September absolute and excess returns
The ETFFixedIncome.com 5-Spot Active Research Model returned −1.21% in September, compared with −2.67% for its 50% AGG / 50% HYG benchmark and −2.61% for AGG. The strategy outperformed those benchmarks by 145.56 and 139.64 basis points, respectively. Its substantial Treasury-bill allocation and limited duration exposure cushioned the fixed income selloff, although corporate credit and senior loans still generated losses.
The primary benchmark rebalances to 50% AGG and 50% HYG quarterly and allows weights to drift between reviews. AGG provides a separate broad bond-market comparison. The model’s static five-ETF policy portfolio provides a third reference for isolating the value of active allocation decisions.
| Period | Model | 50/50 | AGG | vs 50/50 bp | vs AGG bp |
| September | -1.21% | -2.67% | -2.61% | +145.56 | +139.64 |
| Q3 | -0.52% | -2.75% | -3.50% | +222.68 | +297.65 |
| YTD | 0.68% | -1.56% | -2.78% | +224.37 | +345.93 |
| 1Yr | 1.84% | -0.49% | -1.86% | +232.98 | +369.86 |
| 3Yr | 18.56% | 18.80% | 12.70% | −23.37 | +586.16 |
Cumulative total returns through September 30, 2026. September and Q3 are calendar periods. One year and three years use calendar endpoints. Model and rebalanced comparator returns include the workbook’s transaction-cost assumption, but exclude advisory fees. All results are hypothetical.
Strategy level attribution summary
September’s benchmark advantage came from the portfolio’s defensive structure. The active overlay actually detracted 16.24 basis points versus the static five-ETF policy, which returned −1.05%. Retained high-yield exposure and a smaller Treasury-bill allocation offset some of the protection provided by the underlying policy mix.
For Q3, the model declined 0.52%, versus 2.75% for the 50/50 benchmark and 3.50% for AGG. Its 222.68-basis-point advantage over 50/50 included a 6.81-basis-point drag from active allocation. Year to date, the model gained 0.68%, versus −1.56% for 50/50 and −2.78% for AGG. The 224.37-basis-point advantage over the primary benchmark again came from the policy structure, with active allocation detracting 13.16 basis points.
Over one year, the model returned 1.84%, versus −0.49% for 50/50 and −1.86% for AGG. Over three years, it gained 18.56%, trailing 50/50 by 23.37 cumulative basis points while beating AGG by 586.16. Three-year annualized returns were 5.84%, 5.91% and 4.07%, respectively. Active allocation added 45.78 cumulative basis points versus the static policy over three years, but that benefit did not fully close the gap to the credit-heavy primary benchmark.
The live signal has turned defensive, but the October review does not authorize a trade. Current holdings therefore remain the reference allocation. The report separates those holdings from the more defensive signal targets that have not passed the implementation rules.
October position update
The latest verified holdings are the September 30 closing weights. There was no September trade, and the pending October 1 decision is also no trade. Changes from September 1 reflect market drift. The signal targets show the direction the model would prefer if its trading conditions were satisfied; they are not the current portfolio.
| ETF | Policy | Sep 1 held | Sep 30 held | Change pp | Signal target |
| SGOV | 40.00% | 31.09% | 31.57% | +0.47 | 45.63% |
| TLT | 0.00% | 0.00% | 0.00% | +0.00 | 0.00% |
| LQD | 20.00% | 19.12% | 18.70% | −0.42 | 20.28% |
| SRLN | 25.00% | 29.91% | 30.16% | +0.25 | 23.54% |
| HYG | 15.00% | 19.87% | 19.56% | −0.30 | 10.55% |
SGOV holds 31.57%, LQD 18.70%, SRLN 30.16% and HYG 19.56%, with no TLT position. SGOV and SRLN increased by 0.47 and 0.25 percentage point during September through relative performance, while LQD and HYG declined by 0.42 and 0.30 point. The portfolio remains below the 40% SGOV policy weight and above policy in SRLN and HYG.
The unexecuted defensive target would raise SGOV to 45.63%, reduce SRLN to 23.54% and cut HYG to 10.55%. LQD would be 20.28% and TLT would remain zero. Implementing those targets would require 15.64% one-way turnover, but turnover alone does not authorize a trade.
