Statistics Portugal

National Accounts Department / Short Term Statistics Unit

Mrs. Adelina Andrade.  

Instituto Nacional de Estatística (INE)

Avenida António José de Almeida, 5

1000-043 Lisboa, Portugal

Index of industrial production (IPI).

NACE Rev. 2            

Activity coverage:

Sections B to E (excluding Division36, 37 and 39) of NACE Rev. 2.

Size coverage:

No size coverage is applied. In each KAU 85% of turnover by product is selected indenpendently of the size of the  company that produces it.

Definitions correspond to the Commission Regulation Nr. 1503/2006

The production index represents the evolution of value added at constant prices. The definition used is in accordance with the Commission Regulation. 80% of the series measure production in physical quantities, 13% in terms of the number of hours worked and 7% in terms of deflated turnover. Depending on the activity, enterprises report physical production (in volume), Prices, finished products sold and number of hours worked.

Planned changes in information collected

None

Accounting conventions

Calendar month.

Reporting unit: Enterprise

Observation unit(s): KAU when the method of physical quantities is used.
Enterprise when number of hours worked and finished products sold methods are used.

All units in business register of section B to E (excluding Division36, 37 and 39).

Sample Frame: 25500

Units reporting: 1778 on the monthly Prodcom survey, and 864 from annex A turnover survey and 1 KAU (351) with admnistrative source for Base Year 2015

Portugal

Base 2015 series are only comparable since 2005. Data for Production Index is available since 1970

Base year: 2015

Indices

Month

Legal basis:

The general framework is the law”Lei 22/2008” from May 13th and STS regulations.

Obligation on units to provide data:

The survey is mandatory and fines up to EUR 50,000 be applied.

Planned changes in legal basis, obligation to respond and frame used

None

Data are transmitted to other international organizations (Eurostat, UN and OECD).

The National Statistical Law (”Lei 22/2008” from May 13th) establishes the main principles as well as the working methods and composition of the Statistical System.

NACE REV 2 with less than 4 units and below a response rate of 85% are considered confidential.

An advance release calendar that provides notice one-quarter in advance of the precise release dates is disseminated on the Statistics Portugal Internet website [http://www.ine.pt]. In addition, a two-month in advance release calendar is disseminated in Informação Estatística, (in Portuguese). Free copies can be ordered from the contact person shown on the dissemination formats page.

http://www.ine.pt/xportal/xmain?xpid=INE&xpgid=ine_calendarios&xlang=en

Simultaneous release to all interested parties

The data are released simultaneously to all interested parties by issuing the press release (“Destaque” in Portuguese) to the media and by simultaneously placing the data on the INE Internet website [http://www.ine.pt].

Identification of internal government access to data before release

No information is provided in advance to the Ministry before the press release  for information. For more detailed information see the document dissemination policy at Statistics Portugal - Web Portal (ine.pt)

https://www.ine.pt/xportal/xmain?xpid=INE&xpgid=ine_cont_inst&INST=70084023

Transmission to Eurostat and further use of the statistics

Data are transmitted to Eurostat by STADIUM using GESMES version 3.1 at the same time as data is released, or the day before under embargo.

Monthly, 30 days after the end of reference period.

Frequency: monthly, press release available at http://www.ine.pt/xportal/xmain?xpid=INE&xpgid=ine_destaques&xlang=en

Activity Coverage: NACE Rev. 2 Section B to E and their respective divisions. MIGS (capital goods, consumer durable and non-durable consumer goods, intermediate goods, energy). NACE 38 is also covered and published but not sent to eurostat.

Length of time series: from 01/2005 (NACE Rev.2, Base 2015)

Presentation: All of the series are transmitted and published as an index in raw data, calendar adjusted and seasonally adjusted forms.

Identification of ministerial commentary on the occasion of statistical releases: No ministerial commentary is attached to the release of data by the statistical office.

Data for production are published in the “Boletim Mensal de Estatística” (Monthly Statistics Bulletin) available in Portuguese language only and can be purchased by ordering it from the Customer Support or directly at http://www.ine.pt/ > On Line Catalogue > General Statistics.
The “Anuário Estatístico de Portugal” (Statistical Yearbook of Portugal) is a bilingual publication (Portuguese/English) and can be purchased by ordering it from the Customer Support or directly at http://www.ine.pt/ > On Line Catalogue > General Statistics. It can also be viewed and/or downloaded in pdf/xls format, free of charge, at http://www.ine.pt/ > Statistical Data > General Statistics (X) Monthly bulletin.

Data are published in www.ine.pt (press release, excel format file and historical data access on-line database).

http://www.ine.pt/xportal/xmain?xpid=INE&xpgid=ine_base_dados&contexto=bd&selTab=tab2

There is no public dissemination of micro data. Micro data are available for scientific investigations after confidentiality treatment of data.

Data are sent to Eurostat to be used in European aggregates and/or to be released as national data.

