MacroLens: A Multi-Task Benchmark for Contextual Financial Reasoning under Macroeconomic Scenarios

2026-06-23

AI Summary

Existing financial reasoning benchmarks struggle to uniformly integrate price, fundamental, macroeconomic, and textual signals, and often suffer from look-ahead bias, data latency, information redundancy, and cross-temporal leakage. This work proposes MacroLens—the first multitask benchmark covering 4,416 U.S. small- and micro-cap stocks from 2021 to 2026—featuring point-in-time aligned price data, XBRL financial filings, macroeconomic time series, SEC documents, and news articles, along with 1,130 automatically detected macro-event scenarios. We introduce a multimodal financial reasoning framework with a scenario-aware layer and systematically evaluate 19 methods, including six model families and a five-stage ablation study of large language models. Our analysis quantifies the contributions of heterogeneous signals across seven distinct tasks, and we release the full dataset to support reproducible research.

Abstract

Financial decision-making is contextual: forecasting prices, valuing companies, and assessing event exposure weigh price history, accounting fundamentals, macroeconomic regime, and contemporaneous text. A benchmark over these four signals is hard to build because finance violates four assumptions of time-series evaluation: text must be gated by its publication date to prevent look-ahead, quarterly fundamentals are reported with a one- to ninety-day lag, filing text is partly redundant with the numerical statement fields it accompanies, and macroeconomic regimes leak across calendar splits. No public benchmark addresses all four signals jointly. MacroLens covers 4,416 U.S. small- and micro-cap equities over 2021-2026. Seven tasks share one point-in-time panel of prices, 46.8M XBRL accounting facts, 53 macroeconomic series, 295,860 SEC filings, and 215,882 news articles, plus a scenario layer of 1,130 macroeconomic events across 49 types automatically detected and rendered as natural language. Tasks span contextual forecasting, public and private valuation, statement generation from fundamentals and descriptions, scenario-conditioned returns, and real-estate valuation. We evaluate 19 methods across six families spanning naive heuristics through time-series foundation models, fine-tuned LLM-based time-series models, and zero-shot large language models (LLMs), plus a five-step feature-context ablation on two frontier LLMs and a gradient-boosted baseline. MacroLens is released at https://huggingface.co/datasets/DeepAuto-AI/MacroLens.

Problem

• financial reasoning
• macroeconomic scenarios
• time-series evaluation
• contextual decision-making
• benchmark dataset

Innovation

• contextual financial reasoning
• macroeconomic scenarios
• time-series benchmark
• look-ahead bias mitigation
• multi-task evaluation