Price movements and the definitive changeover to the euro

Transcription

1 Price movements and the definitive changeover to the euro Germaine BRUDIEU Marie-Christine ERNOULT Thierry LACROIX Division Prix à a Consommation Phiippe GALLOT Syvain HECK Fabien TOUTLEMONDE Division Synthèse Conjoncturee The prospect of the changeover to the euro as ega tender has been giving rise to fears of infation among French househods. It wi be shown here that their subjective perception of infation is specificay infuenced by the prices of certain types of products that have moved particuary sharpy in recent months, and for which the changeover to the euro has taken pace eariest: foodstuffs, manufactured goods other than cothing and durabes. Even so, this in no way signifies that these fears are grounded. These price increases affect ony a sma part of the housewife s basket of purchases and can be expained by factors other than the changeover to the euro. In fact, the overa impact of the changeover to the euro on prices depends to a great extent on the way in which the so-caed «psychoogica» price schedues in euros wi be drawn up. If one accepts that a agents combine to round prices systematicay up or down, the impact on infation woud be +1% or -1.1%. These two assumptions represent extremes. Besides, in certain cases, the changeover to the euro wi accompany the habitua price-tag revisions, or ony anticipate them by a few weeks. The fina resut is therefore ikey to be coser to the midde of the range. Infation and changeover to the euro: consumer fears In the run-up to the changeover to the euro, a «patform» of 10 additiona questions was introduced into the monthy consumer confidence survey. Most of these questions attempt to measure over time the rate at which the changeover to the singe currency is affecting French peope s daiy ife: use of payment methods denominated in euros, attention given to the practice of dua pricing, etc. Nevertheess, some questions dea more specificay with subjective perceptions reated to the changeover. One of these additiona questions reates to expectations of infation. The repies bring out ceary that the fears of price rises connected to the changeover to the singe currency have increased as the date of the changeover has drawn nearer. In June 1999, 38% of French househods questioned considered that the introduction of the euro woud ead to price increases, compared with 44% who thought it woud have no effect. In November 2001, 72% of househods expected an infationary impact and ony 21% thought prices woud remain stabe. Two reguar questions aso reate to househods views on infation. One of these concerns the recent past, the other the future prospects. Taken over the ong term, the baance of opinion reating to the future prospects is sometimes difficut to interpret, since its variations, which can be quite marked, do not aways refect subsequent actua price movements. In fact, the peaks sometimes recorded seem often to constitute an «over-reaction» on the part of househods, either to the announcement of economic poicy decisions or to the incidence of ceary-identified externa shocks. For exampe, in the recent past, the most pronounced peaks coincided with the Guf War in the autumn of 1990, the sharp rise in taxes on petroeum products and tobacco in the spring of 1993 and the rise in the VAT rate in the Summer of When it comes to the assessment of recent price movements, the repies provide a coser ong-period match with turning points in infation. Even so, the match remains ess than perfect. One expanation coud be that househods, when asked to give their assessment of the overa evoution in prices, tend to focus on certain products ony. In fact, when one ooks at the detai of price movements for par- Do you think that the introduction of the euro wi have an impact on price eves? june 1999 dec june 2000 dec june 2001 sept (in %) No, there wi be no impact on prices You think there is a risk that prices wi rise You think that prices wi fa Don t know Tota Source : Insee househods confidence survey. nov December

2 ticuar sub-items, a simpe cacuation of correation coefficients brings out two points: Evoutions in prices for food, manufactures and energy are particuary cosey correated with househods responses regarding past prices. At a more refined eve, the baance of opinion being examined seems more specificay reated to four groupings in the price index: food ex seasona food, other manufactures and, to a sma extent, energy and other services (1). Instantaneous regression of the baance of opinion regarding recent price movements on the various components of the consumer price index (CPI) arrives at simiar resuts. The baance of opinion seems to be significanty and jointy inked to the four groupings isted earier (see equation). The system of «psychoogica» weighting suggested by the resut of the equation seems to be remote from the weightings actuay used for the construction of the CPI (those that actuay correspond to the «housewife s basket», based on the product structure of French househods consumption). In particuar, the evoution in the prices of foodstuffs seems to be distincty overweighted and that of services underweighted. This suggests that househods opinions regarding prices seem to depend more on the prices of products bought most frequenty. Since Q2 2001, househod opinion has been moving unfavouraby, just when overa infation was decining significanty. The cacuation we have just carried out suggests that at east part of this apparent anomay stems from the fact that househods are psychoogicay more incined to be attentive to changes in certain prices. This is true, notaby, for prices of food ex seasona food, a grouping where price rises in 2001 were faster than for the overa index. Even so, the match remains reativey imper- TABLE 1 : CORRELATION COEFFICIENTS BETWEEN Y-O-Y PRICE CHANGES AND BALANCE OF OPINION CPI weighting (in %) Opinion on past prices Housewives price outook 17.4 FOOD Fresh Seasona food produce TOBACCO MANUFACTURES Cothing and footwear Other manufactures ENERGY SERVICES Housing, water and sewerage coection Heath services Transports and communications Other services OVERALL Estimation period : january 1987-september (1) grouping «other manufactures» contains cars, furniture, househod equipment, persona hygiene products, sporting and eisure goods and stationery. The grouping «other services», for its part, contains ceaning and repair services, services connected with housing, cutura and recreationa services, vehice-reated services, hairdressing, insurance, parking and motorway tos, financia and funera services, socia wefare services and hotes and catering. 