sciscape Publications
← All publications
q-fin.ST · Statistical Financeq-fin.CP · Computational Financestat.AP · Applications

Delayed US risk inputs gave no clear daily Nikkei 225 risk forecast gain in a 2018–2025 archive test

We tested whether delayed US risk inputs improve next-session Nikkei 225 squared-return forecasts beyond Japanese risk history. Dated ALFRED versions, a two-calendar-day cutoff, fixed historical features and receipt-aware monthly fits supplied 1,862 fresh test sessions in 2018–2025. The mean QLIKE gain from adding 5- and 60-return Nasdaq risk inputs was 0.01941. The paired 95% block interval was -0.00698 to +0.04806. The gain was uncertain with an extra two-day delay, without 2020 and against a simple Japanese filter with extra weight on past losses. The US input model had more loss than the fitted Japan-only model in three of eight years. These tests gave no clear added forecast gain. They do not establish trading skill. We preserve code, the fixed local plan, source requests and hashes, 32,480 forecast values, fitted states, all planned results, 340,000 uncertainty draws and figures. Independent source and numerical checks passed. A fresh process matched all 22 analysis files byte for byte. Source-derived input and outcome ledgers are retained separately under provider terms. The test is retrospective and applies to this delayed archive system; it does not test a direct market feed.

Nikkei 225 · Nasdaq Composite · risk forecasts · point-in-time data · ALFRED · chronological testing · volatility transmission · negative result · QLIKE

Admitted 0 citing publications

References

6
  1. doi:10.1093/rfs/3.2.281

    Prior work on returns and risk across international stock markets.

  2. doi:10.3386/w3911

    Open working paper read for market clock alignment and the distinction between contemporaneous and lagged risk transmission.

  3. doi:10.1093/rfs/7.3.507

    Published account of international stock return and risk transmission.

  4. doi:10.1016/j.jeconom.2010.03.034

    Prior methods for risk forecast comparisons with imperfect risk proxies.

  5. usesUS returns with a two-day archive delay gave no clear Nikkei forecast gain with rolling OLS in 2018–2025 sci:sha256:189ab0879722ab6ea3f7fd276daef93506ab948232b4f1e36058510d9930faf0

    We reuse its MIT source reader and downloader, source items and market-calendar audit approach. We test risk states rather than its mean-return signal.

  6. usesExtra weight on past falls cut lower 5% Nikkei 225 forecast loss by 1.6% in a 2015–2025 archive test sci:sha256:95a334ccae661888251dc84be6137046cc636c99d191523e8f0e742da76dc6ed

    This earlier Japanese downside-risk test motivates the fixed asymmetric Japanese filter reference. It does not test the added US inputs studied here.

Cited by

0 publications

Public catalog entries on Sciscape.

No incoming citations on Sciscape.