The Türkiye 2024 World Bank Enterprise Survey
Implementation Report
I. Introduction
This document provides information on the World Bank Enterprise Surveys (WBES)
implemented in Türkiye between December 2023 and February 2025. The WBES collect both
objective data based on firms’ day-to-day experiences, and perceptions of enterprises regarding
the business environment in which they operate. The WBES currently cover over 220,000 firms
in 160 countries with a total of 362 surveys since 2006. The WBES are also used to build a panel
of enterprise data that will make it possible to track changes in the business environment over
time.1
This report describes the sampling design of the survey, the data set structure as well as
additional information that may be useful when using the data, such as information on survey non-
response and the appropriate use of the sampling weights.
II. Sampling Structure
The WBES use stratified random sampling, where the population of establishments is first
separated into non-overlapping groups, called strata, and then respondents are selected through
simple random sampling from each stratum. The detailed methodology is provided in the Sampling
Note.2 Stratified random sampling has several advantages over simple random sampling. In
particular, it:
• produces unbiased estimates of the whole population or universe of inference, as well as
at the levels of stratification
• ensures representativeness by including observations in all of those categories
• produces more precise estimates for a given sample size or budget allocation, and
• may reduce implementation costs by splitting the population into convenient subdivisions.
The WBES typically use three levels of stratification: industry classification, establishment
size, and subnational region (used in combination). Starting in 2022, the WBES bases the industry
classification on ISIC Rev. 4 (with earlier surveys using ISIC Rev. 3.1).3 For regional coverage
within a country, the WBES has national coverage.
1
A “panel interview” refers to an interview with a business that was also interviewed in the previous WBES.
2
The Sampling Note is available at:
https://www.enterprisesurveys.org/content/dam/enterprisesurveys/documents/methodology/Sampling_Note-
Consolidated-2-16-22.pdf. For further methodological background see, Richard L. Scheaffer; Mendenhall, W.;
Lyman, R., “Elementary Survey Sampling”, Fifth Edition, 1996.
3
The WBES universe includes: all manufacturing (ISIC 4.0 codes 10-33), services (ISIC 4.0 codes 41-43, 45-47,
49-56, 58, 61, 62, 69-75, 79, 95). Details on sectoral coverage and the WBES universe of inference can be found in
the Enterprise Surveys Manual and Guide (p. 4).
1
II.1 Stratification Categories
The Türkiye 2024 WBES uses the following stratification categories:
• Industry: 10 categories:
• Within manufacturing: Food, Textiles, Garments, Fabricated Metal Products, Other
manufacturing
• Within services: Construction, Retail, Hotels, Transportation and Other Services
• Size: 3 categories: Small (5 to 19 employees), Medium (20 to 99 employees), Large
(100 or more employees)
• Region: 7 categories: Marmara, Aegean, Mediterranean, Central Anatolia, Eastern
Anatolia, South Eastern Anatolia and Black Sea regions.
Industry strata for the manufacturing sector were selected by their contribution to total
value added, to employment and to total number of establishments, using Turkish Statistical
Institute (TUIK) publications. The manufacturing sectors stratified separately represent 16%, 15%
and 13% of total value added, employment, and number of establishments, respectively, with the
rest of sectors grouped into a residual, "Other Manufacturing", stratum.
The strata for the services sector were selected by contribution to total number of
establishments, to employment and to total value added. The sectors stratified separately represent
38%, 40% and 22% of the total number of establishments, employment and total value added of
the whole ES universe, respectively.
Region strata of the Türkiye 2024 WBES were selected based on administrative divisionsto
achieve minimum required precision of estimates at the level of each stratification region.
II.2 Universe
The universe of inference includes all formal (i.e., registered) private sector businesses
(with at least 1% private ownership) and with at least five employees. In terms of sectoral criteria,
all manufacturing businesses (ISIC Rev 4. codes 10-33) are eligible; for services businesses, those
corresponding to the ISIC Rev 4 codes 41-43, 45-47, 49-53, 55-56, 58, 61-62, 69-75, 79, and 95
are included in the Enterprise Surveys. Cooperatives and collectives are excluded from the
Enterprise Surveys. All eligible establishments must be registered with the tax offices. The
universe table is the total number of eligible establishments, and the table is partitioned by the
stratification groups (industry classification, establishment size, and subnational region) in a
country.
For the Türkiye 2024 WBES, the universe table, shown in Table 1 below, was obtained
from TUIK.
