Optimal price setting with observation and menu costs: Data Appendix Fernando Alvarez

Francesco Lippi

Luigi Paciello

University of Chicago

University of Sassari, EIEF

EIEF

December 6, 2010

1

Accuracy of survey data

This document discusses recent survey evidence on firms’ price setting behavior, and in particular on firms’ activities related to information acquisitions and price adjustments. These recent studies provide measures of two distinct dimensions of the firm’s price setting process: the frequency with which the firm reviews the adequacy of its main produce price, and the frequency of actual price changes. Empirical evidence on the latter abound in the literature, while it is scarce on the former. Given this scarcity, understanding the reliability of available survey evidence becomes important. The typical survey question asks firms: “In general, how often do you review the adequacy of your main product price(without necessarily changing it)?”; with possible choices yearly, semi-yearly, quarterly, monthly, weekly and daily. The same surveys contain questions on frequency of price changes too. Fabiani et al. (2007) survey evidence on frequencies of reviews and adjustments for different countries in the Euro area, and Blinder et al. (1998), Amirault, Kwan, and Wilkinson (2006), and Greenslade and Parker (2008) present similar evidence for US, Canada and UK. The question on the frequency of review asks firms to report their ’average behavior’ in a typical calendar year. The question on the frequency of adjustment typically asks firms to report the number of price changes in the previous calendar year. The survey questions on the frequency of price review find their theoretical motivation in the work of Mankiw and Reis (2002) and Reis (2006), where the review is a costly acquisition of relevant information for the price setting decision. Below we address some of the potential strength and weaknesses of this survey data. From this analysis we conclude that this survey data contains valuable and consistent information on firms’ behavior regarding the processes of reviewing and adjusting the price of their products.

1

1.1

The representativeness of survey samples

A positive aspect of surveys conducted in the Euro area and UK is that they include a relatively large number of observations. In all Euro-area countries, but Italy and Luxembourg, the survey sample is above a thousand firms. The UK sample includes about 700 firms. Surveys within the Euro area have been coordinated by the ECB, so share the same structure, and results are summarized by Fabiani et al. (2006). Surveys have been conducted in a period of low inflation, between 2003 and 2004, on samples of firms obtained from existing pools of firms also used for business cycle analysis by central banks or regional statistical agencies. Surveys were conducted by sending the questionnaire to a senior manager through regular mail. The sectoral coverage is limited to manufacturing in France and Germany. Other countries included also other sectors, aiming at reproducing a sample representative of the underlying economy. Belgium, Italy, Spain and Netherlands also include the retail sector while the remaining countries don’t. The vast majority of countries exclude the construction sector from the analysis. As discussed in Fabiani et al. (2006), chapter 2, ”...despite the above differences, results are pretty similar across countries. This lessens the potential significance of the drawbacks traditionally attached to the use of surveys...” Similarly, the UK survey by Greenslade and Parker (2008) has been conducted by the Bank of England and includes about 700 firms and industrial breakdown of respondents is broadly consistent with shares in UK GDP. Differently from the European surveys, existing surveys for the U.S. and Canada are characterized by a relatively small sample size. For instance, the survey evidence available for the U.S. by Blinder et al. (1998) is based on a relatively small sample of firms, about 200. The relatively small sample size may cast doubts on the representativeness of the sample than the survey for the Euro area and UK. However, comparisons of statistics on the frequency of price changes to other sources of similar information show consistent results.

