Page 176 - Special Topic Session (STS) - Volume 3
P. 176

STS538 Pedro Luis do N. S. et al.
                     IBGE still has no access to scanner data. However, a peculiar feature of the
                  fiscal  system  of  Bra-zil  points  for  another  source  that  can  be  even  more
                  attractive than scanner data. The fiscal legislation in Brazil requires that every
                  legal  sales  transaction  performed  by  a  store  (brick  and  mortar  or  online)
                  generates  an  individual  electronic  record  with  information  that  includes:
                  description and code of each product sold; the price paid per unit of each
                  product sold; the total price paid by product; the total amount paid; the date
                  and time of the sale; the media used for payment (cash, credit or debit card);
                  and  detailed  information  on  the  establishment,  including  its  unique  fiscal
                  identification code (the CNPJ). In some cases, information on the buyer is also
                  available, especially for transactions where buyers are enterprises.
                     The information contained in the e-records is as useful as those contained
                  in scanner data sets and would therefore provide one of the best data sources
                  for CPI compilation. Such information is already provided to government fiscal
                  entities, covering all the legal transactions performed by all retailers in the
                  country. Currently the e-records are collected by the states’ tax authorities and
                  consolidated for the whole country by the Federal Revenue Service.
                     The access to the information contained in the data base of the e-records
                  could provide many improvements for the NSCPI. It should be possible to
                  expand the coverage of the NSCPI to finer geographical levels, and to compile
                  CPIs for different target populations based on consumer baskets constructed
                  for different profiles (finer level of income ranges, for instance). The use of
                  more powerful superlative indices would also be possible since information on
                  price  and  quantities  are  available  in  such  data  sets.  The  e-records  can
                  potentially provide information or the derivation of more frequent updates of
                  the CPI baskets and the weights of the basket components. The number of
                  elements in the basket could also be increased, since the “collection costs”
                  which  scales  with  the  number  of  items  in  the  baskets  and  constrains  the
                  number of elements to be priced are not present here, though processing and
                  storage costs of the e-records should be considered. The indices  could, in
                  principle, be calculated in higher frequencies since the information is updated
                  more frequently.
                     Important challenges are also posed for the use of the e-records for CPI
                  compilation. The first is the access to such data. The e-records are protected
                  by fiscal confidentiality laws, and so far, IBGE has not managed to get legal
                  rights  to  access  to  the  transactions  e-records.  To  try  to  circumvent  this
                  restriction initial negotiations were held with the tax authorities for the access
                  but so far have not succeeded.
                     Another  important  point  to  consider  is  that  product  descriptions  and
                  classifications in the e-records are probably not in straight correspondence
                  with the classification required by the CPI. Such analysis and the integration of
                  the e-records into a classification and unified product description system as

                                                                     165 |I S I   W S C   2 0 1 9
   171   172   173   174   175   176   177   178   179   180   181