Economic indicators of territories (experimental statistics)

To meet the demands of data users for a more detailed territorial breakdown of statistics, the CSB has compiled experimental data on the economic indicators of enterprises by territorial breakdown. 

The indicators are calculated based on the person's home and workplace address, which should be taken into account when interpreting the indicators. For the self-employed, the person's place of residence is taken as the workplace.

New data sources and methods are used in the calculation of experimental statistics, in an attempt to expand the range of statistics or the level of detail according to the needs of users.

It should be noted that experimental statistical methods are not fixed, tested, and internationally harmonized, and can be changed to improve data quality.

The CSP publishes experimental statistics to receive user feedback, evaluate the analytical potential of the data, its relevance to reality and user needs. By publishing experimental statistics, the CSP provides data users with new sources of information for decision-making.

• LKAUs of economically active enterprises in the market sector, their added value and value of output (production value until 2020).
• Persons employed in legal entities, their labour earnings and wage-equivalent earnings. Persons who have received income.

Economically active enterprises – merchants, agricultural holdings and fishermen farms, state and local government budget institutions, funds, foundations and associations and natural persons – performers of economic activity – operating in the territory of Latvia and covering all kinds of economic activity classified in NACE Rev. 2, and persons employed by them. Persons who have received income.

N/A

Statistics database

N/A

Data on output value is available from 2021 and is not comparable with data on output value available until 2020.

N/A

The value added and output value of the LKAU (production value until 2020) are calculated for economically active enterprises in the market sector, using publicly available Annual Reports and Consolidated Annual Reports. The average annual number of employees is calculated using the SRS and statistical reports 1-year “Comprehensive Report on Enterprise Activities” and 2-year “Report on Labor”. It is calculated by counting the number of employees on the last date of each month of the reporting year and dividing the sum by the number of months in the reporting year.

The LKAU is determined based on the data of the structural units of enterprises registered with the SRS, supplemented with information provided by the enterprise itself in the statistical report 1-DVV "Local activity units of an enterprise (institution)", i.e. the address of the LKAU, type of economic activity (industry) and average number of employees.

Information about LKAU is not available:

1. if the company (institution) has not registered it with the SRS and/or has not reported or reported it incompletely/partially in the statistical report 1-DVV (for example, only the main structural units without further decoding);
2. if the company (institution) does not have to submit a 1-DVV report because it has only one type of economic activity and does not have any structural units registered with the SRS;
3. LKAU are not listed for funds, foundations and associations;
4. VUUs are not recorded for agricultural and fishing holdings whose main activity is NACE A “Agriculture, forestry and fishing”.

LKAU for public organizations (except cooperative societies) and farmers and fishermen were added from the data of structural units of enterprises registered with the SRS.

For agricultural and fishing holdings whose main activity is NACE Section A "Agriculture, forestry and fishing", data are available only on the number of persons employed.

To specify the location of the LKAUs for natural persons (self-employed) and agricultural and fishing holdings with one LKAU and NACE Rev. 2 code starting with 01 or 03, data from the Statistical Register of Agricultural Holdings were used.

For NACE K section “Financial and insurance activities”, 2017 data is available only on the number of persons employed.

For auxiliary activities, even if they are located in a separate unit, such as administration, accounting, data processing, process monitoring, purchasing, transportation, repairs, etc., the value of production, value added, is not calculated. Auxiliary activities are carried out so that the unit can produce goods and services for third parties or to facilitate this production. The products of auxiliary activities are not supplied to third parties. All costs of an enterprise's auxiliary activities are attributable to the main and secondary activities, and therefore to the enterprise's GVA.

To determine the reorganizations of legal entities, data from the Business Register (BR) on enterprises, public persons and institutions, reorganizations of legal entities, and branches of foreign merchants were used.

Ministry of Education and Science (MES) data on teachers in educational institutions and State Education Information System (SEIS) data on educational institutions.

