Statistical Journal of the IAOS

This article serves as an introduction to the series of articles included in the SJIAOS special issue about the comprehensive updates made to the System of National Accounts 2025 (2025 SNA) and the Integrated Balance of Payments and International Investment Position Manual, Seventh Edition (BPM7). The article provides an overview of the increasing complexity of economic activity, incorporating new forms of value creation such as digital platforms, cryptocurrencies, and intangible assets. It notes that updated standards enhance consistency and integration between national accounts and external sector statistics, facilitating more accurate measurement and analysis for policymakers. The article outlines the key themes included in the special issue such as the explicit recognition of data as a produced asset, expanded indicators for digitalization, and improved classification of financial instruments and institutional sectors. The article underscores the collaborative nature of these updates, achieved through international cooperation and extensive consultation, ensuring global relevance and acceptance. It concludes by advocating for ongoing revision and harmonization of statistical manuals to maintain relevance and support effective policy decisions.

An important feature of the updated System of National Accounts (SNA) and the updated Balance of Payments and International Investment Position Manual (BPM) frameworks relates to the measurement of sustainability, supporting long-term policy decisions. The new statistical guidance, together with the System of Environmental-Economic Accounting (SEEA), enhances the ability to track the accumulation and use of a more broadly defined set of capital assets. By broadening the scope of macroeconomic statistics to include sustainability dimensions, the updated frameworks provide policymakers with better tools to manage risks, evaluate policy trade-offs, and align short-term decisions with long-term economic objectives.

The System of National Accounts (SNA) is the international standard for countries to compile macroeconomic statistics, covering a full set of interconnected tables that provides users with insights into a country's economy. A lot of the indicators from the SNA are heavily used by policy analysts to obtain insights in various aspects of the economy and its most well-known indicator, GDP, is often used to represent societal progress or the economic well-being of the population. Over the past decade, there has been increasing criticism that policy users are placing too much emphasis on GDP. As it is primarily constructed as a measure of production, it is indeed correct that GDP is not a good indicator to assess overall well-being. In that regard, there are various other indicators in the SNA that are doing a better job in providing insights in household's material well-being, such as household income and wealth, although they also fall short of capturing various elements that would be of relevance in assessing people's well-being. For these reasons, the international statistical community has explored what information and indicators could be added to the framework to provide better insights in aspects related to well-being. Although well-being is a multidimensional phenomenon that covers several objective and subjective elements that go beyond the scope of the SNA, it has been acknowledged that the framework may still provide important insights to users. This article provides an overview of the new elements included in the 2025 SNA to shed more light on various aspects affecting well-being, providing an overview of the conceptual framework and presenting the new elements, including labour, health care, education, unpaid household service work, and household distributional results.

The 2025 System of National Accounts (SNA) introduces significant advancements in the measurement of economic activities in the digital age. This article explores the methodologies developed to capture the impact of digitalization, including the recognition of data as an asset, the role of digital intermediation platforms, and the treatment of free digital products. These updates provide a more accurate and comprehensive understanding of the modern economy, reflecting the true nature of economic transactions today.

By enhancing the measurement of GDP and other key macroeconomic indicators, the 2025 SNA offers new insights into the digital economy's influence on production, consumption, and international trade. This article aims to provide a comprehensive understanding of these developments and their implications for economic measurement and analysis, highlighting the potential extension of analytical possibilities stemming from the innovations related to the digital economy in the 2025 SNA.

Globalization has fundamentally altered the landscape of production and income allocation, placing greater demands on macroeconomic statistical frameworks. The rise of multinational enterprises (MNEs), with complex cross-border operations and financial strategies, including transfer pricing and the use of special purpose entities (SPEs), has blurred the connection between where goods are produced and where income accrues. This disconnect complicates the interpretation of key indicators and can lead to misdiagnosis of domestic economic conditions and risks. To address these challenges, user needs have shifted toward more nuanced measures that reveal who truly benefits from economic activity. These evolving needs have driven significant updates to international statistical standards, including the 2025 SNA and BPM7, which introduce new frameworks and supplementary measures designed to better capture the realities of a globalized economy. This article examines the profound impact of MNEs on macroeconomic statistics, highlights the limitations of existing frameworks, and assesses recent advances in international guidance. Drawing on country experiences, it demonstrates the importance of adapting measurement approaches to ensure that economic data remain meaningful and informative in an interconnected world.

