spb/firc0001_uqo

1% ============================================================================2\section{Related Literature}\label{sec:literature}3% ============================================================================45\paragraph{Text as data in accounting and finance.} The narrative content of SEC filings has become a standard empirical input. \citet{loughran2011} showed that general-purpose sentiment dictionaries misclassify financial text and introduced finance-specific word lists; \citet{loughran2016} survey the subsequent decade of textual analysis in accounting and finance, and \citet{gentzkow2019} provide the general econometric framework for treating text as data, including the measurement-error and validation issues that any text-derived variable inherits. \citet{dyer2017} document how 10-K disclosure has lengthened and become more boilerplate over time, a trend that matters for us because repeated boilerplate is the principal source of false-positive ``projects''. \citet{cohen2020} show that changes in otherwise repetitive filing language carry information, which motivates our emphasis on \emph{repeated} mentions of the same initiative across filings rather than on single occurrences. \citet{grimmer2013} offer the general methodological warning that automated text measures must be validated against human judgment on the specific task at hand, which is the stance adopted in Section~\ref{sec:validation}.67\paragraph{Measuring corporate activity from disclosures.} Closest to our construct are text-based measures of what firms do rather than how they feel. \citet{hoberg2016} build product-market classifications from 10-K business descriptions; \citet{hobergmaksimovic2015} identify financially constrained firms from the language of the liquidity discussion; \citet{hassan2019} measure firm-level political risk from the share of earnings-call text devoted to politics; \citet{li2021} measure corporate culture with word embeddings; and \citet{kelly2021} and \citet{kalyani2025} track the emergence and diffusion of technologies through patents, calls and job postings. SPID differs from these approaches in its unit of observation. Rather than a firm-level score summarizing the intensity of a theme, it targets \emph{discrete initiatives}---objects with a name, a type, a status, cited years, a location, partners and an amount---that can be followed from filing to filing. This is the project-centric view that commercial intelligence products sell, but with open provenance and a documented extraction contract.89\paragraph{Structured information extraction with language models.} The extraction step relies on an instruction-tuned LLM used in a constrained-output mode: a fixed system prompt, a strict JSON object schema, temperature zero, and an explicit instruction to extract only initiatives grounded in the supplied passage. This design follows the general lesson of the text-as-data literature that the measurement instrument must be fixed, versioned and auditable, and it is deliberately conservative: the model is asked to prefer fewer, concrete projects to many vague ones, to return an empty list when no concrete project is described, and to attach a self-reported confidence. Semantic search over projects uses sentence embeddings \citep{reimers2019}. We make no claim that self-reported confidence is calibrated; its calibration is one of the items of the validation protocol.1011\paragraph{Entity resolution and record linkage.} Linking repeated mentions of the same initiative across years is a record-linkage problem in the sense of \citet{fellegi1969}; \citet{christen2012} reviews the modern toolbox of blocking, comparison and classification. The current SPID resolver is a deterministic blocking rule (issuer, project type and a location or name token) rather than a probabilistic linker. We document its behavior and its failure modes in Section~\ref{sec:resolution}, and we treat the migration to an explainable, reversible linkage with human adjudication of ambiguous clusters as part of the validation roadmap.1213\paragraph{Infrastructure.} The database is implemented as a single analytical file in DuckDB \citep{raasveldt2019}, with Parquet exports; acquisition respects the SEC's published fair-access policy \citep{secfairaccess}, and the Russell~3000 is defined following FTSE Russell's published construction methodology \citep{ftserussell2026}, approximated here through a fund-holdings proxy (Section~\ref{sec:sample}).14