PurposeA significant number of published empirical studies of real estate auctions abound in real estate literature. These studies focus on the auction price formation process through the comparison of sales from different auctions to revenue from private negotiations. While these studies have made useful contributions to the literature, they are largely confined to data emanating from auctions in the UK format. Other auction formats have been virtually ignored; thus creating a gap in the literature. This study, therefore seeks to extend research in the auction price formation process through an empirical analysis of transaction prices from the perspective of another auction format: first sealed bid auction.Design/methodology/approachThe data for this study consist of a sample size of 120 residential properties sold at auction of the Federal Government Landed Properties (FGLP) in Ikoyi, Lagos, Nigeria. Transaction data for each of the sampled properties were compiled from Vol. 11, No. 3834 of Thisday newspaper. These include open market values (reserve prices), bid prices, numbers of bids, property specific and bidders' attributes. These were then analysed using descriptive, paired t‐test statistic and regression model.FindingsThe empirical evidence suggests that auction properties sell at an 11.33 per cent premium relative to predicted market values. Also provided, through the regression model, are variables having a significant impact on the auction prices.Originality/valueThe paper provides empirical evidence of the auction price formation process in first sealed bid auction where published empirical studies are still relatively few. Overall, the results of the analysis presented suggest that auctions offer an optimal method of selling property.