This article explores recent advancements in the measurement of cross-border financial flows and positions, as introduced in the latest edition of the International Monetary Fund (IMF)'s Integrated Balance of Payments and International Investment Position Manual (BPM7). It highlights how BPM7 enhances the presentation of foreign direct investment, portfolio investment, and other financial instruments, while emphasizing the policy value of integrating balance of payments (BOP) flows with international investment position (IIP) data through the integrated IIP framework. These innovations offer a more coherent and comprehensive view of global financial linkages, supporting improved analysis of systemic risks and more informed financial stability policymaking.

This paper examines how the updated macroeconomic statistical standards reflect the ongoing financial innovations with regards to new financial instruments, service providers, and activities. It discusses the updated statistical methodologies (recently updated 2025 SNA/BPM7) to capture activities related to fintech, crypto-assets, and other emerging phenomena in digital finance, as well as challenges and aspects still to be addressed. Moreover, the article highlights the significance of data in understanding the evolving financial landscape and assessing macroeconomic implications of fintech and crypto assets.

Islamic finance is based on Shari'ah, or Islamic law, which prohibits the receipt and payment of “riba” (normally translated as interest), “gharar” (excessive uncertainty), “maysir” (gambling), and short sales or financing activities that it considers harmful to society. Rather, it involves risk and reward sharing, transactions with real economic purposes, and fair actions of the participants. Consequently, innovative arrangements such as contractual profit-and-loss sharing and leasing arrangements are common. However, macroeconomic statistics, such as the System of National Accounts (SNA) and the Balance of Payments Manual (BPM) have traditionally focused on the coherent treatment of conventional finance rather than Islamic finance. Thus, comprehensive and coherent internationally-endorsed recommendations to account for Islamic finance in both the national accounts and external sector statistics frameworks were absent. These gaps might have impacted the quality and comparability of economic statistics in countries with prominent Islamic financial activities. They have been addressed by the recently-concluded global efforts to update the System of National Accounts, 2008 (2008 SNA) and the sixth edition of the Balance of Payments and International Investment Position Manual (BPM6). These have resulted in recommendations on Islamic finance in the System of National Accounts, 2025 (2025 SNA) and the seventh edition of the Balance of Payments Manual (BPM7) from which Islamic finance statistics can be compiled to better inform policy makers and other stakeholders.

This paper emphasizes the importance of accurately measuring informal economic activity to support inclusive and sustainable economic growth policies. A major advance in this area is the integration of internationally agreed labor-statistical definitions of informality into the System of National Accounts 2025 (2025 SNA) and the seventh edition of the Balance of Payments and International Investment Positions Manual (BPM7). For the first time, these standards enable consistent measurement of informal production, employment, and external-sector activities within macroeconomic accounts. Recognizing informal activities ensures that the contributions of millions of workers and enterprises are properly reflected in economic analysis, thereby supporting more equitable policy responses and fostering development that leaves no one behind. The paper details recent methodological progress in capturing informal economic activity through integrated frameworks that link informal economic production from both enterprise and employment perspectives. It highlights ongoing challenges, including data gaps, inconsistent national practices, and limited capacity in many countries. Improved statistics can inform a range of policy applications, from poverty reduction and labor market regulation to macroeconomic planning. The paper concludes by identifying priorities for further research, including incorporating data into mainstream economic indicators and improving comparability of informality statistics across countries.

The global economy is in a state of constant transformation, shaped by digital disruption, climate change, shifting demographics, and evolving global value chains. This ongoing change presents challenges not only for researchers, businesses, and governments striving to understand economic realities, but also for macroeconomic statisticians, who must develop and maintain the international statistical standards that support the measurement of economic activity. Recently, the System of National Accounts (SNA) and the Balance of Payments and International Investment Position Manual (BPM) underwent updates; notably, there was a 17-year gap between the previous and current versions, and it is expected to take an additional 5 years for countries to implement the new standards. This article argues that a 25-year cycle from one update's conception to full implementation may no longer be sufficient to meet the needs of users in today's rapidly changing environment. The article contends that the process for updating statistical standards must evolve. It first highlights the SNA/BPM research program over the short to medium term (as it is now understood) and then proposes a path forward for addressing and resolving this research in a more agile and adaptive approach.

The growing availability of online data creates new opportunities to improve the timeliness and detail of official statistics, particularly in domains such as price monitoring and inflation measurement. However, leveraging web-scraped data for official use requires alignment with standardized classification frameworks such as the European Classification of Individual Consumption According to Purpose (ECOICOP). We train two natural-language models, a lightweight convolutional neural network (CNN) and a fine-tuned BERTimbau transformer, to classify Portuguese food and beverage items into ECOICOP categories. Using 100,000 product titles scraped from six national supermarket sites and labeled via a human-in-the-loop workflow, the CNN reaches a macro-F1 of 92.19 % with minimal computing cost, while the transformer attains 94.00 %, the first such result for Portuguese. Both models are published on Hugging Face, enabling reproducible inference at scale while the source data remain confidential. The study delivers the first open-source Portuguese ECOICOP classifiers for food and beverage products, a replicable low-resource labeling workflow, and a benchmark of accuracy-speed trade-offs to guide researchers in similar tasks.

