A Description of Swedish Producer and Import Price Indices PPI, EXPI and IMPI

Transcription

1 STATSTCS SWEDE Rev (10) A Description of Swedish roducer and port rice ndices, EX and M The rice indices in roducer and port stages () ai to show the average change in prices in producer and iport stages for different industries and product groups i.e. in the first distribution stage when products are delivered fro Swedish producers or arrives to Sweden. The price indices are calculated and published onthly with reference year Approxiately price observations are collected onthly fro around enterprises. The data collected are broken down into doestic producer prices, export and iport prices. Monthly price indices are calculated for each of these ain groups. With soe exceptions, the survey's product coverage spans agriculture, forestry and fishing, ining and quarrying, anufacturing, electricity, gas, stea and air conditioning supply and water supply, sewerage, waste anageent and reediation activities (CA 2008 sections A-E). The doestic producer price index, HM, gauges the developent of prices for products that are anufactured and sold in Sweden. The export price index, EX, easures the producer price changes for products produced in Sweden for exports, while the iport price index, M, is a price index for iported products. Doestic and export prices together ake up the total producer price index,. port prices cobined with doestic producer prices for price index for doestic supply, T. Fi gure 1: nterrel ationships between the various price indices RODUCER RCES EXORT MARKET RODUCER RCES DOMESTC MARKET MORT RCES Export rice ndex EX roducer rice ndex roducer rice ndex, Hoe Sales, HM rice ndex for Doestic Supply T port rice ndex M Data are collected roughly between the 1st and the 20th each onth after the statistical onth concerned and are published around the 25th. Approxiately half of the data are collected via telephone (TDE) and the rest by paper questionnaires or e-ail. The price data relate to doestic producers' selling prices and iporters' purchasing prices. Respondents are required to leave a coent about any

2 STATSTCS SWEDE Rev (10) eventual price change and whether there have been any changes to the specifications of the respective products. f this is the case, after consulting the respondent the Deceber base price is adjusted, where appropriate, so that quality changes are factored out of the price coparison. 1 Objective and background The rice indices in roducer and port stages () stage is intended to reflect the average price developent, both overall and for various product groups. rices are easured, respectively, at the first distribution stage, where the products are supplied by the Swedish producer, and at the first purchasing stage, where the products enter Sweden. rice index series with a rough product group breakdown have been calculated since Fro 1920 onwards a wholesale price index has been reported onthly, with ore of a fixed structure and a ore detailed product group breakdown than was previously the case. This set of statistics took on its odern for in 1963, when a ore systeatic, international classification of branches of econoic activity was introduced. As of the publication of ndex for January of 2009 (27 February 2009), the roducer rice ndex for ports adopted the product noenclature of S n connection to this, the total aggregate of the roducer rice ndex, roducer rice ndex, hoe sales, Export rice ndex, port rice ndex, and the rice ndex for doestic supply was redefined. While having earlier been defined as products in sections A - D in accordance with S 2002, they now reflect price developents for products in sections B - E in accordance with S n general ters, this eans that products fro agriculture, forestry and fishing are no longer included in the total aggregate although they are available in Sweden s Statistical Databases. nstead electricity, heating, gas and water are included. The change is in accordance with EU recoendations and contributes to increased coparability between countries. There are also a nuber of differences at the ore detailed level between how the product groups are defined by the two different noenclatures. 2 Observed variables The variables for which the survey collects price data are products transactions relating to Swedish producers' sales of own-anufactured products and products iported into Sweden. 3 Statistical targets The producer and iport price indices are designed to easure on a onthly basis, also within different product groups, the average price developent, in SEK, of Swedish-produced products, both in overall ters and broken down into doestic arket and exports, and iported products. This set of statistics easures price developent at the first arketing stage, i.e.

