Currency Holdings, Bank Savings and Inflation Dynamics in Nigeria: A VAR Analysis of Currency in Circulation, Currency Outside Banks and Savings Deposits

This study examined the dynamic interrelationships and causal relationships among currency in circulation (CIC), currency outside banks (COB), savings deposits (SD), and inflation (INFL) in Nigeria using a Vector Autoregressive (VAR) framework. Monthly data covering January 2010 to May 2025, comprising 182 observations, were analysed. The Augmented Dickey–Fuller test established that all variables were stationary after first differencing. The Johansen cointegration test revealed no significant long-run relationship among the variables, with a trace statistic of 42.2586 (p = 0.1516) and a maximum-eigenvalue statistic of 24.3520 (p = 0.1229). VAR lag-order selection criteria indicated that lag 3 was optimal, as the Akaike Information Criterion (AIC = 39.42), Final Prediction Error (FPE = 1.317 × 10¹⠷), and Hannan–Quinn Information Criterion (HQIC = 39.80) attained their minimum values at lag 3, although the Schwarz/Bayesian Information Criterion (SC/BIC = 40.20) selected lag 2. Based on the majority of the selection criteria, a VAR(3) model was adopted. The results showed significant short-run interactions among the monetary variables. The third lag of CIC significantly influenced CIC (β = 0.8679, t = 3.3614) and COB (β = 0.8292, t = 2.9623), while the first lag of COB significantly affected CIC (β = 1.1285, t = 3.5169) and INFL (β = 7.20 × 10⠻⠷, t = 2.3385). Savings deposits exhibited strong persistence, with their first lag significantly affecting current SD (β = 0.9219, t = 12.1265). Inflation was also highly persistent, with its first lag producing a significant positive effect (β = 1.6025, t = 21.9628). Granger causality tests established bidirectional causality between COB and CIC, with COB Granger-causing CIC (χ² = 15.3970, p = 0.0015) and CIC Granger-causing COB (χ² = 10.0911, p = 0.0178). Variance decomposition showed that at the tenth period, CIC explained 86.28% of its own forecast error variance, while COB accounted for 12.46%. The findings revealed strong short-run interactions among Nigeria's monetary and financial variables, particularly between CIC and COB. The study recommends enhanced monitoring of currency outside banks and policies that promote banking-sector intermediation and savings mobilisation to support effective monetary and inflation management