|
on Big Data |
| By: | Kostiantyn Okhrimenko (University of Warsaw, Faculty of Economic Sciences) |
| Abstract: | This paper investigates the feasibility of predicting art prices using machine learning methods applied to a dataset of 20, 905 paintings and drawings scraped from the Artsper online marketplace. We test tree-based ensemble models (Decision Tree, Random Forest, XGboost) and deep learning architectures (MLP, CNN, Fusion) on both tabular metadata and hand-crafted image features. Results consistently show poor predictive performance across all model types and feature sets. We argue that this outcome is not a methodological failure but a substantive finding: it constitutes a diagnosis of the market structure of contemporary art. Drawing on hedonic pricing theory (Rosen 1974), the sociology of cultural fields (Bourdieu 1993), and the economics of valuation (Velthuis 2005; Beckert and Rössel 2013), we propose a three-layer model of art price determinants: physical attributes (observable and partially captured by models), visual-aesthetic features (observable but poorly quantifiable), and narrative-reputational capital (largely unobservable in cross-sectional platform data). |
| Keywords: | art market, machine learning, hedonic pricing, art valuation, cultural economics, XGboost, CNN |
| JEL: | Z11 C45 C53 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:war:wpaper:2026-18 |
| By: | Talita Greyling (Centre for Well-being, Artificial Intelligence and Social Impact (C.WAIS) and School of Economics, University of Johannesburg, Johannesburg, South Africa); Rangan Gupta (Department of Economics, University of Pretoria, Private Bag X20, Hatfield 0028, South Africa); Christian Pierdzioch (Department of Economics, Helmut Schmidt University, Hamburg, Germany) |
| Abstract: | We develop a social media sentiment index based on tweets extracted from Twitter and use a newspapers-based supply bottlenecks index to study by means of random forests, a machine-learning technique, how the latter affects the former for six European countries, after controlling for a wide array of other macro-finance predictors. We find that the predictive relationship between supply bottlenecks and sentiment is generally negative, but in a nonlinear manner. Supply chain constraints emerge as an important predictor of sentiment relative to the other control variables, and its predictive effect increases over the predictive horizon. |
| Keywords: | Social media sentiment, Supply bottlenecks, Machine learning |
| JEL: | C22 C53 E23 E70 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:pre:wpaper:202616 |
| By: | Tanisa Tawichsri; Suppawong Tuarob; Nuwat Nookhwun; Chinjuta Sangasaeng |
| Abstract: | We develop a news-based inflation expectations index for Thailand using a scalable workflow that integrates topic modeling, LLM-assisted labeling, and fine-tuned BERT classification. Based on 1.1 million Thai-language news articles from 2015–2024, the index leads both headline inflation and firm inflation expectations. Given that inflation narratives in news are inherently subjective and often ambiguous, we show that prompt design can materially affect downstream economic inference. In out-ofsample forecasting, augmenting autoregressive benchmarks with the news index reduces RMSE by up to 32% for headline inflation and 30% for firm inflation expectations, with gains increasing at longer horizons. SHAP-based decomposition reveals a horizon-dependent information structure: price-specific topics drive short-term forecasts, while macroeconomic narratives dominate at longer horizons. Our findings demonstrate that LLM-assisted text analysis can generate economically meaningful inflation indicators in non-English, emerging-economy settings. The index also performs particularly strong during periods of elevated inflation uncertainty. |
| Keywords: | Inflation expectations; Text-based indicators; Online news data; Large language models (LLMs); Machine learning; Sentiment analysis; Nowcasting and forecasting; Emerging economies |
| JEL: | E31 E37 D84 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:pui:dpaper:252 |
| By: | Daniel Coll Sol (Tax Justice Network, London, United Kingdom); Mario Cuenda Garcia (Tax Justice Network, London, United Kingdom); Bathusi Gabanatlhong (Institute of Economic Studies, Faculty of Social Sciences, Charles University, Prague,); Miroslav Palansky (Institute of Economic Studies, Faculty of Social Sciences, Charles University, Prague,); Tijmen Tuinsma (Tax Justice Network, London, United Kingdom) |
| Abstract: | This paper introduces the abnormality method to estimate illicit financial flows (IFFs) using a bilateral gravity model complemented by a machine learning technique to analyse unexplained financial flows to offshore centres. The findings provide robust evidence linking abnormal flows to offshore financial centres with tax avoidance and evasion and offer new estimates of their scale, costs, and geographical distribution. In 2023, abnormal flows to tax havens and secrecy jurisdictions reached US$2.8 trillion, resulting in foregone tax revenues exceeding US$60 billion. These flows originated mainly from Europe, the Americas, and Asia, flowing mostly to European tax havens. Random Forest analysis confirms that tax haven and secrecy jurisdiction status are key determinants of abnormal financial flows. Furthermore, the analysis of the Automatic Exchange of Information (AEOI) regulation indicates an increase in abnormal flows held in secretive jurisdictions. |
| Keywords: | Illicit Financial Flows, Offshore Financial Centres, Revenue Losses, Machine Learning, Automatic exchange of Information |
| JEL: | H26 F21 F23 C45 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:fau:wpaper:wp2026_08 |
| By: | Kohei Maehashi (Bank of Japan); Daisuke Miyakawa (Waseda University); Takatoshi Sasaki (Bank of Japan) |
| Abstract: | In this paper, we develop a model to predict large currency swings using transaction-level data on foreign exchange options, collected by trade repositories (TRs). These data allow us to capture heterogeneous currency risk perceptions of individual market participants. By applying a quantile regression combined with machine learning for variable selection, we find that market participants' views extracted from trade repository data significantly improve the predictions of large currency swings. |
