VERSION OF 14/01/2019

Description of country output generated by the coordinated analysis

Part 1: COVERAGE and DEMOGR Files

1.cccCOVERAGE

This output file is generated by setting DoCov = True in the run control file (RUN_ONSPROD.SAS), which calls the code in COVERAGE.SAS. The output contains basic information on the linked Business Register (BR), Production Survey(PS) and E-Commerce (EC) input datasets.

Output variables

EUK0Industry classification (see notes below)

SZ_CLSSize class

YEARYear to which data pertain

N_BRNumber of firms (from Business Register)

N_PSNumber of firms (from Production Survey)

N_ECNumber of firms (from E-Commerce Survey)

N_BRPSNumber of firms (merged BR, PS)

N_BRECNumber of firms (merged BR, EC)

N_PSECNumber of firms (merged PS, EC)

Emp_BRNumber of employees (from Business Register)

Emp_PSNumber of employees (from Production Survey)

WN_PSIGNORE. NOT USED

WEmp_PSIGNORE. NOT USED

Emp_BRPSNumber of employees (merged BR, PS)

Emp_PSECNumber of employees (merged PS, EC)

SRCCoding for different calls to COVERAGE.SAS (see notes below)

Notes

  • EUK0 should be read as “industry”. The entries in this column are taken from the ‘althier’ dataset provided by the co-ordinator (but note that the industry codes have been sorted alphabetically by SAS, and the ordering differs from that in ‘althier’).
  • Size class codes are common to all output tables and are shown in Table 1.3 of the User Guide.
  • SRC should be read as ‘SAS run control’. For those rows where EUK0=DISTR, Elecom etc (that is, from the ALT1 column of the ‘althier’ dataset), SRC=2 and data are tabulated by industry, size class and year. For rows where EUK0=10t4, 15 etc (from the EUK0 column in ‘althier’), SRC=1 and the data are tabulated by industry and year only.
  • SRC plays a similar role in other output datasets.

Example

Table 1 is a screen shot from gbrCOVERAGE.xls, where the SAS output has been converted to Excel. Note that values of -999 mean that the output has been suppressed for disclosure purposes. Also note that some rows have been hidden, in order to group together results for different size classes in a single year. In the highlighted rows, the number of firms in industry ‘Minter’ (Intermediate manufacturing), in size class 5, in year 2005 is 1458 in the Business Register (BR) dataset, 828 in the Production Survey (PS) dataset and 75 in the E-Commerce (EC) dataset. These numbers are unchanged in the matched BR-PS and BR-EC datasets, but the number of firms reduces to 52 in the matched PS-EC dataset.

Columns J, K, L and M show the employment totals for these sets of firms. For example, the 52 firms in the matched PS-EC datasets employ a total of 8485 workers. This is an average of 163 workers per firm, which is consistent with size class 5 (100-249).

Things to check

  • In the GBR output table there are some rows with sz_cls=0 and some where sz_cls is missing, reflecting zero or missing employment in the input datasets. This is OK, but none of the N_ or Emp_ variables should be negative.
  • In general, one would expect N_EC and N_PS to reflect the sample sizes of the two surveys, while N_BR should reflect the population of firms.
  • The Emp_ statistics are UNWEIGHTED, so EMP_ values based on sample surveys will be lower than the population estimates Emp_BR.

2.cccDEMOGR

This output file contains firm demographic data from the Business Register (BR) input dataset. This output file is currently generated automatically from the main code (ONSPROD.SAS), rather than controlled by the run control file. Thus this output file will be generated each time you execute RUN_ONSPROD.SAS.

Output variables

ALT0Industry classification (see below)

SZ_CLSSize class

YEARYear to which data pertain

STATUSEntrant, Exiter, Continuer, or One-year firm (see below)

COUNTNumber of firms

EMPEmployment

CO_POSAggregate of positive firm-level employment change

CO_NEGAggregate of negative firm-level employment change

Notes

  • Alt0 is the industry classification based on EU-KLEMS (in this case, the alt0 column of the ‘althier’ dataset, again re-ordered alphabetically by SAS)
  • Coding for STATUS: CO (standing for Continuer) if the firm is on the BR dataset in year = t-1, t, t+1; EN (=Entrant) if the firm is in the BRin t,t+1, not t-1; EX (=Exiter) if the firm is in the BR in t-1,t, not t+1; OY (=One Year) if the firm is in the BR only in year t.

Example

An extract from gbrDEMOGR attached as Appendix 2, again converted from SAS to Excel format. The highlighted rows show that, in industry 26 (Other non-metallic mineral products) in year 2004, size class 1, there were 2789 continuing firms on the BR, with total employment of 9828. 340 firms entered the business register in this industry and size class in 2004 (and survived into the register in 2005), with total employment of 713. 372 firms employing 1004 workers exited in 2004 (that is, they did not appear in the business register in 2005), and there were 67 firms (employing 136 workers) in this industry and size class that appeared only in the 2004 business register.

Columns CO_POS and CO_NEG show the sum of employment changes for status CO and EX (that is, for firms in the BR dataset in t and t-1), summed separately for firms with positive and negative employment changes. Not surprisingly, employment changes of exiting firms showed larger negative changes than positive changes (-376 versus +58).

Eric Bartelsman and Mark Franklin

Page 1 of 4

VERSION OF 14/01/2019

Appendix 1 – Screenshot from gbrCOVERAGE.xls

Appendix 2 – Screenshot from gbrDEMOGR.xls

Page 1 of 4