Federal Independent Agencies
Here's a look at documents from federal independent agencies
Federal Independent Agencies
Featured Stories
Inter-American Development Bank: 'Geographic and Sectoral Impact of Productivity'
WASHINGTON, Sept. 11 (TNSLrpt) -- The Inter-American Development Bank issued the following white paper in August 2026 entitled "The Geographic and Sectoral Impact of Productivity."
Here are excerpts:
* * *
Abstract
To understand the geographic and sectoral impact of productivity, we build the first input-output matrix disaggregated at the geographic level within a country using unique administrative data harmonized to match national accounts. We use this data from Chile to calibrate a state-of-the-art general equilibrium quantitative trade model with production networks, labor mobility, firm ... Show Full Article WASHINGTON, Sept. 11 (TNSLrpt) -- The Inter-American Development Bank issued the following white paper in August 2026 entitled "The Geographic and Sectoral Impact of Productivity." Here are excerpts: * * * Abstract To understand the geographic and sectoral impact of productivity, we build the first input-output matrix disaggregated at the geographic level within a country using unique administrative data harmonized to match national accounts. We use this data from Chile to calibrate a state-of-the-art general equilibrium quantitative trade model with production networks, labor mobility, firmselection, international and domestic trade, congestion of fixed factors, and knowledge diffusion. We consider two applications. First, we study the aggregate effects of local productivity shocks.
We show that location-sector interactions are crucial: locations and sectors separately account for less than half of the dispersion in GDP elasticities from location-sectorspecific productivity shocks. Geography-specific input-output linkages explain 16% of the dispersion, due to the role played by small and influential markets. Second, we analyze the exit of a large steel plant. We show that geographically disaggregated production linkages substantially increase the propagation of the plant exit.
* * *
Introduction
Understanding how local productivity shocks propagate through the economy is ubiquitous in policy debates. Some of these shocks arise from policy interventions, such as place-based initiatives aimed at boosting productivity in lagging regions or industrial policies targeting specific sectors. Other shocks are due to adverse events, including natural disasters, social conflicts, or the exit of large firms; or positive events such as increases in global demand for commodities that are geographically concentrated. Each of these episodes initially impacts a specific market defined by its sector and location. More importantly, these shocks can then propagate indirectly across locations and sectors through multiple channels, such as upstream and downstream propagation through supply chains, potentially generating uneven impacts, with winners and losers across locations and sectors. Disaggregated quantitative analysis has been constrained by the lack of geographically disaggregated economic accounts, particularly input-output tables, which are crucial for understanding the propagation of such shocks.
We construct novel geographically and sectorally disaggregated economic accounts that characterize economic activity across location and sector pairs in Chile, including geographic and sectoral input-output matrices, using unique administrative data and survey sources. The resulting disaggregated accounts are disciplined to match national accounts and aggregate identities. We combine this new dataset with a state-of-the-art general equilibrium quantitative trade model (Caliendo et al., 2018) to study how local productivity changes are transmitted across all locations and sectors of the economy.
These shocks operate through multiple mechanisms, including backward and forward input-output linkages, domestic and international trade, labor migration, congestion in fixed factors such as land and infrastructure, selection and extensive-margin adjustments of firms, and knowledge diffusion.
The disaggregated economic accounts link national accounts aggregates to detailed information disaggregated along both sectoral and location dimensions. Specifically, we build on national accounts built by the Central Bank of Chile (CBC), which provides estimates of regional gross domestic product (GDP), and extend them by geographically disaggregating the input-output matrices. The resulting dataset is internally consistent: aggregating sectoral and spatial flows exactly reproduces national totals and satisfies all accounting identities. Such disaggregated measures of economic activity and interconnections through supply chains remain scarce, both for Chile and the rest of the world.
To implement this measurement, we combine a wide range of administrative data covering the universe of formal firms in Chile between 2018 and 2023. First, we use standard administrative tax records from the Chilean Internal Revenue Service to construct firmlevel measures. This data provides direct measures of total, domestic, and international sales; total, domestic, and imported material expenditures; value added; employment; wages; capital stocks and investment. We start from national accounts regional GDP estimates and disaggregate those at the location-sector level using shares from firms' tax records. Second, and more novel, we exploit firm-to-firm value-added tax records to directly measure production linkages across locations and sectors. Existing work typically infers these linkages from surveys or national input-output tables and spatially allocates them using indirect methods, such as gravity equations or proportionality assumptions. In contrast, the Chilean firm-to-firm data report the origin and destination of each transaction for the universe of formal firms, allowing us to observe trade flows across locations and sectors rather than imputing them. This feature enables us to recover the production network at a fine geographic and sectoral resolution without imposing assumptions on the data.
Taken together, this dataset delivers a comprehensive coverage of economic activity across the Chilean territory and provides a direct, measurement-based characterization of geographically and sectorally disaggregated input-output linkages.
Using the model calibrated with the disaggregated economic accounts, we implement two applications to understand how location-sector-specific productivity shocks propagate throughout the economy. The first application studies the aggregate effects of local productivity shocks across all location-sector pairs in the economy. The second application analyzes the exit of a large steel plant in the south of Chile, quantifies its local and aggregate effects, and identifies winners and losers across locations and sectors.
The first application delivers five results. First, when shocks are allowed to vary only by location, or only by sector, we recover the broad qualitative patterns emphasized in the quantitative trade literature. We document the elasticity of GDP to productivity shocks, which we refer to GDP elasticities. We scale these elasticities by the share of GDP that is exposed to the shock, that is, the Domar weight of the shock. We scale by the Domar weight in order to control for mechanical heterogeneity driven by the magnitude of the shock. Regional GDP elasticities are largest in economically central and diversified areas, while sectoral GDP elasticities are strongest in business and social services and comparatively muted in mining. Second, once shocks are disaggregated at the location sector level, what matters is the dispersion rather than the average effects. Average GDP elasticities remain close to those obtained under one-dimensional shocks, but heterogeneity rises sharply. Furthermore, geography accounts for 12% of the dispersion in aggregate GDP elasticities, sectors account for 16%, and 73% is accounted for by the interaction between the two. Where a shock occurs and which sector is hit are therefore not sufficient statistics for understanding the impact of location-sector-specific shocks. What matters is the interaction between sectors and locations. This underscores the relevance of including geography when measuring economic accounts and the propagation of productivity shocks.
Third, standard sufficient statistics (Baqaee & Farhi, 2020) remain informative but incomplete. If Domar weights were sufficient statistics, all GDP elasticities should be equal to one. The fact that there is significant dispersion means that those sufficient statistics are not sufficient in this framework. As shown by Baqaee and Farhi (2019, 2020, 2024), the difference between these sufficient statistics and the model can be driven by both non-linearities and inefficiencies of the model.
Fourth, trade and input-output linkages shape propagation through distinct channels.
Shutting down production networks reduces the dispersion of impacts by 44%. Furthermore, without input-output linkages, sectors become more important. Shutting down trade has an even larger effect, accounting for about 54% of the dispersion of GDP elasticities.
Contrary to input-output linkages, without trade, dispersion of GDP elasticities becomes much more geographic and the location-sector interaction falls sharply. These channels are therefore complementary: trade redistributes shocks across sector, while production networks redistribute shocks across space.
Finally, geographically disaggregated input-output tables matter mainly for heterogeneity among smaller and influential markets rather than for average aggregate effects.
Replacing the geography-specific input-output matrix with the standard national inputoutput table leaves average GDP elasticities largely unchanged, but removes about 16% of the overall dispersion of GDP elasticities across location-sectors. This reduction is driven by small markets that have high GDP elasticities in the benchmark model. Furthermore, eliminating the geography aspect of the input-output matrix increases the relevance of sectors in GDP elasticity dispersion by about 19%, highlighting the relevance of appropriately measuring the spatial dimension of input-output tables.
The second application illustrates these insights in a concrete and policy-relevant setting: the exit of a large plant in a location that relies substantially on the economic activity of that plant. Although the plant that exited, Huachipato, accounts for only a negligible share of national GDP, its exit generates aggregate losses several times larger than its direct contribution. About three-quarters of the national GDP and employment losses occur outside the location-sector of the plant, showing that localized firm exits can have substantial national consequences once labor mobility, trade and production networks are accounted for.
This case study also highlights the value of geographically disaggregated inputoutput data. Replacing geography-specific production linkages with national input-output coefficients lowers the GDP elasticity of the shock from 3.41 to 1.51 This comparison shows that spatial detail in production linkages matters quantitatively for the aggregate strength of propagation. This case study reinforces our central message: localized productivity shocks propagate through the economy not because of geography or sector alone, but because of how the two interact within the economy's spatial and production network structure.
Related Literature. We relate to three strands of the literature. First, we contribute to the literature on the construction of disaggregated economic accounts. A prominent set of contributions constructs cross-region and international trade matrices, often by combining shipment data, international trade statistics, input-output tables, and regional labor or output shares as in Leontief and Strout (1963), Caliendo and Parro (2015), and Caliendo et al. (2018, 2019). Because disaggregated and domestic trade data are typically unavailable, these approaches rely on proportionality assumptions or gravity-based allocations to spatially disaggregate producer flows. For instance, Rodriguez-Clare et al. (2025) recover industry-level trade flows across U.S. region pairs using a gravity-based framework, while Andersen et al. (2026) apply a similar approach to Denmark for firm's transactions. On the other hand, Caliendo et al. (2018) measures the geographical aspect of input-output linkages through proportionality assumptions. While these methods have substantially advanced the measurement of spatial trade linkages, they recover cross-location flows indirectly and do not identify the full geography-sector input-output structure of the economy. A closely related strand uses administrative tax records to measure income flows and production linkages at a high level of granularity as in Card et al. (2013), Huneeus (2018), Dhyne et al. (2021), Adao et al. (2022), Bernard et al. (2022), and Atkin et al. (2025).
In particular, Atkin et al. (2025) exploits Chilean administrative data to measure detailed bilateral transactions between individuals, firms, and the government, highlighting how distortions influence real income across individuals, but they do not focus on the spatial aspect of the production network let alone its role on propagating productivity shocks.
Our paper contributes to this literature by constructing novel geographically and sectorally disaggregated accounts for Chile that jointly characterize economic activity and production linkages at the location-sector level for an entire country using detailed administrative data and without imposing additional assumptions or structure on the data. Furthermore, all bilateral flows are disciplined to exactly match Chile's national accounts and input-output tables, ensuring national aggregation, internal consistency, and compliance with accounting identities.