Historically, decennial population censuses have served as the foundation for official statistics on commuter flows between municipalities. However, to obtain more timely data, Istat, the Italian National Institute for Statistics, among other statistical agencies, has transitioned to conducting annual census surveys with smaller sample sizes. This shift introduces several challenges in deriving accurate commuting statistics. The availability of Mobile Network Operator (MNO) data enables the identification of recurring flows between home and work or study locations. These flows could potentially become the primary source for generating official statistics by properly adjusting the MNO coverage in a quasi randomisation (QR) approach. In this paper, we explore the use of MNO data to estimate the number of commuters between the municipalities of an Italian region, with a focus on the quasi-randomisation approach.

Micro-level bank card data offers statistical offices a timely and granular view of economic activity, enabling deeper insights into the digital economy. The National Statistics Office of Malta, in collaboration with the Central Bank of Malta, has acquired point-of-sale (POS) and online transaction data for the use of official statistics. Methodologies for data standardisation, linkage to the Statistical Business Register and merchant identification via fuzzy matching are outlined. The alignment between traditional survey and bank card indicators, was examined to analyse the bank card data's representativeness of the retail industry. Two metrics were derived: the Merchant Penetration Rate and the Weighted Average Bank Card Usage Rate, to assess the utility of bank card data as a proxy for market activity. Significant limitations remain a concern, such as partial market coverage and reliance on third-party data providers. The paper also discusses policy implications for improving statistical visibility to improve coverage, regional analysis, and reducing tax evasion. This article aims to contribute to the growing need for alternative data in official statistics and the evolving role of national statistics institutes in a digital economy.

Health represents a fundamental dimension of individual well-being, but it is equally significant when examined at the societal level, where individual and collective health outcomes are intrinsically interrelated. This study aims to analyse health across the 27 European Union Member States by investigating the impact of various types of determinants and assessing the potential for generating reliable predictions of health indicators. In the first part of the analysis, we focus on the population's self-perceived health status, exploring how different data processing strategies can enhance the performance of machine learning algorithms, particularly in the context of small sample sizes. In the second part, we replicate the same methodological approach using an objective health indicator—life expectancy at birth—in order to compare and contrast the findings. This comparison offers insights into the effectiveness and robustness of the methodological framework applied, and allows for broader reflections on the interplay between subjective and objective measures of health.

Statistics Netherlands (CBS) is responsible for compiling and publishing reliable and high quality statistics of the Dutch society. Effectively communication of official statistics is important: it indicates the value for society. However, assessing quality aspects with respect to the design of visualizations and correct understanding of the message of statistical communication, have not received much attention yet. In this paper we aim to explore further directions of multi-disciplinary research to evaluate the quality of statistical information communication as a first step towards more cognitive friendly statistics. The central research question is how can statistical information be effectively visualized and communicated so that it is understandable and usable by the intended audience? As first explorations, several user evaluation studies are described were different types of visualizations, graphs, infographics, texts and pictures have been evaluated with various user groups. These findings illustrate relevant elements that indicate the perception, understanding and design quality of statistical information and communication products. Reflections on a holistic, interdisciplinary approach that unified guidelines for designing different types of statistical communication products with evidence based research, contributes to the development of quality framework for producing cognitive friendly statistics, i.e., an effective strategy of statistical information and communication.

Statistical organizations worldwide are increasingly adopting open source technologies for producing official statistics. This shift is motivated by the potential of open source tools to increase transparency, improve efficiency, and enhance reproducibility. Moreover, young professionals in statistics and data science enter the labour market with strong skills in open source tools. The adoption of open source software signifies a change in how statistical organizations operate and collaborate.

This paper provides an overview of the state of open source adoption in official statistics. It details the open source movement among statistical organizations, the experiences of Statistics Netherlands with open source adoption and the creation of R-packages implementing common statistical methods. It also describes the development and use of the “awesome list of official statistics software” and discusses a set of principles for open source in official statistics, derived from best practices across various organizations. These principles have been endorsed (June 2025) by the Conference of European Statisticians (CES).

Furthermore, it explores future directions for maturing this community, including metrics for assessing maturity, such as true independence of software modules, support for uncertainty propagation, and privacy by design. Moreover it presents ideas on redesigning the statistical open source landscape.

