push, Demonetisation, and Pandemic Effects

Articles in Press

Document Type : Original Article

Authors

Suman Das ¹

Kaju Nath ²

Amit Bikram Chowdhury ²

Nabendu Debnath ²

Rajib Mallik ²

¹ NIT Calicut

² NIT Agartala

10.22067/ijaaf.2026.99693.1701

Abstract

India's Unified Payments Interface (UPI) scaled from zero to 22,347 million monthly transactions between April 2016 and April 2026, with a decade-cumulative transaction value of approximately ₹1,099 lakh crore. Using NPCI's monthly series, this paper examines three candidate structural breaks in UPI's growth path, coinciding with Demonetisation (November 2016), COVID-19 onset (March 2020), and the post-lockdown recovery (July 2020), through a sequence of log-linear, structural-dummy, piecewise, and quadratic trend models. Because these dates are drawn from known policy events rather than detected statistically, we treat the resulting break tests with caution and report a supplementary search that does not fully corroborate them, leaving a confirmation-bias risk we do not consider resolved. The piecewise specification fits the data substantially better than a log-linear benchmark, though its residuals show meaningful serial correlation that widens the true uncertainty around individual coefficients; we address this with heteroskedasticity- and autocorrelation-robust standard errors throughout. Rather than treating the estimated post-COVID shift as a fixed multiplier, we compare the fitted growth trajectories directly and find the gap is largest immediately after the shock and narrows substantially thereafter, a pattern we describe as a substantial but non-permanent change in trajectory. A quadratic specification points to decelerating growth with a turning point around April 2024, consistent with, though not sufficient to establish, S-curve-style maturation. We interpret these patterns as temporal associations rather than causal effects, and discuss the resulting identification, serial-correlation, and generalisability limitations for comparable digital payment programmes elsewhere.

Keywords

UPI

digital payments

structural break

piecewise regression

Digital India

Subjects

• Finance

---