Second, we contribute to the literature of propagation through production networks that studies how local productivity shocks impact aggregates and propagate through the economy. A classical starting point is the insight that, first-order aggregate responses can be summarized by Domar weights, as formalized by Hulten (1978). Building on this foundation, a large literature studies shock propagation through input-output linkages and network structure, from early multisector contributions such as Long and Plosser (1983) and Horvath (2000) to modern network-based analyses in Acemoglu et al. (2012, 2016), and nonlinear general-equilibrium frameworks in Baqaee and Farhi (2019, 2020, 2024). This strand of the literature typically abstracts from studying the role of geography.
We contribute to this literature by showing the role of geography for the propagation of shocks. In particular we bring to this literature the frameworks from the quantitative spatial literature (Caliendo et al., 2018) which include trade frictions related to geography and we measure the role of geography by building disaggregated economic accounts across space in Chile.
Finally, our paper contributes to the literature of quantitative spatial economics. This literature has developed general equilibrium frameworks to quantify how productivity and trade shocks impact economic activity across locations and affect aggregate welfare through trade costs and geography, starting from the Ricardian structure in Eaton and Kortum (2002) and its counterfactual implementation in Dekle et al. (2008). This approach is extended to multi-sector models with input-output linkages in Caliendo and Parro (2015) and to fully specified spatial equilibria in Caliendo et al. (2018) and Caliendo et al. (2019).1 Our contribution to this literature is to provide a unified measurement and quantitative framework to study the geographic and sectoral propagation of productivity shocks through empirically identified production networks embedded in a spatial general equilibrium. The closest paper to our approach is Caliendo et al. (2018), which constructs spatial trade linkages by combining data on cross-region shipments, input-output trade between producers, and regional labor or output shares. We depart from this approach by directly measuring bilateral geography-sector input flows using administrative firm-to-firm transaction data, thereby identifying the full origin-destination structure of production without relying on proportional or gravity-based allocations. This richer measurement is especially important for service sectors, whose geographic flows have been largely absent from existing spatial analyses due to their non-physical nature and the lack of survey data. We apply our framework to the Chilean economy, whose pronounced geographic heterogeneity provides a natural setting to study how productivity shocks propagate across sectors and locations.
The remainder of the paper is organized as follows. Section 2 describes the construction of the disaggregated economic accounts, combining state-of-the-art administrative records and survey sources and making them consistent with national accounts. Section 3 presents descriptive evidence on the economic composition of Chile and the geographic structure of the production network. Section 4 introduces the quantitative trade model and defines the sufficient statistics used to measure the aggregate effects of location-sector shocks and the geographic, sectoral, and geography-sector exposures that govern their propagation. Section 5 presents the model calibration. Section 6 reports the results from the quantitative exercises and their applications. Section 7 concludes.
* * *
View full text here: https://publications.iadb.org/en/geographic-and-sectoral-impact-productivity
[Category: IADB]
Here are excerpts:
* * *
Abstract
To understand the geographic and sectoral impact of productivity, we build the first input-output matrix disaggregated at the geographic level within a country using unique administrative data harmonized to match national accounts. We use this data from Chile to calibrate a state-of-the-art general equilibrium quantitative trade model with production networks, labor mobility, firm ... Show Full Article WASHINGTON, Sept. 11 (TNSLrpt) -- The Inter-American Development Bank issued the following white paper in August 2026 entitled "The Geographic and Sectoral Impact of Productivity." Here are excerpts: * * * Abstract To understand the geographic and sectoral impact of productivity, we build the first input-output matrix disaggregated at the geographic level within a country using unique administrative data harmonized to match national accounts. We use this data from Chile to calibrate a state-of-the-art general equilibrium quantitative trade model with production networks, labor mobility, firmselection, international and domestic trade, congestion of fixed factors, and knowledge diffusion. We consider two applications. First, we study the aggregate effects of local productivity shocks.
We show that location-sector interactions are crucial: locations and sectors separately account for less than half of the dispersion in GDP elasticities from location-sectorspecific productivity shocks. Geography-specific input-output linkages explain 16% of the dispersion, due to the role played by small and influential markets. Second, we analyze the exit of a large steel plant. We show that geographically disaggregated production linkages substantially increase the propagation of the plant exit.
* * *
Introduction
Understanding how local productivity shocks propagate through the economy is ubiquitous in policy debates. Some of these shocks arise from policy interventions, such as place-based initiatives aimed at boosting productivity in lagging regions or industrial policies targeting specific sectors. Other shocks are due to adverse events, including natural disasters, social conflicts, or the exit of large firms; or positive events such as increases in global demand for commodities that are geographically concentrated. Each of these episodes initially impacts a specific market defined by its sector and location. More importantly, these shocks can then propagate indirectly across locations and sectors through multiple channels, such as upstream and downstream propagation through supply chains, potentially generating uneven impacts, with winners and losers across locations and sectors. Disaggregated quantitative analysis has been constrained by the lack of geographically disaggregated economic accounts, particularly input-output tables, which are crucial for understanding the propagation of such shocks.
We construct novel geographically and sectorally disaggregated economic accounts that characterize economic activity across location and sector pairs in Chile, including geographic and sectoral input-output matrices, using unique administrative data and survey sources. The resulting disaggregated accounts are disciplined to match national accounts and aggregate identities. We combine this new dataset with a state-of-the-art general equilibrium quantitative trade model (Caliendo et al., 2018) to study how local productivity changes are transmitted across all locations and sectors of the economy.
These shocks operate through multiple mechanisms, including backward and forward input-output linkages, domestic and international trade, labor migration, congestion in fixed factors such as land and infrastructure, selection and extensive-margin adjustments of firms, and knowledge diffusion.
The disaggregated economic accounts link national accounts aggregates to detailed information disaggregated along both sectoral and location dimensions. Specifically, we build on national accounts built by the Central Bank of Chile (CBC), which provides estimates of regional gross domestic product (GDP), and extend them by geographically disaggregating the input-output matrices. The resulting dataset is internally consistent: aggregating sectoral and spatial flows exactly reproduces national totals and satisfies all accounting identities. Such disaggregated measures of economic activity and interconnections through supply chains remain scarce, both for Chile and the rest of the world.
To implement this measurement, we combine a wide range of administrative data covering the universe of formal firms in Chile between 2018 and 2023. First, we use standard administrative tax records from the Chilean Internal Revenue Service to construct firmlevel measures. This data provides direct measures of total, domestic, and international sales; total, domestic, and imported material expenditures; value added; employment; wages; capital stocks and investment. We start from national accounts regional GDP estimates and disaggregate those at the location-sector level using shares from firms' tax records. Second, and more novel, we exploit firm-to-firm value-added tax records to directly measure production linkages across locations and sectors. Existing work typically infers these linkages from surveys or national input-output tables and spatially allocates them using indirect methods, such as gravity equations or proportionality assumptions. In contrast, the Chilean firm-to-firm data report the origin and destination of each transaction for the universe of formal firms, allowing us to observe trade flows across locations and sectors rather than imputing them. This feature enables us to recover the production network at a fine geographic and sectoral resolution without imposing assumptions on the data.
Taken together, this dataset delivers a comprehensive coverage of economic activity across the Chilean territory and provides a direct, measurement-based characterization of geographically and sectorally disaggregated input-output linkages.
Using the model calibrated with the disaggregated economic accounts, we implement two applications to understand how location-sector-specific productivity shocks propagate throughout the economy. The first application studies the aggregate effects of local productivity shocks across all location-sector pairs in the economy. The second application analyzes the exit of a large steel plant in the south of Chile, quantifies its local and aggregate effects, and identifies winners and losers across locations and sectors.
The first application delivers five results. First, when shocks are allowed to vary only by location, or only by sector, we recover the broad qualitative patterns emphasized in the quantitative trade literature. We document the elasticity of GDP to productivity shocks, which we refer to GDP elasticities. We scale these elasticities by the share of GDP that is exposed to the shock, that is, the Domar weight of the shock. We scale by the Domar weight in order to control for mechanical heterogeneity driven by the magnitude of the shock. Regional GDP elasticities are largest in economically central and diversified areas, while sectoral GDP elasticities are strongest in business and social services and comparatively muted in mining. Second, once shocks are disaggregated at the location sector level, what matters is the dispersion rather than the average effects. Average GDP elasticities remain close to those obtained under one-dimensional shocks, but heterogeneity rises sharply. Furthermore, geography accounts for 12% of the dispersion in aggregate GDP elasticities, sectors account for 16%, and 73% is accounted for by the interaction between the two. Where a shock occurs and which sector is hit are therefore not sufficient statistics for understanding the impact of location-sector-specific shocks. What matters is the interaction between sectors and locations. This underscores the relevance of including geography when measuring economic accounts and the propagation of productivity shocks.
Third, standard sufficient statistics (Baqaee & Farhi, 2020) remain informative but incomplete. If Domar weights were sufficient statistics, all GDP elasticities should be equal to one. The fact that there is significant dispersion means that those sufficient statistics are not sufficient in this framework. As shown by Baqaee and Farhi (2019, 2020, 2024), the difference between these sufficient statistics and the model can be driven by both non-linearities and inefficiencies of the model.
Fourth, trade and input-output linkages shape propagation through distinct channels.
Shutting down production networks reduces the dispersion of impacts by 44%. Furthermore, without input-output linkages, sectors become more important. Shutting down trade has an even larger effect, accounting for about 54% of the dispersion of GDP elasticities.
Contrary to input-output linkages, without trade, dispersion of GDP elasticities becomes much more geographic and the location-sector interaction falls sharply. These channels are therefore complementary: trade redistributes shocks across sector, while production networks redistribute shocks across space.
Finally, geographically disaggregated input-output tables matter mainly for heterogeneity among smaller and influential markets rather than for average aggregate effects.
Replacing the geography-specific input-output matrix with the standard national inputoutput table leaves average GDP elasticities largely unchanged, but removes about 16% of the overall dispersion of GDP elasticities across location-sectors. This reduction is driven by small markets that have high GDP elasticities in the benchmark model. Furthermore, eliminating the geography aspect of the input-output matrix increases the relevance of sectors in GDP elasticity dispersion by about 19%, highlighting the relevance of appropriately measuring the spatial dimension of input-output tables.
The second application illustrates these insights in a concrete and policy-relevant setting: the exit of a large plant in a location that relies substantially on the economic activity of that plant. Although the plant that exited, Huachipato, accounts for only a negligible share of national GDP, its exit generates aggregate losses several times larger than its direct contribution. About three-quarters of the national GDP and employment losses occur outside the location-sector of the plant, showing that localized firm exits can have substantial national consequences once labor mobility, trade and production networks are accounted for.
This case study also highlights the value of geographically disaggregated inputoutput data. Replacing geography-specific production linkages with national input-output coefficients lowers the GDP elasticity of the shock from 3.41 to 1.51 This comparison shows that spatial detail in production linkages matters quantitatively for the aggregate strength of propagation. This case study reinforces our central message: localized productivity shocks propagate through the economy not because of geography or sector alone, but because of how the two interact within the economy's spatial and production network structure.