10 Conjoncture in France

3 fect, probaby because the exercise coud ony be carried out for highy aggregated macroeconomic data and because it was not possibe to appy a more refined anaysis. The «sensitive basket»: monitoring focused on mass consumption goods to which househods are particuary attentive As part of the macroeconomic monitoring accompanying the changeover to the euro, INSEE has incuded monitoring of a «sensitive basket» of products, extracted from the norma sampe used for the cacuation of the CPI. The idea of using such a basket is twofod: to identify both the mass consumption goods that are frequenty purchased, for which consumers are iabe to be more sensitive to possibe price movements at the time of the changeover to the euro, and products for which the impact of the changeover coud ead to substantia price changes (ow-priced goods, goods with rounded prices, with psychoogica prices, deivered by vending machines). The function of this sensitive basket is not to repace that of the price index, which aone makes it possibe to measure the monthy evoution of overa infation, but rather to concentrate attention on a sma number of products iabe to have a greater infuence on househods price expectations. As regards foodstuffs, the two series (overa index and sensitive basket) show substantiay the same tendency. This indicates that the recent brisk movements in the prices for this grouping in the sensitive basket refects, on the basis of a few carefuy seected products, a phenomenon that has been more genera over the recent past, namey the acceeration in the prices of foodstuffs ex seasona food. In the case of manufactures, there is a gap between the two series. This phenomenon is concentrated on the other manufactures grouping. In fact, it is the excusion from the sensitive basket of durabe goods, whose price movements are very imited (cars and especiay audio-visua and PC equipment) that argey expains the difference. Finay, the year-on-year movements in the price index for the sensitive basket and the CPI as regards other services were fairy simiar from the beginning of 1999 unti Q2 2000, and again since the end of ast year, so that the major turning points in the «other services» grouping seem to be satisfactoriy captured by the prices of the services picked out for incusion in the sensitive basket. A in a, the anaysis of the evoutions in the index for the sensitive basket brings out notabe price acceerations for two sectors in the recent past: on the one hand, food ex seasona food, in the form of mass consumption goods sod in supermarkets or speciaist stores (bakers, butchers); on the other ÉQUATION OPP = FexSF O _ MANUF ENERGY O _ SERV (. 7 75) (. 3 53) (. 5 64) (. 197) ( ) OPP: Opinion on Past Prices (baance of opinions in %) FexSF: Food ex Seasona Food O_MANUF: Other ManufacturesO_SERV: Other Services FexSF, O_MANUF, ENERGY and O_SERV are expressed as year-on-year price changes. The figures in brackets beneath each coefficient give the Student s t statistic. The assumption of stationarity of the expanatory variabes used is not rejected by the KPSS test. The ony variabe faiing to pass is the baance of opinion reating to recent prices. It is nevertheess regarded as stationary in this cacuation, given that, by definition, any baance of opinion is bounded. hand, severa varieties in the «other manufactures» grouping. It woud seem that the more marked acceeration between May and September 2001 in the prices of the other manufactures in the sensitive basket is expained by the over-representation of persona hygiene and ceaning products and by the absence of durabe goods from this basket. The anaysis based on the sensitive basket therefore gives a better understanding of househods recent fears. This is because the groupings in which there were the argest price increases in the sensitive basket in recent months incude precisey those that seem to have the greatest infuence on that subjective perception of infation: foodstuffs ex seasona food and other manufactures. However, there is nothing at this stage to show that this concentration of price acceerations emerging from the anaysis of the December

