Challenges of Using National Accounts Poster.Pdf
Total Page:16
File Type:pdf, Size:1020Kb
Challenges of Using National Accounts for a More Frequent Update of the Brazilian CPI Weights 16th Ottawa Group Meeting Rio de Janeiro – Brazil Rodrigo V. Ventura May, 2019 [email protected] Background Motivation Overview of the IBGE’s CPI • The upper-level substitution bias • Various sources of data can • The HBS is the main data source The bottom-up approach of the Deriving the weighting pattern: in a fixed basket index is likely to complement each other by for deriving expenditure shares Brazilian CPIs within the SNIPC the Brazilian Household Budget increase with the age of the establishing more reliable and for the goods and services weights. accurate general weights. covered by the official Brazilian Survey – POF CPIs produced by IBGE under • Increasing the frequency of • One of the main sources the scope of the SNIPC. Both the national IPCA and the national INPC updating weights minimizes the recommended and used for this are obtained by aggregating the regional ≈ 1,000 interviewers applied 7 distorting impact that changes in purpose is the System of National indexes through a weighted arithmetic mean. questionnaires during 7 days in 75,000 consumer preferences and item Accounts (NA). • Hence, weights updates depends The 16 areas have their own baskets of households distributed in 1,900 substitutions might have on the • There are two main practical on how often the HBS is products and for each of them separate CPIs municipalities. index. advantages in using NA data for conducted, which unfortunately is are calculated. • It is often recommended in CPI estimating the weights: taking place on a less frequent literature the basket updating of basis (more than seven-year ➢These data are disseminated weights as frequently as intervals) because the survey annually on a calendar basis. possible – at least every 5 years costs and budgets constraints. São Luís/MA Fortaleza/CE Belém/PA – in order to ensure that they do ➢The scope and coverage of CPI MG CPI MA CPI AC CPI DF CPI PA Recife/PE not become unrepresentative Household Final Consumption CPI ES CPI CE Aracaju/SE Rio Branco/AC CPI Brazil Brasília/DF Salvador/BA CPI BA CPI SE and irrelevant. Expenditure (HFCE), adjusted • In this sense, NA can be an (IPCA / INPC) Goiânia/GO Belo Horizonte/M to cover all expenditure by alternative and complementary G CPI RJ CPI PE • The Household Budget Surveys Campo Grande/MS Vitória/ES households within the source for deriving CPI Rio de Janeiro/RJ CPI SP CPI MS CPI PR CPI GO CPI RS (HBS) is in general the primary Curitiba/PR São Paulo/SP economic territory, is expenditure weights using the data source for deriving weights. consistent with the scope and HFCE estimates. Porto Alegre/RS • Updating of weights depends on coverage of the Brazilian CPI POF 2017-18 how often a HBS is conducted. (as defined by the domestic However, a program of annual concept). surveys could be very costful. Challenges of using National Accounts for weight updates Differences in scope and Non-integrated Impact of revisions of definition of consumption classification structures National Accounts data on Producing annual weights for the individual metropolitan areas between CPI and NA CPI weights Use of HFCE definitive Concordance mapping Exclusion of some non- Pro-rata approach: assumptions between HFCE and CPI monetary expenditures of results obtained on a (‘translators’) the HFCE financial year basis • No relative change in the households 1 expenditure proportions of a given Using more detailed data from Imputed rentals (owner-occupied NA data for a given year t are metropolitan area with respect to the NA (at the level in which the housing), Financial Intermediation revised several times (annual rest of the metropolitan areas surveyed 2 works of balance of supply Services Indirectly Measured cyclical revisions) before they can and demand of products are (FISIM), and Not for Profit be considered as final (that • For each metropolitan area, ratios by carried out) Institutions Serving