Related Literature. We relate to three strands of the literature. First, we contribute to the literature on the construction of disaggregated economic accounts. A prominent set of contributions constructs cross-region and international trade matrices, often by combining shipment data, international trade statistics, input-output tables, and regional labor or output shares as in Leontief and Strout (1963), Caliendo and Parro (2015), and Caliendo et al. (2018, 2019). Because disaggregated and domestic trade data are typically unavailable, these approaches rely on proportionality assumptions or gravity-based allocations to spatially disaggregate producer flows. For instance, Rodriguez-Clare et al. (2025) recover industry-level trade flows across U.S. region pairs using a gravity-based framework, while Andersen et al. (2026) apply a similar approach to Denmark for firm's transactions. On the other hand, Caliendo et al. (2018) measures the geographical aspect of input-output linkages through proportionality assumptions. While these methods have substantially advanced the measurement of spatial trade linkages, they recover cross-location flows indirectly and do not identify the full geography-sector input-output structure of the economy. A closely related strand uses administrative tax records to measure income flows and production linkages at a high level of granularity as in Card et al. (2013), Huneeus (2018), Dhyne et al. (2021), Adao et al. (2022), Bernard et al. (2022), and Atkin et al. (2025).
In particular, Atkin et al. (2025) exploits Chilean administrative data to measure detailed bilateral transactions between individuals, firms, and the government, highlighting how distortions influence real income across individuals, but they do not focus on the spatial aspect of the production network let alone its role on propagating productivity shocks.
Our paper contributes to this literature by constructing novel geographically and sectorally disaggregated accounts for Chile that jointly characterize economic activity and production linkages at the location-sector level for an entire country using detailed administrative data and without imposing additional assumptions or structure on the data. Furthermore, all bilateral flows are disciplined to exactly match Chile's national accounts and input-output tables, ensuring national aggregation, internal consistency, and compliance with accounting identities.
Second, we contribute to the literature of propagation through production networks that studies how local productivity shocks impact aggregates and propagate through the economy. A classical starting point is the insight that, first-order aggregate responses can be summarized by Domar weights, as formalized by Hulten (1978). Building on this foundation, a large literature studies shock propagation through input-output linkages and network structure, from early multisector contributions such as Long and Plosser (1983) and Horvath (2000) to modern network-based analyses in Acemoglu et al. (2012, 2016), and nonlinear general-equilibrium frameworks in Baqaee and Farhi (2019, 2020, 2024). This strand of the literature typically abstracts from studying the role of geography.
We contribute to this literature by showing the role of geography for the propagation of shocks. In particular we bring to this literature the frameworks from the quantitative spatial literature (Caliendo et al., 2018) which include trade frictions related to geography and we measure the role of geography by building disaggregated economic accounts across space in Chile.
Finally, our paper contributes to the literature of quantitative spatial economics. This literature has developed general equilibrium frameworks to quantify how productivity and trade shocks impact economic activity across locations and affect aggregate welfare through trade costs and geography, starting from the Ricardian structure in Eaton and Kortum (2002) and its counterfactual implementation in Dekle et al. (2008). This approach is extended to multi-sector models with input-output linkages in Caliendo and Parro (2015) and to fully specified spatial equilibria in Caliendo et al. (2018) and Caliendo et al. (2019).1 Our contribution to this literature is to provide a unified measurement and quantitative framework to study the geographic and sectoral propagation of productivity shocks through empirically identified production networks embedded in a spatial general equilibrium. The closest paper to our approach is Caliendo et al. (2018), which constructs spatial trade linkages by combining data on cross-region shipments, input-output trade between producers, and regional labor or output shares. We depart from this approach by directly measuring bilateral geography-sector input flows using administrative firm-to-firm transaction data, thereby identifying the full origin-destination structure of production without relying on proportional or gravity-based allocations. This richer measurement is especially important for service sectors, whose geographic flows have been largely absent from existing spatial analyses due to their non-physical nature and the lack of survey data. We apply our framework to the Chilean economy, whose pronounced geographic heterogeneity provides a natural setting to study how productivity shocks propagate across sectors and locations.
The remainder of the paper is organized as follows. Section 2 describes the construction of the disaggregated economic accounts, combining state-of-the-art administrative records and survey sources and making them consistent with national accounts. Section 3 presents descriptive evidence on the economic composition of Chile and the geographic structure of the production network. Section 4 introduces the quantitative trade model and defines the sufficient statistics used to measure the aggregate effects of location-sector shocks and the geographic, sectoral, and geography-sector exposures that govern their propagation. Section 5 presents the model calibration. Section 6 reports the results from the quantitative exercises and their applications. Section 7 concludes.
* * *
View full text here: https://publications.iadb.org/en/geographic-and-sectoral-impact-productivity
[Category: IADB]
Inter-American Development Bank: 'Are the Vulnerable Non-Poor Different? Heterogeneous Responses to Macroeconomic Fluctuations in Latin America'
WASHINGTON, Sept. 11 (TNSLrpt) -- The Inter-American Development Bank issued the following white paper in August 2026 entitled "Are the Vulnerable Non-Poor Different? Heterogeneous Responses to Macroeconomic Fluctuations in Latin America."
Here are excerpts:
* * *
Abstract
This paper studies the dynamics of vulnerable non-poor households in Latin America, focusing on how they compare with the poor and how both groups respond to macroeconomic fluctuations. Using harmonized household survey microdata from 15 countries over more than three decades (1992-2024), complemented with longitudinal data ... Show Full Article WASHINGTON, Sept. 11 (TNSLrpt) -- The Inter-American Development Bank issued the following white paper in August 2026 entitled "Are the Vulnerable Non-Poor Different? Heterogeneous Responses to Macroeconomic Fluctuations in Latin America." Here are excerpts: * * * Abstract This paper studies the dynamics of vulnerable non-poor households in Latin America, focusing on how they compare with the poor and how both groups respond to macroeconomic fluctuations. Using harmonized household survey microdata from 15 countries over more than three decades (1992-2024), complemented with longitudinal datafor four of the region's largest economies, we examine how the size of socioeconomic groups evolves with long-run economic growth and cyclical fluctuations and whether labor market responses to macroeconomic conditions differ systematically between poor and vulnerable individuals. We find that the share of vulnerable non-poor individuals has risen modestly, reflecting higher inflows from poverty than outflows to richer groups. While labor market outcomes are strongly procyclical for both poor and vulnerable individuals, a striking pattern emerges across countries: In less developed economies, vulnerable individuals experience significantly stronger cyclical changes in labor market outcomes than the poor, whereas the opposite pattern arises in more developed economies, particularly along employment margins. Exploiting longitudinal data, we further show that transitions into and out of vulnerability are driven primarily by changes in labor income, especially through employment and hourly earnings adjustments. Overall, the results highlight that the relationship between vulnerability and macroeconomic fluctuations depends critically on labor market structure, which shapes the adjustment margins available to different groups of workers
Introduction
Over the last three decades, Latin America has experienced a substantial decline in poverty (Chang et al., 2025; Gasparini et al., 2023), yet a large share of the population remains concentrated just above the poverty line. These vulnerable non-poor households are not classified as poor under conventional definitions, but they still face a substantial risk of falling into poverty in the face of adverse shocks. Understanding the dynamics of this group is particularly relevant in a region characterized by recurrent macroeconomic instability, widespread labor market informality, and limited social protection systems (Levy and Schady, 2013; OECD, 2025).
While a large literature has examined poverty dynamics and labor market adjustment over the business cycle, much less is known about how vulnerable non-poor households respond to macroeconomic fluctuations and whether their responses differ systematically from those of the poor. This distinction is potentially important. Although both groups face economic insecurity, they may occupy different positions within the labor market, differ in their attachment to employment margins, and therefore adjust differently to changes in aggregate economic conditions. Understanding these heterogeneous responses is key to characterizing the nature of vulnerability in Latin America and the mechanisms through which macroeconomic fluctuations translate into changes in socioeconomic status.
This paper studies the dynamics of poor and vulnerable non-poor groups in Latin America, with a particular focus on their relationship with macroeconomic fluctuations. To this end, we draw on the Socio-Economic Database for Latin America and the Caribbean (SEDLAC, 2024), a large database of harmonized household surveys developed by CEDLAS and the World Bank, covering 15 Latin American countries over the period 1992-2024. Throughout the paper, we classify individuals into three groups based on daily household per capita income measured in 2017 PPP US dollars: the poor, with incomes below USD 6.85; the vulnerable non-poor, with incomes between USD 6.85 and USD 14; and the non-vulnerable, with incomes above USD 14.
The analysis combines repeated cross-sectional information for the full sample of countries with longitudinal data for four of the region's largest economies--Argentina, Brazil, Chile, and Peru-- allowing us to examine both aggregate changes in the size of socioeconomic groups and individual labor market responses over the business cycle.
We first examine how the size of these groups evolved over the last three decades and how these changes correlate with long-run economic growth and cyclical economic fluctuations. To distinguish between long-run and cyclical movements, we decompose GDP per capita into trend and cyclical components using a Hodrick-Prescott filter (Hodrick and Prescott, 1997). We document two main patterns. First, the share of vulnerable non-poor individuals increased, driven by a greater incidence of transitions from poverty into vulnerability than from vulnerability into the non-vulnerable. This increase reinforces the view of vulnerability as a quantitatively important and persistent socioeconomic condition in Latin America. Second, while long-run economic growth is associated with a moderate expansion of the vulnerable group, negative cyclical fluctuations tend to increase its size, suggesting that macroeconomic conditions induce substantial movements across socioeconomic groups. These patterns are consistent with the idea that the vulnerable non-poor function as a transition group within the region's income distribution.
We then examine whether labor market outcomes of the poor and the vulnerable non-poor respond differently to macroeconomic conditions. A central empirical challenge is that labor market outcomes and socioeconomic status may be jointly determined by contemporaneous economic conditions. To address this issue, we implement two complementary strategies. First, for a subset of countries with longitudinal data, we classify individuals into socioeconomic groups based on lagged household income and examine how subsequent labor market outcomes evolve with macroeconomic conditions. Second, using repeated cross sections for the full sample of countries, we classify individuals according to predicted household income constructed from predetermined characteristics. Both approaches yield similar patterns. Across all countries, labor market outcomes are strongly procyclical for both groups: Periods of above-trend economic activity are associated with higher employment, lower unemployment and informality, and higher wages and earnings. However, the relative magnitude of these responses differs systematically across countries with different labor market structures.
In economies with higher levels of poverty, informality, and rurality, vulnerable individuals tend to experience stronger labor market improvements during economic expansions than the poor.