Dissemination of documentation on methodology and sources used in preparing statistics

A detailed methodological note in Portuguese only for base year 2015 is available free at www.ine.pt document nº 180 at: http://smi.ine.pt/DocumentacaoMetodologica/Detalhes/1479

Description of standard tables produced

The production index is published for 28 divisions of NACE Rev. 2, for the Sections B, C, D and E and also for the Main Industrial Groups. There is no regional breakdown.

Not available.

Monthly micro data are analyzed, checking the major changes on monthly and annual growth rates at enterprise level and also cross checking with other related variables of the unit’s answer. If necessary, units are contacted by telephone or email. Indices are analyzed to check the coherence of the results and, if necessary, the procedure is restarted.

All these procedures respect the quality policy defined in INE’s Code of Practice 

https://www.ine.pt/xportal/xmain?xpid=INE&xpgid=ine_cont_inst&INST=278796088

Indices are based on reliable data; Publication of results is always punctual; Further methodological details and other information can be found in point 14; multiple plausibility checks with different sources are done.

Short-term statistics are used by different national (such as Central Bank, Ministry of Finance and researchers) and international institutions (Eurostat) and non-institutional users, according to the needs expressed in the Statistical Superior Council (SSC). The SSC is the body where are analyzed the user’s needs and defined the Statics Program.

Apart from occasional user surveys aimed at specific target groups, Statistics Portugal conducts 6 permanent user satisfaction surveys targeted at the users of 6 different SP services: the User Support Service (responding to information requests), SP Portal (website), SP Libraries, Study visits hosted by SP, Training sessions to basic and secondary school teachers aiming at improving statistical literacy and Training sessions in the context of Statistics Portugal Information Network in Higher Education Institutions Libraries.

User satisfaction measured in each of the permanent user satisfaction surveys is summarised into a user satisfaction index using the extreme balance responses (EBR). This approach values the extreme answers in detriment of the “central” valuations which tend to represent a less expressive dissatisfaction or satisfaction. The scale weights are as follows: EBR = F1 * (-1) + F2 * (-0,5) + F3 * (-0,25) + F4 * (0,25) + F5 * (0,5) + F6, where Fi means the relative frequency of observed values for each one of the categories i (i= 1,…,6). Data is referenced in a metric scale between -1 and +1, where values next to -1 mean full dissatisfaction and values near +1 mean full satisfaction.

Results have been steady since 2014 and the user satisfaction index reached 0,64 EBR in 2018.

The Indices are fully compliant regarding STS-Regulation. There are no requests of additional information.

It is not possible to indicate a classical sampling error.

Response Rate

A threshold of 85% is defined for dissemination. Nevertheless, usually data is published with a response rate of more than 90%. Data for MIG and for sections of NACE can be published with a response rate bellow 85% since users are informed.

Actions to speed up or increase the rate of response

A lot of checks are carried out at the micro and macro data level. For example the following processing errors may occur:

A) Treatment on micro-data-level

• Human errors during the data editing process: these errors are often recognized with automated checks – online validation of chain and YOY rates of the unit;
• Plausibility errors: a lot of checks are carried out automatically (but under possibility of human interaction): e.g. check of completeness/ sums/ horizontal logical dependences/ vertical relations/ average rates/ minimum/maximum values, etc..
• Imputation errors: the imputed value is not within a plausible range of the per capita values by unit

The qualified checking of original respondent data, contacts to reporting units for corrections and the usage of sophisticated plausibility routines shall minimize the processing errors on micro-data-level.

B) Treatment on macro-data-level

• Final checks with analyze-tables (aggregates / more complex relations). These tables include all variables of the STS-survey. The macro data tabled are verified by computer programs but also by the staff whether these macro-data as publication figures look plausible. The persons responsible for the index calculation have their own computer programs to test the index-plausibility of the different data at 3 and 2 digit level and for all aggregates. If inconsistencies are detected the data must be verified at micro-level again by using the raw data of the questionnaire.

•  Model effects, such as:

- Imputation errors: the usage of previous period values tends to result in under or overestimating the real development (especially if big enterprises' values are missing). Therefore specific imputation methods for missing data are used to correct these errors.

- Overcoverage / Undercoverage:  Classification errors and Overcoverage are observed especially at the reference periods of January due to changes of the respondent units (i.e. deaths of enterprises, changes in enterprise structures, new classifications). Undercoverage:

One of the reasons is the loss of responding units due to false classifications and the difficulty to find alternative enterprises during the reference year, yet it is not possible to quantify the under coverage rate.

Recommendations previewed in the Final Report of the Task Force Retail Trade Quality  are applied to all indicators, mainly book 3.

The response rate is normally very high for the first release, nevertheless in case of non-response the largest units are contacted by telephone and units may also be contacted by fax. Ultimately a fine may be imposed.

This indicator is released 30days after the end of the reference period. The monthly data is considered to be definitive one year after the end of the calendar year in which the reference month falls. Revisions are made to the two previous periods.

Timetable of data collection:

Questionnaires are sent around the 25th of the reference month. The enterprises are requested to answer within 10 days (calendar days) of the end of the reference month.