1.2

Relationship to the transaction data on price changes

One criticism of survey evidence on firms’ price setting behavior has to do with the possibility that the employee answering the survey may not be informed on the relevant aspects of the price setting process, or may misinterpret survey questions. A test of the accuracy of survey evidence can be obtained by comparing it to evidence coming from other more direct sources of information about similar aspects of the price setting process. In particular, we can compare survey evidence on the frequency of price changes to similar statistics computed from posted prices and collected in the construction of the CPI. First, survey evidence on the frequency of price changes in the U.S. by Blinder (1998) is consistent with similar statistics computed from posted prices by Bils and Klenow (2004) and Nakamura and Steinsson (2008). In fact, while the median frequency of adjustment reported by Blinder (1998) is substantially smaller than the median frequency of price changes in the CPI (1.4 adjustments a year in the former versus 2.8 in the latter), this discrepancy roughly disappears after conditioning these statistics on the same types of industries. In fact, Bils and Klenow (2004) suggest that ”..A possible contributor to the difference in findings is that firms in the Blinder et al. survey sell mostly intermediate goods and services (79 percent of their sales) rather than consumer items..”. In particular, one key distinction between consumer 2

goods and intermediate goods and services is that the former display much higher frequency of price adjustment, mainly due to the presence of sales as documented by Nakamura and Steinsson (2008). Along these lines, Nakamura and Steinsson argue that the exclusion of sales from the CPI data can reconcile estimates of the frequency of price changes obtained from posted prices with estimates obtained by Blider et al. : ”...The finding that finished-goods producer prices exhibit a substantial degree of rigidity confirms for a broader set of products the results of a number of previous studies (e.g., Blinder et al. [1998];Carlton [1986])...”. To be more specific, Nakamura and Steinsson find that the median frequency of nonsale price changes for consumer goods is between 1.1-1.4 adjustments per year, and the median frequency of price change for finished-goods producer prices is comparable to that of consumer prices excluding sales. Fabiani et al. (2006) reach similar conclusions when comparing CPI and PPI data with survey data on the frequency of price changes in the Euro area: estimates of the frequency of price changes from CPI data in Dhyne et al. (2005) indicate that the median firm adjusts prices every 4/5 quarters, while estimates obtained from survey evidence suggest that the median firm adjusts every 4 quarters. This analysis suggests that survey evidence may contain accurate information on price setting behavior. The accuracy of evidence on the frequency of price changes also suggests that surveys may provide accurate information also on other aspects of the price setting process.

1.3

Cross-Sector Evidence

Another dimension along which we can compare survey data with CPI data on the frequency of price adjustment is by looking at the cross-section of industries. There is high heterogeneity across sectors in the frequency of adjustment. In fact, Nakamura and Steinsson (2008) report that sectors such as transportation goods, travel and unprocessed food have a much higher frequency of price adjustment than sectors such as furnishing, apparel, recreation and services (excluding travel). In Figure 2 we report the sector average frequencies of adjustment and reviews for the four countries for which we have detailed information. Each circle in each plot represents one sector/industry in a given country. The larger the circle, the larger the number of observations for that sector/industry. Two facts are worth mentioning. First, we obtain similar rankings in terms of which sectors have the higher/lower frequency of adjustment from the survey data in the European countries relatively to the US evidence from CPI data. For instance, furnishing (Nace 31) is consistently ranked among the sectors with the lowest frequency of price adjustment in France, Italy and Germany; textile (Nace 13) has a relatively low frequency of price changes in Italy; recreation (Nace S) has the lowest frequency of price changes in the UK; transportation vehicles (Nace G) have the second largest frequency of price changes in the UK. Moreover, if we exclude sectors with relatively small number of respondents and rank the remaining ones according to their frequencies of adjustment/review, we can see that these rankings are consistently similar across the different countries. More specifically, Spearman’s rank correlation coefficients between France and Germany are equal to 0.69 and 0.61 for the average frequencies of review and adjustment respectively. Second, with the exception of the retail sector in the UK, all the other (few) cases in which the mean frequency of adjustment is larger than the mean frequency of reviews are characterized by a 3

small sample size. So these are statistics which may not be very informative on those sectors. Equivalently, all industries for which we have a relatively high number of observations are characterized by higher mean frequency of review than adjustment.