In contrast, in more developed economies, the poor often exhibit larger cyclical responses, particularly along employment margins. These patterns suggest that the relationship between vulnerability and macroeconomic fluctuations depends critically on the structure of labor markets. In less developed economies, poor individuals are more likely to be concentrated in low-productivity activities with limited scope for cyclical adjustment, while vulnerable individuals appear better positioned to benefit from economic expansions. By contrast, in more developed economies, poor individuals exhibit stronger attachment to the employment and unemployment margins, making their labor market outcomes more responsive to cyclical fluctuations.
Finally, exploiting the longitudinal data available for a subset of countries, we examine transitions into and out of vulnerability, thereby linking the cyclical labor market responses with changes in socioeconomic status. We document substantial movement into and out of vulnerability over the economic cycle and show that transitions are driven primarily by changes in labor income rather than other income sources or household composition. We further show that, within labor income, both the employment and earnings margins matter, with changes in hourly labor income playing a particularly important role in transitions into and out of vulnerability.
This paper relates to several strands of the literature. First, it contributes to work on vulnerability and the vulnerable non-poor in developing countries, which emphasizes that a large share of households above the poverty line remain exposed to adverse shocks and cannot be considered economically secure (Banerjee and Duflo, 2008; Birdsall et al., 2014; Lopez-Calva and Ortiz-Juarez, 2014; OECD, 2019). Second, it relates to the literature examining labor market adjustment over the business cycle in Latin America, particularly the role of the informality and employment margins in shaping responses to macroeconomic fluctuations (Bosch and Maloney, 2010; Fernandez and Meza, 2015; Coskun, 2022). Our contribution is to bring these strands of the literature together by examining how labor market responses to cyclical fluctuations differ between poor and vulnerable non-poor individuals, how these differences vary across labor market structures in the region, and how they translate into transitions across socioeconomic groups over the cycle.
The rest of the paper is organized as follows. Section 2 describes the main data sources and the construction of the harmonized database used throughout the analysis. Section 3 discusses the definition and operationalization of the vulnerable non-poor group and presents its main demographic, educational, and labor market characteristics. Section 4 examines the evolution of socioeconomic groups and their relationship with long-run economic growth and cyclical fluctuations. Section 5 analyzes how labor market outcomes of poor and vulnerable individuals respond to macroeconomic conditions and whether these responses differ systematically across socioeconomic groups and countries. Section 6 exploits longitudinal data to study transitions across socioeconomic groups over the economic cycle and the labor market mechanisms underlying those transitions. Section 7 concludes.
* * *
View full text here: https://publications.iadb.org/en/are-vulnerable-non-poor-different-heterogeneous-responses-macroeconomic-fluctuations-latin-america
[Category: IADB]
Here are excerpts:
* * *
Abstract
This paper studies the dynamics of vulnerable non-poor households in Latin America, focusing on how they compare with the poor and how both groups respond to macroeconomic fluctuations. Using harmonized household survey microdata from 15 countries over more than three decades (1992-2024), complemented with longitudinal data ... Show Full Article WASHINGTON, Sept. 11 (TNSLrpt) -- The Inter-American Development Bank issued the following white paper in August 2026 entitled "Are the Vulnerable Non-Poor Different? Heterogeneous Responses to Macroeconomic Fluctuations in Latin America." Here are excerpts: * * * Abstract This paper studies the dynamics of vulnerable non-poor households in Latin America, focusing on how they compare with the poor and how both groups respond to macroeconomic fluctuations. Using harmonized household survey microdata from 15 countries over more than three decades (1992-2024), complemented with longitudinal datafor four of the region's largest economies, we examine how the size of socioeconomic groups evolves with long-run economic growth and cyclical fluctuations and whether labor market responses to macroeconomic conditions differ systematically between poor and vulnerable individuals. We find that the share of vulnerable non-poor individuals has risen modestly, reflecting higher inflows from poverty than outflows to richer groups. While labor market outcomes are strongly procyclical for both poor and vulnerable individuals, a striking pattern emerges across countries: In less developed economies, vulnerable individuals experience significantly stronger cyclical changes in labor market outcomes than the poor, whereas the opposite pattern arises in more developed economies, particularly along employment margins. Exploiting longitudinal data, we further show that transitions into and out of vulnerability are driven primarily by changes in labor income, especially through employment and hourly earnings adjustments. Overall, the results highlight that the relationship between vulnerability and macroeconomic fluctuations depends critically on labor market structure, which shapes the adjustment margins available to different groups of workers
Introduction
Over the last three decades, Latin America has experienced a substantial decline in poverty (Chang et al., 2025; Gasparini et al., 2023), yet a large share of the population remains concentrated just above the poverty line. These vulnerable non-poor households are not classified as poor under conventional definitions, but they still face a substantial risk of falling into poverty in the face of adverse shocks. Understanding the dynamics of this group is particularly relevant in a region characterized by recurrent macroeconomic instability, widespread labor market informality, and limited social protection systems (Levy and Schady, 2013; OECD, 2025).
While a large literature has examined poverty dynamics and labor market adjustment over the business cycle, much less is known about how vulnerable non-poor households respond to macroeconomic fluctuations and whether their responses differ systematically from those of the poor. This distinction is potentially important. Although both groups face economic insecurity, they may occupy different positions within the labor market, differ in their attachment to employment margins, and therefore adjust differently to changes in aggregate economic conditions. Understanding these heterogeneous responses is key to characterizing the nature of vulnerability in Latin America and the mechanisms through which macroeconomic fluctuations translate into changes in socioeconomic status.
This paper studies the dynamics of poor and vulnerable non-poor groups in Latin America, with a particular focus on their relationship with macroeconomic fluctuations. To this end, we draw on the Socio-Economic Database for Latin America and the Caribbean (SEDLAC, 2024), a large database of harmonized household surveys developed by CEDLAS and the World Bank, covering 15 Latin American countries over the period 1992-2024. Throughout the paper, we classify individuals into three groups based on daily household per capita income measured in 2017 PPP US dollars: the poor, with incomes below USD 6.85; the vulnerable non-poor, with incomes between USD 6.85 and USD 14; and the non-vulnerable, with incomes above USD 14.
The analysis combines repeated cross-sectional information for the full sample of countries with longitudinal data for four of the region's largest economies--Argentina, Brazil, Chile, and Peru-- allowing us to examine both aggregate changes in the size of socioeconomic groups and individual labor market responses over the business cycle.
We first examine how the size of these groups evolved over the last three decades and how these changes correlate with long-run economic growth and cyclical economic fluctuations. To distinguish between long-run and cyclical movements, we decompose GDP per capita into trend and cyclical components using a Hodrick-Prescott filter (Hodrick and Prescott, 1997). We document two main patterns. First, the share of vulnerable non-poor individuals increased, driven by a greater incidence of transitions from poverty into vulnerability than from vulnerability into the non-vulnerable. This increase reinforces the view of vulnerability as a quantitatively important and persistent socioeconomic condition in Latin America. Second, while long-run economic growth is associated with a moderate expansion of the vulnerable group, negative cyclical fluctuations tend to increase its size, suggesting that macroeconomic conditions induce substantial movements across socioeconomic groups. These patterns are consistent with the idea that the vulnerable non-poor function as a transition group within the region's income distribution.
We then examine whether labor market outcomes of the poor and the vulnerable non-poor respond differently to macroeconomic conditions. A central empirical challenge is that labor market outcomes and socioeconomic status may be jointly determined by contemporaneous economic conditions. To address this issue, we implement two complementary strategies. First, for a subset of countries with longitudinal data, we classify individuals into socioeconomic groups based on lagged household income and examine how subsequent labor market outcomes evolve with macroeconomic conditions. Second, using repeated cross sections for the full sample of countries, we classify individuals according to predicted household income constructed from predetermined characteristics. Both approaches yield similar patterns. Across all countries, labor market outcomes are strongly procyclical for both groups: Periods of above-trend economic activity are associated with higher employment, lower unemployment and informality, and higher wages and earnings. However, the relative magnitude of these responses differs systematically across countries with different labor market structures.
In economies with higher levels of poverty, informality, and rurality, vulnerable individuals tend to experience stronger labor market improvements during economic expansions than the poor.
In contrast, in more developed economies, the poor often exhibit larger cyclical responses, particularly along employment margins. These patterns suggest that the relationship between vulnerability and macroeconomic fluctuations depends critically on the structure of labor markets. In less developed economies, poor individuals are more likely to be concentrated in low-productivity activities with limited scope for cyclical adjustment, while vulnerable individuals appear better positioned to benefit from economic expansions. By contrast, in more developed economies, poor individuals exhibit stronger attachment to the employment and unemployment margins, making their labor market outcomes more responsive to cyclical fluctuations.
Finally, exploiting the longitudinal data available for a subset of countries, we examine transitions into and out of vulnerability, thereby linking the cyclical labor market responses with changes in socioeconomic status. We document substantial movement into and out of vulnerability over the economic cycle and show that transitions are driven primarily by changes in labor income rather than other income sources or household composition. We further show that, within labor income, both the employment and earnings margins matter, with changes in hourly labor income playing a particularly important role in transitions into and out of vulnerability.
This paper relates to several strands of the literature. First, it contributes to work on vulnerability and the vulnerable non-poor in developing countries, which emphasizes that a large share of households above the poverty line remain exposed to adverse shocks and cannot be considered economically secure (Banerjee and Duflo, 2008; Birdsall et al., 2014; Lopez-Calva and Ortiz-Juarez, 2014; OECD, 2019). Second, it relates to the literature examining labor market adjustment over the business cycle in Latin America, particularly the role of the informality and employment margins in shaping responses to macroeconomic fluctuations (Bosch and Maloney, 2010; Fernandez and Meza, 2015; Coskun, 2022). Our contribution is to bring these strands of the literature together by examining how labor market responses to cyclical fluctuations differ between poor and vulnerable non-poor individuals, how these differences vary across labor market structures in the region, and how they translate into transitions across socioeconomic groups over the cycle.
The rest of the paper is organized as follows. Section 2 describes the main data sources and the construction of the harmonized database used throughout the analysis. Section 3 discusses the definition and operationalization of the vulnerable non-poor group and presents its main demographic, educational, and labor market characteristics. Section 4 examines the evolution of socioeconomic groups and their relationship with long-run economic growth and cyclical fluctuations. Section 5 analyzes how labor market outcomes of poor and vulnerable individuals respond to macroeconomic conditions and whether these responses differ systematically across socioeconomic groups and countries. Section 6 exploits longitudinal data to study transitions across socioeconomic groups over the economic cycle and the labor market mechanisms underlying those transitions. Section 7 concludes.