Why the October review remains on hold
| Trade condition | Latest reading | Required | Result |
| Regime persistence | 5 sessions | At least 10 | Fail |
| Forecast alpha | 6.82 bp | At least 10 bp | Fail |
| One-way turnover | 15.64% | At least 3% | Pass |
| Quarterly review | October 1 | Quarter start | Scheduled |
The model requires agreement across the persistence, turnover and forecast-alpha conditions. The defensive regime began September 24 and had persisted for only five sessions by month-end. The forecast-alpha estimate also fell short of its hurdle. The strategy therefore retains the existing allocation despite the change in live signals.
This discipline limits turnover and avoids acting on short-lived regime changes, but it also creates a real exposure lag. HYG and SRLN still account for 49.73% of the portfolio even though credit signals have deteriorated. Under the configured quarterly schedule, a failed October review would normally leave the next scheduled opportunity at the first trading session of the following quarter.
Positioning control note: the summary tab labels July 1, 2026 as the last rebalance and displays an “executed target.” The implementation log records no trade on that date. Its last trade marked executed is April 1, 2024. This report uses the implementation log and reconciled drifted holdings, rather than treating the summary’s target or suggested drift trades as completed transactions.
Indicator and signal recap
The model’s live regime moved from Balanced income at August-end to Defensive cash at September-end. Confidence increased from 54.82% to 80.08% and the live overlay setting rose from zero to 100%. Those readings inform the target allocation; the quarterly trade conditions determine whether the portfolio implements it. The held regime remains Risk-on credit.
| Signal | Aug 31 | Sep 30 | Change |
| Credit | +0.07 | -0.64 | -0.71 |
| Duration | -1.00 | -1.00 | +0.00 |
| Inflation | -1.00 | -1.00 | +0.00 |
| Defensive | +0.02 | +1.00 | +0.98 |
| Breadth | +0.20 | +0.05 | -0.15 |
| Floating Rate | +0.15 | -0.11 | -0.27 |
| Credit Sleeve | +0.12 | -0.80 | -0.92 |
| Floating Sleeve | +0.05 | -0.04 | -0.09 |
| Defense Sleeve | -0.11 | +1.30 | +1.41 |
The first six rows are the adjusted signals supplied to the 5-Spot model. Composite sleeve readings can exceed the −1 to +1 range. Higher credit and floating-rate sleeve readings support HYG and SRLN; higher defense readings favor SGOV.
The credit signal weakened from +0.07 to −0.64, and the credit sleeve fell from +0.12 to −0.80. The sleeve combines credit and equity breadth with the defensive signal, so a sharp increase in defensive pressure compounds the deterioration in credit. The defense sleeve rose from −0.11 to +1.30, explaining the higher desired Treasury-bill allocation.
Duration remained at −1.00 and the TLT confirmation test failed. The model only adds long Treasuries when the relevant duration conditions and TLT’s trailing 63-session performance support the position. Defensive conditions alone do not create a TLT allocation. The zero weight avoids adding long-duration exposure into a period of rising real yields.
The floating-rate signal declined from +0.15 to −0.11 and its composite sleeve moved from +0.05 to −0.04. This weakens the case for increasing SRLN even though floating-rate coupons provide less direct rate sensitivity than long fixed-rate bonds. Credit risk remains relevant to senior loans.
The inflation input stayed at −1.00. That is a portfolio signal, not a forecast that inflation is falling. In the inherited Core Aggressive framework, higher real yields and weak inflation-sensitive asset momentum can produce a negative investment signal while reported inflation remains elevated.
Signal quality and interpretation
The 5-Spot feed inherits the Core Aggressive model’s inputs. The accompanying Core workbook lacks August 31 credit-spread and advance–decline-line readings, which its formula guards treat as neutral. The month-end comparison should therefore be read with that limitation. We use the numeric signals and trade rules here; a generic “more supportive for credit risk” description in the macro-factor sheet does not match the negative September credit reading.
September market recap and October outlook
September penalized duration and credit risk. In the supplied Core market series, the two-year Treasury yield rose 53 basis points to 4.87%, the 10-year rose 53 basis points to 5.27% and the 30-year rose 38 basis points to 5.62%. The 10-year real yield climbed from 2.44% to 2.93%, while the 10-year breakeven increased only 5 basis points to 2.36%. The dominant real-yield move favored limited duration exposure.