3 STATSTCS SWEDE Rev (10) doestic producers' selling prices and iporters' purchase prices, respectively. For Swedish-anufactured products, the price should be exworks for the doestic arket and free on board (FOB) for export sales. On the iport side, prices should be CF (cost, insurance, freight). VAT and other taxes on products are not taken into account. Any change of the prices charged/paid that stes fro a change in quality or other price-deterining characteristics should be adjusted for. n principle, only pure price changes that are reflected in the pricing of coparable transactions should have an ipact on the price index. The price concerned is that which the purchaser actually pays after deduction of any discounts. Under Council Regulation (EC) o 1165/98 concerning short-ter statistics, the index has to easure the developent of prices, including taxes on products (apart fro VAT) for orders received during the onth, priarily on a certain day in the onth. The Swedish index, however, continues ainly to reflect the developent of the average price for the onth's invoices, excluding taxes on products. 4 Object and population The price index at the producer and iport stage is concerned with transactions. The population coprises, respectively, all transactions relating to sales at the producer stage, and purchases at the iport stage, of products belonging to the following sections according to S 2007 (CA 2008). Section A B C D E Designation agriculture, forestry and fishing products fro ining and quarrying anufacturing electricity, gas, stea and air conditioning supply water supply, sewerage, waste anageent and reediation activities S 2007 is a classification of products based on activities fro the Swedish Standard ndustrial Classification, S S 2007 replaced the old classification of S 2002 in January of S 2007 corresponds to ACE rev 2 (Classification of Econoic Activities in the European Counity) which corresponds to SC rev 2 (nternational Standard ndustrial Classification of all Econoic Activities) and CA Sapling and calculation of weights rice data is collected for a saple of typical transactions. Each year the saple is revised and updated, particularly in areas with coodity categories of increasing iportance. The saple is however not reexained copletely. Therefore enterprises and ite categories ay reain in the saple although they are not any longer aong the ost iportant ones. Each year also the respondents are asked to review the specifications

4 STATSTCS SWEDE Rev (10) and replace those that are no longer typical for the transactions ade within each category. As fro 2004 probability sapling is used: cobinations of enterprises and C 8-digit categories are selected within CA strata s with probabilities proportional to their value. Strata are defined according to CA, typically at the 5-digit level. The nuber of observations for iports, exports and the doestic produce and sales are approxiately 1800, 1400 and 1800 respectively. The observations are supplied by a total of 1900 enterprises. The weights, or the value aounts, that are used to aggregate single price ratios are allocated fro the saple fraes according to the following procedure: All C 8-digit categories (Cobined oenclature) containing that are larger than, or equal to, the CA total divided by the nuber of C 8- digit categories included in the saple, will receive a weight that equals its share of the CA category. Hence, relatively large C 8-digit categories ay represent theselves only, whereas saller categories included in the saple also represent those other sall categories not included in the saple. Before 2004 the distribution of value weights was proportional to the value of each included C 8-digit category. Within C 8-digit categories the values are distributed to producers/iporters according to analogous principles. Within enterprises the values are distributed equally between the specifications, or according to any inforation that ay have been received fro the respondent. Exaple: Within a category three units have been selected representing 40, 9 and 1 per cent respectively. Since 40>100/3 the largest will receive the weight 40 whereas the rest of the weight will be split equally between the other two, (100-40)/2=30. The weights are updated annually as fro January each year, based on year y-2 data. Values are then updated to represent the Deceber value of previous year, by ultiplying the value by the corresponding index for the specific product group of Deceber y-1 with base year y-2. For ore inforation about sapling in the Swedish see pdf 6 Variables The ain variable is the price, as described above. The price, however, can be reported in any currency. rices reported in foreign currency are then converted into SEK. The invoice currency and the reporting currency are, therefore, copleentary variables.

