Production Of Official Statistics Research Articles

Machine learning and data augmentation in the proxy means test for poverty targeting

Recent years have seen increased interest in the use of alternative data sources in the definition and production of official statistics and indicators for the UN Sustainable Development Goals. In this paper, we consider the application of data science to the production of official statistics, illustrating our perspective through the use of poverty targeting as an application. We show that machine learning can play a central role in the generation of official statistics, combining a variety of types of data (survey, administrative and alternative). We focus on the problem of poverty targeting using the Proxy Means Test in Indonesia, comparing a number of existing statistical and machine learning methods, then introducing new approaches in the spirit of small area estimation that utilize area-level features and data augmentation at the subdistrict level to develop more refined models at the district level, evaluating the methods on three districts in Indonesia on the problem of estimating 2020 per capita household expenditure using data from 2016–2019. The best performing method, XGBoost, is able to reduce inclusion/exclusion errors on the problem of identifying the poorest 40% of the population in comparison to the commonly used Ridge Regression method by between 4.5% and 13.9% in the districts studied.

Answering Current Challenges of and Changes in Producing Official Time Use Statistics Using the Data Collection Platform MOTUS

Abstract The modernization of the production of official statistics faces challenges related to technological developments, budget cuts, and growing privacy concerns. At the same time, there is a need for shareable and scalable platforms to support comparable data, leading to several online data collection strategies being rolled out. Time Use Surveys (TUS) are particularly affected by these challenges and needs as they (while producing rich data) are complex, time-intensive studies (because they include multiple tasks and are administered at the household level). This article introduces the Modular Online Time Use Survey (MOTUS) data collection platform and explains how it accommodates the challenges of and changes in the production of a TUS that is carried out in line with the Harmonized European Time Use Survey guidelines. It argues that MOTUS supports a shift in the methodological paradigm of conducting TUS by being timelier and more cost efficient, by lowering respondent burden, and by improving the reliability of the data collected. Importantly, the modular structure allows MOTUS to be easily deployed for various TUS configurations. Moreover, this versatile structure allows comparable, complex diary surveys (such as the household budget survey) to be performed on the same platform and with the same applications.

Building ethical guidelines to produce official statistics: the statistical ethics system (SETE) for the national administrative department of statistics (DANE) in Colombia

ABSTRACT This article describes and analyzes the design and functioning of the Statistical Ethics System (SETE) in Colombia’s National Statistical Office. It presents the methodology and general process of planning and implementation of the System, supported by a conceptual analysis of the requirements for an ethical functioning of official statistics. The general objective of the article is to make a practical contribution to the understanding of conceptual and practical features that ought to be considered in the implementation of an ethical system in the operation of the public sector, particularly in the area of the production of official statistics. The text is divided into five parts. In the first part, we make a general introduction to the ethical value of official statistics. In the second and third parts, we make a presentation and analysis of the principles and risks that were selected as the basic ethical framework for the production of official statistics at the National Administrative Department of Statistics (DANE). The fourth section presents the conceptual and practical bases of how SETE operates. The fifth section we present some concluding remarks.

Compilations of official statistics in Kenya: Assessing their adequacy for urban policy and governance

Statistics are useful indicators that may inform urban development and management decisions. This article reviews production of official statistics in Kenya to bring up their implications for urban policy and governance. It approaches the issue from the point of view that official statistics are always inadequate in developing countries. In this context it considers the compilations of urban statistics mainly in the annual statistical digest and finds that there were useful urban statistics compilations that could be improved for governance prior to the implementation of the 2010 Constitution. However, after the implementation of the 2010 constitution, urban statistics have been abandoned for devolved governance units in terms of counties. Therefore, it recommends the reinstatement and improvement of urban statistics in the annual digest.

Implementation of Administrative Data in the Production of Official Statistical Information in the Field of Public-Legal Relations

The article examines the issue of statistics on the resolution of public legal disputes and the possibility of its implementation in the production of official statistics. The problem of the formation of statistical reporting by the judicial bodies of Ukraine in the field of public-legalrelations as administrative data is considered. Methodological regulations on the organization of state statistical surveillance on the consideration of cases of administrative offenses and persons brought to administrative responsibility in Ukraine were analyzed. It was established that the existing tools for committing administrative offenses do not reflect the current state and dynamics. Conclusions were drawn on the relevance of developing and approving such methodological provisions to replace those that have lost their validity.
 Particular attention is focused on the study of the experience of the Federal Republic of Germany in the field of administrative proceedings, in particular, the methodology of statisti­cal observation of administrative proceedings, introduced by the Federal Statistical Service of the country, is considered. The regulatory framework, on the basis of which the statistical observation of the resolution of public legal disputes by administrative courts functions in the Federal Republic of Germany, is considered. A conclusion was formed about the expediency of borrowing similar experience and introducing similar statistical observation in Ukraine.
 The research substantiates the need to develop methodological provisions for the new state statistical observation “Execution of administrative proceedings”, which will contribute to ensuring access and transparency to indicators of the procedural activity of administra­tive courts, reflecting the problems of law enforcement practice and the effectiveness of the judicial form of protection against violations committed by state authorities.

