Knowledge that Transforms

While financial textual analysis increasingly relies on complex machine learning, we propose a simpler, data-driven alternative. Using elastic net regressions on a massive panel of 10-K n-grams, we construct a specialized dictionary that weights phrases by their marginal predictive power. This bottom-up methodology effectively forecasts expected stock returns, with a spread portfolio generating significant average returns. Our approach outperforms prominent financial dictionaries, off-the-shelf large language models, and machine learning algorithms. These results demonstrate the value of identifying financial meaning from the bottom up, highlighting the need for domain- specific models trained on relevant financial contexts.

Tick sizes affect market quality through a tradeoff between pricing fidelity and undercutting. The U.S. Tick Size Pilot (TSP), which raised the minimum tick from 1¢ to 5¢, provides a natural experiment to study this tradeoff. We find that the TSP harmed liquidity for stocks with spreads below 10¢ but improved liquidity for stocks with spreads above 15¢. These opposing effects explain the mixed results across prior studies which pool together stocks with very different prevailing spreads. We recommend researchers using the TSP for causal inference should, at minimum, split samples at 10¢-spreads to account for these heterogeneous liquidity effects.

High-frequency traders (HFTs) mainly operate on public exchanges. Since the European regulatory changes in 2018 (Markets in Financial Instruments Directive II), some HFTs began operating Systematic Internalizers (SIs), i.e., single-dealer platforms where clients trade against the dealer's inventory. Using Swedish equity market data, we show that higher HFT dealer-platform activity reduces displayed exchange liquidity: quoted spreads widen and depth falls. Effective spreads are largely unchanged for HFTs but increase for non-HFT traders. Price efficiency improves as return autocorrelations and excess variance ratios move closer to random-walk benchmarks. Evidence suggests HFT dealers' inventory management is the main channel behind these results.