SEARCH

SEARCH BY CITATION

Keywords:

  • E3;
  • E31;
  • E5

Sudo et al. (2014) use point of sales (POS) data to investigate micro price dynamics. POS data are highly disaggregated, high-frequency data on retail prices. Macroeconomists have been paying attention to micro price data not only because they are useful in testing theories about the pricing decisions made by firms, but also because micro price data are considered useful for understanding important macroeconomic questions, such as the transmission mechanism of monetary policy and inflation dynamics.

Standard textbooks explain that, when prices are sticky, monetary policy can affect the real interest rate, which in turn affects aggregate expenditure through the intertemporal substitution of demand. This transmission mechanism is represented by, for example, the Euler equation for consumption. Since micro price data are, by definition, disaggregated, they may provide rich information about agents' behavior that lies behind aggregate price data. For example, if one is interested in the degree of price stickiness, it would be very useful to measure how frequently each firm changes its price. For this reason, the measures of the frequency of price changes are useful for investigating the degree of price stickiness. Another reason for looking at micro price data is that one can study the heterogeneity of pricing decisions across firms.

Sudo et al. (2014) provide a number of interesting facts about Japanese micro price data. In this comment, I focus my discussion on how to interpret two of their findings from the viewpoint of macroeconomic theory. The first finding is that posted prices in Japan are ten times as flexible as those in the USA. What is the implication of this for macroeconomics? One might be tempted to conclude that monetary policy in Japan is less effective since prices are more flexible. However, from a theoretical point of view, what matters for consumption decisions is the cost per unit of consumption flow not the price of unit consumption expenditure per se. As Aguiar and Hurst (2007) argue, consumption flow is a product of purchased goods and home production. By spending more time on home production, a consumer can save spending while achieving the same level of consumption flow. Furthermore, purchasing goods requires shopping time. A consumer can find a cheaper price by spending more time on shopping. This implies that he/she can substitute his/her time (spent on shopping and home production) for the amount of money spent on purchasing goods. A change in the price of the consumption flow consists not only of a change in the price of purchased goods, but also a change in the shadow value of shopping time and home production. Therefore, it is not theoretically obvious that the degree of stickiness in the price of purchased goods has a one-to-one mapping to the degree of stickiness in the price of the consumption flow.

Furthermore, what matters for aggregate demand (and the monetary transmission mechanism) is whether purchase prices of each consumer are sticky or not. It does not necessarily matter how sticky the prices of goods sold at a particular shop are. Consider the following example. Household A goes shopping only on the weekend and household B goes shopping only on weekdays, but both of them go to the same shop. Suppose the shop charges different prices on weekends and weekdays, but that these two prices are kept fixed for a year. In this example, even though the prices are considered as being flexible, the prices facing each household are sticky. This implies that the analysis using POS data should be complemented with information provided by household-level data. Abe and Shiotani (2014) investigate household-level scanner data and provide interesting findings. However, a limitation of their household-level data is the fact that the data are not available over a long time horizon. In contrast, POS data are available for long time horizon, which makes POS data more useful for investigating price dynamics.

Second, Sudo et al. (2014) find that the frequency of bargain sales are correlated with labor market conditions. This is consistent with the observation that the consumption flow is a product of goods purchases, shopping, and home production. When the labor market is slack, a consumer may have more time for shopping and home production. In response to this, retailers may increase the frequency of sales. From a macroeconomic point of view, this can be represented by a negative markup shock in the New Keynesian Phillips curve. It has been argued that the negative markup shock has contributed to Japanese deflation. Sudo et al.'s findings are consistent with this view. However, again, their result does not necessarily mean that the nominal price of consumption flow has declined because more bargain hunting may just mean that a consumer substitutes money spent on goods for shopping time.

References

  1. Top of page
  2. References