MCH ADVISORY RESEARCHMain site ↗
Macro & Regime OS

Mid-cycle · Cautious / Deteriorating

Positioning verdict: HALT Authoritative · 09 Sep 2026

9 Sep 2026 · engine mrg-1.0.0

Two layers. Regime Gate v1 Authoritative is the rule that governs new positions. Everything marked Shadow · experimental is the Macro & Regime OS: descriptive, unvalidated, and driving no decision until its evidence gates pass. Dotted terms carry a definition.

Three horizons

structural Shadow · experimental
constructive
4 supportive · 1 neutral · 1 adverse — a tally of named dimension states, not a weighted score
tactical Shadow · experimental
cautious
3 supportive · 1 neutral · 2 adverse — a tally of named dimension states, not a weighted score
trading Authoritative
halt
Regime Gate v1 verdict, verbatim — the authoritative layer. This shadow engine does not compute a trading call of its own.

These may disagree without contradicting one another: a constructive multi-year backdrop and a cautious next-fortnight stance are different questions. Composite score: unvalidated — none published.

What changed Shadow · experimental

The six dimensions read credit VERY_LOOSE, growth EXPANDING, inflation STABLE_ELEVATED, internals DETERIORATING, policy RESTRICTIVE, volatility NORMAL. Changed since the last snapshot: internals moved from BROAD_RISK_ON to DETERIORATING. Largest twenty-session moves: Initial jobless claims 206,000 (2nd percentile), -2,000 over twenty sessions; SPY realized vol (20d, annualized) 8.32 (20th percentile), -5.66 over twenty sessions. Data quality: degraded (1 stale, 0 unresolved).

Every change behind that summary (6)
  • regime changeinternals moved from BROAD_RISK_ON to DETERIORATING
  • signal moveInitial jobless claims 206,000 (2nd percentile), -2,000 over twenty sessions — deteriorating
  • signal moveSPY realized vol (20d, annualized) 8.32 (20th percentile), -5.66 over twenty sessions — improving
  • signal moveVIX (daily close history) 14.32 (28th percentile), -0.83 over twenty sessions — stable
  • signal moveICE BofA CCC & lower OAS 10.51 (98th percentile), +0.34 over twenty sessions — deteriorating
  • data qualitywti_momentum is stale and its dimension is marked degraded

Today’s positioning Authoritative

This panel, and only this panel, governs new positions. It is Regime Gate v1: four signals under a fixed rule, unchanged by the shadow engine below.

⚠ HALT REGIME — review checklist Part A3 before any action.

Verdict: HALTNo new LEAPS today; equity only; hedge sleeve 1.5×.

  • New LEAPS allowed? NO
  • New speculative names allowed? No
  • hedge sleeve multiplier: 1.5×
  • Manual drawdown review needed? Yes — run checklist Section A3 if NLV within 5% of any drawdown tier
SignalReadingStatus
SPX vs 200-day MA+7.82% vs 200-day; 200-day risingGREEN
VIX15.72GREEN
Fed funds trajectory3.50–3.75% · neutralYELLOW
NDX breadth (% above 50-day MA)48.5%YELLOW

The rule is fixed and applied without discretion: all four green is normal sizing; one yellow and no red is caution; two or more yellow, or any red, halts new LEAPS. Today: 2 green, 2 yellow, 0 red.

Sources and caveats
  • SPX vs 200-day MAAlpha Vantage (SPY as S&P 500 proxy — AV carries no ^GSPC) · as of 2026-09-08 · SPY 765.96 · 200d MA 710.43 · MA slope +1.47%/month
  • VIXyfinance (^VIX) — documented exception to AV-only; AV carries no ^VIX · as of 2026-09-08 · VIXY rejected as a proxy: futures ETF, roll decay, misclassifies the level
  • Fed funds trajectoryauthored 2026-09-02 — FOMC statement 29 Jul 2026 (federalreserve.gov/newsevents/pressreleases/monetary · as of 2026-09-02 · A hold with an unusually hawkish minority — three hike-dissents and elevated-inflation language. Not a 'turn' (no action, no majority guidance shift), but decidedly not dovish; the cut path is inflation-dependent. Assumes no inter-meeting action Aug 2026 (none in evidence). Goes stale 2026-09-16 — re-author that evening after the SEP/dot-plot meeting.
  • NDX breadth (% above 50-day MA)computed from Alpha Vantage closes · list authored 2026-07-29 · 50 of 103 constituents above their 50-day
  • MCH Pre-Trade & Portfolio Review Checklist, Part A1 (MCH Regime Rule).

The six dimensions Shadow · experimental

Six independent readings. There is deliberately no combined score: a composite would have to earn its authority through registered validation, and none has.