The State Revenue Service (SRS) data for calculating personal income was used from annual income declarations (including declaration types such as D1, D1¹, D2, D2¹, D3, D3¹ and income from capital gains), notifications on amounts paid to individuals, employer reports (DDZ), self-employed or employee reports, reports of seasonal agricultural workers' income tax payers and personal income tax reports on the income earned and tax paid by an individual - non-resident - in the Republic of Latvia, on amounts paid to employees of micro-enterprises and patent payers. Information on employees and the full list of employees prepared by the SRS were used to determine professions.

The places of residence were determined using the Register of Natural Persons (FPR, until 28.06.2021. Population Register (IR)) maintained by the Office of Citizenship and Migration Affairs, the population estimate provided by the CSB (see https://stat.gov.lv/en/metadata/5911-population-and-key-demographic-indicators/det#stat_process Data collection → Population determination), the results of the 2011 Population Census (TS) and the Real Estate State Cadastre Information System (NĪVKIS) maintained by the State Land Service.

State Address Register, including spatial data of territorial subdivision units, as well as the boundaries of neighbourhoods and densely populated areas published on the Latvian Open Data Portal.

Enterprise data structure

The input data used for companies has a four-level hierarchy:

UUK

An enterprise is the smallest aggregate of legal units, which is an organizational unit that produces goods or services and has a certain degree of autonomy in decision-making, in particular regarding the allocation of its current resources. An enterprise carries out one or more activities in one or more locations. An enterprise may be a single legal unit.

Includes only those enterprises that produced goods, provided services or employed people during the reference period, regardless of whether they were active for the entire reference period or only part of it. Passive (currently inactive) and inactive enterprises are not included. It also includes local units (branches) that do not constitute a separate legal entity and that are dependent on foreign enterprises.

All of the above economic indicators are known for each company.

NMK

Each company can have one or more legal entities (taxpayer registration code).

From the SRS DDZ data, a list of employed persons (with personal identification codes) is known for each legal entity each month.

The annual average number of jobs in the NMK is calculated by determining the number of employees in the NMK each day and calculating the annual average number of employees per day. The calculation does not include persons who are on long-term absence (for example, on parental leave), if the period of absence is longer than 60 days.

The number of employees for other types of employment, such as self-employed, patent payers, seasonal agricultural workers, is calculated according to the tax payment periods made by the taxpayer. A person is considered employed if there were income or patent payments made during the relevant period.

Long-term employment is calculated as the arithmetic average of the number of employees per calendar day of the year with an employment period longer than 150 days.

Short-term employment is calculated as the average of the fourth quintile of the number of employees on a calendar day of the year with an employment period shorter than 150 days. This calculation reduces the impact of one or a few days of employment (for example, election commissions) on the number of short-term employees, while not affecting those legal entities where short-term employment is seasonal in nature.

In order to reduce the multiple recruitment of employees to jobs in reorganized NMKs (both before and after the reorganization), reorganizations were identified and NMKs that were active for most of the year were selected for further data processing.

First, cases were selected where the NMK either has no employees in the following years, or the NMK has been divided by splitting or separation according to the UR data (the first NMK continues to exist) during the year before or after the transfer of the majority of employees. The reorganizations of UR legal entities were excluded if the legal entity registration number, which concludes the reorganization process, is the parent company ID of the foreign merchant branch, which is not indicated in the data on foreign merchant branches.

Then, in both cases, a pair of dates was determined when the majority of employees left the first NMK and started working in the second NMK. If at least 10 employees did not transfer from the first NMK to the second on the day when the majority of employees made the transfer, and later (it may be earlier if the NMK is divided by division or separation), the NMK was not considered to be reorganized. The number of transferring employees must either constitute more than half of all employees of the first NMK, or the NMK must be divided by division or separation.

In the 2021 enterprise data, the following public entities and public institutions (legal units) with more than 10 employees are not separately identified and are excluded from data redistribution:

• Jūrmalas pilsētas Lielupes pamatskola (NMK 90000051576),
• Rēzeknes novada pašvaldības iestāde Veselības un sociālās aprūpes centrs “Malta” (90001863245),
• Rēzeknes novada veco ļaužu pansionāts (90000931826),
• Vārkavas apvienības pārvalde (40900036931).