4 sensitive basket, is inked to the changeover to the singe currency. What one needs to know is whether the items whose prices have acceerated are among those for which the changeover to the euro has taken pace at an earier stage. Measuring the buid-up of the changeover to the euro in the sensitive basket The sensitive basket constitutes a fexibe and sensitive instrument from which it is possibe to work directy on price records. In order to attempt to identify the infuence of the changeover to the euro on pricing practices and infation, a certain number of indicators have been introduced. A first group concerns the price schedues used, whie the remainder reate to the rate at which price tags are updated. A first indicator records, month by month, the percentage of series whose prices in francs are not in round figures (in other words, those that do not end in 0 or 5 centimes). Athough this is not systematicay true, a price in francs that is not in round figures can indicate that the pricing is carried out directy in euros and then converted for the purpose of the price dispay in francs. In fact, observation of the percentage of series invoved suggests a steady buid-up of the changeover to the euro in two sectors in particuar: food ex seasona food (seasona food is excuded from the sensitive basket) and manufactures other than cothing and durabes (2). For these two items, the proportion of price records not in round figures in francs rose from practicay 0% to amost more than 40% between January and November In the case of food ex seasona food, these were mainy packaged products (chocoate-fied «after-schoo» biscuits, canned beans, orange juice). Among manufactures ex cothing and durabes it appied particuary to certain ceaning and persona care and hygiene products (kitchen paper, razor bades, detergents, disposabe nappies, soap, shampoo). (2)It shoud be recaed that the «manufactures» grouping in the sensitive basket incudes no durabe goods, given that such purchases are by definition very infrequent. 12 Conjoncture in France

5 On the other hand, cothing, energy and services seem for the present to be itte affected by this phenomenon. A second interesting indicator is provided by the percentage of series whose price in francs is stabe to dua conversion (conversion into euros, foowed by conversion into francs). Prices set in euros before being converted for posting in francs are necessariy in this situation, whereas a price that is initiay set in francs has, a priori, ony a 15% chance of being stabe to dua conversion (3). Here again, it is food ex seasona food and manufactures ex cothing and durabes that have seen a considerabe increase in this percentage in the recent past, providing a further indirect indication that the products in these groupings have been the first to be affected by the changeover to the euro. As is the case for prices in francs, a majority of prices in euros wi probaby be in round figures, after a transition period. Taking into account the scae difference between the two currencies, but aso the fact that round prices are aso psychoogica prices, the notion of round figures in euros has been enarged as compared to the one in francs. Round figures in euros end in 0,5 or 9 cents. Unike the two preceding indicators, no notabe tendency in the proportion of series whose price is round in euros was recorded in H In Q4, a faint tendency in this direction coud be seen in cothing and services, with the percentage of series carrying round prices increasing sighty. (3) To be more precise, the probabiity of stabiity to dua conversion for a randomy-seected price in francs in a uniform distribution is 1 / %. Possibe distortion occurs at the time of the changeover from francs to the euro, since rounding to eurocents is ess precise than rounding to franc centimes. For exampe, a price of francs, after rounding, comes to euros, which is equivaent to francs. This is therefore a price that is not stabe to dua conversion. On the other hand, a price of francs is equivaent to euros and euros is equivaent to francs, meaning that it is a price that is stabe to dua conversion. TABLE 2 : PROPORTION OF SERIES WITH PRICES NOT IN ROUND FRANCS Jan. Feb. March Apri May June Juy Aug. Sept. Oct. Nov. Food Cothing Other manufactures ex. durabes Energy Services Tota (in %) TABLE 3 : STABILITY TO DUAL CONVERSION (F F) Jan. Feb. March Apri May June Juy Aug. Sept. Oct. Nov. Food Cothing Other manufactures Energy Services Tota TABLE 4 : PROPORTION OF SERIES WHOSE PRICES ARE ROUND IN EUROS (in %) Jan. Feb. March Apri May June Juy Aug. Sept. Oct. Nov. Food ,1 33,2 33,4 Cothing ,9 30,3 30,6 Other manufactures ex. durabes ,2 35,6 36,2 Energy ,9 33,0 28,4 Services ,1 33,3 35,3 Tota ,8 33,1 33,2 How to read the tabe : Round prices in francs end in 0 or 5 centimes. Round prices in euros end in 0,5 or 9 cents. For tabe 3, if prices were randomy set in a uniform distribution, the proportion of prices stabe to dua conversion woud be about 15%. For tabe 4, if prices were randomy set in a uniform distribution, the proportion of prices round in euros woud be 30%. December

6 The intersection of the various indicators confirms the genera diagnosis. Series whose prices shift to «non-round» prices in francs do so to prices set in euros, being stabe to dua conversion. However, the new prices in euros correspond ony marginay to round prices. This resut suggests either that retaiers are for the moment giving priority to the cacuation of prices in euros, without worrying about the possibe psychoogica virtues of the prices obtained as a resut, or that in the changeover priority is being given to products that are not particuary concerned by round price schedues. However, the tone has tended to change in the most recent months, with a somewhat more appreciabe rise in the proportion of products whose prices are psychoogica in euros among those whose prices are stabe to dua conversion. This is particuary the case for cothing and services. In these sectors, «psychoogica» pricing pays a fundamenta roe, to which we sha be returning ater. The changeover from the franc to the euro in many cases eads to a change in the eve of prices (and the price tag). In this case, a particuar evoution in the statistics reating to changes in price tags from one month to another coud be another way of making it possibe to apprehend the changeover. Foowing the eimination of seasona effects from this phenomenon (4), three indicators have been cacuated, as percentages of the number of price records: unchanged, move to a higher price tag, move to a ower price tag. The proportion of records with