Timeliness reduction on industrial turnover index based on machine learning algorithms

The modernisation of the production of official statistics should make use not only of new data sources but also of novel statistical methods applied to traditional survey and administrative data. This improves the traditional quality standards. Here we present an application of statistical learning algorithms to improve the timeliness under a controlled compromise of accuracy of the Spanish Industrial Turnover Index (ITI). The methodology has been developed based on a modular and standardized approach that could be easily extended to other surveys. Our advanced index allows us to predict the ITI 31 days before publication with a median error of 0.5 points over the period Mar 2016–Apr 21, in an index with large oscillations. The results are promising and support the idea of the use of these techniques in improving the quality dimension of timeliness while accuracy is kept under control.

Ethical considerations in the use of Machine Learning for research and statistics.

This paper, based upon new guidance created in collaboration with researchers from several national statistical institutes, explores the main ethical considerations associated with the use of machine learning techniques for aggregate statistics. The aim of this paper is to provide applied, practical ethical guidance for researchers using machine learning techniques.
 Following an extensive literature review, alongside discussion and collaboration with a number of national statistical institutes, it was identified that there was a need for applied guidance on the use of machine learning for the production of official statistics by the international research and statistical community. Feedback was gathered from interested stakeholders, which found that whilst there were resources available to researchers relating to the ethical considerations of machine learning projects, these focus mainly on operational uses of machine learning, and furthermore, lacked advice on how to practically mitigate ethical issues that arise throughout the project lifecycle.
 The guidance focuses on four main ethical considerations, found to be prevalent within machine learning research, and offers ways to mitigate these issues should they arise. These are: the importance of minimising and mitigating social bias and discrimination within machine learning research, and clearly communicating these, and the limitations of our research; the need to consider the transparency and explainability of machine learning research, and the implications this has for reproducibility; the importance of maintaining accountability throughout machine learning processes, ensuring that models are used only for their intended purposes, and that different stakeholders are aware of their responsibilities; the need to consider the confidentiality and privacy risks arising from the data used, both in relation to training data which is fed into the machine, and outputs resulting from the machine learning’s findings.
 The guidance has been well-received following its release, and feedback from the wider user community to date has been positive. The IPDLN conference provides an opportunity for further feedback to be collated to ensure that the guidance continues to be valuable to its intended audience.

Масштаби та наслідки вимушеної міграції населення України внаслідок збройної агресії Російської Федерації

The article is devoted to the assessment of the scale and consequences of forced external and internal migration due to the active phase of the Russia-Ukraine war. The urgency of the work is due to the need for scientific understanding of migration trends formed after February 24, 2022 and the lack of analytical research on migration in this period. The purpose of the article is to assess the scale of forced migration of the population of Ukraine both within the country and abroad. The novelty of the study is to identify the consequences of large-scale migratory movements, which began on February 24, 2022, and develop recommendations for their regulation. The research methodology is based on the use of rather eclectic information from open sources, expert opinions and assessments, etc., in conditions when the production of official statistics is stopped, conducting sample surveys, at least “face to face”, is extremely difficult. The trends of forced migration within Ukraine and abroad have been assessed. It was established that the active departure of the population from Ukraine was observed in the first weeks of the open war phase, after that the flow of refugees slowed down. According to Ptoukha Institute for Demography and Social Studies of the NAS of Ukraine, based on data from the State Border Guard Service, the number of war refugees from Ukraine is 3.1-3.2 million people, but there is a risk of significant increase. Possible amounts of irreversible migration losses, depending on the impact of military and economic factors, will range from 600-700 thousand people to 5.0-5.5 million. The number of those who were forced to change their place of residence within Ukraine is estimated at 5.0-5.5 million. The factors of return of forced migrants to Ukraine are identified, the consequences of large-scale forced migrations are assessed, and recommendations for their regulation are proposed. It is concluded that a more accurate estimate of the number of both IDPs and forced external migrants requires the development of comprehensive methodology that will take into account data from various sources and adequate information provision (in particular, the introduction of sample surveys in Ukraine and abroad on a regular basis). Validation of performed assessments data is possible after the census.