Paddy data is essential for policymakers to formulate Indonesia's national food security strategies, especially regarding harvested area. Currently, estimates are generated monthly using ground truth data from sampled locations through the Area Sampling Frames (ASF) process. However, the high cost and various field obstacles highlight the need for alternative methods. Satellite Imagery Time Series (SITS) data, mainly Sentinel-1 historical imagery, offers a promising alternative for detecting phenological stages. However, SITS data require machine learning modeling—for example, XGBoost—which has proven successful in several classification tasks. This study presents an alternative approach in West Java province, a major region for paddy production. The workflow includes preliminary analysis, data preprocessing, region-specific modelling, prediction, and estimation. Most regional clusters demonstrate high accuracy in classifying phenological stages, and harvested area patterns closely align with official statistics, demonstrating the effectiveness and potential of this approach.

Machine Learning (ML) models often achieve high accuracy, but fail to meet the reliability and robustness standards required for Official Statistics (OS). Neural networks, in particular, function as black-box predictors prone to overconfidence, offering no direct method to measure true uncertainty in predictions and estimates of population parameters. Non-rigorous approaches include treating ML predictions as gold standard data and heuristic notions of uncertainty, like softmax scores in classification problems, as valid measures of confidence. This can easily lead to unreliable uncertainty quantification. This paper handles two distinct problems: (1) quantifying prediction-level uncertainty for new observations and (2) quantifying noise-free uncertainty for estimates of population parameters. We propose handling the former via conformal prediction (CP) and the latter using prediction-powered inference (PPI). Both are model-agnostic statistical frameworks for uncertainty quantification. Finally, we present real-world use cases for OS, applying both techniques.

National statistical offices (NSOs) increasingly rely on record linkage to link census data, administrative sources, and survey responses. However, conventional string-similarity methods often struggle with free-text fields. To address these challenges, this paper systematically benchmarks modern open-source large language models (LLMs) against classic string-based comparators for record linkage. Building on these findings, this paper introduces a hybrid approach that retains well-established probabilistic frameworks yet integrates an LLM-based classifier for ambiguous record pairs. A Bayesian update is applied to combine the LLM's output with the prior probability, with the aim of reducing the burden on manual clerical review. The experiments show that selectively deploying open-source LLMs for the most uncertain pairs can significantly reduce manual effort by refining decisions through Bayesian updating. As NSOs must ensure transparency, explainability, and adherence to official statistical standards, this paper systematically addresses these concerns while evaluating the potential of LLMs for record linkage. Practical considerations including secure on-premises deployment, computational cost, human-in-the-loop review, and calibration are discussed to support responsible adoption in official statistics.

National Statistical Offices (NSOs) are increasingly adopting new analytical practices to address growing demands for timely, granular and policy-relevant statistics. Among these, Non-Specialist Data Science (NSDS) has emerged as a strategic approach to expanding analytical capacity by enabling professionals without formal data-science training to conduct advanced analysis within secure institutional environments. This paper presents a comparative study of governance models for NSDS in three NSOs: the UK Office for National Statistics (ONS), France's National Institute of Statistics and Economic Studies (INSEE) in partnership with the Centre d’Accès Sécurisé aux Données (CASD), and Statistics New Zealand (Stats NZ). Drawing on documentary analysis and semi-structured interviews, the study explores how these organisations embed NSDS within the Generic Statistical Business Process Model (GSBPM) and complementary governance mechanisms such as the Five Safes. The findings highlight the importance of secure analytical environments, tiered access regimes, methodological oversight and structured capacity building in supporting expanded analytical participation without undermining statistical quality, confidentiality or institutional credibility. The paper argues that NSDS should be understood primarily as a governance transformation rather than merely a technical innovation, with implications for workforce development, organisational design and international standard-setting in official statistics.

The Swedish Labour Force Survey (LFS) is the foundation for official statistics on the Swedish labour market. As from 2021, it must comply with the new EU framework regulation on social statistics. To this end, several changes were made to the survey, both regarding the population and the questionnaire. Moreover, the auxiliary information used in the estimation was revised at the same time. The combined effect of these changes caused breaks in many of the LFS time series.

To estimate and correct for the breaks in the LFS time series, Statistics Sweden performed a parallel run with both the old and the new questionnaires during 2021. In this paper, we describe how the parallel run was complemented with other available data to analyse the breaks in the time series. Moreover, using a combination of imputation models for the micro data and calibration, we derive recalibrated weights for all the LFS respondents during 2005–2020. These recalibrated weights constitute a link at the micro level between the old and new procedures. Time series derived using the recalibrated weights and the imputed microdata are fully comparable over time with time series calculated using the new procedure from 2021 and on.