The Federal Reserve raised its target range by 25 basis points to 3.75%–4.00% on September 16. August CPI increased 0.4% month over month and 3.4% year over year. The September 30 PCE release showed headline inflation at 3.4% and core inflation at 3.0%, with monthly increases of 0.3% and 0.2%. The combination of persistent inflation and resilient spending constrained the case for near-term monetary relief. [3–5]
High-yield spreads widened from 265 basis points on September 1 to 312 at month-end, while investment-grade spreads increased from 81 to 84 basis points. MOVE rose from 75.32 at August-end to 110.46. The spread comparison uses September 1 because August-end readings are missing from the source workbook.
The 5-Spot portfolio benefited from having no TLT exposure as the ETF lost 5.38%. SGOV gained 0.30% and SRLN declined 0.38%, outperforming LQD’s 3.39% decline and HYG’s 2.73% loss. The model still lost money because corporate bonds and loans accounted for more than two-thirds of the portfolio, but the bill allocation substantially reduced the drawdown relative to both benchmarks.
October outlook
Our October assessment favors liquidity and restraint on duration until real yields and bond volatility stabilize. That view aligns with the model’s defensive signal target. The actual strategy, however, carries its existing credit overweights into the review because persistence and forecast alpha have not met the trading thresholds. Portfolio exposure should be judged from the held weights rather than the live regime label.
Inflation and household demand remain in tension. August real consumption rose 0.6% even though real disposable income was unchanged, and the saving rate stood at 4.1%. Strong spending limits the urgency for easing, while income pressure increases the risk of weaker demand later. Energy and tariff-related price pressure could further complicate disinflation. [5]
If credit spreads stabilize and volatility retreats, retaining SRLN and HYG could preserve income and participation in a recovery. If spreads continue to widen, the delay in implementing the defensive target could become more costly. The 31.57% bill allocation provides a cushion, but it is materially below the unexecuted 45.63% target. A sustained Treasury rally is another opportunity cost because TLT remains unheld.
October’s scheduled checkpoints include the employment report on October 2, CPI on October 14 and the FOMC meeting on October 27–28. New data can change live signals, but the model’s quarter-start implementation rule governs trading. Passing a persistence threshold later in October does not, by itself, create an authorized intramonth rebalance. [4,6,7]
Performance attribution versus the benchmarks
The table separates active allocation from the structural contribution of the static policy portfolio. ETF rows attribute the active model’s return difference versus the drifting, quarterly rebalanced policy mix. The benchmark bridge then reconciles performance versus 50% AGG / 50% HYG and AGG.
| Contribution bp | Sep | Q3 | YTD | 1Yr | 3Yr |
| SGOV | −2.61 | −8.11 | −23.86 | −33.35 | −132.22 |
| TLT | +0.00 | +0.00 | +0.00 | +0.00 | −10.50 |
| LQD | +1.27 | +1.99 | +1.76 | +1.76 | −16.42 |
| SRLN | −1.72 | +8.97 | +10.74 | +18.82 | +92.80 |
| HYG | −13.17 | −9.67 | −1.83 | +4.06 | +112.10 |
| Trading cost difference | +0.00 | +0.01 | +0.03 | +0.04 | +0.02 |
| Active minus static | −16.24 | −6.81 | −13.16 | −8.68 | +45.78 |
| Static minus 50/50 | +161.80 | +229.50 | +237.53 | +241.66 | −69.15 |
| Active minus 50/50 | +145.56 | +222.68 | +224.37 | +232.98 | −23.37 |
| 50/50 minus AGG | −5.92 | +74.97 | +121.56 | +136.88 | +609.53 |
| Active minus AGG | +139.64 | +297.65 | +345.93 | +369.86 | +586.16 |
Basis points of cumulative return difference. Totals use unrounded values. A positive cost difference means the active portfolio incurred less modeled cost than the static policy. Displayed 0.00 values can include immaterial amounts.
What drove the results
September’s 145.56-basis-point advantage over 50/50 combined 161.80 basis points from the static policy structure with a 16.24-basis-point active drag. HYG positioning was the largest active detractor at −13.17 basis points. Lower SGOV exposure detracted 2.61 basis points, and the SRLN overweight cost 1.72. Lighter LQD exposure offset 1.27 basis points.
For Q3, SRLN added 8.97 basis points versus static policy, but HYG and SGOV positioning detracted 9.67 and 8.11 basis points. Year to date, the SGOV underweight cost 23.86 basis points, exceeding SRLN’s 10.74-basis-point benefit. The strategy’s benchmark outperformance during these periods should therefore be credited mainly to its policy allocation.