Mean log−Frequency of Price Revision

Germany 2.5 23 2021 25 22 26 15 1719 3224 28

2

30

27

1.5 36 33

1

29 34 31

35 18

0.5 16

0 −0.5 −0.5

0

0.5

1

1.5

2

2.5

Mean log−Frequency of Price Adjustment

Mean log−Frequency of Price Revision

Mean log−Frequency of Price Revision

Mean log−Frequency of Price Revision

Figure 1: Average industry frequency of price changes vs. adjustments

Italy 5 45 4

55 01 46

22

29

3

24 10 98

2 13 23

81

1

32 31

17 25 26 20 28

30 27 11

0 −1 −1

0

1

2

3

4

5

Mean log−Frequency of Price Adjustment

France 5 Sectors 45 degree line

22 21 17 24 27 15 26 29 28 253430 32 35 20 36 33 31

4 3 2

19 18 1

14

0 −1 −1

0

1

2

3

4

5

Mean log−Frequency of Price Adjustment UK 5 4.5

G F

D

4

g

K 3.5

I

H

S

L

C

3 2.5 2 1.5 1

1

2

3

4

5

Mean log−Frequency of Price Adjustment

Note: Sources: Stahl (2005), Loupias and Ricart (2004) and Greenslade and Parker (2008). Each point in the scatter plot refers to a NACE 2 digits sector. 1: agriculture; 10: food products; 13: textile; 15: leather products; 16: wood products; 17: paper products; 18: recorded media; 19: coke and petroleum products; 20: chemical products; 21: pharmaceutical products; 22: plastic products; 23: non-metallic mineral products; 24: basic metals; 25: fabricated metal products, except machinery and equipment; 26: computer, electronic and optical products; 27: electric motors, generators, transformers; 28: machinery and equipment; 29: motor vehicles; 30: other transport equipment; 31: furniture; 32: other manufacturing; 33: installation of machinery and equipment; 45-46: wholesale and retail trade; 55: accommodation; 81: services to buildings; 98: undifferentiated goods- and service. For the UK data refers to NACE 1 digit: C: manufacturing; F: Construction; g: Retail; H: Transportation/communication; L: Real estate; D: Electricity/Gas; G: Wholesale vehicles; I: Hotels/Restaurants; K: Financial Intermediation; S: Recreational.

4

1.4

Do survey respondents interpret a price review correctly?

A possible critique to the survey evidence is that survey respondents do not interpret the corresponding survey question in the same way as our theory would suggest. In other words, given the absence of a detailed description of the review process, the respondent may refer to some other activity rather than collection of information when responding the survey. So even if the respondent is perfectly informed on the price setting process of the firm, and the sample is perfectly representative of the economy, it is still possible that we get a biased estimate of the average frequency of review. Unfortunately, there is no direct evidence on the frequency of review to which we can compare the survey data, as done for the frequency of price changes. The only exception is Zbaracki et al. (2004) who measure the review process from an accounting perspective, and identify it as the managerial cost of changing prices at one large manufacturing firm. One argument in favor of the view that respondents know about the review process is that Zbaracki et al. (2004) find the managerial cost, i.e. collecting and gathering the information, of changing prices is a very important component of the price setting process, accounting for about 5 percent of the company’ s net margin, and about six times as large as the physical cost of changing prices. Therefore, given this evidence, one could argue that the review process is a rather important aspect of firm management, unlikely to be confused with some other activity. Moreover, we can run a more direct test of the accuracy of the survey answers about the frequency of reviews. We can use the predictions of our theory to test the consistency of survey answers on the review process. In particular, beyond the questions on the frequency of review and adjustment, German firms are asked to report whether they adjust prices mostly with a time-dependent rule through a yes/no type of question. According to our theory firms use a time-dependent rule of price adjustment only if the menu cost is zero. In such a case the theory predicts that reviews coincide with adjustment, so that the average frequencies of review/adjustment are identical. If survey respondents interpret the survey question according to our theory, then we should find that in the subsample of firms reporting of using a time-dependent rule of price adjustment the difference between the average frequency of review and adjustment should be much closer to zero than in the whole sample (the difference need not be necessarily zero if we allow for some noise/measurement error). We can test this prediction by computing the mean of the difference between the average frequency of review and adjustment in the subsample of firms reporting of mostly changing prices with a time dependent rule (subsample 1), and comparing it to the mean obtained on the sample of remaining firms (subsample 2). Table 1 reports results from this exercise: the mean obtained in subsample 2 is substantially higher than the mean obtained in subsample 1, 3.1 versus 0.9 respectively; in addition according to the confidence intervals the mean in subsample 2 is larger than in subsample 1 with at least 95 % probability. We interpret this evidence as supportive of our theory in the analysis of the survey data on the price review activity. At the same time, we wee this result as supportive of the thesis that this survey evidence contains valuable information on the review process.