stabe prices seems to have been headed very sighty downward in the recent past in a sectors, indicating an increased rate of updating of price tags, which is a priori consistent with a gradua changeover to pricing in euros. However, this tendency is difficut to capture, because of the distortion introduced by the change in TABLE 5 : SUMMARY OF CERTAIN INTERSECTIONS price tags brought about by the cut in VAT decided in Apri Food ex seasona food stand out. The proportion of price-tag changes increased more markedy in 2001 for this grouping, because of an increase in the proportion of upward price changes. It shoud be recaed, however, that the acceeration in the prices for this grouping resuts party from ceary-identified exogenous factors that have nothing to do with the changeover to the euro (impact of anima heath crises on the price of meat, prospect of the introduction of the 35-hour working week for the sma traders). This sti eaves the possibiity that the sight increase in price-tag changes seen recenty at overa eve coud aso stem in part from the changeover to the euro, notaby in two particuar ways: % of series with «non-round» in F....whose prices are stabe to dua conversion among these «non-round» in F. the setting of prices directy in euros instead of francs, which has an impact of a few centimes at most; eary impementation of certain price rises that woud in any case have occurred a few months ater, in order to respect the stabiisation agreement concuded between producers and distributors for the period running from 1 November 2001 to 31 March In this case, the price changes woud not ead to a genuine acceeration in prices, inasmuch as the impact on the... whose prices are «round» in euros among the «non-round» in F. whose prices are «round» in euros among the series stabes to dua conversion January February March Apri May June Juy August Sept Oct Nov pattern of prices over time woud be transitory ony, consisting of this simpe shift in time. To sum up, in view of the anaysis carried out on the sensitive basket, the shift in prices from a franc to a euro basis seems so far to have been carried out at rates differing widey from one sector to another. It has affected mainy mass distribution products, essentiay food ex seasona food and manufactures ex cothing and durabes. And it is precisey these groupings that are, on the one hand, among those on which househods concentrate in order to assess the evoution of past prices and, on the other, those whose prices have tended to acceerate in recent months. This combination of phenomena woud seem to expain the buid-up of infationary fears inked to the changeover to the euro. Conversey, cothing, durabe goods and services are sectors for which the shift to pricing in euros seems to have been sower (athough apparenty gathering pace (4)The updating of price tags is a highy seasona phenomenon, especiay in the cothing sector. (5)The anaysis of the distribution of prices shown here was carried out for a the price records for the index in the months of Apri 2001, with the excusion of seasona food. It therefore does not cover the tariff schedues, which are not coected through price records on the ground. In tota, it covers 61.7% of the overa price index weighting. 14 Conjoncture in France

7 INDICATOR OF UPDATING OF PRICE TAGS 8 9 Methodoogica note: these indicators are computed not ony on the sensitive basket, but on the whoe CPI prices excuding seasona food and tariff schedues. They represent the percentage of records as compared to the average on the ast four years. For exampe, there is 71% of no change records in Apri 2000 as opposed to a reference mean in Apri of 82.2%, so that the indicator no change equas 11.4% somewhat in this fina quarter) and where a sower acceeration in prices has been recorded. At this stage, it becomes interesting to try to caibrate the impact of the changeover to the euro on the pricing of these items so as to be abe to appreciate the goba infuence, positive or negative, that the changeover to the euro might have in the short term on the rate of increase in the CPI. Simuation of rounding effects: the importance of psychoogica price schedues The setting of prices is done according to psychoogica criteria for certain products. For exampe, the price of a product wi be set to 990 francs rather than 1000 francs ; besides, prices often end in 0, 5 or 9. Anaysis of the distributions of prices in each of the main CPI groupings (5) suggests a need to make a distinction between items where psychoogica prices occupy a predominant pace and the rest. It happens that this distinction coincides with the sectora dividing ine that emerges from the previous section: cothing, durabe goods and services, on the one hand; food ex seasona food and manufactures ex cothing and durabes, on the other. In the case of food ex seasona food, the distribution appears to be reativey continuous for eves beow 20 francs. A psychoogica effect seems to emerge for higher prices. However the frequency peaks observed (29, 39, 49 francs etc.) remain imited in scae (ess than 0.7% of the distribution). In fact, it is rather at the centimes eve that the search for a psychoogica effect becomes reevant. The setting of psychoogica prices is more widespread in services. Unike foodstuffs, this mainy concerns the ast digit in francs and ony to a sma extent the centimes, as seems ogica for a grouping where the average price is higher. For eves between 10 and 100 francs, haf the prices in francs end in 0, 5 or 9. For eves between 100 and 300 francs, in more than six cases out of ten, prices in francs end in 0, 5 or 9. In the case of manufactures, psychoogica prices pay an important roe ony for cothing and durabe goods. In cothing, for eves between 10 and 100 francs, 81% of prices end in 9, 0 or 5 (63% for those