http://orcid.org/0000-0002-8392-0536
![Sample our Global Development journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5929/TOC-Subject-Sample-2021-Global-Development.jpg"})
ABSTRACT
This paper analyzes regional price differentials in Poland at the NUTS-2 and NUTS-3 levels. It applies unique raw-price data and calculates regional purchasing power parity (PPP) deflators for the 16 NUTS-2 regions. It then estimates PPP deflators for the 66 NUTS-3-level regions by applying the multiple imputation approach. Finally, it verifies whether these are intra- or interregional price inequalities that have a greater influence on the overall price inequality level. It is found that the price levels are significantly higher than the average in the better-developed regions and lower in the lagging ones. It is also found that it is the intra- rather than the interregion differentials that influence more the overall inequality level.
DISCLOSURE STATEMENT
No potential conflict of interest was reported by the authors.
SUPPLEMENTAL DATA
Supplemental data for this article can be accessed at http://dx.doi.org/10.1080/17421772.2018.1503705
ORCID
Bartlomiej Rokicki http://orcid.org/0000-0002-8392-0536
Geoffrey J. D. Hewings http://orcid.org/0000-0003-2560-3273
Notes
1 See Roos (Citation2006) for a more extensive review of the geographical price differentials literature.
2 NUTS = Nomenclature of Territorial Units for Statistics.
3 In NUTS-3-level regions, the data on prices of representative goods and services are not available. Hence, we cannot simply follow the EUROSTAT/OECD methodology.
4 LAU = local administrative unit.
5 The small number of representatives means that estimated deflators may be not accurate, especially for particular categories of expenditure.
6 The representative products are chosen by the Polish Central Statistical Office. The centrally fixed list of representative items of goods and services remains obligatory and unchanged for all regions covered by the price surveys throughout a given year. From the description’s point of view, the list covers two types of representative items: (1) products precisely described including their substantial parameters, determining their features; and (2) products representing narrow assortment groups such as clothes, underwear, footwear etc., which are supplied onto the market in short series. In the second case one might face a potential bias if qualitatively different varieties are included as representatives. However, in the Polish case this should not be a problem. First, because price collectors are required to survey the most popular shops in a given location. In this sense, possible differences in the quality of products between regions should be a consequence of their availability (and not price, a collector’s taste or choice). Second, the data on representative items are being collected from up to 26 locations in each NUTS-2 region. As a result, we have no reason to believe that our regional data on representative items prices are biased. Therefore, we did not perform any raw data adjustment, as in the case by Kocourek et al. (Citation2016).
7 We have verified the structure of consumption within the main expenditure categories available from the publications of Polish Central Statistical Office (e.g., Household Budget Survey in 2012), and the consumption patterns are fairly similar across regions. We do not have access to the raw data from the Household Budget Survey for the whole period. However, we have used individual data from the 2010 survey to compare the results based on homogeneity assumption with those that allow for heterogeneity of spending within different expenditure categories. We find that both are almost identical (see the supplemental data online). Other studies also show that the inclusion of lower-level consumption heterogeneity does not necessarily influence the results at the aggregate level significantly (e.g., Kim, Hewings, & Kratena, Citation2015). Hence, the assumption concerning homogeneity of spending within the main expenditure categories across regions appears to be reasonable and should not lead to the significant bias of our price deflators.
8 This is due to the structure of the housing market in Poland. Here, the rental applies almost exclusively to the apartments in the biggest cities. At the same time, most Polish families own their own houses or apartments (either under the private or collective ownership). Apart from historical reasons (most of the families who were renting apartments became owners after 1989), this is also the consequence of taxation policy. Property tax is low and does not take into account the value of the property. Among the representatives we use to calculate indices related to house rental, we include services such as the rental of a parking lot, administrative costs in apartments, costs related to the use and service of elevators, cleaning service etc. Furthermore, these are energy sources that triple the share of rental in whole spending within the COICOP04 – in their case, the description of representatives is very detailed and guarantees full comparability across regions. Given the above, we believe there are no reasons to assume that deflators estimated for the Housing category do not reveal real differentials in regional prices.
9 The detailed time-series of the deflators are available from the authors upon request.
10 Going from north to south, these are Zachodniopomorskie, Lubuskie and Dolnośląskie.
11 A category representing typical manufacturing activity.
12 On average, better developed regions are found to have higher prices. These regions have higher regional GDP, a lower share of agriculture, higher wages, higher population density and lower unemployment. The dummies for the western and eastern border regions were included in order to account for the impact of cross-border retail trade (higher prices close to Germany; lower prices close to the former Soviet Union). All the main variables are statistically significant at 1%. Owing to lack of space, we do not report the regression results, but they are available from the authors upon request.
13 Although we compared different data sets based on 5, 10, 20, 50 and 100 imputations, we still did not find significant differences in terms of the imputed PPP values.
14 Owing to lack of space, we do not report calculated inflation rates, but they are available from the authors upon request.
15 In accordance to the New Economic Geography models, this could be due to the faster agglomeration process triggered by lower trade costs.
16 Detailed results are available from the authors upon request.
17 A family of inequality measures that meets the criteria of mean independence, population size independence, symmetry, Pigou–Dalton transfer sensitivity, decomposability and statistical testability (e.g., World Bank, Citation2005).
18 The Gini index cannot be decomposed into within and between-group inequalities. Owing to lack of space, we do not present figures for the remaining generalized entropy indices, but they are available from the authors upon request.