Households occurs only in year t+2) SNIPC aggregates are fixed along the 3 (NPISH) period between HBSs HFCE data HFCE weights Stability of the CPI conceptually aligned disaggregated for the weights series obtained CPI weights at the regional level (metropolitan areas) partially updated via HFCE data for CPI weighting 19 SNIPC subgroups via NA database purpose First empirical results Conclusions 5,0 Communication 5,8 12,0 5,1 Courses, reading and stationery 7,1 10,8 2,9 10,7 Recreation, smoking and photography 4,7 7,1 • The results show the Alternative IPCA (NA weighted) Personal services 5,5 10,0 2,3 Personal cares 2,3 tracking closely to the Original IPCA (as published). 5,4 Health 5,2 3,4 Pharmaceutical and optical products 3,3 • Average annual inflation rate: 6,9% (Alternative IPCA) x 6,8% 17,9 8,0 Transport 19,3 0,1 (Original IPCA) Fabrics 0,1 6,5 6,6 Jewelery and bijouterie 0,8 6,4 6,4 6,3 0,3 6,1 6,1 6,2 5,9 6,0 5,9 1,9 5,8 • Cumulative inflation rate: 59,2% (Alternative IPCA) x 58,3% Footwear and accessories 1,9 6,0 SNIPC Subgroups 4,4 Clothing 4,5 (Original IPCA) 0,3 Repair and maintenance services 0,4 Average annual inflation Electrical and electronic devices 4,1 1,6 4,0 3,2 Furniture 2,2 4,2 • The upper bias of the index – inherent to the calculation Fuel and energy 4,6 5,5 Housing charges and maintenance 8,7 2,0 formula based on the measurement of a fixed basket over 6,0 Food outside home 7,9 Food at home 20,3 time – was relatively counterbalanced by the impact 14,7 0 5 10 15 20 25 weights % 0,0 that the variation of the prices of Food at home had POF National Accounts 2010 2011 2012 2013 2014 2015 2016 on the general IPCA. Official IPCA (as published) Alternative IPCA (NA weighted) • The subgroup gained weight in the alternative structure and recorded average annual inflation higher than the average Average IPCA weight compared to the average weight Official IPCA (as published) compared to an alternative IPCA rate of the total economy. calculated using National Accounts data as the primary calculated using National Accounts data as the expenditure expenditure source (%) - Brazil, 2010 source for the weights - Brazil, 2010-2016 • The international literature does not recommend replacing the HBS with the National Accounts. The HBS at regular intervals (less than 5 years) remains a 160,0 fundamental element for the general updating of 4,6 Communication 3,9 5,0 index no. weights, especially in the more disaggregated Courses, reading and stationery 4,5 2,8 Recreation, smoking and photography 4,1 150,0 classification levels for which there is no data availability 7,1 Personal services 6,5 2,2 Personal cares 2,6 via National Accounts. 6,9 Health 5,0 3,0 140,0 Pharmaceutical and optical products 3,5 15,4 Transport 18,4 0,0 • In a scenario of budgetary constraint, the Fabrics 0,1 1,0 Jewelery and bijouterie 0,4 130,0 complementary use of new data sources (National 1,7 Footwear and accessories 1,8 SNIPC Subgroups 4,0 Clothing 3,9 Accounts, for example) would be an advantageous 0,3 Repair and maintenance services 0,4 120,0 2,8 alternative to ensure the most frequent updating of Electrical and electronic devices 1,7 2,8 Furniture 2,2 the CPI weights and, therefore, to ensure its credibility in 4,2 Fuel and energy 5,3 6,1 110,0 Housing charges and maintenance 10,4 the midst of an economy in constant transformation. 6,9 Food outside home 8,8 23,1 Food at home 16,5 0 5 10 15 20 25 100,0 weights % 2009 2010 2011 2012 2013 2014 2015 2016 • This is a question of great importance, especially when POF National Accounts Base: 2009=100 considering the economic and social costs caused by Official IPCA (as published) Alternative IPCA (NA weighted) a bias in the price index, caused by an obsolete basket of products and outdated weights. Average IPCA weight compared to the average weight Index comparison between official IPCA (as published) calculated using National Accounts data as the compared to an alternative IPCA calculated using National primary expenditure source (%) - Brazil, 2016 Accounts data as the expenditure source for the weights.