5

Table 1: Difference between frequency of price review and adjustment, and time-dependent review/adjustment rule

Whole Sample Subsample 1 Subsample 2

Coefficient p-value 95 percent confidence interval 2.8 0.00 [ 2.49, 3.14] 0.9 0.02 [ 0.18, 1.57] 3.1 0.00 [ 2.75, 3.45]

Sources: Stahl (2005); mean of the difference between the frequency of review and adjustment; Subsample 1 includes only those firm reporting of changing prices mainly according to a timedependent reviewing/adjusting rule; Subsample 2 includes firms not included in Subsample 1.

1.5

Comparing frequencies of review and adjustment

After assessing the accuracy of survey data, in this section we discuss the gap between the average frequency of review and adjustment. As discussed by Fabiani et al. (2007), there is overwhelming evidence of a positive gap: the median firm reviews more often than it adjusts its price. However, the way the survey is constructed makes hard to draw conclusions about the size of this gap. In particular, the fact that in most countries firms are asked to report the average frequency of review into predetermined ”bins” makes hard a quantitative comparison with measures of average frequency of price changes. The upper panel of Table 2 reports the median frequency of price reviews and the median frequency of price adjustments across all firms in surveys taken from various countries. The median firm in the Euro area reviews its price a bit less than three times a year, but changes its price only about once a year, and similar for UK and US. Canadian firms report higher median frequencies of adjustment and review. This can be explained by the fact that the Canadian sample includes a higher fraction of large firms. Table 2: Price-reviews and price-changes per year AT BE FR GE IT Review Change

4 1

Review Change

54 11

1 1

4 1 Mass 12 53 8 9

NL PT SP EURO CAN Medians 3 1 4 2 1 2.7 12 1 1 1 1 1 1 4 of firms (%) with at least 4 reviews/changes 47 43 56 28 14 43 78 21 11 11 12 14 14 44

UK

US

4 2

2 1.4

52 35

40 15

Number of price changes and reviews per year. The sources for the medians are Fabiani et al. (2007) 2003 Euro area survey, Amirault, Kwan, and Wilkinson (2006) 2003 Canadian survey, Greenslade and Parker (2008) 2008 UK survey.

Figure 2 plots the average number of price reviews against the average number of price adjustments across a number of industries in six countries. This figure shows that in the six countries the vast majority of the industry observations lies above the 45 degree line, where 6

Frequency of review (log scale) .5 1 5 10 20 40 80

Figure 2: Average industry frequency of price changes vs. adjustments

.5 France Belgium

1 5 10 20 Frequency of Adjustment (log scale) Germany UK

Italy Rev=Adj

40

80

Spain Rev=4*Adj

Note: data for each dot are the mean number of price changes and reviews in industry j in country i.

