March 7, 2016

Prices are sticky – but does it matter?

New evidence that menu costs do hold firms back

Tim Hyde

Menu at a sidewalk cafe in Paris. One source of sticky prices may be the cost of actually communicating price changes to customers.

Bigstock

It may seem that prices are always changing (usually in the wrong direction), but economists actually wonder why prices seem to be so stable. The conditions that drive supply and demand are constantly shifting with changes in technology, the whims of consumers, regulations and laws, and even the weather. On some level, it is surprising that prices aren’t changing every day, or even every minute, to keep up.

In practice though, prices tend to hold pretty steady – data collected by the Bureau of Labor Statistics show that the average product sold by U.S. companies sees a permanent price change only once or twice a year. Some products are sold at the same price for ages, including the famous example of the 6.5 oz. bottle of Coke that cost 5¢ for decades, a price streak that persisted through the Great Depression and two world wars.

Clearly there is some reason why businesses don’t find it sensible to change prices in response to every possible economic impulse, and one strain of economic thinking holds that a business must pay a cost to change the price of one of its products. These costs can take many forms: the market analysis required to find the right price, the time to meet with executives to persuade them to change course, and the physical cost of changing the price on a menu or website. In many cases, these "menu costs" may be so high that a business decides not to change its price until the old price becomes completely untenable.

The reasons why a retailer might delay raising the price of a shirt from $18.99 to $19.99 may seem esoteric, but the debate over why prices tend to be "sticky" has played a role in the larger debate surrounding New Keynesian macroeconomics. Some theoretical New Keynesian models show that sticky prices can be hugely costly to the economy, and could even cause or exacerbate a recession. Researchers generally concur that prices are sticky, but they haven’t been able to determine which explanations are correct and, more importantly, if these constraints really matter to the extent that New Keynesian models would predict.

Some researchers have pored over cost data from individual companies to try to trace sticky prices to their source. Others have taken a more sociological approach, conducting interviews about price-setting practices with dozens of mid-level executives across the northeastern U.S. An article appearing in the January issue of the American Economic Review applies a new methodology to measure the impact of sticky prices across a very large number of firms using data from stock market returns.

In Are Sticky Prices Costly? Evidence from the Stock Market (PDF), authors Yuriy Gorodnichenko and Michael Weber argue that a close examination of stock prices in the minutes after major economic news breaks can tell us about the true costs of sticky prices, at least in the eyes of market participants. By comparing stock returns for companies that have less sticky and more sticky prices, they can determine if price stickiness matters in a substantial way for companies.

To figure out which companies are most afflicted by menu costs, the authors access confidential data from a massive ongoing survey of prices undertaken by the U.S. Bureau of Labor Statistics that is used to calculate the monthly Producer Price Index. The authors track price changes for tens of thousands of items sold by hundreds of large companies, and calculate the frequency of price adjustment for each company.

Rather than delve into the minutiae of each company’s price-setting process, the authors simply compare the rate of price changes at different companies and infer that the companies with more rapidly changing prices are more flexible (with lower menu costs) than other firms. These differences could arise for a number of reasons, including different supply chains, management philosophies, or even the use of technology that makes price changes easier.

The authors create a model to predict how flexible and inflexible firms will respond differentially to economic news. In the model, flexible firms can adjust prices frequently at very low cost, so they are almost always very close to their preferred price. There is no reason for them to stray very far from the price that makes the most profit, even if market conditions are changing quickly and even if there is only minimal gain to be had from the new, better price. Whether it is responding to a sale by competitors, a disruption in the supply chain, or shifting tastes of customers, these firms can quite easily adapt.

The inflexible firms have to pay much higher menu costs to change their prices, so they do not have the liberty to adjust their prices whenever a minor tweak might bring in more profits. Between the cost of calculating and implementing a new price, it usually isn’t worth it to make a change.

It is impossible to track how firms respond to every stimulus, so the authors use periodic announcements from the U.S. Federal Reserve Board as a widely-observed and easily-quantified source of economic news. The authors collect data on stock market returns in 30-minute windows covering the time that each Fed press release about monetary policy was issued (the study covers 137 such announcements from 1994 to 2009).

To quantify the degree of surprise contained in each release, the authors look to data from futures trading on the Chicago Mercantile Exchange. These futures markets provide an indication about where traders think the federal funds rate (one of the main policy instruments of the Fed) will be in the future. When Fed announcements are released, these markets tend to respond very quickly to incorporate the new information about which way interest rates are going. The authors use the size of the jump that happens at the moment each Fed press release comes out to measure how "surprising" it is.

The authors' hypothesis is that, if sticky prices are actually significant for firms, stock returns for flexible firms and inflexible firms will react differently in the minutes after each Fed announcement. Specifically, stock prices for the inflexible firms should show greater volatility after more surprising announcements. This is because large changes can either be very bad for inflexible firms – if they are stranded with inopportune prices and can't easily change them – or very good – if the change is so great that they finally decide to adjust to a new price.

The stock return data shows that this is indeed the case. Stock prices for the most flexible firms in the data (those whose average price spell lasts about 2 months) respond only about a third as much to Fed surprises as the most inflexible firms' stock prices.

The authors repeat this analysis on subsets of firms that might be more susceptible to fed funds rate shocks in general (like firms that sell long-lasting durable goods and firms with high labor costs) and the basic pattern holds: inflexible firms show greater stock market volatility in the minutes after a surprising Fed announcement.

This is evidence that market participants collectively recognize the significance of menu costs, even if they do not say so in so many words. If sticky prices were simply a choice made by some firms that had little economic consequence, the stock market returns would not behave so differently for the different groups of firms.

The findings here lend credence to the hypothesis that sticky prices are costly for firms, and cast some doubt on inflation targeting as a way to combat deflationary spirals during recessions. The authors see the potential for more studies of stock market data combined with microdata on price changes from firms to help answer big-picture questions about the way that monetary policy can help and hinder economic growth. ♦

"Are Sticky Prices Costly? Evidence from the Stock Market" appears in the January 2016 issue of the American Economic Review.