5 STATSTCS SWEDE Rev (10) Where prices are reported in foreign currencies, the Custos and Excise Departent's exchange rates are used for converting into SEK. These are based on the exchange rate on the third Wednesday of the onth preceding the easureent onth, with soe adjustent being ade for exchange rates that deviate by ore than 5% within the easureent onth. This ethod is used instead of, for exaple, the Swedish Central Bank's average exchange rates in order to proote the usability of the index as a deflator for estiating export and iport values at constant prices in foreign trade statistics. Overall, approxiately 50% of all export price data and approxiately 60% of all iport price data are provided in foreign currency, whereas other prices are in SEK. 7 ndex forulas The price index at the producer and iport stage can be described as a Laspeyres-type chain index with annual links. A Laspeyres index is defined as: Q Q t; k 0; k t k 0; k 0; k 0 = = 0; kq0; k k 0; kq0; k k k t; k 0; k, where 0 and t are the price per unit in the base period (0) and the coparison period (t), respectively, and Q 0 is the quantity in the base period. Suation is across products (indicated by k). n the centre eleent above, the index is expressed as the ratio between the suated value of the baseperiod quantities at the coparison-period and base-period price level, respectively. The index can also be expressed, as in the final eleent, as an average of the products' price relatives weighted with the base-period values. The index links in the price index at the producer and iport stage are calculated with the current onth in year y as the coparison period, and with Deceber y-1 as price reference period. The weighting period is the year y-2. A link is thus defined as: V y, s y 1, dec s Vs s p p y, ; s =, y 1, dec; s where the value weight Vs is the value of the transactions in year y-2 that are represented by the specification s, adjusted by the estiated price developent fro year y-2 to Deceber y-1, V y 1, = V, s g. dec s y 2; s y 2; g

6 STATSTCS SWEDE Rev (10) g is typically the C 8-digit category for the selected specification. See section 5 above "Sapling and calculation of weights" for a description of how the value aounts V y-2;s are calculated. 8 Reference periods The chained index nuber is priarily calculated for onths and indicates the particular onth's price level copared with the average price level in 2005 (2005=100). 100 y 1 y, Y, dec y, 2005 = Y 1, dec 1,. DEC y dec , M Y = , DEC M = JA As set out above, the index nuber ainly reflects the developent of an average price for the onth. The annual average index is the unweighted arithetical ean of the onthly indices.

7 STATSTCS SWEDE Rev (10) Appendix 1: Quality evaluation Changes of the prices charged/paid that are the result of changed specifications regarding quality, quantity or other price-deterining factors should not influence the price index. Only pure price changes that are reflected in the pricing of coparable transactions should have an ipact on the price coparison. n the price index at the producer and iport stages, distinctions are ade between the following types of specification changes. A ew product specifications: an explicit evaluation is ade in which the new product is given a higher value than the iediate predecessor; C ew product specifications: an explicit evaluation is ade in which the new product is given a lower value than the iediate predecessor; E ew product specifications: an explicit evaluation is ade in which the new product is given the sae value as the iediate predecessor; ew product specifications which result erely in a change in volue or variety, e.g. a inor change in the size of the packaging; ew product specifications: overlap ethod applied by obtaining the real base price for the new product (for Deceber of the previous year); K ew product specifications: overlap ethod applied by obtaining the new product's price for the previous onth; M roduct no longer sold for rest of year, index iputed fro other aggregate; O ew product specifications: valuation ipossible, index assued to be unchanged fro previous onth; R ew product specifications: valuation ipossible, index assued to be unchanged fro two onths earlier or ore; T Changed definition of price, or price calculated on a different basis fro previously; Z Other. Calculation of an adjusted reference price (price as at Deceber of previous year) n cases A and C, a new base price is calculated fro the quality evaluation ade in the particular instance. The question arises here as to what ipact the change in specifications has had on the price. The effect on price can be gauged either in ters of the changed specifications' ipact on the cost of supplying the product or in ters of how the custoer is expected to appraise the change. A new reference price (for onth ) is calculated as ˆ ˆ F =, (1) = 0 1 1

