Opinion, Berkeley Blogs

Is there a border in online markets?

By Yuriy Gorodnichenko


Eric Schmidt, a former CEO of Google, said in 2012 that the Web will dissolve national barriers. While we wait to see what is going to be left of national borders in the future, the rising importance of e-commerce can already help economists understand the nature of frictions in international trade.

Indeed, online markets have unique qualities. For example, the physical cost of changing prices is negligible for internet stores and therefore internet prices could be fluctuating every instant (minute, day, week) in response to shifting demand and supply conditions. Search for best online prices for very narrowly defined goods is particularly cheap and simple as consumers do not need to travel anywhere, buyers can establish the distribution of prices with just a few clicks, and pressure for price convergence is especially strong with ubiquitous price comparison websites. More generally, the geographical location of consumers and stores is largely irrelevant in e-commerce and therefore administrative borders and similar frictions are likely to play a much more limited role.

These special qualities of online markets can help understand why, for example, pass-through of exchange rate fluctuations (i.e., how changes of the nominal exchange rate translate into changes of domestic prices) and reversion to the law of one price (i.e., a given good should be sold at the same price in different locations) are generally weak in international data and thus constitute one of the central puzzles (“the border effect”) in international economics. Strikingly, while the law of one price is an appealing concept, the vast majority of previous research has emphasized various frictions that prevent the law from holding over relative long periods. These frictions can take a variety of forms but the most popular barriers leading to violations of the law are search costs, costs of nominal price adjustment, and transportation costs. Assessing the contribution of these frictions has been remarkably difficult as these frictions are ubiquitous in standard markets with brick-and-mortar stores.

Fortunately, online markets are a great laboratory: in these highly integrated markets with low frictions of price adjustment, easy search and price comparisons, and limited influence of geographical barriers, one can rule out some popular explanations of the puzzle and narrow down the set of plausible theories. In a recent paper , Oleksandr Talavera and I use internet prices in the U.S. and Canada for a broad array of products to provide new evidence on the nature and sources of frictions in price adjustment and departures from the law of one price.

For almost 5 years (2008-2013) we scraped prices and other relevant information from a leading price comparison website. A typical look of a price comparison website is illustrated by Figure 1. The data include each good’s unique identifier (similar to barcodes in the scanner price data), each good’s description, prices for each seller, each seller's unique identifier, the number of seller reviews, the ranking of seller quality, reviews of goods, etc. The dataset covers a broad range of goods sold online including software, computer parts, electronics, and others. We have collected information for more than 140,000 goods and more than 11 million price quotes.


This dataset is unique in many ways. First, the time span (almost 5 years) is considerably longer than the time span usually available for researchers studying online prices (typically a year or less). Second, the coverage of goods is much broader than in previous analyses of online prices which typically focused on books and CDs. Third, we collected prices for identical goods in the U.S. and Canada so that comparison of prices is direct and simple. Fourth, our data include information on important attributes such as the reputation of sellers and goods as revealed by ratings of sellers and products. Fifth, our data include many sellers rather than one retail chain and therefore we can assess the relative importance of different sources of price variation (e.g., store-level, country-level, etc.). Finally, the high frequency of our data allows us to time reactions of prices to other high frequency events such as changes in the exchange rate or natural experiments thus making identification more clear-cut.

We find that the size of price changes in online stores (approximately 4 percent) is less than half the size of price changes in regular stores (approximately 10 percent). Price changes occur much more frequently in online stores (approximately once every 3 weeks or less) than in regular stores (once every 4-5 months or more). We also document that price dispersion is substantial and persistent, even for very narrowly defined goods. For example, the average standard deviation of log prices in a given week for a precisely defined good at the bar-code level is approximately 0.12. Thus, the evidence is consistent with the view that online prices are much more flexible than prices in regular stores but there are still some imperfections.

Once these basic facts are established, we study the sensitivity of online prices to fluctuations of the nominal exchange rate. Since adjustment of online prices is unlikely to have any physical costs and, with easy shipping, the physical location of the seller is much less important, the pass-through could be quick and nearly complete while it can be slow and partial in the prices of regular stores because of the frictions associated with trade flows and mobility of buyers. We find that, on average, the pass-through in online markets is incomplete but large and amounts to approximately 60-75 percent, which is greater than the 20-40 percent pass-through documented for regular markets. The speed of price adjustment to equilibrium levels is substantially faster in online markets (half-life is about 2-2.5 months) than in regular markets (half-life varies from 3 quarters to a few years). Multiple margins of adjustment (frequency of price changes, direction of price changes, size of price changes, exit of sellers) are active in the process of responding to nominal exchange rate shocks.

There is significant heterogeneity in pass-through and the speed of price adjustment across goods. Using the richness of our data, we show that for goods with certain characteristics the pass-through can be close to 100 percent. We also document that the size of the pass-through and the speed of price adjustment are systematically associated with the degree of price stickiness, turnover of sellers, returns to search, synchronization of price changes, reputation of sellers, and the degree of competition. These results help reconcile the heterogeneity of estimated pass-throughs and speeds of adjustment across studies and provide new facts for theoretical models to match.

In summary, the stylized facts we document about prices in online stores relative to prices in regular, brick-and-mortar stores are: 1) the duration of price spells is shorter; 2) the size of price change is smaller; 3) pass-through is larger; 4) the speed of price adjustment is faster; 5) dramatic heterogeneity of the pass-through and the speed of adjustment is systematically related to fundamental properties of the goods and the markets in which these goods are sold. These findings are consistent with reduced frictions and increased integration in online markets and thus can inform policymakers and researchers about what one may expect to observe when frictions in regular markets are reduced. To the extent future retail will shift to the internet, one can therefore expect that cross-country price differentials are going to be smaller and less persistent, bringing the law of one price closer to reality.

Maybe Eric Schmidt is right and the border effects in international trade will vanish.