II.3 Sampling Frame
The WBES requires the most complete and up-to-date Sampling Frame, the list of eligible
establishments with information on industry classification, size, address and other contact
information that will be used to randomly select the sample. In countries where a previous round
2
of the WBES exists, the Sampling Frame also includes the Panel Sampling Frame, which provides
information about all the establishments that participated in the previous round of the WBES in
the country.
The Sampling Frame for the Türkiye 2024 WBES was constructed from the following
sources (see counts of establishments in the frame in Table 2). The Panel Sampling Frame was
constructed using information on all the establishments that participated in the Türkiye 2019
WBES. The Fresh Sampling Frame, i.e., list of establishments that are in the WBES universe and
that have not participated in the Türkiye 2019 WBES, was obtained from TUIK.
For every WBES, necessary measures are taken to ensure the quality of the frame; however,
the sample frames are not immune to the typical problems found in establishment surveys: positive
rates of non-eligibility, repetition, non-existent units, etc. Given the impact that non-eligible units
included in the sample universe may have on the results, eligibility adjustments may be needed
when computing the appropriate sampling weights for individual observations (if and only if the
weights are computed the sampling frame that is also the universe). Table 4 reports response
outcomes.
II.4 Sample Design
The WBES sample design, i.e., target number of interviews in each combination of
stratification categories (cells), is generated using the two primary criteria: 1) minimize the
difference from the purely proportional sample within each cell; and 2) achieve a sufficient sample
size by stratification category to allow for estimates of a given level of precision. 4 Additional
information on the criteria for determining the sample size by stratification category is given in
the Sampling Note, and additional information on the sample design is given in the Enterprise
Surveys Manual and Guide.5 The original survey design for the Türkiye 2024 WBES is given in
Table 3.
III. Data Collection
The detailed information on the WBES methodology and data collection is provided in the
Enterprise Surveys Manual and Guide. The interviews for the Türkiye 2024 WBES were
conducted between December 2023 and February 2025. The interviews were conducted in
Turkish. For monetary variables, the currency was Turkish Lira.
III.1 Questionnaire
The standard WBES questionnaire covers several topics regarding the business
environment and business performance. These topics include general firm characteristics,
infrastructure, sales and supplies, trade, management practices, competition, innovation, capacity,
4
Additional constraints are also considered in the design stage. These are generally practical and include, for
example, having a sufficient number of available contacts in the sample.
5
The Enterprise Surveys Manual and Guide is available at:
https://www.enterprisesurveys.org/content/dam/enterprisesurveys/documents/methodology/Enterprise-Surveys-
Manual-and-Guide.pdf
3
land and permits, finance, business-government relations, exposure to bribery, labor, and
performance. Information about the general structure of the questionnaire is available in the
Enterprise Surveys Manual and Guide.
III.2 Contractor
The fieldwork for the Türkiye 2024 WBES was implemented by Yontem Arastirma and
Ipsos Turkiye. The selection for the implementing agency followed the standard World Bank
procurement practices that are described in more detail in the Enterprise Surveys Manual and
Guide.
III.3 Sampling and screening
Samples are drawn by the Enterprise Surveys team in batches, following the stratification
and sample design. The contractor conducted a thorough screening process before scheduling the
ES interviews. Results of the screening are provided using the eligibility and status codes as listed
in Table 4. In cases of unit non-response (either a refusal or an inability to obtain an interview after
exhaustive attempts), the contractor proceeded with the contact that appeared next in the list drawn
in the respective cell. The process of sampling and screening is described in more detail in the
Enterprise Surveys Manual and Guide.
III.4 Survey response
In all surveys, including the WBES, some respondents choose not to participate. The
Enterprise Analysis team and the contractor take all necessary measures to boost participation,
through various methods of recruitment. The proper management of the screening process and
sample replacement ensures that the resulting sample remains random.