* * *
View full text here: https://publications.iadb.org/en/are-vulnerable-non-poor-different-heterogeneous-responses-macroeconomic-fluctuations-latin-america
[Category: IADB]
Inter-American Development Bank: 'AI and Judicial Productivity: The Impact of MIDAS on the Courts of Fortaleza, Brazil'
WASHINGTON, Sept. 11 (TNSLrpt) -- The Inter-American Development Bank issued the following white paper in August 2026 entitled "AI and Judicial Productivity: The Impact of MIDAS on the Courts of Fortaleza, Brazil."
Here are excerpts:
* * *
Abstract
This paper presents preliminary results from a pilot study conducted in the courts of Ceara, Brazil. The study evaluates the impact of introducing a tool that uses natural language processing and machine learning techniques to cluster judicial acts by textual similarity on clerk productivity, measured as the number of case files a clerk can produce ... Show Full Article WASHINGTON, Sept. 11 (TNSLrpt) -- The Inter-American Development Bank issued the following white paper in August 2026 entitled "AI and Judicial Productivity: The Impact of MIDAS on the Courts of Fortaleza, Brazil." Here are excerpts: * * * Abstract This paper presents preliminary results from a pilot study conducted in the courts of Ceara, Brazil. The study evaluates the impact of introducing a tool that uses natural language processing and machine learning techniques to cluster judicial acts by textual similarity on clerk productivity, measured as the number of case files a clerk can producein a day.
Estimates indicate that treatment-group clerks produced approximately 10 more case files per day than control-group clerks, a statistically significant difference equivalent to a 37% increase relative to the control group mean. The results are robust to the exclusion of outlier observations and exceptionally productive clerks.
* * *
Introduction
Estimates from 2019 indicated that around 1.5 billion individuals were unable to resolve their legal problems, despite living in contexts with a functioning justice system and institutions (World Justice Project, 2019). In Brazil, the State Justice system closed 2025 with nearly 58 million pending judicial processes, despite having adjudicated and closed more than 30 million during that year1. In the state of Ceara alone, the state Court of Justice (TJCE) received 692,630 new judicial processes and ended the year with more than one million pending.2
These backlogs are not merely an administrative problem, as the efficiency of the judicial system has significant economic repercussions. The literature shows that delays in resolving judicial proceedings hamper contract enforcement and insolvency resolution, generating a negative effect on firms' credit and investment (Djankov et al., 2008; Ponticelli and Alencar, 2016; Visaria, 2009; Chemin, 2012). Similarly, reductions in judicial processing times have been associated with greater entrepreneurship (Chemin, 2009), firm performance (Chakraborty, 2016), and the country's economic growth and development (Amirapu, 2021; Djankov et al., 2025). The need to increase judicial productivity becomes even more relevant in a context in which artificial intelligence tools are also reducing the costs of access to justice and potentially increasing the volume of litigation faced by courts (Shah and Levy, 2026).
This paper analyzes whether providing judicial clerks with a tool based on natural language processing (NLP) that clusters similar rulings, allowing them to process cases in batches rather than one by one, increases their daily productivity relative to the traditional workflow. To this end, we carried out a pilot study to examine the consequences of introducing the MIDAS system (Mecanismo Identificador de Actos Similares -- Similar Acts Identification Mechanism) at the First-Degree Judicial Secretariat (SEJUD) of the Court of Justice of Ceara (TJCE), Brazil3 .
MIDAS is a tool that uses NLP and machine learning techniques to cluster judicial rulings by textual similarity. Whereas in the traditional workflow clerks must prepare each case file one by one, with MIDAS clerks can prepare multiple case files at once, thereby reducing the average time per case and increasing total output volume. This increase in productivity should, in principle, contribute to reducing the backlog, to the extent that it does not simply shift the bottleneck to another stage of the judicial process.
The study was conducted between June and August 2025 within the framework of the Modernization Program for the Judiciary of the State of Ceara (PROMOJUD), with support from the Inter-American Development Bank (IDB). The study compared the daily productivity of 62 clerks assigned to two groups: a treatment group that processed case files with the support of MIDAS, and a control group that followed the traditional workflow. The central element of the design was randomization at the level of the ruling or judicial act: within each cluster generated by the system, rulings were randomly reordered using a Python script, and a fraction was extracted to be processed individually by the control group. This mechanism ensures that the judicial acts assigned to both groups come from the same pool of rulings and are comparable in their observable and unobservable characteristics in expectation.
The results indicate that treatment-group clerks produced on average 10 more case files per day than control-group clerks, a statistically significant difference equivalent to a 37% increase relative to the control group mean. Furthermore, these results are robust to the exclusion of potentially outlying observations, which reinforces the robustness of the conclusions. In terms of workload, our estimates suggest that access to the tool reduces case file preparation time from approximately 16 to 12 minutes per case file, a gain of 4 minutes per case that, accumulated over the workday, allows clerks to process one-third more case files in the same amount of time.
Furthermore, the tool appears to expand the right tail of the productivity distribution: days on which a clerk analyzes more than 100 case files are notably more frequent among those who used MIDAS. However, since the treatment group included both clerks with full adherence to the system and clerks with partial exposure, the estimated effects should be interpreted as a lower bound on the tool's true impact.
This paper contributes to the existing literature on the use of artificial intelligence to improve efficiency in the judicial sector (Aidid and Alarie, 2023; Casey and Niblett, 2019; Volokh, 2019).
Prior studies have shown how these tools can be used to automate routine tasks, such as analyzing and processing evidence, conducting investigations, and classifying legal documents (de Oliveira and Nascimento, 2021; Oliveira and Sperandio Nascimento, 2025; Razmetaeva and Razmetaev, 2021; Solovey et al., 2025), allowing court staff to devote more time to other activities (Alarie et al., 2018; Borgesano et al., 2025). The literature also highlights their potential to assist in judicial decision-making and, in some cases, to partially substitute judges' functions (Chen et al., 2022; Volokh, 2019). Some countries already report productivity gains derived from their use, including Argentina and Brazil (de Sousa et al., 2022; OECD, 2025). Our study contributes to this growing body of research by being, to our knowledge, the first to estimate the impact of an NLP-based clustering tool on judicial staff productivity through a pilot study with randomization at the level of the judicial act.
The remainder of the paper is organized as follows: Section 2 describes how the intervention works, the MIDAS system, and its position within the Judicial Secretariat's workflow. Section 3 then presents the study design, while Section 4 describes the data used. Section 5 presents the preliminary results of the pilot study, as well as the lessons that should be considered in future studies. Finally, Section 6 discusses the study's main methodological limitations and concludes the paper.
* * *
View full text here: https://publications.iadb.org/en/ai-and-judicial-productivity-impact-midas-courts-fortaleza-brazil
[Category: IADB]
Here are excerpts:
* * *
Abstract
This paper presents preliminary results from a pilot study conducted in the courts of Ceara, Brazil. The study evaluates the impact of introducing a tool that uses natural language processing and machine learning techniques to cluster judicial acts by textual similarity on clerk productivity, measured as the number of case files a clerk can produce ... Show Full Article WASHINGTON, Sept. 11 (TNSLrpt) -- The Inter-American Development Bank issued the following white paper in August 2026 entitled "AI and Judicial Productivity: The Impact of MIDAS on the Courts of Fortaleza, Brazil." Here are excerpts: * * * Abstract This paper presents preliminary results from a pilot study conducted in the courts of Ceara, Brazil. The study evaluates the impact of introducing a tool that uses natural language processing and machine learning techniques to cluster judicial acts by textual similarity on clerk productivity, measured as the number of case files a clerk can producein a day.
Estimates indicate that treatment-group clerks produced approximately 10 more case files per day than control-group clerks, a statistically significant difference equivalent to a 37% increase relative to the control group mean. The results are robust to the exclusion of outlier observations and exceptionally productive clerks.
* * *
Introduction
Estimates from 2019 indicated that around 1.5 billion individuals were unable to resolve their legal problems, despite living in contexts with a functioning justice system and institutions (World Justice Project, 2019). In Brazil, the State Justice system closed 2025 with nearly 58 million pending judicial processes, despite having adjudicated and closed more than 30 million during that year1. In the state of Ceara alone, the state Court of Justice (TJCE) received 692,630 new judicial processes and ended the year with more than one million pending.2
These backlogs are not merely an administrative problem, as the efficiency of the judicial system has significant economic repercussions. The literature shows that delays in resolving judicial proceedings hamper contract enforcement and insolvency resolution, generating a negative effect on firms' credit and investment (Djankov et al., 2008; Ponticelli and Alencar, 2016; Visaria, 2009; Chemin, 2012). Similarly, reductions in judicial processing times have been associated with greater entrepreneurship (Chemin, 2009), firm performance (Chakraborty, 2016), and the country's economic growth and development (Amirapu, 2021; Djankov et al., 2025). The need to increase judicial productivity becomes even more relevant in a context in which artificial intelligence tools are also reducing the costs of access to justice and potentially increasing the volume of litigation faced by courts (Shah and Levy, 2026).
This paper analyzes whether providing judicial clerks with a tool based on natural language processing (NLP) that clusters similar rulings, allowing them to process cases in batches rather than one by one, increases their daily productivity relative to the traditional workflow. To this end, we carried out a pilot study to examine the consequences of introducing the MIDAS system (Mecanismo Identificador de Actos Similares -- Similar Acts Identification Mechanism) at the First-Degree Judicial Secretariat (SEJUD) of the Court of Justice of Ceara (TJCE), Brazil3 .
MIDAS is a tool that uses NLP and machine learning techniques to cluster judicial rulings by textual similarity. Whereas in the traditional workflow clerks must prepare each case file one by one, with MIDAS clerks can prepare multiple case files at once, thereby reducing the average time per case and increasing total output volume. This increase in productivity should, in principle, contribute to reducing the backlog, to the extent that it does not simply shift the bottleneck to another stage of the judicial process.
The study was conducted between June and August 2025 within the framework of the Modernization Program for the Judiciary of the State of Ceara (PROMOJUD), with support from the Inter-American Development Bank (IDB). The study compared the daily productivity of 62 clerks assigned to two groups: a treatment group that processed case files with the support of MIDAS, and a control group that followed the traditional workflow. The central element of the design was randomization at the level of the ruling or judicial act: within each cluster generated by the system, rulings were randomly reordered using a Python script, and a fraction was extracted to be processed individually by the control group. This mechanism ensures that the judicial acts assigned to both groups come from the same pool of rulings and are comparable in their observable and unobservable characteristics in expectation.
The results indicate that treatment-group clerks produced on average 10 more case files per day than control-group clerks, a statistically significant difference equivalent to a 37% increase relative to the control group mean. Furthermore, these results are robust to the exclusion of potentially outlying observations, which reinforces the robustness of the conclusions. In terms of workload, our estimates suggest that access to the tool reduces case file preparation time from approximately 16 to 12 minutes per case file, a gain of 4 minutes per case that, accumulated over the workday, allows clerks to process one-third more case files in the same amount of time.
