“…Also, other than by taking 1-year lags, we do not address possible endogeneity problems, due to the difficulty of finding good instruments. Our results should hence not be interpreted as demonstrating causality, a caveat that also applies to earlier studies of LCBM development in SSA (Berensmann et al, 2015;Essers et al, 2016;Mu et al, 2013) and other regions (Bhattacharyay, 2013;Burger & Warnock, 2006;Claessens, Klingebiel, & Schmukler, 2007;Eichengreen, Panizza, & Borensztein, 2008). Nonetheless, we believe the econometric analysis that follows contributes to our understanding of SSA LCBMs and helps to lay the groundwork for better identification of causal relations as wider samples and longer time series become available.…”