Does easy availability of cash effect corruption? Evidence from panel of countries

Using annual panel data of 54 countries for the period 2005-13, we examine whether cash in circulation, both aggregate and large denominated banknotes, affects the level of corruption in a country. Standard panel data models like pooled OLS, random effect and system GMM suggest that the ratios of (i) aggregate currency in circulation to M1 and, (ii) large denominated banknotes to M1 are both statistically significant determinants of corruption. Tests for reverse causality within a panel Granger framework reveal unidirectional causality of the first variable with corruption, but a bi-directional one with the second variable. These findings suggest that the central banks should try to limit the supply of banknotes of large denomination.