Less liquid markets for government bonds (LLMs) are characterized by many well recognized challenges which reduce the reliability of the classic Nelson-Siegel-Svensson (NSS) parsimonious approach. We document key stylized facts about government bond markets concerning liquidity, diversity of maturities available, bid-ask spread in price quotes, as well as price distortion in the very short end of the curve due to switch auctions. Based on these facts, we augment the NSS approach with model- and data-driven endogenous system of weights which permits reliable estimation of yield curves in LLMs. We apply our approach to the data for one of the largest European emerging markets: Poland. Through a battery of sensitivity analyses we show that there exists a class of weights that systematically gives better results than the classic NSS approach. The best fit weights have at least the same weight for the short end of the curve as a sum for all other tenors of bonds. It proves that inferring from the liquidity in particular maturities raises the information content and quality of yield curve estimation, which links our results to the expectation hypotheses. Unlike findings for most mature markets (e.g. US), for Poland there is a limited domain where pure expectations hypothesis (PEH) cannot be ruled out. Moreover, expectations hypothesis (EH) holds in Poland for almost all horizons. The existence of term premia structure explains the differences between compounded rates of returns from shorter investments and longer term zero-coupon yields of corresponding maturity.