This dissertation consists of two applied studies of price formation. In the first chapter of this dissertation I study the bidding strategies in on-line auctions, under assumption that bidders are asymmetric. In the analysis I use the unique data set of the used notebooks auctions. In my data set one can distinguish several types of bidders according to their bidding activity: some bidders participate repeatedly even after winning one or more items, while others stop actively participating in the auctions after winning one item or even without succeeding. Frequent auction participants may buy items for resale purposes, while occasional buyers purchase items mostly for own consumption and hence care more about private use value. I conjecture that bidding behavior of players who participate frequently and demand multiple units of the good is better described by common-values model, while valuations of people with single unit demand and lower participation rate have stronger private component. Reduced-form empirical analysis uncovers significant asymmetries in the bidding behavior of these two groups. An increase in the number of participants in an auction increases the winning bid, but the increase is smaller if the additional bidder has high participation rate. I also find that the bidders who participate in the auctions more often pay lower price upon winning. Next, I use simulated maximum likelihood estimator to estimate the parameters of unobserved distributions of valuations in a structural model. The common value component is found to be important in the valuation structure as the variance of its distribution is significantly different from zero. I also find that the mean of the distribution of the common value component is positively related to the qualities of the items. The parameters of the private component distribution differ across the groups of bidders, with more successful bidders (more wins at fewer attempts) having less uncertainty about idiosyncratic component of their value, leading to the greater relative magnitude of the common component in their valuation. In the second chapter of the dissertation pricing behavior of gasoline stations is studied. One feature of the gasoline market is that retail prices exhibit downward stickiness: the prices tend to increase quickly in response to rising wholesale prices but decline more slowly in response to decreasing wholesale price. I build an incomplete markets model and compute its equilibrium using numerical methods. The data generated using computed policy functions mimics the asymmetric patterns of real-world data. The key assumption that generates such price behavior is that a gas station owner is credit constrained and must pay for deliveries in advance. The gas station owner lets the price increase quickly to generate enough revenue to cover next delivery cost. The prices decrease slowly as this is the owner's chance to accumulate buffer stock of assets for future purchases.