The Use of Registries Data in Statistical Studies of Employment

According to the Ministry of Digital Transformation there are more than 350 state registries in Ukraine of which about 200 contain personal information about the population. The creation of registries was based on a narrow departmental approach without due attention to the information value of this resource. The State Statistics Development Programme until 2023 identifies expanding the access of state statistics bodies to administrative data and registries as one of the key areas for reducing the reporting load on respondents. The draft Law of Ukraine “On Official Statistics” (registered by the Verkhovna Rada of Ukraine on September 2, 2021 № 5886) provides an institutional basis for expanding such access for producers of official statistics including to individual registries data.
 The article aims to reveal the possibilities of using registries data in statistical studies of employment, in particular the data from the registry of insured persons which is administered by the Pension Fund of Ukraine. The scale of information about insured persons is constantly growing but their use for analytical and statistical purposes is still limited. The scale of information about insured persons is constantly growing but their use for analytical and statistical purposes is still limited.
 Since the beginning of 2021 the quarterly distribution of employees by the number of gross wages has been formed by the registry of insured persons which has significantly reduced the load on respondents of the state statistical observation “Survey of Enterprises on Labor Statistics”. The transition to an alternative data source did not break the general proportions and trends in the distribution of employees by wages groups but increased the dependence of obtaining statistical information on regulatory changes in administrative procedures. Thus, the changes in the form and frequency of companies' fiscal reporting led to an increase in the waiting period for the publication of the distribution of employees by wage groups by two months. The intensification of the use of registries data in the production of official statistics requires the expansion of the powers of the State Statistics Service of Ukraine for practical participation in the policy for the development of public registries.
 The registries data of the Pension Fund of Ukraine can significantly expand the list of short-term and additional statistical characteristics of employment and wages especially in terms of structural distributions for the analysis of differentiation and inequality. For in-depth research it is proposed to develop standardized procedures aimed at forming a representative sample of the registries data which will be available in the form of depersonalized microdata to a wide range of researchers.

Innovation during the COVID-19 crisis: Why it was more critical for official statistics than ever1

During the COVID-19 pandemic, quantitative evidence was desperately needed in order to understand and manage an unprecedented situation, and to make important national, European and international decisions. Eurostat and the national statistical institutes (NSIs) played a key role in managing the pandemic, providing society with a high quality statistical information service. In particular, the crisis accelerated innovation in statistical production, steered complex processes of change towards the use of new data sources and privately held data for official statistics, enhanced the adoption of new statistical methods, and consequently the production of experimental statistics and dashboards. While the new data ecosystem provided opportunities for the production of official statistics, the importance of privacy preservation, data security and the development of adequate data quality frameworks remained a priority. Important strands of work for the future would be: satisfying the increased needs of users, as well as supporting decision-making and the delivery of government services in emergencies. NSIs would also do well to invest in innovation, collaborate and establish partnerships with the data providers and research communities that have worked closely with them since the beginning of the pandemic. This article is based on the experiences of six NSIs in the Netherlands, Spain, France, Italy, Germany and Finland during the pandemic.

Automatic mapping of Italian forest disturbances between 1985 and 2019 using Landsat imagery and Google Earth Engine

Forests play a key role in the carbon cycle and the fight against climate change. Long-term monitoring of forest dynamics represents a key element for understanding forests transformations due to forest harvestings and disturbances including fires, wind storms, frost or drought events, and pathogen attacks. This work aims at mapping and evaluating the forest disturbances that have occurred in Italy since 1985, using Landsat satellite imagery and apposite algorithms. We predicted about 1.8 million forest disturbances occurring during the observation period. Disturbances ranged between 27.923 ha in 2014 and 261.733 ha in 1985. Most of the forest disturbances have been identified in Sicilia and Calabria. Commission errors fluctuated between 29% in 2012 and 65% in 2001 while omission errors were between 8% in 2014 and 88% in 2003. The results that we present in this work can increase our understanding of Italian forests, and serve as basis for future research, while the methodology we applied can support the production of official statistics on forest disturbances.

A quality framework for statistical algorithms

As national statistical offices (NSOs) modernize, interest in integrating machine learning (ML) into official statisticians’ toolbox is growing. Two challenges to such an integration are the potential loss of transparency from using “black-boxes” and the need to develop a quality framework. In 2019, the High-Level Group for the Modernisation of Official Statistics (HLG-MOS) launched a project on machine learning with one of the objectives being to address these two challenges. One of the outputs of the HLG-MOS project is a Quality Framework for Statistical Algorithms (QF4SA). While many quality frameworks exist, they have been conceived with traditional methods in mind, and they tend to target statistical outputs. Currently, machine learning methods are being looked at for use in processes producing intermediate outputs, which lead to a final statistical output. Therefore, the QF4SA does not replace existing quality frameworks; it complements them. As the QF4SA targets intermediate outputs and not necessarily the final statistical output, it should be used in conjunction with existing quality frameworks to ensure that high-quality outputs are produced. This paper presents the QF4SA, as well as some recommendations for NSOs considering the use of machine learning in the production of official statistics.