ending in 9). In durabe goods, psychoogica prices are aso particuary frequent: for eves between 1000 and 6000 francs, as many as 92% of the prices in francs end in 90, 99 or 95 (incuding 85% for those ending in 90). In other manufactures, on the other hand, psychoogica prices occupy a ess centra position, even though for price eves between 100 and 6000 francs, certain recurrent price endings are to be found (73% for prices in francs ending in 0, 9 or 5). Since September, a certain shift from pricing in francs to pricing in euros has admittedy emerged in cothing and services. However this sti seems to be margina. The changeover to the euro took pace eariest in sectors where psychoogica pricing is ess widespread: food ex seasona food and manufactures ex cothing and durabes. In order to caibrate the potentia impact of the changeover to the euro, it is therefore necessary to carry out two exercises, making it possibe to capture the twin aspects December

8 of the phenomenon. The first exercise consists of re-cacuating, on the basis of a the price records, the eve of the index after simuating various assumed conversions and roundings of prices from francs to euros: the officia rue - i.e. rounding to the nearest eurocentime - but aso aternative «biased» practices - systematic rounding up or down, or to the tenth of a euro rather than a eurocentime. A second exercise, in order to take into account the existence of psychoogica prices, consists of simuating the impact of the shift to various types of price schedues invoving what are taken to be psychoogica prices in euros. other products (seasona food, products whose prices are recorded in the French Overseas Departments, domestic staff, etc.) wi probaby find their 0,800 0,686 0,571 0,457 0,343 0,229 0,114 prices affected by the changeover to the euro but they account for ess than a fifth of the grouping excuded from our anaysis. DISTRIBUTION OF PRICES FOR FOOD EX SEASONAL FOOD (IN %) For technica reasons, these two rounding simuation exercises are carried out for a products whose prices are the subject of direct recording on the whoe of the French territory with the excusion of seasona food. These cover 61.75% of the tota weighting of the price index. The products not covered can be grouped into three categories: 0,000 1,40 1,20 1, Prices in francs DISTRIBUTION OF PRICES FOR SERVICES (IN %) products with reguated prices (tobacco, rents, garbage disposa, eectricity, gas, rai transport, posta services, heathcare, TV icences, chid carers, etc.) to which the officia rounding rue wi be appied at the time of the changeover to the euro (or has aready taken pace in some cases); 0,80 0,60 0,40 0,20 0, Prices in francs products with compex pricing (cars, air transport, teecommunications, heath insurance, financia services, etc.) whose prices wi probaby ony make a very sma impact as a resut of the changeover to the euro. The diversity of the products on offer, the wide range of pricing systems and the undertaking made by some operators in these sectors, which are often highy competitive, point to a certain overa neutraity regarding infation resuting from the changeover to the euro; 4,00 3,43 2,86 2,29 1,71 1,14 0,57 0,00 DISTRIBUTION OF PRICES FOR CLOTHING (IN %) Prices in francs 16 Conjoncture in France

9 A in a, the effects of rounding on the products not incuded in the simuation exercise are ikey to be sma and have been ignored here. For the sake of easier comprehension, the potentia impact cacuated wi be provided both for the basket which is the subject of the simuation, and for the entire coverage of the CPI, using the assumptions set out earier. 5,00 4,29 3,57 2,86 2,14 1,43 DISTRIBUTION OF PRICES FOR DURABLES (IN %) In the first exercise, six systematic rounding practices were defined: to the nearest higher eurocentime, to the nearest ower eurocentime, to the nearest eurocentime (the officia rue), to the nearest higher tenth of a euro, to the nearest ower tenth of a euro and finay to the nearest tenth of a euro. Rounding to the nearest eurocentime, and even to the nearest tenth of a euro, has a virtuay neutra impact on the overa index. The impact in absoute vaue on infation of a systematic practice of rounding to the nearest higher or ower eurocentime does not exceed 0.14% for the set of prices considered here (in other words, roughy 0.1% for tota infation). In the case of foodstuffs ex acoho and tobacco, where the impact is potentiay greater because average price eves are ower than in other sectors, the potentia price rise remains imited to 0.2%. The potentia effect of systematic rounding to the higher or ower tenth of a euro is naturay greater, being of the order of 1.3% for the set of prices being considered 0,71 0,00 0,70 0,60 0,50 0,40 0,30 0,20 0, Prices in francs DISTRIBUTION OF PRICES FOR OTHER MANUFACTURES (IN %) 0, Prices in francs (0.8% for tota infation), up or down. However, the rounding assumptions just described may seem simpistic and hence impausibe, particuary for items for which psychoogica pricing is widespread. A second exercise was therefore carried out, aiming to take into account aso the future existence of psychoogica price schedues in euros. The precise definition of these schedues for TABLE 6 : SUMMARY OF DISTRIBUTION OF PSYCHOLOGICAL PRICES Price range (in francs) Two ast figures in francs Last figure in francs Centimes Last figure in centimes Food ex fresh produce Cothing Durabes Other manufactures Services % 59 3% 99 16% 99 23% 99 32% 90 85% 7 7% 90 10% 15 4% 0 7% 8 18% 99 6% 79 25% 49 36% 90 56% 99 89% 6 12% 99 18% 25 7% 50 11% 9 26% 49 9% 89 33% 29 44% 49 66% 95 92% 5 16% 50 24% 20 10% 20 15% 4 12% 9 22% 9 63% 9 67% 9 58% 0 90% 9 18% 0 33% 0 20% 0 33% 8 24% 4 34% 5 75% 0 81% 0 87% 9 95% 5 31% 9 58% 5 39% 5 50% 7 36% 2 44% 0 82% 5 96% 5 93% 5 99% 7 42% 5 73% 9 50% 9 61% 95 14% 90 29% 0 55% 0 86% 0 96% 0 99% 0 27% 0 75% 0 81% 0 65% 90 25% 0 43% 90 74% 90 92% 90 97% 97 99% 90 37% 95 78% 50 87% 90 67% 50 33% 80 53% 95 89% 95 98% 20 97% 98 99% 50 46% 90 81% 90 89% 50 68% 0 51% 0 75% 0 82% 0 94% 0 98% 0 99% 0 63% 0 82% 0 95% 0 74% 5 91% 5 90% 5 98% 5 100% 5 99% 8 99% 5 84% 5 87% 5 96% 5 79% 2 92% 3 92% 8 99% 6 100% 2 99% 7 99% 9 88% 8 89% 1 97% 8 82% How to read the tabe : In the durabe goods sector, for eves between and francs, 85% of prices end in 90 ; 89% end in 90 or 99 and 92% end in 90, 99 or 95. December