LKAU

A local kind-of-activity unit (LKAU) is an economic activity carried out at a specific address (LKAUs are also allocated to auxiliary sectors). If there were two or more different economic activities at the same address (for example, a shop and a café), then each of them was considered a local kind-of-activity unit. 

For each LKAU, the average number of employees per year is known. Each LKAU has a known link to an enterprise and a legal entity. A LKAU can only be linked to one legal entity.

LKAU for public organizations (except cooperatives) and farmers and fishermen's farms, for which they are not listed, were added from the data of structural units of enterprises registered with the SRS. Since the addresses of structural units are in text form, address codes or coordinates were assigned either by processing address records or manually.

For natural persons (self-employed), the place of residence of the person is taken as the place of work, while for natural persons (self-employed) and agricultural and fishing holdings with one LKAU and NACE Rev. 2 code starting with 01 or 03, the location of the LKAU is determined from the address codes of land units or buildings or cadastral designations added to the Statistical Register of Agricultural Holdings, if fewer inhabitants lived within a one-kilometer radius around their coordinates than around the original LKAU coordinates. Thus, the number of LKAUs in cities has been reduced, since the main place of production is indicated in the Statistical Register of Agricultural Holdings.

In order to link data on teachers in educational institutions and data on the institutions themselves from the Ministry of Education and Training with the LKAU data, transition tables were created, as educational institutions in the Ministry of Education and Training data lack common identifiers with any other information system. Only institutions with more than one employee were used, thus excluding those with only managers (mostly higher education institutions). There is a many-to-many relationship between educational institutions and their structural units in the Ministry of Education and LKAU data. Transition tables were created semi-automatically, mainly based on addresses and similarities in institution names. The sector (NACE Rev. 2) indicated for educational institutions in the LKAU data was corrected to correspond to the actual type of economic activity and to be stable over time.

PK

Employed persons are identified by their personal identification number (PK). To determine their place of residence, the residence information update events from the monthly data downloads of the FPR are taken as a basis (if several changes have occurred during the month, information is available only for the last ones). Only those records in which the residence information has changed compared to the previous period are used for further processing. In most cases (>98%) these records contain the address object code, the coordinates of which are available. If the reason for updating the residence information is the cancellation of the residence registration (false information was provided or the legal basis was lost), the record is deleted, but the previous record is marked as cancelled, using only the administrative territory (county or city that forms the municipality) instead of the address object code in further processing.

For persons in collective housing (except for night shelters), the place of residence is assumed to be based on the addresses of these institutions and dated to the beginning of the year. Data source – population estimate provided by the CSB, in which data on collective housing is obtained from several data sources.

In case of missing address object codes, records from the 2011 TS are added, if the period and territorial unit (parish or city) match, but without taking into account collective housing.

All three data sources are combined and the periods during which each place of residence was active are determined. For example, if a person was in a social care institution between two residence declaration events, the first declared place of residence was active until the transfer to a social care institution, where they stayed either until the next residence declaration event, if it occurred in the same year, or until the beginning of the following year, when the first declared place of residence was renewed. For deceased persons, the end of the period is taken as the date of death.

Missing address object codes were added from experimental population statistics and based on the similarity of the record with addresses in the State Address Register, including linking property names in NĪVKIS with the relevant parts of the address record, if the property has only one address and the property name is unique in the territorial unit (linking by address). In the case of missing apartment address codes, building address codes were used. If several address object codes correspond to one record, the one that is used for the majority of records in the FPR was used. Address object codes without coordinates were replaced with codes that have coordinates, if the record coincides with the State Address Register and only one pair of coordinates corresponds to it. When reviewing data on places of residence, missing address object codes were found manually for those persons who are in the SRS data, as far as possible.

To determine the distances from the place of residence to all possible places of work, coordinates were added to street and village address codes as average coordinates from buildings on the street or in the village. If it was not possible to add coordinates from addresses to places of residence in Latvia, the centers of population were taken as coordinates . Approximate coordinates were not used to summarize data in geospatial units.

The place of residence within the year is chosen as the one that has the longest overlapping period with the period worked at the workplace. Accordingly, one person may have several places of residence during the year, which may also overlap.