Robust official business statistics methodology during COVID-19-related and other economic downturns

Official statistics has not properly researched and understood how its methods and models behave at times of downturns (and potentially in the corresponding situation of similarly paced (unpredictable and fast) growths). There is generally a wish to make methods robust to unusual changes, but these are often tackled situation by situation. Production of official statistics during COVID-19 has necessitated some radical changes in both data collection and statistical methods; these have been introduced with admirable speed and dedication, but this process would have been made easier with a body of research already in place to draw from. We discuss the issues with the robustness of statistical methods during downturns, and highlight the opportunity to gather data which can be analysed to give evidence for the most robust methods to use as protection against poor measurement during future downturns.

Measuring the Accuracy of Aggregates Computed from a Statistical Register

AbstractThe Italian National Statistical Institute (Istat) is currently engaged in a modernization programme that foresees a significant revision of the methods traditionally used for the production of official statistics. The main concept behind this transformation is the use of the Integrated System Statistical Registers, created by a massive integration of administrative archives and survey data. In this article, we focus on how to measure the accuracy of register estimates of a population total from measurements calculated at the unit level. We propose the global mean squared error (GMSE) as a statistical quantity suitable for measuring accuracy in the context of the production of official statistics. It can be defined to explicitly consider the main sources of uncertainty that may affect registers. The article suggests a feasible calculation strategy for the GMSE that allows National Statistical Institutes to build algorithms that can promptly be applied for each user request, thus improving the relevance, transparency and confidence of official statistics. Through a simulation study, we verified the efficacy of the proposed strategy.

An end-to-end statistical process with mobile network data for official statistics

Mobile network data has been proven to provide a rich source of information in multiple statistical domains such as demography, tourism, urban planning, etc. However, the incorporation of this data source to the routinely production of official statistics is taking many efforts since a diversity of highly entangled issues (access, methodology, IT tools, quality, skills) must be solved beforehand. To do this, one-off studies with concrete data sets are not enough and a standard statistical production process must be put in place. We propose a concrete modular process structured into evolvable modules detaching the strongly technological layer underlying this data source from the necessary statistical analysis producing outputs of interest. This architecture follows the principles of the so-called ESS Reference Methodological Framework for Mobile Network Data. Each of these modules deals with a different aspect of this data source. We apply hidden Markov models for the geolocation of mobile devices, use a Bayesian approach on this model to disambiguate devices belonging to the same individual, compute aggregate numbers of individuals detected by a telecommunication network using probability theory, and model hierarchically the integration of auxiliary information from the telco market and official data to produce final estimates of the number of individuals across different territorial regions in the target population. A first simple illustrative proposal has been applied to synthetic data providing preliminary software tools and accuracy indicators monitoring the performance of the process. Currently, this exercise has been applied to the estimation of present population and origin-destination matrices. We present an illustrative example of the execution of these production modules comparing results with the simulated ground truth, thus assessing the performance of each production module.

Mapping the United Nations Fundamental Principles of Official Statistics against new and big data sources

Never has there been a more exciting time to be an official statistician. The data revolution is responding to the demands of the CoVID-19 pandemic and a complex sustainable development agenda to improve how data is produced and used, to close data gaps to prevent discrimination, to build capacity and data literacy, to modernize data collection systems and to liberate data to promote transparency and accountability. But can all data be liberated in the production and communication of official statistics? This paper explores the UN Fundamental Principles of Official Statistics in the context of eight new and big data sources. The paper concludes each data source can be used for the production of official statistics in adherence with the Fundamental Principles and argues these data sources should be used if National Statistical Systems are to adhere to the first Fundamental Principle of compiling and making available official statistics that honor citizen’s entitlement to public information.

