EUROSTAT
Directorate G: Business Statistics
Unit G-1: Structural business statistics /
Prodcom
Report on Eurostat Quality visit to Statistics Hungary
Purpose of the visit
The Eurostat Prodcom team is conducting a number of "Quality Visits" to the Prodcom teams in the NSIs, in order to discuss the methodology used, detect potential problems and identify best practice that can be shared with other countries. The visit to Statistics Hungary is the second of these visits.
After each visit, Eurostat will prepare a report on the discussions and, after checking it with the team visited, will load it to the Prodcom CIRCA website, together with any material such as Powerpoint slides provided by the team in the NSI.
Institution
Hungarian Central Statistical Office (HCSO)
The official Hungarian statistics is organized and coordinated by the "Hungarian Central Statistical Office" and its regional offices. Currently the number of permanent staff is about 1,050 in the central office, while the total number of people employed at regional offices is about 450.
Survey
The Hungarian Statistics of Industrial Production surveys the industrial commoditiesproduced and the industrial services that havebeen provided in Hungary during the reference period (i.e. the calendar year). The Hungarian Statistics of Industrial Production are produced by two distinct but interlinked surveys:
- An annual survey which covers all industrial products and services, and
- A monthly survey which provides monthly data for about 200 of the most important commodities produced in Hungary.
Relevance
The main users are:
- National Accounts, Producer Price Index
- Governmental authorities and public administration
- Trade associations and enterprises
- Research institutions and students
- Mass media
- Eurostat (PRODCOM)
- Other international organisations (e.g. UN, OECD)
Definition of the target population
The target population consists of all the industrial commodities that are produced and the industrial services that have been provided onHungarian territory during the reference year.
Sampling Frame
The sampling frame is drawn up from the Hungarian Business Register and includes all enterprises having their main or secondary activity within the manufacturing sector according to the NACE classification, plus hypermarkets regardless of their declared main or secondary activity.
The sampling and observational unit is the enterprise. For enterprises outside the manufacturing sector the observational unit may be the KAU.
Sampling design
The HCSO uses a cut-off sample. All information related to the units over a threshold is collected via the survey (annual and/or monthly survey). The information related to the units under the threshold is not compensated for.
The threshold of the annual survey (about 12,000 enterprises): all enterprises having an activity (main and/or secondary) within the manufacturing sector (e.g. NACE 10-37) with 50 or more employees for that activity and if that activity produces an annual turnover of 500 Million HUF or more.If it is necessary in order to achieve 90% coverage, enterprises with fewer that 50 employees may be surveyed.
The threshold of the monthly survey (about 2000 enterprises): all enterprises which manufacture at least one of the products included on the list of the most important products for the Hungarian economy (about 200 products) with 50 employees or more for an industrial activity and industrial output valued at 500 million HUF or more. The cut-off sample for the monthly survey is updated each month.
In the energy sector, enterprises with more than 5 employees are included in both the annual and the monthly survey.
Method of data collection
Statistics Hungary uses paper and on-line questionnaires. The HCSO uses customised questionnaires by pre-printing Prodcom codes as far as is possible based on the information of the previous year or the information coming from other statistics such as statistics on "Performances and expenses of enterprises".
The production involving subcontracting is reported separately. Also, the questionnaire design foresees a separate column for the turnover due to industrial activities (which is an approximation of the value of production) and another one for the turnover due to other activities (e.g. trade, transport, taxes and so on). This approach allows additional validation checks on erroneous reporting of the value of production by including non-industrial activities. Meanwhile the total turnover is compared with the turnover reported in other business surveys (e.g. SBS, STS).
Particular attention is paid to providing professional guidance to the respondents and keeping a close contact with them.
Product nomenclature
The HCSO uses a national version of the Prodcom List, namely the Annual List of Industrial Products and Services (BTO) compiled based on the Hungarian Product Classification(HPC).
The users have free-of-charge access to a CD version of the Hungarian product classification including a smart way of searching and scrutinising the correspondence with other classifications (e.g. CPA, CN, Prodcom). A similar electronic tool is available on-line on the official website of the HCSO.
Taking into account that Hungarian enterprises use the CN for their internal purposes (e.g. accountancy), a link table for CN-BTO conversion is created and released free of charge with the CD and on the website.
Currently, the BTO has about 12,000 headings. Therefore, it is a more detailed classification than the Prodcom List but, as from 2008, it has been replaced by a new classification (ITO) based on NACE rev 2 and including about 4000 headings. The first 8 digits of the ITO are identical with the Prodcom headings and the 9th digit defines the national products (about 400). Therefore, no ITO-Prodcom conversion table is needed any more.
Measurement errors
The main problems encountered during data collection are
- Some missing data;
- The products may be classified using the wrong product codes.
The following actions are taken in order to minimise measurement errors:
- During the whole data collection process, the HCSO staff maintains a direct and close contact with the surveyed units;
- The most important enterprises are individually checked by the statisticians on the basis of their experience and information about the particular unit and economic branch referred to.
Processing errors
A lot of micro and macro plausibility checks are made at different aggregation levels, between historical data and with data from other surveys (e.g. SBS, STS, foreign trade statistics).
The unit/item non-response units are compensated for using imputation as follows:
- The missing volume data are imputed using the reported value of production and the average price of the product based on the reported data of a minimum of three other enterprises.
- Unit non-response units are compensated for as follows:
–Imputation is based on the turnover provided in Structural Business Survey and using the structure of products for the activity referred at national level. Thus, the value of production is estimated from the share of turnover and the volume of production is estimated using the average price of the products.
–If no SBS data is available, the imputation is based on the data reported in the previous year, if any.
–The unit non-response units which are active industrial enterprises (e.g. having their main activity in the manufacturing sector) and for which the imputation based on SBS or previous year data is not possible, are compensated for using the turnover reported in the Business Register using the same approach as the imputation based on SBS data.
The national data with three contributors or fewer are suppressed from publication (i.e. the data are confidential).
ORACLE is used to perform all data processing.
Non-response errors
The average response rate for both the annual and monthly surveys is above 95%.
The main problem encountered by enterprises is the classification of their products to a correct code. Some difficulties are encountered in deciding if a product is industrial or agricultural (e.g. milk, chips, and potatoes).
In order to minimise the non-response rate, different reminders are foreseen(e.g. letters of demand, telephone contact) at different stages in the data collection process.
Timeliness, revision policy and practice:
Thepreliminary annual data are published on 30 June (t+180) and the final annual production data are released on 30 September (t+270).
The use of Prodcom statistics in National Accounts (NA):
- Prodcom provides useful information about the relevant products for NA, which is the output structure of industrial production at national level;
- The Prodcom statistics are used to allocate the domestic sales to the main types of economic expenditure (e.g. intermediary/final use);
- The Prodcom statistics are an important source to apply additional cross-checking;
The main disadvantages of Prodcom (from the NA point of view) are:
- It is based on too detailed a classification compared with NA requirements;
- No information about the small units (i.e. the units under the threshold limit which are not compensated for);
- Sometimes, there is poor consistency of Prodcom statistics with other statistics such as SBS or Foreign Trade Statistics.
Conclusions
The visit was very useful for the Eurostat team which learnt a lot about the way the Prodcom survey is conducted in Hungary and about the national uses of Prodcom.
M. Agafitei
B.D. Williams
1