Time-based surveys are important tools in collecting data over time, but the presence of measurement errors (MEs) can significantly compromise the accuracy of the survey results. This article evaluates the impact of MEs in time-based surveys and introduces a new class of memory-type estimators designed to mitigate these errors using exponentially weighted moving average (EWMA) statistics. Unlike conventional estimators, which often assume constant error structures, the proposed memory-type estimator accounts for dynamic error processes and past and present observations in form of EWMA statistics, thus improving the precision of estimates over time. Through both theoretical analysis and empirical simulations, we demonstrate that the memory-type estimator provides superior performance in reducing bias and variance compared to the existing estimation methods under various error scenarios. This research contributes to the field of survey methodology by providing an advanced tool for more accurate time-based data analysis, particularly in contexts where MEs cannot be ignored. The findings highlight the potential for improving data reliability and informing policy decisions based on time-series survey data.

In response to a resolution by the United Nations Economic and Social Council (ECOSOC), the Productive Capacities Index (PCI) was developed by UNCTAD as a multidimensional statistical tool to measure, monitor, and benchmark productive capacities through a comprehensive framework, encompassing human capital, natural capital, energy, transport, ICT, institutions, private sector, and structural change. This paper outlines the conceptual foundation of the PCI, the methodological approach used for indicator selection and data transformation, and the statistical processes, particularly Principal Component Analysis, used to derive the composite index. Emphasis is placed on how the PCI supports evidence-based policymaking, facilitates cross-country comparisons, and contributes to statistical development agendas aligned with the 2030 Agenda for Sustainable Development. The PCI serves as a valuable complement to traditional indicators such as GDP, offering insights into the underlying capabilities necessary for economic and social development.

The widespread reliance on agriculture in developing countries necessitates efficient methods for collecting and processing reliable, timely agricultural data. This paper documents over 25 years of experience from national agricultural surveys and censuses in Mozambique, reviewing data collection procedures and outcomes since 1999/2000. It identifies key challenges in data collection, management, and analysis, and proposes solutions for improvement. Key lessons derived include: the critical role of pilot surveys in refining instruments; the continuous need to revise and upgrade questionnaires for clarity and programming; the importance of crop-specific conversion factors for accurate production estimates; leveraging historical data for improved comparability; strengthening data management protocols and server oversight; developing tailored survey weights for improved area estimation; and ensuring continuous data analyst involvement throughout the survey lifecycle. These practices collectively enhance methodological rigor, data quality, and the overall utility of agricultural statistics.

This study used data from the 2011 Bangladesh Demographic and Health Survey and the 2011 Bangladesh Population and Housing Census. The Fay-Herriot model was applied to estimate the districts and sub-districts (upazilas) level indicators of undernutrition (stunting, wasting and underweight) for children under five years of age in Bangladesh. At the district level, Narsingdi (63.9%), Feni (19.8%) and Bandarban (62.5%) had the highest prevalence of stunting, wasting and underweight respectively. Among the 64 districts, 28 had stunting prevalence above 40%, wasting prevalence exceeded 15% in 21 districts and 16 districts had underweight prevalence exceeding 40%. Furthermore, all sub-districts within the Sylhet Division were identified as the most severely affected areas for all three indicators of undernutrition. Overall, the Sylhet Division emerged as one of the regions facing the greatest challenges in terms of the three undernutrition indicators for children under five.

Poverty in Thailand shows strong spatial dependence that existing administrative boundaries fail to capture, leading to policies that overlook local socioeconomic realities. This study proposes a data-driven regionalization framework to infer geographically coherent “policy regions” that better represent poverty dynamics. Using household-level data from the Thai People Map and Analytics Platform (TPMAP), we analyze spatial autocorrelation across multiple poverty factors through Moran’s statistics and principal component analysis, followed by spatially constrained hierarchical clustering to delineate coherent regions. Bayesian hierarchical and geographically weighted regression models are then employed to examine how education influences household income at provincial, regional, and national levels. Our results identified six regions that reflect more accurately poverty structures than official divisions. Northern and Northeastern Thailand emerge as the regions most affected by low education, income, and savings, while Central Thailand shows higher inequality. The inferred regions demonstrate that spatially contiguous provinces often share similar socioeconomic structures, suggesting that policy targeting should align with these patterns rather than provincial borders. Our findings provide a quantitative foundation for evidence-based regional planning, enabling policymakers to design differentiated yet regionally coordinated interventions. The approach illustrates how spatial statistical modeling can bridge the gap between data analysis and effective poverty-alleviation policy.