The place of residence for determining the average annual income and the location of the LKAU for natural persons (self-employed) is the longest place of residence during the year. In case it is located abroad or has been cancelled, the next longest place that is not abroad or cancelled is used. If the places of residence in all periods are located abroad or have been cancelled, the longest of the periods is used. In the case of LKAU, only places of residence located in Latvia are used.

The employment period was used from SRS employer reports, notifications on amounts paid to individuals, Appendix D3 of the annual income tax return, and data on patent payers.

The level 3 occupation code (six to seven characters) is used from the SRS data on employees and the full list of employees. Both sources are used because in many cases the occupation code can only be found in the full list of employees from previous years or, in rarer cases, only in the data on employees.

In calculating wages and income equivalent to them, several types of income are used, which, by their economic nature, are comparable to remuneration for work: wages, income from company contracts, patent fees (equal to the minimum wage), remuneration of employees and owners of micro-enterprises, income of seasonal agricultural workers, and income earned abroad.

Total personal income

When calculating a person's total income, the total income indicated in the annual income declaration is used. If the annual income declaration has not been submitted, then the calculation is performed using all the previously indicated SRS data sources. Annual income also includes occasional income, such as gambling winnings or inheritance received. This makes the analysis difficult, therefore the Local OutlierFactorwas usedto identify and exclude outliers.

A person's total income and gross salary are calculated by summing the annual:

• gross salary and wages,

consisting of a basic salary/wages (monthly salary, position salary), remuneration for time worked or the amount of work performed, regular and irregular additional payments and bonuses, income from overtime, payments for annual and additional leave, additional leave payments, holiday benefits, compensation for unused leave, payments for work on public holidays and other days that are not working days, as well as mandatory social insurance contributions paid by employees and personal income tax, supplemented by labour remuneration subsidies;

• gross earnings from self-employment,

which include copyright payments;

• benefits and pensions,

consisting of unemployment, survivors', sickness, education, family/child and housing benefits, as well as old-age, sickness and disability pensions;

• income from capital,

which consists of interest, dividends, and profits from capital investments.

The average annual gross income in the territories of residence was calculated in two ways – by referring to the number of persons receiving income in the specific territory and by referring to the number of inhabitants in the specific territory in the relevant year (average annual gross income calculated per capita). If a person has lived in several territories in the relevant year, the territory of residence was considered to be the one in which they stayed for the longest number of days.

Determining distances between residence and all possible workplaces

To determine the distances between the place of residence and all possible workplaces, from which to then determine the workplace closest to the place of residence, data from the topographic map of the Latvian Geospatial Information Agency at a scale of 1:10,000 was used (data from the 3^rd and 4^th editions combined with manual corrections and additions, since 2022, a separate, updated data set has been used for each year). The calculation was performed with OSRM both by road and on foot. Distances by foot were used if they did not exceed two kilometers. The input data was prepared using the transition table topographic maps and the OpenStreetMap road classification. Information on one-way roads and turn restrictions was not taken into account, as the topographic map does not contain such data.

Data redistribution

Data redistribution occurs in two steps:

1. Data redistribution from UUK to NMK;
2. Data redistribution from NMK to LKAU or PK.

Redistribution from UUK to NMK

Of all UUKs, only a fraction have more than one NMK. The following indicators are calculated for each NMK:

1. The number of persons employed by the NMK is calculated by adding up the number of persons employed by the LKAUs belonging to the NMK.
2. The salary and equivalent income of the NMK are calculated by adding the salary and equivalent income of the PKs belonging to the NMK.

The economic indicators of the NMK, except for the number of employed persons, are calculated according to the following algorithm:

1. If the UUK has only one NMK, then the NMK data is the same as the UUK data.
2. If the wages and equivalent income for all NMKs of one company are known, then the economic indicators are redistributed in proportion to the NMK wages and equivalent income.
3. If the wages and comparable income for all NMKs of one company are not known, then the economic indicators are redistributed proportionally to the NMK's employment period.
4. If the salary and equivalent income of some employed persons are unknown, then the period of employment of the PK is converted into salary and equivalent income, based on the distribution of the PK with salary and equivalent income.