the two frequencies coincide. Most of the industries for Belgium, Spain and the UK have a ratio of number of reviews per adjustment between 1 and 2 (i.e. lies between the two lower straight lines). The data for France, Italy and Germany has much higher dispersion in this ratio. We believe that the reason of the higher dispersion for Italy, Germany and France is due to the measurement error discussed above. Our belief is based on the fact that the questionnaire in the surveys for Belgium and Spain treat price reviews and price changes symmetrically and they record the average frequencies as an integer as opposed to a coarse bin. For four countries Table 3 classifies the answers of each firm on three mutually exclusive categories: 1) those that change their prices more frequently than they review them, 2) those that change and review their prices at the same frequency, 3) and those that change their prices less frequently than they adjust them. Table 3 shows that most of the firms respond that they review their prices at frequencies greater or equal than the one in which change their prices. We conjecture that the percentage of firms in category 1, i.e. those changing the price more frequently than reviewing it, is actually even smaller than what is displayed in the table due to measurement error. To sum up, survey evidence consistently implies that the typical firm reviews prices at least as many times as it adjusts them, and in most sector the frequency of review is larger than the frequency of adjustment. However, while this evidence strongly support the thesis of a typical firm characterized by positive gap between the average frequency of review and adjustment, the coarseness of the survey questions (in particular about the frequency of reviews) prevents obtaining accurate quantitative measures of this gap. Nevertheless, the 7

Table 3: Frequency of Price Changes and Reviews at Firm Level Spain∗

Belgium France Germany Italy Percentage of Firms with: 1) Change > Review 2) Change = Review 3) Change < Review N of Observations (firms)

3 80 17 890

5 38 57 1126

19 11 70 835

16 38 46 141

0 89 11 194

∗

For Spain is only for firms that review four or more times a year. Sources: Table 17 in Aucremanne and Druant (2005) for Belgium, and our calculations based on the individual data described in Loupias and Ricart (2004), Stahl (2009), and Fabiani, Gattulli, and Sabbatini (2004) for France, Germany, and Italy. For Spain from section 4.4 of Alvarez and Hernando (2005).

evidence of a positive gap provides valuable information as, mixed with our theory, suggests the presence of some menu cost and the absence of substantial inflation indexation.

1.6

More detailed information on country surveys

Figure 3 plots the CDF for the frequencies of review and adjustment. The source of the data are Stahl (2009) for Germany, Loupias and Ricart (2004) for France, Fabiani, Gattulli, and Sabbatini (2004) for Italy and Greenslade and Parker (2008) for UK. We tossed those observation that were either missing or reporting an irregular frequency of review or adjustment. It is interesting to notice that the distribution of frequency of reviews first order Figure 3: Cumulative distribution of frequency of price adjustment and review 1

1.4 1.2 1

CDF France

CDF Germany

0.8

0.6

0.4

0.8 0.6 0.4

adjustment revision

0.2

0

0

1

2

4

6

12

52

adjustment revision

0.2 0

365

0

1

2

Yearly Frequency

4

6

12

52

365

Yearly Frequency

1

1

0.9 0.8

0.7

CDF Italy

CDF UK

0.8

0.6 0.5

0.6

0.4

0.4

0.2

0.2

adjustment revision

0.3 1

2

4

12

52

0

365

Yearly Frequency

adjustment revision 0

1

4

12

Yearly Frequency

Note: Frequencies are measured on a per-year basis.

8

52

365

stochastically dominates the distribution of frequency of adjustments. Table 4 reports the mean frequencies of price reviews and adjustment, computed across all firms. While noisy, the mean confirms results from the median. In all countries the average frequency of review is larger than the average frequency of adjustment. In addition, the ranking of countries obtained from ratio of the average frequency of review to adjustment is roughly comparable to the ranking obtained from the medians. Table 4: Price-reviews and price-changes per year: medians and means

Review Change

AT

BE

FR

GE

IT

4 1

1 1

4 1

3 1

1 1

1.2 0.9

23.2 3.5

4.9 2.1

27.7 5.1

Review 12.3 Change 2.6

NL PT SP EURO CAN UK US Medians 4 2 1 2.7 12 4 2 1 1 1 1 4 2 1.4 Means 52.4 3.6 1.89 16.9 99.7 39.2 30 2.3 1.9 1.85 3.3 61.3 33.5 27

Number of changes and reviews per year. The sources for the medians are Fabiani et al. (2007) 2003 Euro area survey, Amirault, Kwan, and Wilkinson (2006) 2003 Canadian survey, Greenslade and Parker (2008) 2008 UK survey. The sources for the means are discussed below.