The manipulation of official statistics as corruption and ways of understanding it

We argue that the manipulation of official statistics is corruption and indeed it is grand corruption and political corruption. We note the steps that have been taken in recent decades, which inter alia help make corruption in official statistics more difficult, but argue that instances of corruption persist. To decisively address the problem, it is important to have an analytical tool for understanding it. In this context, we provide a typology of the various broad manifestations or phenomena of corruption in official statistics. We then proceed to identify the process or mechanism that gives rise to these manifestations/phenomena, the elements/component parts of this process/mechanism and the ways these parts interact. This constitutes the schematic model that we propose for understanding corruption in official statistics. In this context, we provide a discussion of the nature of the individual elements/component parts of the process of corruption of official statistics, i.e., we discuss the drivers of the phenomena, the enabling conditions of the phenomena, the modalities and methods used to arrive at the phenomena, and the vectors or agents that execute/propagate the phenomena. We believe all cases of corruption in official statistics can be analyzed using this schematic model. The benefit of having such a model is that it enables one to identify what institutional or legal setting, action, institution or person presents a problem, vulnerability or source of risk in a given system of production of a specific official statistical product in a given country, and address it. The model can also inform a discussion about what needs to be changed at the level of international/supranational arrangements, whether concerning institutional settings, processes, or legal and ethical frameworks, affecting the production of official statistics.

Towards a methodological framework for estimating present population density from mobile network operator data

The concept of present population is gaining increasing attention in official statistics. One possible approach to measure present population exploits data collected by Mobile Network Operators (MNO), from simple Call Detail Records (CDR) to more informative and complex signalling records. Such data, collected primarily for network operation processes, can be repurposed to infer patterns of human mobility. Two decades of research literature have produced several case studies, mostly focused on to CDR data, and a variety of ad-hoc methodologies tailored to specific datasets. Moving beyond the stage of explorative research, the regular production of official statistics across different MNO requires a more systematic approach to methodological development. Towards this aim, Eurostat and other members of the European Statistical System are working towards the definition of a general Reference Methodological Framework for processing MNO data for official statistics. In this contribution we report on the methodological aspects related to the estimation of present population density, for which we present a general and modular methodological structure that generalises previous proposals found in the academic literature. Along the way, we define a number of specific research problems requiring further attention by the research community. We stress the importance of comparing different methodological options at various points in the data workflow, e.g. in the geolocation of individual observations and in the inference method. Finally, we present illustrative results from a case-study based on real signalling data from a European operational network, complemented by numerical results from a simple simulation scenario.

The System for Statistical Measurement of Digital Economy Parameters: Development Trends

The main challenges faced by statistical activities under the pressure of the global dissemination of digital technologies are outlined. The need for constructing a comprehensive system for measurement of digital economy parameters is demonstrated, and the requirements with which it has to comply are set. The principal trends of the process involved in the development of the system for measurement of digital economy parameters are determined. 
 The need for elaboration, implementation and formalization of the metadata laying the ground for constructing the system for statistical measurement of digital transformations, and for making the relevant regulatory base is substantiated; the difficulties involved in this process are outlined. A comparative description of the views held by proponents and opponents of change in the traditional conceptual framework for measurement of economic parameters in the new conditions is given. Problems are highlighted that are associated with constructing a new model of statistics for the digital economy, supposed to integrate all the existing methodical and methodological elaborations in the field of information society’s statistics, on the one hand, and offer new approaches to statistical measurement of various aspects of digitalization, on the other. Main problems related with creating an appropriate information base for constructing the indicators that would report on digital economy components are determined, with outlining ways for their solution. The essential types of efforts on modernization of technological facilities in the official statistics system are shown. Change in the pattern of the demand for statistical information is described, and ways to improve feedback of statistical information users are proposed. It is justified that the staff of statistical offices needs to build the competencies required for design of digital technologies and their use in the professional activities; recommendations on the content of training and professional development programs are given. It is demonstrated that practices of digital economy measurement and dissemination of its results need to be globally standardized; the role of international organizations in optimization and standardization of various processes involved in the production of official statistics at national and international level is highlighted. Potential areas of future efforts in constructing the system for measurement of digital economy are outlined.

Coordination and quality assurance through a programme for official statistics: The Norwegian case

During the last few years, coordination and cooperation in the production of official statistics have been emphasised in the international statistical community. Updated versions of quality frameworks strengthen coordination as a key for improving the quality of official statistics. Quality requirements for such statistics have been clarified. This is also reflected in amended or new statistical legislation. In Norway a new statistics act was adopted in 2019, prescribing a multiannual programme for official statistics. The first draft programme has been prepared by Statistics Norway in cooperation with other producers of such statistics. It was adopted by the Government at the end of 2020. The programme defines requirements for and the content of official statistics, and which authorities will be responsible for the production of these statistics. This paper considers how a programme for official statistics can promote coordination, cooperation and quality improvements of official statistics. Preparing the programme for Norway has revealed some challenges which are discussed. These comprise issues linked to professional independence and impartiality of producers of official statistics and the relationship between official statistics and management information. Reflections on these challenges and how they are resolved are important to establish a good understanding of the concept “official statistics”.