Different design - different cost: An empirical analysis of combinatorial public procurement bidding of road maintenance

This paper is an empirical analysis of first-price sealed-bid procurement auctions in Sweden, with and without combinatorial bidding. The data comprises procurement auctions of identical contracts (road resurfacing) with identical bidders conducted under the same time period (2009-2011) in two different regions in Sweden. Given the comparison of the suppliersʼ offered price per tons of asphalt, we cannot reject the hypothesis of identical distribution of standalone bids generated in both types of auction. The distribution of package bids within the combinatorial format is significantly lower than the distribution of standalone bids within the non-combinatorial format, suggesting substantial cost reduction of allowing package bidding. Also, within the combinatorial format, our analysis of data indicates higher costs when packages are predetermined by the purchaser rather than chosen freely by the suppliers.