Redistribution from NMK to LKAU or PK

Every year, more than 10 thousand enterprises have more than one LKAU. For each LKAU, the average number of persons employed per year is known (mostly rounded to whole numbers).

If data on persons employed in the NMK are available (PK list), then the economic indicators are redistributed to the PK level. PK data can be aggregated to obtain LKAU data. If data on persons employed in the NMK are not available, then the economic indicators are redistributed to the LKAU level.

PK list NMK

The selection of PKs in the NMK is based on SRS data on employed NMKs. PKs ​​are grouped by the number of days worked: long-term (more than 150 days) and short-term (less than 150 days). The groups are sorted by the product of income and days worked, selecting a certain number according to the NMK's average long-term and short-term employment.

PK recruitment LKAU

The direct connection of PK to LKAU is not known directly from administrative data. PK connection to LKAU is estimated in the following way, based on several assumptions:

• If the NMK has only one LKAU, then all persons employed by the NMK are assigned to the single LKAU.
• If the NMK has more than one LKAU, then the assignment of the PK to the LKAU is carried out based on the following conditions:
  • Using the Ministry of Education and Science data on teachers in educational institutions and the created transition tables between educational institutions in the VIIS data and the LKAU, the PK is linked to the corresponding LKAU, which can also be several. In other cases, within the NMK, each PK is linked to one LKAU.
  • The distribution of the number of PKs by LKAU is controlled by the known number of employed persons in each LKAU.
    • If the number of PKs is equal to the number of persons employed in the NMK by the total number of persons employed in the LKAU, then the distribution of the number of PKs by LKAU coincides with the known number of persons employed in the LKAU.
    • If the number of PK is greater or less than the number of persons employed in the NMK by the total number of employees of the LKAU, then the distribution of the number of PK by LKAU is determined in proportion to the number of employees of the LKAU. Taking into account that the distribution of the number of PK must be expressed in whole numbers, the lower and upper limits of the distribution of the number of PK are determined, which are calculated by rounding down and up (to whole numbers) the proportionally redistributed number of PK.
  • Using the calculated distances from the place of residence to all possible workplaces (LKAU), the connection of the PK to the LKAU is chosen in such a way as to minimize the total distance of all PK from the place of residence to the workplace within the framework of the NMK.
    • For persons with unknown residence, the distance is assumed to be 0 to all possible LKAUs.
    • The distance is adjusted (reduced) according to the correspondence between the PK profession and the LKAU sector. Pairs of compatible professions and sectors are defined, for example, a PK with the profession “teacher” will most likely be employed in a LKAU with the economic activity (sector) “school”.

The PK assignment to the LKAU is determined for each NMK by defining and solving a linear programming problem (Hornik, Meyer, Schwendinger, & Theussl, 2019; R Core Team, 2019; Schumacher, 2018a, 2018b; Theussl, 2017).

After attaching the PK to the LKAU, it is possible to calculate the amount of PK wages and equivalent income attached to each LKAU.

Distribution of economic indicators by LKAU

Economic indicators (except for the number of employees) are divided as follows:

• If a PK has been linked to the LKAU:
  • If all PKs have a known salary and equivalent income in all UUK NMKs, then economic indicators (except for the number of employees) are distributed to each PK in proportion to the PK's salary and equivalent income.
  • If wages and comparable income in the UUK NMK are not known for all PKs, then economic indicators (except for the number of employees) are distributed to each PK in proportion to the PK's employment period.
  • If the salary and equivalent income of some employees are not known in the UUK NMK, then the employment period of the employee is converted into salary and equivalent income, based on the distribution of employees with salary and equivalent income.
  • The LKAU data can then be calculated by summing the data of the PKs attached to the LKAU.
• If the PK has not been linked to the LKAU (for example, there was no data on employees (PK) in the SRS DDZ or the total salary of all PKs is 0):
  • If the number of employees in the NMK is greater than 0, then economic indicators (except for the number of employees) are distributed to each LKAU in proportion to the number of persons employed in the LKAU;
  • If the number of employees in the NMK is 0, then economic indicators (except for the number of employees) are distributed equally to each VDVU (proportionate to the number of VDVUs).