10 SIMULATIONS OF ROUNDING TO THE PSYCHOLOGICAL PRICE SCHEDULES Being by definition sti unknown, the psychoogica price schedues in euros were determined by reference to what can now be found in francs. For exampe, the so -caed broad psychoogica price schedue in euros for durabe goods in the price range from 100 to 1000 euros was determined on the basis of what can currenty be observed for durabe goods whose prices fa in the range between 100 and 6000 francs. For these, it turned out (cf. Tabe 6) that 66% of the prices end in 49, 90 or 99. This same hypothesis was then adopted for the prices in euros. In the case of durabe goods whose prices converted into euros fa in the range euros, having no corresponding reference in francs (very few durabe goods are sod at prices between 10 and 100 francs), it was decided, for ack of anything better, to appy the same broad price schedue as the one worked out in the case of cothing. Simiary, in the case of cothing products whose converted price fas in the range 1-10 euros, the broad price schedue worked out for foodstuffs was appied by anaogy. The simuation of the rounding to the broad psychoogica price was carried out in three stages. First stage: a price records were converted into euros and rounded to the nearest psychoogica price in the fine schedue (it is known that the overa impact is neutra, see third coumn of tabe 7). Second stage: for each grouping and each price range, account was taken of the expected frequency of a rounding of the price to the broad psychoogica schedue (this frequency corresponding to that observed in francs). For exampe, given that 66% of the prices of durabe goods in the range euros are expected to end in 49, 90 or 99, their prices were modified ony two times in three. Third stage: once the frequency has been taken into account, for prices which have to be rounded in the direction of the broad psychoogica schedue, the price variation resuting from this operation was cacuated. If the variation exceeded 20%, up or down, the choice was made, for the sake of reaism, to round the price to a different psychoogica price, what might be caed the refined broad price, appying, not the initia broad price schedue, but the schedue corresponding to the next ower price range. In the event that the price aready fe in this bottom range, the fine psychoogica schedue was appied. For exampe, in the exercise consisting of rounding to the next higher broad psychoogica price, et us suppose that the conversion is appied to a bookcase priced at 699 francs. Conversion into euros gives euros, rounded initiay to 107 euros, the nearest fine psychoogica price. In the second stage ony 66% of these prices are rounded to the broad psychoogica price (this being the frequency for durabe goods in the price range euros). A number between 0 and 1 was then drawn at random and if this number was beow 0.66 the rounding was carried out (if not, nothing was done). In this case, the price rounded to the broad psychoogica price becomes 149 euros, corresponding to an increase of 39.25%. This was judged to be unreaistic because it exceeds 20%. One therefore then rounded the price to the broad refined psychoogica price corresponding to the rounding practice appied for the ower price range of euros for durabe goods. This meant that the price was rounded to 109 euros (giving a tota price increase of 2.3%, since the initia converted price is euros). n EFFECT OF ROUNDING TO PSYCHOLOGICAL PRICES (in %) Type of rounding Fine schedue assumption Broad schedue assumption Sector Upper psy Lower psy Nearest psy Upper psy Lower psy Nearest psy Food Acoho - tobacco Cothing - footwear Housing and utiities Furniture and househod equipment, daiy maintenance Transport Leisure and cuture Hotes and catering Other goods and services Tota ex tariff schedues and seasona food Tota Conjoncture in France

11 TABLE 7 : EFFECTS OF ROUNDING ON THE PRICES Sector Average price in francs weight (in %) Higher euro centime Higher tenth of euro Type of rounding (en %) Lower euro centime Lower tenth of euro Nearest euro centime Nearest tenth of euro Food Acoho and Tobacco Cothing and footwear Housing - water - gas eectricity Furniture househod equipment and daiy maintenance Transport Leisure and cuture Hotes and catering Other goods and services Tota ex specia tariffs and fresh produce Tota n.d the purposes of the exercise is naturay open to discussion, since the detais of the actua psychoogica prices in euros are sti unknown. The choice was therefore made on the basis of observation of the distribution of prices in francs. Moreover, since psychoogica price schedues vary from one sector to another and, within a given sector, from one price range to another, the schedues in euros were defined for intersections of sectors and price ranges. For each intersection, two schedues were chosen (see Tabe 8), one «fine» and one «broad». This choice reies on the assumption that the practice of retaiors wi adopt a two stage process. Rounding to a psychoogica price in the fine schedue invoves ony sma variations and coud therefore take pace rapidy after the introduction of the euro. Psychoogica prices in the broad schedue woud be the subject of introduction at a second stage by retaiers. The fine schedue is based on