Over one year, the active overlay detracted 8.68 basis points despite positive SRLN and HYG effects. Over three years, HYG added 112.10 basis points and SRLN 92.80, offsetting much of the SGOV underweight’s 132.22-basis-point drag. Net active allocation added 45.78 basis points, while the static policy lagged the 50/50 benchmark by 69.15. The resulting strategy shortfall versus the primary benchmark was 23.37 basis points.
Attribution method
Daily ETF allocation effects compare beginning active and static-policy weights, multiplied by each ETF’s daily total return. Carino logarithmic linking reconciles those effects, plus modeled transaction-cost differences, to the compounded active-minus-static return. Benchmark bridge rows separate the static policy mix from active decisions. They are portfolio-level comparisons rather than security-level attribution against AGG.
Contributions to total model return
Absolute contributions identify which holdings generated gains or losses. Values are percentage points of cumulative model return and include a separate trading-cost row.
| Contribution pp | Sep | Q3 | YTD | 1Yr | 3Yr |
| SGOV | +0.09 | +0.29 | +0.84 | +1.16 | +4.65 |
| TLT | +0.00 | +0.00 | +0.00 | +0.00 | −0.09 |
| LQD | −0.65 | −1.02 | −0.86 | −0.80 | +2.79 |
| SRLN | −0.11 | +0.61 | +0.77 | +1.30 | +6.51 |
| HYG | −0.54 | −0.40 | −0.07 | +0.17 | +4.70 |
| Trading costs | +0.00 | +0.00 | −0.00 | −0.00 | −0.00 |
| Model total | −1.21 | −0.52 | +0.68 | +1.84 | +18.56 |
LQD and HYG cost approximately 0.65 and 0.54 percentage point in September, while SRLN detracted 0.11 and SGOV contributed 0.09. Year-to-date income exposure in SGOV and SRLN contributed 0.84 and 0.77 point, offsetting losses in corporate bonds. Over three years, SRLN was the largest absolute contributor at 6.51 points, followed by HYG at 4.70 and SGOV at 4.65.
Calculation and data notes
The workbook’s tabs named Daily Performance and Daily Signals contain month-end snapshots. Calendar-period returns were calculated from the stored NAV endpoints. For ETF attribution and September 1 holdings, the full daily path was reconstructed from Source Data, recorded quarterly trade decisions and ETF returns. The reconstruction matches every available month-end active, static, 50/50 and AGG NAV to numerical precision.
Static policy is 40% SGOV, 0% TLT, 20% LQD, 25% SRLN and 15% HYG, rebalanced quarterly. The primary benchmark also rebalances quarterly. Both drift between resets. Model transaction costs equal 3 basis points times one-way turnover and reduce the gross daily growth factor. Returns exclude advisory fees and taxes. ETF expenses may already be reflected in the supplied total-return series.
One year covers October 1, 2025–September 30, 2026; three years covers October 1, 2023–September 30, 2026. September and Q3 are calendar periods. Current allocations are verified through September 30 only. The pending October review indicates no trade, and its signal targets should not be presented as executed holdings. Summary-label inconsistencies are explained in the positioning section.
Sources
[1] Supplied ETFFI_5Spot_Model_Updated.xlsx, through September 30, 2026: Source Data, Quarterly Implementation, current positions, signals, NAVs and configuration. FactSet total-return data as supplied in the workbook.
[2] Supplied ETFFI Core Aggressive Model Working Copy (3), through September 30, 2026, for Treasury yields, credit spreads, volatility and inherited input limitations.
[3] Federal Reserve, September 16, 2026 FOMC statement. federalreserve.gov/newsevents/pressreleases/monetary20260916a.htm
[4] BLS, August CPI release and October release schedule. bls.gov/news.release/archives/cpi_09112026.htm
[5] BEA, Personal Income and Outlays, August 2026, released September 30. bea.gov/news/2026/personal-income-and-outlays-august-2026
[6] BLS, August employment release and next-release schedule. bls.gov/news.release/archives/empsit_09042026.htm
[7] Federal Reserve, 2026 FOMC calendar. federalreserve.gov/monetarypolicy/fomccalendars.htm
For research and educational purposes. Results are hypothetical and do not guarantee future performance. Credit and interest-rate risks can cause losses. Portfolio suitability depends on investor objectives and constraints.