Furthermore, the tool appears to expand the right tail of the productivity distribution: days on which a clerk analyzes more than 100 case files are notably more frequent among those who used MIDAS. However, since the treatment group included both clerks with full adherence to the system and clerks with partial exposure, the estimated effects should be interpreted as a lower bound on the tool's true impact.
This paper contributes to the existing literature on the use of artificial intelligence to improve efficiency in the judicial sector (Aidid and Alarie, 2023; Casey and Niblett, 2019; Volokh, 2019).
Prior studies have shown how these tools can be used to automate routine tasks, such as analyzing and processing evidence, conducting investigations, and classifying legal documents (de Oliveira and Nascimento, 2021; Oliveira and Sperandio Nascimento, 2025; Razmetaeva and Razmetaev, 2021; Solovey et al., 2025), allowing court staff to devote more time to other activities (Alarie et al., 2018; Borgesano et al., 2025). The literature also highlights their potential to assist in judicial decision-making and, in some cases, to partially substitute judges' functions (Chen et al., 2022; Volokh, 2019). Some countries already report productivity gains derived from their use, including Argentina and Brazil (de Sousa et al., 2022; OECD, 2025). Our study contributes to this growing body of research by being, to our knowledge, the first to estimate the impact of an NLP-based clustering tool on judicial staff productivity through a pilot study with randomization at the level of the judicial act.
The remainder of the paper is organized as follows: Section 2 describes how the intervention works, the MIDAS system, and its position within the Judicial Secretariat's workflow. Section 3 then presents the study design, while Section 4 describes the data used. Section 5 presents the preliminary results of the pilot study, as well as the lessons that should be considered in future studies. Finally, Section 6 discusses the study's main methodological limitations and concludes the paper.
* * *
View full text here: https://publications.iadb.org/en/ai-and-judicial-productivity-impact-midas-courts-fortaleza-brazil
[Category: IADB]
IDB Lab Announces Winners of WeXchange Women STEMpreneurs 2026
WASHINGTON, Sept. 11 -- The Inter-American Development Bank issued the following news release on Sept. 10, 2026:
* * *
IDB Lab Announces Winners of WeXchange Women STEMpreneurs 2026
Five selected startups will present their solutions to investors in October during the IDB Group's Global Entrepreneurship and Technology Forum for Latin America and the Caribbean - GET Forum.
WASHINGTON -- IDB Lab, the innovation and venture arm of the Inter-American Development Bank Group (IDB Group), today announced the five women-led businesses selected as winners of the 2026 WeXchange Women STEMpreneurs competition. ... Show Full Article WASHINGTON, Sept. 11 -- The Inter-American Development Bank issued the following news release on Sept. 10, 2026: * * * IDB Lab Announces Winners of WeXchange Women STEMpreneurs 2026 Five selected startups will present their solutions to investors in October during the IDB Group's Global Entrepreneurship and Technology Forum for Latin America and the Caribbean - GET Forum. WASHINGTON -- IDB Lab, the innovation and venture arm of the Inter-American Development Bank Group (IDB Group), today announced the five women-led businesses selected as winners of the 2026 WeXchange Women STEMpreneurs competition.WeXchange is a platform that connects women entrepreneurs in science, technology, engineering, and mathematics (STEM) across Latin America and the Caribbean with investors and growth opportunities.
The 2026 edition received more than 350 applications from 35 countries, featuring solutions that promote economic or social development through edtech, healthtech, fintech, software as a service (SaaS), biotech, climatetech, agtech, and e-commerce, among other sectors. Colombia-based venture firm, EWA Capital led the selection process, which involved more than 80 judges from over 60 organizations, including 50 investment funds from across Latin America and the Caribbean.
The selected women-led startups are currently raising capital or planning to do so in the coming months. Their solutions will be showcased to venture capital investors and other innovation ecosystem stakeholders from Latin America and the Caribbean during WeXchange Demo Day 2026, which will take place on October 7 in Quito, Ecuador, as part of GET Forum.
The selected startups are:
Kresko RNAtech (Argentina): A biotechnology company developing a new generation of functional ingredients based on bioactive ribonucleic acid (RNA). The company combines artificial intelligence tools, a proprietary stabilization method, and biological validation to identify and develop compounds found in foods and plants with applications in health, nutrition, and beauty.
GOW Credit (Mexico): A technology company developing AI-powered infrastructure to finance small and medium-sized enterprises (SMEs) in Latin America. Its technology enables financial institutions to identify creditworthy SMEs that are often excluded from traditional approval processes, creating new financing opportunities aligned with their risk policies.
Ciudata (Bolivia): A business intelligence platform that supports decision-making for companies and public-sector entities. Using vehicles equipped with cameras and computer vision technology, the company collects territorial information and transforms it into data on commercial establishments and economic activity.
Pilou (Mexico): A wealth management platform that enables individuals and businesses across Latin America to invest in the United States and Mexico. It currently serves more than 1,500 clients in 17 countries.
Salva Health (Colombia): A technology company that developed Julieta, a portable, fast, and noninvasive medical device that facilitates access to breast cancer screening. The solution aims to expand access to early detection tools, particularly in communities with limited access to these services.
In addition to presenting their solutions during WeXchange Demo Day, the winning entrepreneurs will participate in networking activities with investors and other stakeholders in the entrepreneurship and innovation ecosystem during GET Forum, expanding their opportunities to connect with networks across Latin America and the Caribbean.
WeXchange's activities are made possible by the support of the Women Entrepreneurs Finance Initiative (We-Fi).
* * *
About IDB Lab
IDB Lab is the innovation and venture arm of the Inter-American Development Bank Group. We focus on entrepreneurship and technology to co-create solutions to development challenges and activate new industries for growth in Latin America and the Caribbean. IDB Lab supports high-impact startups and the innovation ecosystems they need to thrive through flexible financing, practical knowledge, and global connections. www.idblab.org
* * *
About EWA Capital
EWA Capital is a gender-lens venture capital fund focused on backing early-stage technology companies that are building the future of Latin America. It invests in scalable and disruptive business models that drive progress, accelerate regional transformation, and generate sustainable impact. EWA Capital supports companies from the seed stage through Series A. www.ewa.capital
* * *
About We-Fi
The Women Entrepreneurs Finance Initiative (We-Fi) is an unprecedented partnership dedicated to mobilizing funding for women-owned and women-led businesses in developing countries. We-Fi includes 14 governments and six multilateral development banks as implementing partners, as well as public- and private-sector collaborators from around the world. www.we-fi.org
* * *
Original text here: https://www.iadb.org/en/news/idb-lab-announces-winners-wexchange-women-stempreneurs-2026
* * *
IDB Lab Announces Winners of WeXchange Women STEMpreneurs 2026
Five selected startups will present their solutions to investors in October during the IDB Group's Global Entrepreneurship and Technology Forum for Latin America and the Caribbean - GET Forum.
WASHINGTON -- IDB Lab, the innovation and venture arm of the Inter-American Development Bank Group (IDB Group), today announced the five women-led businesses selected as winners of the 2026 WeXchange Women STEMpreneurs competition. ... Show Full Article WASHINGTON, Sept. 11 -- The Inter-American Development Bank issued the following news release on Sept. 10, 2026: * * * IDB Lab Announces Winners of WeXchange Women STEMpreneurs 2026 Five selected startups will present their solutions to investors in October during the IDB Group's Global Entrepreneurship and Technology Forum for Latin America and the Caribbean - GET Forum. WASHINGTON -- IDB Lab, the innovation and venture arm of the Inter-American Development Bank Group (IDB Group), today announced the five women-led businesses selected as winners of the 2026 WeXchange Women STEMpreneurs competition.WeXchange is a platform that connects women entrepreneurs in science, technology, engineering, and mathematics (STEM) across Latin America and the Caribbean with investors and growth opportunities.
The 2026 edition received more than 350 applications from 35 countries, featuring solutions that promote economic or social development through edtech, healthtech, fintech, software as a service (SaaS), biotech, climatetech, agtech, and e-commerce, among other sectors. Colombia-based venture firm, EWA Capital led the selection process, which involved more than 80 judges from over 60 organizations, including 50 investment funds from across Latin America and the Caribbean.
The selected women-led startups are currently raising capital or planning to do so in the coming months. Their solutions will be showcased to venture capital investors and other innovation ecosystem stakeholders from Latin America and the Caribbean during WeXchange Demo Day 2026, which will take place on October 7 in Quito, Ecuador, as part of GET Forum.
The selected startups are:
Kresko RNAtech (Argentina): A biotechnology company developing a new generation of functional ingredients based on bioactive ribonucleic acid (RNA). The company combines artificial intelligence tools, a proprietary stabilization method, and biological validation to identify and develop compounds found in foods and plants with applications in health, nutrition, and beauty.
GOW Credit (Mexico): A technology company developing AI-powered infrastructure to finance small and medium-sized enterprises (SMEs) in Latin America. Its technology enables financial institutions to identify creditworthy SMEs that are often excluded from traditional approval processes, creating new financing opportunities aligned with their risk policies.
Ciudata (Bolivia): A business intelligence platform that supports decision-making for companies and public-sector entities. Using vehicles equipped with cameras and computer vision technology, the company collects territorial information and transforms it into data on commercial establishments and economic activity.
Pilou (Mexico): A wealth management platform that enables individuals and businesses across Latin America to invest in the United States and Mexico. It currently serves more than 1,500 clients in 17 countries.
Salva Health (Colombia): A technology company that developed Julieta, a portable, fast, and noninvasive medical device that facilitates access to breast cancer screening. The solution aims to expand access to early detection tools, particularly in communities with limited access to these services.
In addition to presenting their solutions during WeXchange Demo Day, the winning entrepreneurs will participate in networking activities with investors and other stakeholders in the entrepreneurship and innovation ecosystem during GET Forum, expanding their opportunities to connect with networks across Latin America and the Caribbean.
WeXchange's activities are made possible by the support of the Women Entrepreneurs Finance Initiative (We-Fi).
