Universal pneumococcal conjugate vaccine (PCV) programs began in Indigenous Australian children in 2001 and all children in 2005, changing to PCV13 in 2011. We used laboratory data for invasive pneumococcal disease (IPD) and coded hospitalisations for non-invasive pneumococcal community acquired pneumonia (PnCAP) to evaluate long-term impact. Annual incidence (per 100,000 population) was calculated for age-specific total IPD, PCV13-non7v serotypes and PnCAP by Indigenous status. Incidence in the pre-universal PCV7 (2002-2004), early-PCV7 (2005-2007), pre-PCV13 (2008-mid 2011) and post-PCV13 (mid 2011-2016) periods was used to calculate incidence rate ratios (IRRs). In the total population, all-age incidence of IPD declined from 11.8 pre-PCV7 to 7.1 post-PCV13 (IRR 0.61, 95% CI 0.59-0.63) but for PnCAP declined <1 year (IRR 0.34, 95% CI 0.25-0.45) and 1-4 years (IRR 0.50, 95% CI 0.43-0.57) but increased significantly ≥5 years (IRRs 1.08 to 1.14). In Indigenous people, baseline PCV13-non7v IPD incidence was 3-fold higher, amplified by a serotype 1 epidemic in 2011. By 2015-6, although incidence of IPD and PnCAP in children <5 years decreased by 38%, neither decreased in people ≥5 years. 15 years post PCV and 5 years post PCV13, direct and indirect impact on IPD and PnCAP differed by age and between Indigenous and non-Indigenous people, with potential implications for long-term PCV impact in comparable settings.