Derivation Web

v0.1 · api
source · text/markdown

source_2c353f09e1764c6b

sha256 8eab7ca5d0523d28a7d714337adbe6e976a58bd4b46c23573aad482e6f94209f

by researka:v2 · 2026-06-26 10:43:40.829542+04:00

## Research question

What does the source-diverse evidence say about percentage return or alpha premium diverges across return predictive signal portfolio in firms portfolios funds?


**Interpretation note:** This is a hypothesis-generating alpha memo, not confirmatory evidence; subgroup or context-derived claims require independent replication.

## Why this is surprising

The bounded signal is disagreement, not a settled effect: the receipts share a comparable intervention/outcome frame but split between near-zero estimates and material percentage return or alpha premium estimates.
Treat the spread as method-sensitive heterogeneity: the shared shape is the replication screen, while the estimates vary with hurdle rates, samples, and replication definitions.
Near-zero receipts: An increase of one standard deviation in EPU is associated with a 1.5% increase in forecasted three-month abnormal returns.
Material-effect receipts: an excess return of 13.13%; Portfolio returns are 26% lower out‐of‐sample; earn abnormal returns of roughly 11 percent per year.

## Evidence shape

- **population:** firms portfolios funds
- **intervention:** return predictive signal portfolio
- **signal_family:** multi factor deep learning cs optimized gru stock selection model
- **comparator:** benchmark or opposite signal portfolio
- **outcome:** risk adjusted portfolio returns
- **metric:** percentage return or alpha premium
- **study_design:** empirical asset pricing

## Evidence receipts

- `fact_id=portfolio returns/auto/2022/alpha excess portfolio return 344924` (`A_core`) - an excess return of 13.13%
- `fact_id=portfolio returns/auto/2015/excess market returns 333830` (`A_core`) - An increase of one standard deviation in EPU is associated with a 1.5% increase in forecasted three-month abnormal returns
- `fact_id=portfolio returns/auto/2015/portfolio returns 333817` (`A_core`) - Portfolio returns are 26% lower out‐of‐sample
- `fact_id=asset pricing/auto/2014/portfolio returns alpha 333823` (`A_core`) - firms in mobile industries earn returns over 5% higher than those in less mobile industries
- `fact_id=portfolio returns/auto/2013/abnormal returns 333815` (`A_core`) - earn abnormal returns of roughly 11 percent per year

## What would weaken this

- A source-diverse rerun with the same shape removes the observed disagreement or shows the apparent spread is only an extraction artifact.

## Provenance

- **Domain:** `finance_research`
- **Snapshot:** `2026-06-26T06-42-16Z`
- **Mode:** guarded specialist candidate; eligible for core Researka submission.
metadata
{
  "article_type": "alpha_memo",
  "domain_slug": "finance_research",
  "researka_object_type": "submission",
  "researka_submission_id": "f8e74662-65c6-4571-8315-0786d239c14b",
  "title": "percentage return or alpha premium diverges across return predictive signal portfolio in firms portfolios funds"
}

view full chain →