| Keywords: | Currency swing; Granular data; Trade repository; Foreign exchange option; Quantile regression; Variable selection |
| JEL: | C22 C55 F31 G17 |
| Date: | 2026–05–29 |
| URL: | https://d.repec.org/n?u=RePEc:boj:bojwps:wp26e10 |
| By: | Robert D. Lieberthal (Lieberthal and Associates, LLC); Richard Tran (MDSight, LLC); Vietbao Phan (Thomas Jefferson University); Jawand Singh (Lieberthal and Associates, LLC; William and Mary University); Elizabeth Sottung (Thomas Jefferson University) |
| Abstract: | Actuaries rely primarily on structured numerical data for reserving and ratemaking, while valuable predictive information in unstructured text including medical records, adjuster notes, and call transcripts remains largely unused. Manual processing of these documents is time-consuming, inconsistent across reviewers, and unscalable. We present a proof-of-concept framework using large language models (LLMs) to extract structured actuarial variables from unstructured claims data. We implement a two-stage processing architecture separating document-level extraction (Stage 1) from claim-level synthesis (Stage 2). A modular four-script Python pipeline processes synthetic FHIR-based claims data and real claims documents, extracting 36 actuarial variables across reserving, ratemaking, and claims management categories. We validate 14 core variables using two independent clinical expert reviewers scoring 20 synthetic claims on a five-point Likert rubric, achieving mean scores above 4.0 and a weighted kappa of 0.53. Integration with chain ladder reserving demonstrates practical actuarial value: severity-segmented analysis reduced reserve estimation error from 6.5% to 4.0%. The open-source implementation includes audit trails and confidence scoring, providing a replicable foundation for LLM-based actuarial variable extraction in property-casualty insurance. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.06089 |
| By: | Omar Abdel Haq (Harvard Business School); Amitabh Chandra (Harvard Business School & Harvard Kennedy School); Tomáš Jagelka (University of Bonn, Dartmouth College, & CREST-Ensae); Erzo F.P. Luttmer (Dartmouth College); Joshua Schwartzstein (Harvard Business School) |
| Abstract: | Large Language Models (LLMs) are trained on a prodigious corpus of human writing and may reveal human preferences over characteristics of life courses, such as income, longevity, and working conditions. We present OpenAI's GPT-5.4 and a broadly representative sample of Americans with pairs of life stories and ask them to choose the life they would prefer for themselves. A person's choice is better predicted by the LLM's choice than by another person’s choice over the same stories, and LLM valuations of several life attributes are similar to those derived from human responses. Our results suggest that LLM responses offer a scalable and cost-effective complement to existing methods for studying human preferences. |
| Keywords: | Life preferences, LLMs, LLM revelation conjecture, life stories, essential life attributes, life attribute valuations |
| JEL: | D90 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:ajk:ajkdps:410 |
| By: | Gaurav Datt; John Nguyen; Pedro Salas-Rojo; Francisco H.G. Ferreira; Paolo Brunori; Vito Peragine; Albert Park; Arturo Martinez Jr.; Joseph Albert Nino Bulan |
| Abstract: | This paper aims to contribute to an understanding of the extent, nature and persistence of unfair inequality in the Asia Pacific region, building on a rich literature on the measurement of inequality of opportunity (IOp). As part of a project to build a global database of IOp, the paper uses microdata from 39 nationally representative household surveys to present IOp estimates for 14 countries that account for about three-quarters of the region’s population. We use consistent data protocols to ensure a high degree of cross-country comparability of IOp estimates. A distinguishing feature of the exercise is the use of machine learning methods to construct IOp estimates, which efficiently balances the risks of potential under- or over-fitting. The resultsshow that, on average, nearly two-fifths of income or consumption inequality across the Asia-Pacific region represents inequality of opportunity attributable to inherited circumstances, though with wide variation across countries, ranging from about a quarter to over half. The cross-country variation in IOp is consistent with a Great Gatsby curve for the Asia-Pacific. A decomposition analysis assesses the relative contributions of different circumstances to IOp. |
| Keywords: | inequality of opportunity, economic mobility, Asia-Pacific, machine learning |
| JEL: | D31 D63 O15 |
| Date: | 2026–04–01 |
| URL: | https://d.repec.org/n?u=RePEc:mos:moswps:paper_1775627852525_205 |
| By: | Sai Ma |
| Abstract: | Using microdata from the Michigan Survey of Consumers, we study how within-household reallocations of attention across news affect inflation expectation bias, measured relative to a real-time, machine-learning full-information benchmark. Shifting attention toward unfavorable (favorable) economic news increases (decreases) forecast bias substantially, while dropping attention to an unfavorable topic has little effect. The largest bias increases come not from inflation news itself, but from attention to unfavorable social, political, and geopolitical narratives. Aggregate news sentiment has no effect on bias when a household's reported attention allocation is unchanged. In aggregate, these effects are amplified when the attention network is dominated by an unfavorable focal hub: bias-reducing favorable narratives are crowded out of limited attention sets, and respondents closer to the hub exhibit larger bias increases. We find that past and present attention to news together account for up to 70 percent of observed forecast bias, with the current attention component rising sharply during recessions and large negative news events. Results are robust to a battery of specification checks and external validation. |
| Keywords: | inflation expectations; limited attention; forecast bias; sentiment; networks |
| JEL: | E31 E52 D83 D84 |
| Date: | 2026–04–17 |
| URL: | https://d.repec.org/n?u=RePEc:fip:fedgif:103344 |