* * *
About IDB Lab
IDB Lab is the innovation and venture arm of the Inter-American Development Bank Group. We focus on entrepreneurship and technology to co-create solutions to development challenges and activate new industries for growth in Latin America and the Caribbean. IDB Lab supports high-impact startups and the innovation ecosystems they need to thrive through flexible financing, practical knowledge, and global connections. www.idblab.org
* * *
About EWA Capital
EWA Capital is a gender-lens venture capital fund focused on backing early-stage technology companies that are building the future of Latin America. It invests in scalable and disruptive business models that drive progress, accelerate regional transformation, and generate sustainable impact. EWA Capital supports companies from the seed stage through Series A. www.ewa.capital
* * *
About We-Fi
The Women Entrepreneurs Finance Initiative (We-Fi) is an unprecedented partnership dedicated to mobilizing funding for women-owned and women-led businesses in developing countries. We-Fi includes 14 governments and six multilateral development banks as implementing partners, as well as public- and private-sector collaborators from around the world. www.we-fi.org
* * *
Original text here: https://www.iadb.org/en/news/idb-lab-announces-winners-wexchange-women-stempreneurs-2026
FHLBanks Price $1 Billion 2-Year Global on September 10, 2026
WASHINGTON, Sept. 11 -- The Federal Home Loan Banks Office of Finance issued the following news release:
* * *
FHLBanks Price $1 Billion 2-Year Global on September 10, 2026
The FHLBanks have priced the $1 billion 2-year Global as follows:
Leads: Barclays Capital Inc., Deutsche Bank Securities Inc., Wells Fargo Securities, LLC
Distribution Group: 7 Firms
CUSIP: 3130BC4W0
Pricing Date: 9/10/2026
Settlement: 9/11/2026
Maturity: 9/1/2028
Treasury Yield: 4.535%
Spread: 3.0 bps
Yield: 4.565%
Price: 99.880
Coupon: 4.5%
Interest Payments: March 1 and September 1, beginning on March ... Show Full Article WASHINGTON, Sept. 11 -- The Federal Home Loan Banks Office of Finance issued the following news release: * * * FHLBanks Price $1 Billion 2-Year Global on September 10, 2026 The FHLBanks have priced the $1 billion 2-year Global as follows: Leads: Barclays Capital Inc., Deutsche Bank Securities Inc., Wells Fargo Securities, LLC Distribution Group: 7 Firms CUSIP: 3130BC4W0 Pricing Date: 9/10/2026 Settlement: 9/11/2026 Maturity: 9/1/2028 Treasury Yield: 4.535% Spread: 3.0 bps Yield: 4.565% Price: 99.880 Coupon: 4.5% Interest Payments: March 1 and September 1, beginning on March1, 2027 (short first coupon)
Preliminary distribution information will be available on www.fhlb-of.com.
This announcement is neither an offer to sell, nor a solicitation of offers to buy, these securities. Distribution estimates are based on order book composition for syndicated issues furnished to the Office of Finance by underwriters at the time of securities pricing and may not reflect the current distribution of securities. The Office of Finance does not independently validate submitted data.
* * *
Original text here: https://www.fhlb-of.com/news/fhlbanks-price-1-billion-2-year-global-on-september-10-2026/
* * *
FHLBanks Price $1 Billion 2-Year Global on September 10, 2026
The FHLBanks have priced the $1 billion 2-year Global as follows:
Leads: Barclays Capital Inc., Deutsche Bank Securities Inc., Wells Fargo Securities, LLC
Distribution Group: 7 Firms
CUSIP: 3130BC4W0
Pricing Date: 9/10/2026
Settlement: 9/11/2026
Maturity: 9/1/2028
Treasury Yield: 4.535%
Spread: 3.0 bps
Yield: 4.565%
Price: 99.880
Coupon: 4.5%
Interest Payments: March 1 and September 1, beginning on March ... Show Full Article WASHINGTON, Sept. 11 -- The Federal Home Loan Banks Office of Finance issued the following news release: * * * FHLBanks Price $1 Billion 2-Year Global on September 10, 2026 The FHLBanks have priced the $1 billion 2-year Global as follows: Leads: Barclays Capital Inc., Deutsche Bank Securities Inc., Wells Fargo Securities, LLC Distribution Group: 7 Firms CUSIP: 3130BC4W0 Pricing Date: 9/10/2026 Settlement: 9/11/2026 Maturity: 9/1/2028 Treasury Yield: 4.535% Spread: 3.0 bps Yield: 4.565% Price: 99.880 Coupon: 4.5% Interest Payments: March 1 and September 1, beginning on March1, 2027 (short first coupon)
Preliminary distribution information will be available on www.fhlb-of.com.
This announcement is neither an offer to sell, nor a solicitation of offers to buy, these securities. Distribution estimates are based on order book composition for syndicated issues furnished to the Office of Finance by underwriters at the time of securities pricing and may not reflect the current distribution of securities. The Office of Finance does not independently validate submitted data.
* * *
Original text here: https://www.fhlb-of.com/news/fhlbanks-price-1-billion-2-year-global-on-september-10-2026/
EPA begins cleanup at Rumsey Mill in Montana
WASHINGTON, Sept. 11 -- The Environmental Protection Agency issued the following news release:
* * *
EPA begins cleanup at Rumsey Mill in Montana
*
HELENA, Mont. - Today, U.S. Environmental Protection Agency (EPA) is announcing the start of a time-critical removal action at the Rumsey Mill site in Philipsburg, Montana, to address heavy metals contamination from historical mining and milling activities. The cleanup will reduce the risk of contaminated material spreading.
"The cleanup at Rumsey Mill demonstrates how targeted action can make a real difference for communities impacted by contamination," ... Show Full Article WASHINGTON, Sept. 11 -- The Environmental Protection Agency issued the following news release: * * * EPA begins cleanup at Rumsey Mill in Montana * HELENA, Mont. - Today, U.S. Environmental Protection Agency (EPA) is announcing the start of a time-critical removal action at the Rumsey Mill site in Philipsburg, Montana, to address heavy metals contamination from historical mining and milling activities. The cleanup will reduce the risk of contaminated material spreading. "The cleanup at Rumsey Mill demonstrates how targeted action can make a real difference for communities impacted by contamination,"said EPA Regional Administrator Cyrus Western. "EPA is proud to begin this important cleanup at Rumsey Mill and to work alongside local partners to address historic mining contamination, reducing human health risks today and improving conditions for the future."
"Trout Unlimited (TU) has worked with representatives of Granite County for years to study metals contamination on Fred Burr Creek and has been invested in fisheries restoration and reclamation in the Upper Clark Fork River watershed for more than two decades," said Trout Unlimited Senior Project Manager Rob Roberts. "This cleanup project at the Rumsey Mill site is an important first step for Fred Burr Creek and should help reduce the level of mercury and other metals at the source. TU looks forward to seeing these necessary improvements completed by the EPA and working with the local community to evaluate if future projects would further improve water quality, fisheries, and impacts to public health."
EPA is committed to accelerating and completing Superfund cleanup work, as part of the Superfund Solutions initiative. This forward-looking effort renews focus on EPA's core mission of protecting human health and the environment, while getting more sites to the finish line faster. The time-critical removal at the Rumsey Mill site will reduce risks to nearby residents, domestic wells and Fred Burr Creek, while also improving long-term water quality, habitat conditions and the overall stability of the site. EPA's Site Assessment program performed an expanded site inspection in 2021, and additional sampling identified elevated levels of arsenic, lead and mercury at the former mill site as well as in tailings material that had been deposited downstream. EPA determined that contaminated material continues to migrate off site, creating potential risks to human health and the environment.
The cleanup will include excavation of contaminated areas containing elevated levels of arsenic, lead and mercury. The excavated material will be transported to a nearby repository at the former mill site, where it will be capped, armored and protected with erosion controls. EPA will also backfill, stabilize and revegetate excavated areas with native seed to improve drainage and support long-term site stability. Additional work will include mine waste consolidation, surface water controls and restoration of the riparian corridor along Fred Burr Creek.
These actions are intended to improve the health of Fred Burr Creek by substantially reducing the ongoing release of mercury and arsenic into Fred Burr Creek. Additional work will include mine waste consolidation, surface water controls and restoration of the riparian corridor along Fred Burr Creek. Together, these measures are expected to reduce the spread of contamination and improve ecological conditions in and around the watershed.
Portions of the cleanup area will occur along the Fred Burr Creek bank. Final restoration of these areas will be coordinated with Trout Unlimited, a non-profit organization that is committed to mine reclamation, stream restoration and other activities that improve fisheries, water quality and watershed health.
EPA expects cleanup work to be completed by November 30, 2026.
Background
The Rumsey Mill Site, an abandoned mill site in the Philipsburg Mining District, is located within the Fred Burr Creek watershed in Philipsburg, Montana. Historical mining and milling operations associated with the site released metals including lead, arsenic and mercury into the surrounding area.
EPA's Site Assessment program performed an expanded site inspection (ESI) in 2021, and EPA's Removal program completed additional sampling in 2023 and 2024, which identified elevated levels of arsenic, lead and mercury at the former mill site as well as an ongoing release to Fred Burr Creek. Sampling also found that the contaminated material has been migrating downslope into and along Fred Burr Creek. Based on these results, EPA's Removal Program is taking action to secure the ongoing release to protect human health and the environment.
***
Original text here: https://www.epa.gov/newsreleases/epa-begins-cleanup-rumsey-mill-montana
* * *
EPA begins cleanup at Rumsey Mill in Montana
*
HELENA, Mont. - Today, U.S. Environmental Protection Agency (EPA) is announcing the start of a time-critical removal action at the Rumsey Mill site in Philipsburg, Montana, to address heavy metals contamination from historical mining and milling activities. The cleanup will reduce the risk of contaminated material spreading.
"The cleanup at Rumsey Mill demonstrates how targeted action can make a real difference for communities impacted by contamination," ... Show Full Article WASHINGTON, Sept. 11 -- The Environmental Protection Agency issued the following news release: * * * EPA begins cleanup at Rumsey Mill in Montana * HELENA, Mont. - Today, U.S. Environmental Protection Agency (EPA) is announcing the start of a time-critical removal action at the Rumsey Mill site in Philipsburg, Montana, to address heavy metals contamination from historical mining and milling activities. The cleanup will reduce the risk of contaminated material spreading. "The cleanup at Rumsey Mill demonstrates how targeted action can make a real difference for communities impacted by contamination,"said EPA Regional Administrator Cyrus Western. "EPA is proud to begin this important cleanup at Rumsey Mill and to work alongside local partners to address historic mining contamination, reducing human health risks today and improving conditions for the future."
"Trout Unlimited (TU) has worked with representatives of Granite County for years to study metals contamination on Fred Burr Creek and has been invested in fisheries restoration and reclamation in the Upper Clark Fork River watershed for more than two decades," said Trout Unlimited Senior Project Manager Rob Roberts. "This cleanup project at the Rumsey Mill site is an important first step for Fred Burr Creek and should help reduce the level of mercury and other metals at the source. TU looks forward to seeing these necessary improvements completed by the EPA and working with the local community to evaluate if future projects would further improve water quality, fisheries, and impacts to public health."
