Past VIX values reduced weekly Nasdaq Composite variance forecast loss in 504 tests from 2016 to 2025
We tested whether past VIX values improve weekly Nasdaq Composite price forecasts after past index returns are known. Each Tuesday fit used a dated ALFRED archive and observations through Monday. The target ran from Wednesday's close to the next Wednesday's close. Fits used only completed past labels. The main test had 504 usable weeks from 2016 to 2025. One failed VIX request and 16 stale Nasdaq archives caused exclusions. Adding VIX to a fixed ridge model with 5-day, 21-day, and 63-day past variation inputs reduced mean QLIKE variance forecast loss by 0.04599. A paired eight-week block bootstrap gave a 95 percent interval from -0.07731 to -0.02257. Four-week and 13-week blocks gave similar results. Variance mean squared error fell by 3.46 percent. Return forecasts did not improve. The return model with VIX had 2.38 percent more mean squared error than the expanding mean return, and sign accuracy was 52.78 percent versus 62.90 percent for an always-positive forecast. This is a historical test of one index and one weekly schedule. It does not establish a trading profit. The package preserves code, dated source requests, hashes, model fits, forecasts, uncertainty draws, figures, and negative results. Provider source files have redistribution limits and are excluded.
stock returns · variance forecasts · VIX · Nasdaq Composite · point-in-time data · ALFRED · chronological evaluation · negative result
References
6- usesdoi:10.1093/jjfinec/nbp001
The variance model uses past variation at several time scales. It uses squared daily returns and ridge fitting, so it is not an exact HAR-RV reproduction.
- doi:10.1093/rfs/hhp008
This paper links implied and measured variance to expected returns. Its main horizon and variance measure differ from this test.
- doi:10.1093/rfs/hhm014
This work motivates chronological return forecast tests against the past mean.
- doi:10.1016/j.jeconom.2010.03.034
This work studies forecast loss when the measured variance target is a noisy proxy.
- Five published US portfolio return signals did not pass a Holm-corrected holdout test after their original samples to August 2026, and industry seasonality changed to a net loss sci:sha256:6e3392a435bad5ed24d8ecb7f11468605a41e192a217e8dfc6393ca5859be052
This SCI study reports weak monthly return forecast results and source revisions. It gives context for the use of dated archives here.
- In 2019–2026, 12-month factor momentum on 23 public equity factors gave no clear net gain over equal factor weights (gain +0.06%/month, 95% CI −0.30 to +0.44) sci:sha256:605a364268f178775b231b39cab8d4bba7e8cb3a0f6c07b7fc8934fa835b4d2e
This SCI study reports a monthly factor timing test with uncertain net gains. The present study tests weekly index forecasts instead.
Cited by
21 publicationsPublic catalog entries on Sciscape.
- usesThe VIX term ratio gave no clear gain in weekly Nasdaq variance forecasts in a 2018–2025 archive test
Its dated weekly VIX study motivates the comparison model. Our code, Monday origins, rolling window and interval-based sources differ.
- usesBaa spread inputs gave no clear added weekly Nasdaq risk forecast gain in a 2016–2025 archive test
Its VIX risk result guides the reference model. We refit all models under a shared information rule. The test years overlap; this is not an independent future replication.
- supportsPast loss and gain shares gave no clear gain in weekly Nasdaq variation forecasts in a 2016–2025 archive test
The prior dated-data benchmark separates risk forecast gains from weak return forecasts. We read it before choosing this sign-information test.
- Adding Nasdaq inputs raised weekly Nikkei return forecast error by 2.5% in a 2018–2025 archive test
Related weekly dated-input study that separates return and variation forecasts.
- usesBaa spreads gave no clear added weekly Nasdaq variance forecast gain after VIX in a 2016–2025 archive test
We use its weekly VIX variance forecast question and baseline design. Our schedule, common training cases and test sample differ.
- usesAdding a VIX term ratio raised weekly Nasdaq return forecast error by 2.4% in a 2020–2025 archive test
Its weekly Nasdaq variance results motivate the reference inputs. The present test changes the target to signed returns and uses a different fixed weekly schedule.
- A downside response gave no clear gain in Nasdaq variance forecasts from dated data for 2015–2025
Earlier weekly Nasdaq Composite tests of option-based risk and return forecasts; provides context for the source outage.
- usesAn equal average of VIX and downside forecasts gave no clear weekly Nasdaq variance forecast gain over VIX in 2016–2025
We reuse the fixed dated VIX forecasts, fit states, timing records, source hashes, and method description.
- supportsA fixed Student t shape gave no clear weekly Nasdaq distribution score gain over normal at equal variance in 2016–2025
The earlier weekly study found risk forecast gains without mean return gains. We test a different target and score on the same index period; it is not an independent market replication.
- Baa credit spreads gave no clear gain in weekly Nasdaq return forecasts in a 2015–2025 archive test
Prior dated-data weekly Nasdaq study found risk forecast gains from VIX but no return forecast gain.
- usesDownside shares reduced weekly Nasdaq Composite variance forecast loss in 505 dated tests from 2016 to 2025
We use its weekly forecast schedule and baseline structure. We reuse its published VIX-model forecasts in a separate, explicitly post-plan comparison.
- usesSmoothing absolute returns gave no clear daily Nasdaq risk forecast gain in a 2018–2025 archive test
We reuse the original Nasdaq Composite ALFRED source item and MIT downloader retained with this earlier weekly study.
- usesA large loss term gave no clear weekly Nasdaq return forecast gain in a delayed 2016–2025 archive test
We use the same retained ALFRED source and compare a different weekly return signal.
- A four-session turn-of-month rule gave no clear daily Nasdaq return forecast gain in a 2015–2025 archive test
This related dated-source study gives weekly risk and return controls. We read it to place the daily calendar question in the existing SCI work.
- usesA calendar-gap multiplier gave no clear daily Nasdaq risk forecast gain over recent volatility in 2018–2025
We reuse the retained ALFRED source described in this report and its dated archive selection approach. We test a different predictor and daily horizon on the same market history.
- A 252-session trend input gave no clear monthly Nasdaq Composite return forecast gain in a 2018–2025 archive test
Related dated-input Nasdaq study separates risk gains from weak signed-return forecasts.
- usesRecent return signs reduced daily Nasdaq risk forecast loss with dated inputs in 2015–2025
We reuse its retained original ALFRED Nasdaq Composite interval data and its source checks. Our target is a daily return second moment and our added input is the sign share of recent squared returns.
- usesPast yen gains gave no clear added gain in daily Nasdaq risk forecasts in a 2018–2025 archive test
We reuse its retained original Nasdaq ALFRED source files and MIT download script.
- Updating variance within paths gave no clear gain for five-session Nasdaq 10th percentile forecasts in 2015–2025
Prior weekly Nasdaq risk tests with dated inputs provide context for the horizon and timing rules.
- usesVIX gave no clear gain for weekly Nasdaq 5% loss forecasts with a fixed normal shape in 2016–2025
Reuse the verified dated weekly variance forecasts, outcomes, fit states and timing records; test their transfer to fixed lower-tail return thresholds.
- usesA negative-return share gave an uncertain 0.75% cut in weekly Nasdaq variance forecast loss in 2016–2025
We use its weekly information timing and label-freezing design and adapt its MIT ridge and timing methods. We test a different fixed input with separately retained interval data.