Why are prices sticky?

Why are prices sticky? I think it is crucial to understand this as it is the defining assumption of the whole New-Keynesian literature. Understanding this may have important implications on those models, as this can lead to differing rates of stickiness with changing circumstances, something the New-Keynesians conveniently choose to ignore. Oh, and there needs to be evidence of relevant stickiness, which is far from being firmly established.

Arthur Fishman finds one way to create price stickiness. Suppose that costs vary from time to time, the exact moment being random. Consumers must search for goods and learn about prices. Sellers can advertise, but it is costly. With some heterogeneity across consumers, some are going to be uninformed. Price rigidity then arises for two reasons: first, some sellers may forgo advertising if the expected duration of the current cost is too short; second, it is not worth advertising if a sufficient number of consumers does not search. This means that for price rigidity to arise, one needs cost flexibility. Kind of strange, isn't it? Even though there is some evidence that costs are more flexible than prices, the model has the implication that when costs become more flexible, prices become more rigid. Find that in the data.