The main measure of survey participation is the yield, which is the ratio of the total number
of achieved interviews to the total number of contacted establishments. There are two main
elements that boost yields. First is the survey participation rate, measured as the share of
establishments that participated among those that can be assumed to have been eligible. The second
element is the quality of frame. If only a small fraction of the contacted establishments is actually
eligible to participate in the survey, then the sampling frame is far from ideal. This quality is
measured by the ratio of the total number of establishments that can be assumed to be eligible for
the WBES with the total number of contacted establishments in the frame. In other words:
𝑌𝑖𝑒𝑙𝑑 = 𝑆𝑢𝑟𝑣𝑒𝑦 𝑟𝑒𝑠𝑝𝑜𝑛𝑠𝑒 𝑟𝑎𝑡𝑒 ∗ 𝑅𝑎𝑡𝑒 𝑜𝑓 𝑞𝑢𝑎𝑙𝑖𝑡𝑦 𝑜𝑓 𝑡ℎ𝑒 𝑓𝑟𝑎𝑚𝑒
which can rewritten as follows:
𝑇𝑜𝑡𝑎𝑙 𝑖𝑛𝑡𝑒𝑟𝑣𝑖𝑒𝑤𝑒𝑑 𝑇𝑜𝑡𝑎𝑙 𝑖𝑛𝑡𝑒𝑟𝑣𝑖𝑒𝑤𝑒𝑑 𝐴𝑠𝑠𝑢𝑚𝑒𝑑 𝑒𝑙𝑖𝑔𝑖𝑏𝑙𝑒
= ∗
𝑇𝑜𝑡𝑎𝑙 𝑐𝑜𝑛𝑡𝑎𝑐𝑡𝑒𝑑 𝐴𝑠𝑠𝑢𝑚𝑒𝑑 𝑒𝑙𝑖𝑔𝑖𝑏𝑙𝑒 𝑇𝑜𝑡𝑎𝑙 𝑐𝑜𝑛𝑡𝑎𝑐𝑡𝑒𝑑
Table 5 provides these measures for the Türkiye 2024 WBES and across its stratification
levels.
4
III.5 Achieved Sample
Tables 6 and 7 provide counts of the WBES interviews collected for each stratification cell,
i.e., broken down by industry, establishment size, and region. Table 6 reports full sample, while
Table 7 shows counts of only panel interviews.
III.6 Sampling Weights
Since the WBES uses stratified random sampling, individual observations should be
properly weighted when making inferences about the population, since unweighted estimates are
biased unless sample sizes are proportional to the size of each stratum. For each WBES, special
care is given to correctly compute sampling weights. Whenever the Universe is used to draw the
sample (i.e., Sampling Frame is the same as the Universe) it is imperative to accurately adjust the
universe within each stratum to account for the presence of ineligible establishments (e.g., the firm
discontinued businesses, or is deemed ineligible due to its business activity or having fewer than
five employees). Proper treatment of panel establishments is also crucial. Details about how the
WBES sampling weights are calculated are given in the Sampling Note.
Three versions of sampling weights are calculated, depending on the assumptions that
determine eligibility of establishments to be counted towards the WBES universe. These
assumptions are called weak, medium, and strong; and are defined as given in the table below.
Prevalence of each of these outcomes for the Türkiye 2024 WBES is given in Table 4. All
indicators and analysis conducted by the Enterprise Surveys team use the sampling weights based
on the median assumption. Tables 8-10 report estimated universe based on the respective
assumption.
Assumption Eligibility codes for inclusion in the WBES universe
Strict 1,2,3,4
Median 1,2,3,4,10,11,13, 133
Weak 1,2,3,4,10,11,13,133,91,92,93,94,12
For descriptions of each eligibility code, see Table 4.
III.7 Item response rates
Item response rate must be differentiated from survey response rates. The latter refers to
participation in the survey itself (see Section III.4) whereas the former refers to the absence of
responses to specific survey questions.6 The WBES, as any survey, suffer from item non-response;
and different strategies are used by the data collection team to address this. In particular:
• For sensitive questions, such as on corruption or tax evasion, enumerators were instructed
to collect the refusal to respond (-8) as a separate response category from don’t know (-9).
• Establishments with incomplete information were re-contacted to fill gaps.
6
The WBES questionnaire is organized so that there is always some entry in the database when the question was
posed. An empty entry means that the question was not asked to the corresponding respondent, typically, due to skip
patterns, or lack of applicability of that question in general.
5
Table 11 provides item response rates for several key variables, broken down across the
stratification levels.
III.8 Database Structure
The WBES datafiles are organized in ways that reflect the corresponding questionnaire.
The variables that are standard across countries have the first letter in their name correspond to the
questionnaire section where the variable belongs in the questionnaire, i.e., a1 denotes section A.
All variables are numeric with the exception of those variables with an “x” at the end of their
names. The suffix “x” denotes that the variable is alpha-numeric.