8 STATSTCS SWEDE Rev (10) where represents prices, F the value of the change, and the old and new product, respectively, and 0, -1 and a particular onth. 0 is thus the reference onth. ˆ can be understood as an estiated or iputed price for 1 the new product in the previous onth. Let us assue, by way of exaple, that the price for a product in the original specifications was SEK 38 in the reference onth (Deceber), that the price rose to SEK 40 in March and that the price for the product according to the new specifications is SEK 45 in April. Meeting the new specifications involves a greater use of resources, so that the cost of supplying the product rises. The price of SEK 45 is thus not coparable with the reference-onth price of SEK 38. Let us further assue that the change in specifications has led to a price increase of SEK 3 per unit. A new reference price that is coparable with the price for the new specifications can thus be calculated as = 40,85 40, The April index (Dec =100) for this quotation can thus be calculated as = 110,16. 40,85 n case a proportional conversion of the reference price is carried out. Case is straightforward, as the new reference price is genuine and it can siply be substituted for the old one. t ay be debatable, however, whether the previously recorded price developent up to and including the previous onth should be superseded in this way. The overlap case K requires ore careful treatent. Here, the index has to be calculated as =, (2) 0, 0 with the new, iputed reference price being correspondingly calculated as =, (3) ˆ0 0 where represents prices, and are the old and new products, respectively, and 0, -1 and each represent a particular onth. 0 is thus the reference onth.

9 STATSTCS SWEDE Rev (10) Let us say that, in the above exaple, the product eeting the new specifications was being sold as early as March at a price of SEK 44. A new basic price that is coparable with the price for the new specifications can then be calculated as = 41,8 40 The observed price change in case M, before the product disappeared, should not be eliinated. Rather, the index should in principle be calculated as 1 A 0, =,, (4) 0 i.e. iputation should be applied only for the tie after the product's disappearance. Where only the results up to and including the onth before the current onth can be used for iputation, the last factor in the expression A becoes the first iputation onth equal to one ( 1, ). The ost practical procedure is now to introduce a new "reference price" (reference index), which is the ratio between the aggregate and the product's index nuber for onth -1 as set out below: A 0, ˆ 0 = 100, where 0, 1 = (5) 0, 0 Thereafter, one siply inserts the index nuber for each new onth (Deceber = 100) for the series which represents the outgoing product. This works because 1 A 0, =, t 0 =, (6) A A 0, 0, 0, 1 = A 0, 1 ˆ 0 Cases O and R are siilar to case as far as the calculation of the new basic price is concerned. We can suarise the index calculation as follows; s 0, =, (7) 0 f where s designates the last onth with the old specification, and f represents the first onth with the new one. The new, iputed reference price is correspondingly calculated as ˆ, (8) 0 = 0 f s We note that, in onth f=, which is the new product's first observation onth, the index according to (7) is unchanged vis-à-vis onth -1.

10 STATSTCS SWEDE Rev (10) Appendix 2: putations The syste uses iputations for four different types of non-response: - Copensation for non-response: The price does not coe in by the deadline and is estiated using the last price to coe in, which is norally the previous onth's price. This ite is flagged, and the nuber of iputations, norally 1-3% of the saple, is docuented each onth. - o transactions: The price is estiated using the last observed price. The ites are flagged in the database. - roducts no longer sold: n cases where a copany has ceased trading or no longer iports/anufactures products under the particular tariff heading, the price developent is iputed using the price developent of a closely related product. This is flagged and is docuented at year's end. orally, 2-4% of the saple is iputed using this ethod at the end of the year. - Larger product groups not in the saple: Where easureents are copletely issing for product groups under S 2007, 5-digit level, with values over SEK 250 illion, or for a C heading with values over SEK 350 illion, these are generally included in the saple with their weights, not grossed up to the price reference onth level. The price developent for these is assued to be in line with that of another, closely related product group for which price easureents are carried out, or alternatively with that of the sae product group on another arket (hoes sales, export, iport). The absence of easureents ay ste, for exaple, fro the fact that the group of products concerned is characterised by unique or "tailorade" products or transactions. This eans that it is ipossible to easure prices for transactions which are coparable and recurring over tie. Another possible reason why the easureents ay be issing is that the basis for the weighting figure or the sapling frae is clearly out of date in relation to the new statistical year.