1.6.1

Computation of the means

Austria: The source of the data is Table 2 and Table 3 in Kwapil, Baumgartner, and Scharler (2005). In order to compute the means, we assigned a yearly frequency of 0.5 to frequency smaller than a year, and took the midpoint of all intervals; we assigned a value of 75 to the group of firms reporting a frequency of price adjustment higher than 50. Then we averaged across the different frequencies, using the fraction of firms at each frequency as a weight. Belgium: The source of the data is Section IV in Aucremanne and Druant (2005). From Aucremanne and Druant (2005), we have information on the average time between consecutive price reviews to be 13 months and the average number of consecutive price changes to be about 10 months. From section IV.1.2 “Overall, the average duration between two consecutive price reviews is 10 months.” From section IV.2, “...This implies that the average duration between two consecutive price changes is almost 13 months...”The following table is from section IV.3. counts the number of firms in the sample that review and adjust prices in a given pair of durations. Below we copy Table 17 - Duration of prices from these authors: France and Italy: The source is the raw data from Loupias and Ricart (2004) and Fabiani, Gattulli, and Sabbatini (2004). We removed missing observations from both series of the frequencies of adjustment and review and averaged across the remaining observations. Notice that we are keeping those firms for which we only have observation of one of the two frequencies. Germany: The source is the raw data from Stahl (2009). We removed missing observations from both series of the frequencies of adjustment and review. Then we averaged across 9

Table 5: Belgium: Duration of prices (number of firms in each bin)

<= 1 Price review <= 1 > 1 and < 12 12 > 12

31 1 2 0

Price change > 1 and < 12 12 197 15 1

12

> 12

8 72 436 5

1 21 37 51

Source: NBB, Aucremanne and Druant (2005). duration <= 1: price is changed/reviewed monthly or more frequently.duration > 1 and < 12 : price is changed/reviewed with a frequency from one month up to one year.duration = 12: price is changed /reviewed once a year.duration > 12 : price is changed/reviewed less than once a year.

the different frequencies, using the fraction of firms at each frequency as a weight. Notice that we are keeping those firms for which we only have observation of one of the two frequencies. In addition, the highest frequency of observation for price adjustment is monthly, while the highest frequency of price review is daily. In order to make the data comparable, we assigned a monthly frequency to all observations at a frequency higher than monthly. Netherlands: The source of the data is Tables 4A-B in Hoeberichts and Stokman (2006). In order to compute the means, we assigned a yearly frequency of 0.5 to firms reporting of adjusting occasionally. Then we averaged across the different frequencies, using the fraction of firms at each frequency as a weight. Portugal: The source of the data is Char 15-16 in Martins (2005). In order to compute the means, we assigned a yearly frequency of 0.5 to firms reporting of adjusting/reviewing less than once a year, and a yearly frequency of 18 to firm reporting of adjusting/reviewing more than twelve times a year. Then we averaged across the different frequencies, using the fraction of firms at each frequency as a weight. Spain: I order to compute the mean, we used data in Tables A10-A11 in Alvarez and Hernando (2005). We assigned a frequency of 6, 2.5 and 0.5 to firms reviewing/adjusting more than four times, between two and three times and less than once a year respectively. Then we averaged across the different frequencies, using the fraction of firms at each frequency as a weight. Moreover, Alvarez and Hernando (2005) show that for Spain, quoting from their section 4.4: When we compare the frequencies of price reviews and of changes, restricting the comparison to those firms that responded to both questions we observe that price changes occur only slightly less frequently than price reviews. The correlation between both frequencies is very high. For instance, among those firms reviewing their prices four or more times a year, 89% declare changing their prices at least four times a year, 4% change them two or three times a year, 6% once a year and 1% less than once a year. The following table constructed from the first row of tables A10 and A11 in Alvarez and Hernando (2005) 10

Table 6: Spain: Frequency of price reviews and price changes (% of firms in each bin)

Price change Price review

At least 4 times per year 13.9 14.0

2 or 3 times per year 15.1 15.6

once a year 56.8 63.1

< once a year 14.3 7.4

Source: Alvarez and Hernando (2005)