EPA is committed to accelerating and completing Superfund cleanup work, as part of the Superfund Solutions initiative. This forward-looking effort renews focus on EPA's core mission of protecting human health and the environment, while getting more sites to the finish line faster. The time-critical removal at the Rumsey Mill site will reduce risks to nearby residents, domestic wells and Fred Burr Creek, while also improving long-term water quality, habitat conditions and the overall stability of the site. EPA's Site Assessment program performed an expanded site inspection in 2021, and additional sampling identified elevated levels of arsenic, lead and mercury at the former mill site as well as in tailings material that had been deposited downstream. EPA determined that contaminated material continues to migrate off site, creating potential risks to human health and the environment.
The cleanup will include excavation of contaminated areas containing elevated levels of arsenic, lead and mercury. The excavated material will be transported to a nearby repository at the former mill site, where it will be capped, armored and protected with erosion controls. EPA will also backfill, stabilize and revegetate excavated areas with native seed to improve drainage and support long-term site stability. Additional work will include mine waste consolidation, surface water controls and restoration of the riparian corridor along Fred Burr Creek.
These actions are intended to improve the health of Fred Burr Creek by substantially reducing the ongoing release of mercury and arsenic into Fred Burr Creek. Additional work will include mine waste consolidation, surface water controls and restoration of the riparian corridor along Fred Burr Creek. Together, these measures are expected to reduce the spread of contamination and improve ecological conditions in and around the watershed.
Portions of the cleanup area will occur along the Fred Burr Creek bank. Final restoration of these areas will be coordinated with Trout Unlimited, a non-profit organization that is committed to mine reclamation, stream restoration and other activities that improve fisheries, water quality and watershed health.
EPA expects cleanup work to be completed by November 30, 2026.
Background
The Rumsey Mill Site, an abandoned mill site in the Philipsburg Mining District, is located within the Fred Burr Creek watershed in Philipsburg, Montana. Historical mining and milling operations associated with the site released metals including lead, arsenic and mercury into the surrounding area.
EPA's Site Assessment program performed an expanded site inspection (ESI) in 2021, and EPA's Removal program completed additional sampling in 2023 and 2024, which identified elevated levels of arsenic, lead and mercury at the former mill site as well as an ongoing release to Fred Burr Creek. Sampling also found that the contaminated material has been migrating downslope into and along Fred Burr Creek. Based on these results, EPA's Removal Program is taking action to secure the ongoing release to protect human health and the environment.
***
Original text here: https://www.epa.gov/newsreleases/epa-begins-cleanup-rumsey-mill-montana
EPA Deputy Administrator Fotouhi Advances Cooperative Federalism in Wyoming
WASHINGTON, Sept. 11 -- The Environmental Protection Agency issued the following news release:
* * *
EPA Deputy Administrator Fotouhi Advances Cooperative Federalism in Wyoming
*
Jackson, Wyoming - Last week, U.S. Environmental Protection Agency (EPA) Deputy Administrator David Fotouhi traveled to Wyoming to advance cooperative federalism and highlight the Trump EPA's commitment to achieving measurable environmental results while supporting economic growth. During the visit, Deputy Administrator Fotouhi addressed state environmental leaders at the Fall 2026 Environmental Council of the States ... Show Full Article WASHINGTON, Sept. 11 -- The Environmental Protection Agency issued the following news release: * * * EPA Deputy Administrator Fotouhi Advances Cooperative Federalism in Wyoming * Jackson, Wyoming - Last week, U.S. Environmental Protection Agency (EPA) Deputy Administrator David Fotouhi traveled to Wyoming to advance cooperative federalism and highlight the Trump EPA's commitment to achieving measurable environmental results while supporting economic growth. During the visit, Deputy Administrator Fotouhi addressed state environmental leaders at the Fall 2026 Environmental Council of the States(ECOS) Meeting and saw firsthand how Wyoming is using innovative approaches to improve water quality and expand American energy production.
"Wyoming is showing what is possible when federal, state, and local leaders work together to achieve real environmental results while supporting economic growth," said EPA Deputy Administrator Fotouhi. "It was great to be on the ground seeing this firsthand in Wyoming, from innovative stormwater solutions to leveraging technology for strong domestic energy production, but it was encouraging to hear this is happening across the nation from a variety of state leaders. EPA is committed to working with our state partners to advance practical solutions that protect our nation's resources and unleash American innovation."
Deputy Administrator Fotouhi began his visit with a tour of the Karns Meadow Stormwater Treatment Wetland in Jackson, Wyoming, alongside EPA Region 8 Administrator Cyrus Western, Wyoming Department of Environmental Quality (WDEQ) Water Quality Administrator Jennifer Zygmunt and local officials. The 40-acre wetland system was constructed with a combination of EPA and local funding to address longstanding water quality challenges in nearby Flat Creek. Today, the wetland captures and treats stormwater runoff from Jackson before it enters Flat Creek within the Greater Yellowstone Ecosystem. Post-implementation monitoring has demonstrated significant improvements in runoff quality, including an average 89% reduction in total settleable solids, a 95% reduction in nitrate, nitrite, and ammonia concentrations, and an 85% reduction in E. coli bacteria. The project demonstrates how targeted investments and innovative solutions can deliver measurable improvements in water quality.
The following day, Deputy Administrator Fotouhi delivered the keynote address at the Fall 2026 ECOS Meeting in Jackson, where he highlighted the Trump Administration's environmental accomplishments during its first 18 months and EPA's commitment to working with states to return ownership of environmental programs to state and local leaders. He was joined by regional administrators from all 10 EPA regions and numerous assistant administrators as they engaged with environmental regulators from all 50 states.
Deputy Administrator Fotouhi concluded his Wyoming visit with a tour of the Pinedale Anticline Natural Gas Field, one of the largest natural gas fields in the United States and a major contributor to Wyoming's energy production for more than two decades. Supplying millions of homes with energy each day, the Pinedale Anticline is an important part of America's domestic energy production and supports reliable, affordable power for American families. The field also demonstrates how technological innovation, combined with federal and state leadership and industry collaboration, can support environmental progress while strengthening domestic energy production. Deputy Administrator Fotouhi was joined on the tour by members of WDEQ, including Air Division Administrator Amber Potts.
Additionally, while in the area, Deputy Administrator Fotouhi joinedExit EPA's website Wake Up Wyoming to discuss the Trump EPA's commitment to energy dominance, including the June proposal to save the Dave Johnston power plant in Converse County, Wyoming. The Dave Johnston power plant uses local Wyoming coal, supports hundreds of Wyoming mining and energy jobs, and is essential to delivering reliable, affordable energy to families across the state.
***
Original text here: https://www.epa.gov/newsreleases/epa-deputy-administrator-fotouhi-advances-cooperative-federalism-wyoming
* * *
EPA Deputy Administrator Fotouhi Advances Cooperative Federalism in Wyoming
*
Jackson, Wyoming - Last week, U.S. Environmental Protection Agency (EPA) Deputy Administrator David Fotouhi traveled to Wyoming to advance cooperative federalism and highlight the Trump EPA's commitment to achieving measurable environmental results while supporting economic growth. During the visit, Deputy Administrator Fotouhi addressed state environmental leaders at the Fall 2026 Environmental Council of the States ... Show Full Article WASHINGTON, Sept. 11 -- The Environmental Protection Agency issued the following news release: * * * EPA Deputy Administrator Fotouhi Advances Cooperative Federalism in Wyoming * Jackson, Wyoming - Last week, U.S. Environmental Protection Agency (EPA) Deputy Administrator David Fotouhi traveled to Wyoming to advance cooperative federalism and highlight the Trump EPA's commitment to achieving measurable environmental results while supporting economic growth. During the visit, Deputy Administrator Fotouhi addressed state environmental leaders at the Fall 2026 Environmental Council of the States(ECOS) Meeting and saw firsthand how Wyoming is using innovative approaches to improve water quality and expand American energy production.
"Wyoming is showing what is possible when federal, state, and local leaders work together to achieve real environmental results while supporting economic growth," said EPA Deputy Administrator Fotouhi. "It was great to be on the ground seeing this firsthand in Wyoming, from innovative stormwater solutions to leveraging technology for strong domestic energy production, but it was encouraging to hear this is happening across the nation from a variety of state leaders. EPA is committed to working with our state partners to advance practical solutions that protect our nation's resources and unleash American innovation."
Deputy Administrator Fotouhi began his visit with a tour of the Karns Meadow Stormwater Treatment Wetland in Jackson, Wyoming, alongside EPA Region 8 Administrator Cyrus Western, Wyoming Department of Environmental Quality (WDEQ) Water Quality Administrator Jennifer Zygmunt and local officials. The 40-acre wetland system was constructed with a combination of EPA and local funding to address longstanding water quality challenges in nearby Flat Creek. Today, the wetland captures and treats stormwater runoff from Jackson before it enters Flat Creek within the Greater Yellowstone Ecosystem. Post-implementation monitoring has demonstrated significant improvements in runoff quality, including an average 89% reduction in total settleable solids, a 95% reduction in nitrate, nitrite, and ammonia concentrations, and an 85% reduction in E. coli bacteria. The project demonstrates how targeted investments and innovative solutions can deliver measurable improvements in water quality.
The following day, Deputy Administrator Fotouhi delivered the keynote address at the Fall 2026 ECOS Meeting in Jackson, where he highlighted the Trump Administration's environmental accomplishments during its first 18 months and EPA's commitment to working with states to return ownership of environmental programs to state and local leaders. He was joined by regional administrators from all 10 EPA regions and numerous assistant administrators as they engaged with environmental regulators from all 50 states.
Deputy Administrator Fotouhi concluded his Wyoming visit with a tour of the Pinedale Anticline Natural Gas Field, one of the largest natural gas fields in the United States and a major contributor to Wyoming's energy production for more than two decades. Supplying millions of homes with energy each day, the Pinedale Anticline is an important part of America's domestic energy production and supports reliable, affordable power for American families. The field also demonstrates how technological innovation, combined with federal and state leadership and industry collaboration, can support environmental progress while strengthening domestic energy production. Deputy Administrator Fotouhi was joined on the tour by members of WDEQ, including Air Division Administrator Amber Potts.
Additionally, while in the area, Deputy Administrator Fotouhi joinedExit EPA's website Wake Up Wyoming to discuss the Trump EPA's commitment to energy dominance, including the June proposal to save the Dave Johnston power plant in Converse County, Wyoming. The Dave Johnston power plant uses local Wyoming coal, supports hundreds of Wyoming mining and energy jobs, and is essential to delivering reliable, affordable energy to families across the state.
***
Original text here: https://www.epa.gov/newsreleases/epa-deputy-administrator-fotouhi-advances-cooperative-federalism-wyoming