STS deadlines have been respected.

Due to harmonized STS indices in the European Union, it is possible to compare calculated Portuguese aggregates with those of other European countries. There is no regional desegregation, but all national territory is covered, including the AR’s (Azores and Madeira).

To ensure comparability over time, it is tried to keep the data base as steady as possible. STS indices are compared over time (previous month and year). Comparisons with NA accounts (quarter and annual) are made every 2 years. Base 2010 series have been back casted until 2005. Base 2005 series have been back casted until 2000. Comparisons with annual prodcom survey are done on a regular basis.

Cross-checks with absolute STS-values (previous month and year).  At the first release moment no other data is available for comparisons.

Confrontations are made, on a monthly base, with turnover surveys and with external trade. This comparison with external trade data is only possible for the first revision.

Raw Data: Aggregates are consistent with all sub aggregates  (Sections,  Migs and 2-digit level). 

Calendar and seasonal adjustment  use direct approach, which means that only Total and Migs are consistent, and all other sub levels are inconsisitent.

Cost 10249 hours, burden 2062 hours per month, acordding to 2016 survey.

The document “Politica de Revisões” (revision policy) contains the guidelines to follow. This document can be found following the link:

https://www.ine.pt/xportal/xmain?xpid=INE&xpgid=ine_inst_politrevisao

Data sent to Eurostat is according to the same revision policy describes above.

Type of source: Data are obtained through statistical surveys and administrative source

Frame on which the source is based: The General Register of Statistical Units (FGUE – Ficheiro Geral de Unidades Estatísticas) for enterprises and Prodcom questionnaire for products (IAPI).

Sample of census: Cut-off sample survey for monthly Prodcom and Stratified Sample Survey for NACEs from Turnover Industry Survey. Exhaustive for administartive source.

Criteria for stratification: No stratification, besides NACE. The sample is drawn by cut off method from the whole population.

Threshold values and percentages:

1) Prodcom survey:

 -The most representative products in each 3 digit heading are selected. On average, the products in the basket account for 90% of total output in the group.

 -The sample is 1889 KAU corresponding to 56 3 digit level NACE.

2) Series from other monthly surveys (Turnover, Prices and Hours Worked) follow the rules defined in each of these variables, therefore originating 27 3 digit level NACE indices.

3) Administrative source is exhaustive.

Frequency of updating the sample: every month entrerprises are updated for demographic purposes. The baskest of products is analysed and compared with prodcom survey every 2 years. Every five years a complete new basket of products and entreprises is redesigned.

Data are collected on a monthly basis.

Questionnaires used in the survey: The questionnaires that are used to collected data for the method of physical quantities or production prices use the Prodcom list as reference. For finished products sold and the number of hours worked the same questionnaire is used as for the turnover, labour input.

Planned changes in national questionnaires: None

Data collection media: The survey is carried out by an electronic questionnaire.

Planned changes in data collection methodology: none

Comparisons on line with previous month and n-12. Off line comparisons within the stratum of companies are done. Data is validated off line with previous and n-12 months.

For Gesmes, data comes from specific software devloped intenally, that validates the data and sends it according to what is published.

Estimates for non-response: Non-response is imputed with the data available for the same month of the previous year, taking into account a trend coefficient of the same enterprise between n-1 and n-13, for the Turnover and Hours Worked.

For Prodcom survey Non-response is imputed with the previous 6 month average of quantities produced by the same unit. In some special cases expert intervention for estimaded values is possible.

If the non-response is due to the closure or temporary suspension of activity, no imputation is made.

Estimates for grossing-up to population levels: The sample results are not grossed-up; weighted with the reference population weights.

Type of index: The production index is expressed in a Laspeyres index and refers to a rate of change in the latest period compared to the previous period.

Method of weighting and chaining: The weighting factor is the gross value added at factor cost for the 3-digit level of NACE Rev. 2, whereas below the 5 digit level of CAE the weights are the value of production, obtained from the PRODCOM survey.

Planned changes in production methods: none

Frequency of weight update: Every five years.

WDA, SA and other calculations:

The monthly series are working days and seasonally adjusted, using JDemetra+, Tramo-Seats and/or ARIMA X13.

 CALENDAR ADJUSTMENT

- - regressor used: country-specific (national holidays); no specific calendar for detailed series is used.

- moving holiday effects are adjusted for easter

- leap year adjustement

- carnival is considered holiday except for years 2011 till 2016

- type of calendar adjustment: regARIMA

- special attention was made for the 4 national holidays supressed during the period of Troika's intervention in Portugal

 other adjustments

When finished products sold are used the values are deflated using the producer price index to produce a volume index.

Detection and replacement of outliers - YES

Seasonal Adjustment

the 84 series area analysed manual, 30 series used additive factors and 54 multiplicative.

Direct adjustment / indirect adjustment via components

- indirect adjustment is used. Divisions and sections of NACE rev2, as well as MIGS are obtained from 3 digit level adjusted series.

no comment