Euro: We used the 2003 nominal GDP to compute the weights and averaged across the countries. Canada: The source of the data is Figure 1 and Table 14 in Amirault, Kwan, and Wilkinson (2006). We assigned a frequency of 0.5 to firms reporting to review sporadically. We took the midpoint in each closed interval for the frequency of price changes (e.g. 3 for firms reporting between 2 and 4), and assigned 547.5=365*1.5 frequency of price changes to firms reporting to adjust prices more than 365 times a year. Then we averaged across the different frequencies, using the fraction of firms at each frequency as a weight. UK: The source of the data is Table C on page 406 and chart 5 on page 407 in Greenslade and Parker (2008). In computing the mean, we excluded firms reporting ”irregularly” and ”other”. Then we averaged across the different frequencies, using the fraction of firms at each frequency as a weight. US: Source data from Blinder et al. (1998) 1992 US survey. To compute the means we use Table 4.1 and Table 4.7, interpolating the bins. Both means and medians are based on a small number of responses (186 and 121), and both are sensitive to details used for the interpolation.

References Alvarez, Luis J. and Ignacio Hernando. 2005. “The price setting behaviour of Spanish firms - evidence from survey data.” Working Paper Series 538, European Central Bank. Amirault, David, Carolyn Kwan, and Gordon Wilkinson. 2006. “Survey of Price-Setting Behaviour of Canadian Companies.” Working Papers 06-35, Bank of Canada. Aucremanne, Luc and Martine Druant. 2005. “Price-setting behaviour in Belgium - what can be learned from an ad hoc survey?” Working Paper Series 448, European Central Bank. Blinder, A. S., E.R.D. Canetti, D.E. Lebow, and J.B. Rudd. 1998. Asking About Prices: A New Approach to Understanding Price Stickiness. New York: Russell Sage Foundation. Fabiani, S., C.S. Loupias, F.M.M. Martins, and R. Sabbatini. 2007. Pricing decisions in the euro area: how firms set prices and why. Oxford University Press, USA.

11

Fabiani, Silvia, Angela Gattulli, and Roberto Sabbatini. 2004. “New Survey Evidence on Price Stickiness.” Working paper, Bank of Italy - Research Department. Greenslade, Jennifer and Miles Parker. 2008. “Price-setting behaviour in the United Kingdom.” Quarterly Bulletin Q4, Bank of England. Hoeberichts, Marco and Ad Stokman. 2006. “Price setting behaviour in the Netherlands results of a survey.” Working Paper Series 607, European Central Bank. Kwapil, Claudia, Josef Baumgartner, and Johann Scharler. 2005. “The price-setting behavior of Austrian firms - some survey evidence.” Working Paper Series 464, European Central Bank. Loupias, Claire and Roland Ricart. 2004. “Price setting in France: new evidence from survey data.” Working Paper Series 423, European Central Bank. Mankiw, N. Gregory and Ricardo Reis. 2002. “Sticky Information Versus Sticky Prices: A Proposal To Replace The New Keynesian Phillips Curve.” The Quarterly Journal of Economics 117 (4):1295–1328. Martins, Fernando. 2005. “The price setting behaviour of Portuguese firms - evidence from survey data .” Working Paper Series 562, European Central Bank. Reis, Ricardo. 2006. “Inattentive producers.” Review of Economic Studies 73 (3):793–821. Stahl, Harald. 2005. “Price setting in German manufacturing: new evidence from new survey data.” Discussion Paper Series 1: Economic Studies 2005,43, Deutsche Bundesbank, Research Centre. ———. 2009. “Price adjustment in German manufacturing: evidence from two merged surveys.” Managerial and Decision Economics 31 (2-3):67–92. Zbaracki, Mark J., Mark Ritson, Daniel Levy, Shantanu Dutta, and Mark Bergen. 2004. “Managerial and Customer Costs of Price Adjustment: Direct Evidence from Industrial Markets.” The Review of Economics and Statistics 86 (2):514–533.

12