BACKGROUND: Within Ulaanbaatar, Mongolia, risk factors for pneumonia are concentrated among children living in informal settlements comprised of temporary shelters (gers). We used pneumococcal carriage surveillance among children from formal and informal settlements hospitalised with pneumonia to evaluate the direct and indirect effects of 13-valent pneumococcal conjugate vaccine (PCV13) against vaccine-type (VT) pneumococcal carriage following a phased introduction of PCV13. METHODS: We enrolled and collected nasopharyngeal swabs from children 2-59 months of age presenting to hospital. Pneumococci were detected using lytA qPCR and serotyped using microarray on a random monthly selection of swabs between November 2015 and March 2019 from two districts in Ulaanbaatar. PCV13 status was determined using written records. We quantified the associations between individual PCV13 status (direct effects) and district-level PCV13 coverage (indirect effects) and VT carriage using generalised estimating equations and explored interactions by settlement type. FINDINGS: A total of 1 292 swabs from 6 046 participants were tested for pneumococci. Receipt of PCV13 and increasing PCV13 coverage independently reduced the risk of VT carriage. For each percent increase in PCV13 coverage, the adjusted odds of VT carriage decreased by 1•0% (OR 95% CI 0•983-0•996; p=0•001), with a predicted decrease in VT carriage rate from 29•1% to 13•1% as coverage reached 100%. There was a trend towards a slower decline within informal settlements (p=0•100). Adjusted PCV13 vaccine effectiveness against VT carriage was 39•1% (95% CI 11•4-58•1%, p=0•009). INTERPRETATION: Substantial indirect effects were observed following PCV13 introduction, including among children living within informal settlements. FUNDING: Bill & Melinda Gates Foundation; Gavi, the Vaccine Alliance.