the fact that, most of the time, prices end in 0, 5 or 9 centimes and that, for higher price eves, the prices are often whoe numbers. The broad schedue is based on the fact that within the prices in the fine schedue, some appear sti more frequenty (ow prices ending in 50 or 90 centimes, whoe numbers ending in 49, 90 or 99 francs for the higher price eves). TABLE 8 : HYPOTHETICAL PSYCHOLOGICAL PRICE SCHEDULES IN EUROS Sector Ranges Fine Broad Food ex fresh produce Cothing Durabes Other manufactures Services In the case of the fine schedue, systematic rounding to the higher psychoogica price or the ower psychoogica price gives rise to a variation of ±0.9% for the coverage considered here (giving ±0.6% for tota infation). The impact of rounding to the nearest psychoogica price is competey neutra. Based on the atter simuation ( fine schedue and rounding to the nearest psychoogica price), simuations based on a broader schedue were carried out in the foowing manner. Observation of price records in francs, whie it made it possibe to bring out the existence of psychoogica price schedues, aso showed that the practice was not 0-1 c1=0 ou c1=0 ou 5 c2=50, 90 ou c1=0 ou 5 c2=50, 90 ou c1=0 ou 5 c2=50, 90 ou c2=0, 90 ou 95 d3=9, 9,90 ou 9, c2=0, 90 ou 95 d3=9, 9,90 ou 9, c2=0 e2=49, 90 ou c1=0, 5 ou 9 c2=0, 50, 80, 90, 95 ou c1=0 ou 5 c2=0, 50, 90 ou c2=0 e1=0, 5 ou c1=0 c2=0 ou c2=0 e1=0 ou c2=0 e1=0 ou 5 How to read the tabe : c1 represents the fina figure in euro centime, c2 the eurocentimes, e1 the fina figure in euros, e2 the fina two figures in euros and d3 the fina three figures, centimes incuded. For exampe the psychoogica prices used in the fine assumption for the intersection of the cothing sector and the 1-10 price range are those whose fina figure in centimes is 0 or 5. systematic. The frequency varies in particuar very widey from item to item. The frequencies of use of psychoogica prices reative to intersections of sector and price range were therefore taken into account (6). Moreover, again in the interests of reaism, when the practice of rounding to a broad psychoogica price ed to a price variation exceeding 20%, it was decided to round the price to a more «refined», and hence more credibe, psychoogica price (for more detais on the method used, see box). In the fina anaysis, systematic rounding to the higher and ower psychoogica price of the broad schedue ed respectivey to a rise of 1.0% and a fa of 1.1% for tota December

12 CHANGEOVER TO THE EURO: LARGE-SCALE DISTRIBUTION TAKING THE LEAD The changeover to the euro is not taking pace at the same pace in the different types of retaiing. Whie at an advanced stage in the case of arge-scae retaiing, it is sti uneven for the smaer stores. This concusion is the same for both the practice of dua pricing and the method of price-setting, but the disparities between the distribution circuits are not on the same scae for a the indicators. According to the monitoring carried out by DGCCRF (1) at the end of October, dua pricing was genera in hypermarkets and supermarkets. It was a itte ess widespread in the other sectors (75.8% of the prices observed), especiay in services, athough there was a tendency for this eeway to be made up. Generay speaking, the conversion rues were being propery respected (in 98.3% of cases). Here too, the service sector has agged somewhat behind, athough sti with percentages above 90%. In addition, faiure to respect the rounding rues seemed to be more connected with an attempt to set prices in round figures in the two currencies than any deiberate attempt to infringe the rues rounding errors are in fact more frequenty in favour of the consumer than at his expense. INSEE s sensitive basket makes it possibe to monitor, for the 20,300 products whose prices are recorded monthy in traditiona saes outets (2), the changeover from francs to euros in the pricing adopted in the various distribution circuits. The indicator adopted here is the proportion of prices that are stabe to dua conversion (F i F), which is a good indicator of the proportion of prices in euros. The variety stores were the first to aunch the changeover, starting in March 2001, with 35% of prices stabe to dua conversion as against 17% for a types of retaiing. In November, with the indicator standing at 84%, this form of retaiing was we in advance of the others. It shoud be remembered, however, that ony 3% of the price records for the sensitive basket, especiay the food items, are observed in the variety stores and that there is probaby a brand effect since most of the saes outets of this type beong to the same group. As regards the hypermarkets and supermarkets, progress has occurred ater. There were a few signs of changeover recorded up to Juy (the proportion of prices stabe to dua conversion rose from ony 16% to 22% in (1)Direction Générae de a concurrence, de a consommation et de a repression des fraudes, the department responsibe to the Finance Minister for consumer affairs and trading standards. (2)By opposition to the prices, most often based on specia tariff schedues, whose observation is carried out centray (eectricity, heathcare, taxis, rai transport, posta services, etc.). (3)The number of observations actuay carried out for this form of retaiing in the sensitive basket is nevertheess imited (2%). infation. It was in the case of hotes and catering that the impact seemed potentiay argest, foowed by cothing and footwear, a sector where the practice of setting psychoogica price schedues is virtuay systematic, and then by eisure and cutura goods, which incude certain «eectronic» durabe goods that are aso concerned in psychoogica pricing. The orders of magnitude for the individua items shown here must in a cases be considered as ower and upper bounds, neither of which is at a ikey to be reached in