anomaly-atlas
Honesty doctrine. Every candidate anomaly is an artifact until proven otherwise; in-sample results are never findings; past statistical regularity does not imply future returns. This is research on statistical properties of market data — not investment advice, not a trading system.

Research / research/notes/transaction_costs.md

Phase 1 notes — transaction cost realism

reviewed

Phase 1 notes — transaction-cost realism · created Tue Aug 11 2026 20:00:00 GMT-0400 (heure avancée de l’Est) · Simon-Pierre Boucher

Transaction-cost realism (§4.5 — Q5, the cost frontier)#

Phase 1 literature notes. Citations verified via OpenAlex, accessed 2026-08-12. Constraint: our data has NO quotes — every spread must be estimated from OHLCV bars. That makes the low-frequency-estimator literature load-bearing.

1. Spread estimation from bar data (our only instruments)#

  • Roll (1984) — see microstructure notes; autocovariance-based, undefined when autocov ≥ 0 (expB: mega-caps).
  • Hasbrouck, J. (2009). Trading Costs and Returns for U.S. Equities: Estimating Effective Costs from Daily Data. Journal of Finance 64(3), 1445–1477. https://doi.org/10.1111/j.1540-6261.2009.01469.x — Gibbs-sampled Roll model; corr 0.965 with TAQ benchmarks.
  • Corwin, S. A. & Schultz, P. (2012). A Simple Way to Estimate Bid-Ask Spreads from Daily High and Low Prices. Journal of Finance 67(2), 719–760. https://doi.org/10.1111/j.1540-6261.2012.01729.x — high-low estimator; we have highs/lows at every timeframe.
  • Abdi, F. & Ranaldo, A. (2017). A Simple Estimation of Bid-Ask Spreads from Daily Close, High, and Low Prices. Review of Financial Studies 30(12), 4437–4480. https://doi.org/10.1093/rfs/hhx084 — CHL estimator; best for illiquid names.
  • Ardia, D., Guidotti, E. & Kroencke, T. A. (2024). Efficient estimation of bid–ask spreads from open, high, low, and close prices. Journal of Financial Economics 161, 103916. https://doi.org/10.1016/j.jfineco.2024.103916 — state-of-the-art OHLC estimator (EDGE), asymptotically unbiased, minimal variance. Primary estimator for expG; Roll/CS/CHL as cross-checks.
  • Fong, K. Y. L., Holden, C. W. & Trzcinka, C. A. (2017). What Are the Best Liquidity Proxies for Global Research? Review of Finance 21(4), 1355–1401. https://doi.org/10.1093/rof/rfx003 — horse race of low-frequency proxies vs intraday benchmarks; validates picking 2–3 complementary proxies.
  • Lesmond, D. A., Ogden, J. P. & Trzcinka, C. A. (1999). A New Estimate of Transaction Costs. Review of Financial Studies 12(5), 1113–1141. https://doi.org/10.1093/rfs/12.5.1113 — LOT: costs from the incidence of zero returns (1.2 %–10.3 % across deciles). Our no-trade minutes are the intraday analogue.
  • Bessembinder, H. (2003). Issues in Assessing Trade Execution Costs. Journal of Financial Markets 6(3), 233–257. https://doi.org/10.1016/s1386-4181(02)00064-2 — measurement pitfalls (trade signing, timing conventions) — reminder that even "measured" costs carry convention risk.

2. Do anomalies survive costs?#

  • Novy-Marx, R. & Velikov, M. (2016). A Taxonomy of Anomalies and Their Trading Costs. Review of Financial Studies 29(1), 104–147. https://doi.org/10.1093/rfs/hhv063 — low-turnover anomalies mostly survive (with mitigation); high-turnover ones mostly do not. Short-horizon effects (ours) are the highest-turnover class → strong prior that Q1/Q2 effects die at the cost frontier.
  • Frazzini, A., Israel, R. & Moskowitz, T. J. (2012). Trading Costs of Asset Pricing Anomalies. SSRN. https://doi.org/10.2139/ssrn.2294498 — real institutional executions: realized costs are far below TAQ-implied for patient flow. Gives the lower bound of the cost sweep.
  • Chen, A. Y. & Velikov, M. (2022). Zeroing In on the Expected Returns of Anomalies. Journal of Financial and Quantitative Analysis 58(3), 968–1004. https://doi.org/10.1017/s0022109022000874 — net of spreads + post-publication decay + data mining, the average anomaly earns ~4 bp/month. The sobering calibration for our whole atlas.
  • Detzel, A., Novy-Marx, R. & Velikov, M. (2023). Model Comparison with Transaction Costs. Journal of Finance 78(3), 1743–1775. https://doi.org/10.1111/jofi.13225 — ignoring costs biases even model comparisons; cost-awareness is not optional at any stage.

3. Design of expG (cost frontier)#

  1. Spread panel: EDGE (Ardia et al.) per ticker-month from daily OHLC + Corwin–Schultz and CHL cross-checks + Roll where defined; validate the three against each other (Fong et al. protocol).
  2. Options as auxiliary evidence: our options chains carry real bid/ask — the only quoted spreads in the dataset; usable as a sanity anchor for the underlying's cost regime (with care).
  3. Sweep, don't point-estimate: report each surviving effect's net value across cost multipliers 0.25×–2× the estimated half-spread (Frazzini lower bound ↔ retail-taker upper bound), and the crossing point where net effect = 0.
  4. Capacity is out of scope (no volume-at-quote data) — state it as a limitation rather than pretend.