Price Setting Before and During the Pandemic: Evidence from Swiss Consumer Prices

How firms set the prices of their goods and services is of great importance for macroeconomic policy. Price-setting behavior determines how shocks propagate through the economy. As prices tend to respond sluggishly to economic conditions, monetary policy has real effects over short horizons. Therefore, a large empirical literature has investigated the extent and causes of sluggish price adjustment.1

In Rudolf and Seiler (2022), we provide new evidence on price-setting behavior based on CPI microdata for Switzerland covering the period 2008—2020. Two noticeable developments fall into this period.

One is the rise of e-commerce. The Swiss Federal Statistical Office (FSO) responded to this trend by collecting prices increasingly online. Earlier studies used web-scraped prices to show that online prices are more flexible than prices collected in stores (e.g., Cavallo, 2018). By uncovering similar trends in Swiss CPI microdata, we document the role of online retailing as a source of greater price flexibility in the data underlying a national CPI.

The other is the COVID-19 pandemic. The pandemic and the measures to contain it greatly restricted business operations and consumer expenditures. Earlier studies on the response of firms’ price-setting behavior to the COVID-19 shock resorted to survey data (Balleer et al., 2020) or online price data (Alvarez and Lein, 2020). Access to up-to-date CPI microdata allows us to study the impact of the pandemic on the full spectrum of the CPI.

The CPI microdata used in our study are the monthly price quotes collected by the FSO to construct the Swiss CPI. We identify 160,000 individual products at the outlet level for 1,300 different goods and services for which prices were collected in more than 3,000 stores across Switzerland. After excluding administered prices, prices based on unit value indices, and a small group of other prices, the sample consists of 7.4 million price quotes from January 2008 to December 2020 and covers up to 60 percent of CPI expenditures.

We establish four stylized facts based on the analysis of the price quotes underlying the Swiss CPI — the first three focusing on price setting before and the fourth focusing on price setting during the pandemic.

In line with earlier studies, we find evidence of considerable price rigidity. Figure 1 illustrates the cross-sectional distributions of the frequency of price changes from January 2008 to December 2019. For each month, the figure shows the mean, median, interquartile range (dark-shaded area), and 15th-85th percentile range (light-shaded area) of the frequencies calculated across varieties.

Three observations stand out. First, the average frequency of price changes is 26.9 percent over the sample period. Hence, around one in four prices changes every month. This result suggests that price rigidity is a pervasive feature of consumer prices in Switzerland. Prices are sticky and do not change continuously. By international comparison, consumer prices in Switzerland are more flexible than in the euro area (Gautier et al., 2022) and, if we disregard price changes related to temporary sales, about as flexible as in the United States (Nakamura and Steinsson, 2008).

Second, the frequency of price changes varies considerably over time and shows distinct peaks and troughs. Some of these periods have straightforward explanations. For example, the substantial increase in the frequency of price changes in early 2015 reflects the effect of the marked Swiss franc appreciation following the decision of the Swiss National Bank to discontinue the minimum exchange rate for the Swiss franc against the euro. This suggests that the frequency of price changes is not fixed exogenously but depends on the state of the economy.

Third, we observe that price rigidity has weakened over time. The average frequency of price changes has increased slightly but steadily from 23.2 percent in 2008 to 31.5 percent in 2019, implying that prices have become more flexible on average. This development is not the result of a broad-based tendency toward more frequent price adjustments but reflects a significant increase in the frequency of price changes in a subset of the CPI. We show that expenditure items for which the FSO has moved to collect prices exclusively online are among the items with the largest increase in the price adjustment frequency. The greater price transparency and competition in online markets provide a possible explanation. Estimating a fixed-effects model relating the frequency of price changes to average profit margins supports this interpretation: the rise of e-commerce has increased the frequency of price changes by increasing competition and reducing profits.

Aggregate inflation is the result of individual price adjustment decisions: inflation may change because more firms change their prices or because firms change their prices by a larger amount. To analyze how fluctuations in these dimensions of price adjustment contribute to inflation dynamics, we decompose inflation into an extensive margin (frequency of price increases and decreases) and an intensive margin (size of price increases and decreases). We find that the variations in the frequency of price increases and decreases contribute more than the variations in the size of price increases and decreases to inflation dynamics. This finding is in line with evidence presented by Gautier et al. (2022) for the euro area.

The synchronization of price changes can have important macroeconomic consequences. Whether all firms change their prices simultaneously or only a fraction adjust prices in every period has implications for the propagation of nominal shocks and the persistence of their effects. We distinguish between two types of synchronization: synchronization within and across outlets. Synchronization within outlets measures the extent to which firms coordinate price changes across their assortment. Synchronization across outlets measures the extent to which stores simultaneously change the prices of the same goods.

The estimation results from a multinomial logit model suggest that price changes are synchronized within rather than across stores. As firms decide to adjust their prices, they tend to change them for multiple products simultaneously. This result is consistent with the view that price adjustments are costly and therefore tend to be synchronized within stores.

The response of price adjustment frequencies to the pandemic was small overall, while the frequency of temporary sales responded countercyclically to the respective demand conditions across sectors.

The COVID-19 pandemic and the measures taken to contain it severely restricted consumer spending. According to Swiss debit card transaction volumes, private consumption declined sharply in spring 2020 and, to a lesser extent, in late autumn 2020. However, the effect of the pandemic varied strongly across sectors. Spending in service sectors such as entertainment and sports, accommodation, and restaurants took the largest hit and recovered only slowly after that. By contrast, spending on food, beverages, and tobacco recorded higher sales throughout 2020 than in the previous year.

Regarding temporary sales, we observe noticeable patterns at the sectoral level. Figure 2 shows the frequency of temporary sales in the food, non-energy industrial goods (NEIG), and services sectors. The results suggest that firms used their sales policies to respond to the demand conditions in the various markets. In the food sector (where demand increased in the first year of the pandemic), the frequency of sales decreased. Conversely, in the NEIG sector (where demand decreased or was temporarily restricted by containment measures), the frequency of sales increased substantially. Temporary sales more than doubled in this sector and amounted to 14 percent on average between March and December 2020. These results are of particular interest for the debate in the literature about whether firms use their temporary-sales policy to adjust to aggregate shocks (Anderson et al., 2017; Kryvtsov and Vincent, 2021).

The empirical findings of our recent work on Swiss CPI microdata are relevant for macroeconomic models with nominal price stickiness. They provide updated and new moments to calibrate such models and highlight recent changes in how prices are set. In particular, they emphasize the effect of the rise of e-commerce and the importance of sectoral heterogeneity in analyzing price-setting behavior.