practice. It is in fact most unikey, given the state of competition, that a economic agents woud co-ordinate their rounding actions in the same direction. It can aso be noted that the impact of systematic rounding to the nearest broad psychoogica price is not entirey negigibe, being sighty negative for overa infation (- 0.1%) (7). n (6) It shoud again be stressed that the exercise carried out here is based on a bod assumption. For ack of being abe to observe for the present the choice and frequency of appication of psychoogica price schedues in euros, these have been defined on the basis of what is currenty found for the distribution of prices in francs of the products in question. (7) This means that the «broad» conversion to the nearest psychoogica figure in francs in fact eads to price fas. For exampe, 199 francs makes amost euros, but euros is equivaent in reaity to ony francs. 20 Conjoncture in France

13 CHANGEOVER TO THE EURO: LARGE-SCALE DISTRIBUTION TAKING THE LEAD (CONTINUATION) seven months for the hypermarkets and from 17% to 26% for the supermarkets), but the changeover then acceerated as the months passed to reach 55% and 57%, respectivey, in November. The movement began with food products and mass consumption manufactures in August, foowed by cothing in September. The convenience stores, concerned amost excusivey with food products, began the process in Apri. In November, more than haf the prices recorded in this type of retaiing were stabe to dua conversion. Traditiona retaiing and services, for their part, showed no significant movement before October (20% of prices were stabe to dua conversion in that month). Here again, the progress since October has been due mainy to fiing stations beonging to the oi majors, which triggered off the conversion of the pumps to euros, at a time when the movement had sti not begun in the hypermarkets. Finay, the hard-discounters bring up the rear, with ony 21% of prices stabe to dua conversion in November (3). The contrast between arge-scae distribution and smaer stores is particuary marked for food products and manufactures ex cothing and energy. It is ess distinct in the other sectors: cothing, where the speciaist chains occupy an intermediate position between hypermarkets and traditiona stores; and energy, where, as has been seen, the situation is the reverse. The objection coud nevertheess be raised that the cassification used for these two sectors does not make it possibe to distinguish the saes outets beonging to distribution networks from those that are independent, the former probaby being ess autonomous in their price-setting poicy. A in a, it is indeed the more concentrated (arge-scae) forms of retaiing, foowed by the non-independent traders (chains, networks) that seem to be furthest ahead in the changeover to the euro, even though in November 2001 the franc seemed sti to be in majority use as regards the setting of price tags (4). The sma independent traders and the saes outets for services, for their part, have concentrated unti now mainy on dua pricing, with the price schedues themseves sti mainy in francs. n (4)The proportion of prices stabe to dua pricing nevertheess exceeded 50% in the variety stores, the convenience stores and, for food and manufactures ex cothing and energy, in the hypermarkets and supermarkets. Concusion The main point emerging from the anaysis set out in this artice is that the items for which the changeover to the euro has taken pace eariest (food ex seasona food and manufactures ex cothing and durabes) beong both to those groups of product whose prices househods are particuary attentive to and those whose prices have tended to be more dynamic in the recent past. This probaby expains why househods fear that the changeover to the euro is iabe to push infation up. Nevertheess, these fears need to be aayed, since these price rises are sufficienty occasiona in nature not to have ed to a genuine sippage in the overa index. On the contrary, against a background of faing oi prices and a deteriorating economic situation, the outook for French infation is downward. Moreover, the coincidence in timing between the changeover to the euro and the rise in prices of certain products does not prove that the changeover itsef is infationary, since there can be other economic reasons expaining the sight acceeration in underying infation in recent months (for exampe, second-round effects inked to the rise in oi prices and the depreciation in the euro that took pace ast year). In these circumstances, the changeover to the euro woud merey be the occasion to proceed to adjustments that were in any case in the pipeine. Certain price rises have for this reason simpy been brought forward a few months, but the infationary impact is in reaity negigibe, being very temporary and due to a simpe shift in the timetabe for updating of price tags. Finay, it turns out that many sectors are sti very itte invoved in the changeover to the singe currency. These incude, first of a, those where psychoogica pricing is widespread (cothing, durabes, services). For these, the impact of the changeover coud we be more substantia, but the caibration exercises show that the fina resut cannot be predicted at this stage, since it depends very cosey on future rounding practices. Even if one were to assume that a agents combine to appy systematic rounding up or down, the impact on infation woud be +1% or -1.1%. Since these are extreme cases, the fina resut is ikey to be coser to the midde of the range. The introduction of new price schedues wi probaby be spread over time, sometimes in stages, in other cases ony when product ranges are renewed. This means that definitive concusions regarding the changeover to the euro and its effects on prices cannot be drawn up unti some time has passed. n December