The WBES datafiles contain two establishment identifiers, idstd and id. The former is a
global unique identifier of each establishment, while the latter is unique identifier within each
survey. The variable idstd can be used to match the WBES establishment one-to-one across
databases. The variables wweak, wmedian, and wstrong corresponding to sampling weights based
on, respectively, weak, median, and strong assumptions about eligibility (see Section III.6). The
variable strata corresponds to the stratum of each observation.7
Additionally, the WBES datafiles contain many standard variables. The variable d1a2_v4
denotes the main activity of the establishment, as obtained during the WBES interview, coded in
the four-digit ISIC Rev. 4. Users should note that this activity may differ from the industry
classification given in the Sampling Frame, a4a. Users are generally advised to use industry
categories based on the realized information in d1a2_v4. Additional sampling information is
contained in variables a2 (region) and a6a (size). The variable panel identifies panel
establishments, i.e., those that participated in the Türkiye 2019 WBES. The combination of a4a,
a2, a6a, and panel forms stratum of each establishment, which is contained in variable strata.
The last complete fiscal year for each establishment is contained in variables a20m (last
month of last complete fiscal year) and a20y (last complete fiscal year).
Note that when an entry in the WBES database is empty, this means that the question was
not asked to the corresponding respondent. This happens when the question is deemed
inapplicable, due to skip patterns or other reasons. In cases when the question was posed, some
entry is provided, including don't know, which is an example of item non-response.
7
Note that the sampling weights may vary within strata for panel establishments due to the procedure used for
projecting the universe from the previous round of the survey.
6
IV. Useful Links
The users of the WBES data may find the following links useful:
• Sampling Note is available at:
https://www.enterprisesurveys.org/content/dam/enterprisesurveys/documents/methodolog
y/Sampling_Note-Consolidated-2-16-22.pdf
• The Enterprise Surveys Manual and Guide is available at:
https://www.enterprisesurveys.org/content/dam/enterprisesurveys/documents/methodolog
y/Enterprise-Surveys-Manual-and-Guide.pdf
• The WBES global questionnaires are available at:
https://www.enterprisesurveys.org/en/methodology
• The projects that are currently being implemented by the Enterprise Surveys team are
available at: https://www.enterprisesurveys.org/en/methodology/current-projects
• The list of all WBES databases and detailed information about each is available here:
https://www.enterprisesurveys.org/content/dam/enterprisesurveys/documents/methodolog
y/DataDetails.xls
• The description of the WBES indicators is available at:
https://www.enterprisesurveys.org/content/dam/enterprisesurveys/documents/Indicator-
Description.pdf
7
Fact Sheet
Source of Universe Table TUIK, 2021 Business Registers
Source of Sampling TUIK, 2021 Business Registers
Frame
Levels of Universe Table Universe Table: Firm-level
and Sampling Frame Sampling Frame: Firm-level
Registration agency Tax Office (Vergi Dairesi)
Stratification sectors Manufacturing of: Food, Textiles, Garments, Fabricated Metal
Products, Other manufacturing; Retail, Construction, Hotels,
Transportation and Other Services
Stratification sizes Small (5 to 19 employees), Medium (20 to 99 employees), Large
(100 or more)
Stratification regions Marmara, Aegean, Mediterranean, Central Anatolia, Eastern
Anatolia, South Eastern Anatolia and Black Sea regions.
Contractor Ipsos Türkiye and Yöntem Araştırma
Fieldwork dates December 2023 – February 2025
Interview languages Turkish
Survey software Survey Solutions
Currency for nominal Turkish Lira
variables
Reference fiscal year 2023 (2 obs.), 2024 (1,083 obs.), and 2025 (331 obs.)
VAT Applicable YES
Additional definitions “Consumption Tax” is referred to as KDV. “Occupancy Permit” is
referred to as “İskan Belgesi”. Ankara is the capital city (a3b);
İstanbul is reffered as the main business city (a3c)ç
Sample Size Total: 1,416 Fresh: 1,124 Panel: 292
Survey response rates Yield: 16% Response rate: 22.7% Frame quality: 70.6%
Item response rates d2: 94.4%; n2a: 95.9%; l1: 99.5%; all TFP vars.: 60.3%
Additional topics covered N/A
in the questionnaire
Additional surveys N/A
available (if any)
Tables
Due to large size of the tables for the Türkiye 2024 WBES, they are provided in the excel file
“ES Implementation Report Tables Turkiye 2024.xlsx”, which is available along with this
Implementation Report in the zipfile that contains all the documentation for this survey.