growthexpanding industrial production +1.1% year-on-year; 3-month rate +0.8pp above the 12-month (accelerating), but the level has not reached the 3.0% bar ACCELERATING requires
the rest of the evidence (8)
  • initial claims 206,000, -2,000 over twenty sessions — the only weekly input here and the fastest read of a labour turn
  • unemployment 4.1% at the 20th percentile — tight
  • payrolls +0.3% year-on-year
  • retail sales +5.1% year-on-year
  • housing starts -13.5% year-on-year — the most rate-sensitive component
  • durable goods +13.0% year-on-year
  • PMI/ISM has no authorised source and is not proxied (recorded exclusion); new orders would be the leading indicator this dimension most wants
  • monthly cadence: this state is not expected to turn on a single print
inflationstable elevated core PCE 3.3% year-on-year; 3-month rate -0.3pp against the 12-month — inside the 0.5pp band either way. The level decides, and it sits above the 2.5% line (target 2.0%) · degraded
the rest of the evidence (4)
  • wages +3.1% year-on-year, -0.3pp against core — not providing wage support
  • 10-year breakeven 2.35% — the market's forward expectation and the only daily input in this dimension
  • core CPI 2.5% year-on-year (the faster-published cousin)
  • WTI +14.2 over twenty sessions — the headline channel, which feeds through with a lag and does not move core directly
policyrestrictive 10-year real yield +2.42% at the 99th percentile of its window — the measure of whether policy actually bites
the rest of the evidence (4)
  • real-yield 20-day change -1bp — equity duration pressure broadly flat
  • 2s10s +43bp — positively sloped and broadly unchanged
  • authored Fed stance 'neutral' (target 3.50–3.75%), as of 2026-09-02, 7 days old — carried as evidence; the measured series determine this state.
  • no market-implied policy path exists (no futures/OIS source); any path reading here is INFERRED from front-end moves, not a market probability
creditvery loose CCC 10.51% vs HY 2.65% — ratio 3.97x (the distressed tail moves first and furthest)
the rest of the evidence (4)
  • NFCI -0.56 — financial conditions looser than average (Fed-published, unfitted, 1971-> : the long-history check the three-year OAS window cannot provide)
  • SLOOS: net +0.0% of banks tightening C&I standards — the direct credit-SUPPLY read (quarterly; spreads price credit, this says whether banks extend it)
  • C&I loans +8.7% year-on-year
  • percentile window: 787 obs (≤10y) — FRED serves the ICE BofA series from 2023-08-29 only; this rank is shallow by construction
internalsdeteriorating index +7.8% above its 200-day but participation fell -10pp over twenty sessions — eroding beneath a firm index
the rest of the evidence (6)
  • net new 52-week highs -1.4% of the index — the extremes are contracting
  • NDX 48% above their 50-day (the megacap complex, tactical window)
  • equal weight vs cap: -0.0019 over twenty sessions (favouring safety)
  • small vs large: -0.0034 over twenty sessions (favouring safety)
  • cyclicals vs defensives: -0.0520 over twenty sessions (favouring safety)
  • DESCRIPTIVE ONLY: participation is reported, not claimed predictive — the behaviour engine's participation model was wrong-signed against its pre-registered direction, so no forward claim is made here without its own registered hypothesis
volatilitynormal realized 8.3% vs implied — variance premium +6.0 points (sellers paid)
the rest of the evidence (3)
  • percentile window: 2520 obs (≤10y)
  • VIX term structure, VVIX and MOVE have no authorised source and are not proxied (registered limitation)
  • TESTED AND DESCRIPTIVE: this dimension reports what volatility conditions ARE, not what follows. A registered test (H-VOL-2, 2026-08-31) found the VIX percentile more informative about an equally distant PAST than about the future, even after skipping the 30-day horizon it prices — so no forecasting claim is made here, and none should be read into it.

Growth and inflation Shadow · experimental

GROWTH MOMENTUM →INFLATION MOMENTUM →REFLATIONGOLDILOCKSSTAGFLATIONCONTRACTION
Goldilocks — growth firming as price momentum eases. Growth momentum 5.37, inflation momentum -0.56 percentage points, tracked over 6 snapshots. Coordinates are median three-month-versus-twelve-month momentum across the registered features; the picture describes where conditions have been and forecasts nothing.

Rates and policy Shadow · experimental

Authored Fed stance: on hold, as of 2026-09-02 (7 days old). It is carried as evidence; the measured series determine the reading.

SignalLevel1D5D20DPctlReadData
2s10s curve0.43-0.04-0.0151neutral 2026-09-03
3m10y curve0.880.050.09 2026-09-03
Effective fed funds3.63-0.85 · 2026-08-01
Authored Fed stance · 2026-09-02
10Y real yield (TIPS)2.42-0.030.08-0.0199bad 2026-09-03
5Y real yield (TIPS)2.150.08-0.02 2026-09-03
10Y Treasury4.77-0.020.100.0899bad 2026-09-03
2Y Treasury4.34-0.050.140.0986bad 2026-09-03
30Y Treasury5.250.060.03 2026-09-03
3M Treasury3.890.05-0.0169bad 2026-09-03

Credit and liquidity Shadow · experimental

SignalLevel1D5D20DPctlReadData
Bank credit, all commercial banks5.96%0.25 2026-08-26
HYG/LQD relative strength0.7501-0.00040.0009 2026-09-08
Banks tightening C&I lending standards (SLOOS, net %)0.00%51neutral · 2026-07-01
Commercial & industrial loans outstanding8.66%1.16 2026-07-01
Chicago Fed NFCI-0.56-0.0832good · 2026-08-28
ICE BofA CCC & lower OAS10.51%0.200.3499bad 2026-09-03
ICE BofA US High Yield OAS2.65%-0.010.02-0.064good 2026-09-03
ICE BofA US IG OAS0.81%0.020.0337neutral 2026-09-03

Market internals Shadow · experimental

SignalLevel1D5D20DPctlReadData
XLY/XLP (cyclicals vs defensives)1.3567-0.0520 · 2026-09-08
IWM/SPY (small vs large)0.3847-0.0034 · 2026-09-08
NDX % above 50DMA48.50%-2.00-7.90 2026-09-08
S&P 500 net new 52-week highs (highs minus lows, %)-1.43%-0.41-5.51 2026-09-08
RSP/SPY (equal-weight vs cap)0.2829-0.0019 · 2026-09-08
S&P 500 % above 200DMA63.30%-5.90-10.00 2026-09-08
SPY vs 200-day MA7.82%-0.54 2026-09-08

Volatility and stress Shadow · experimental

  • VIX term structure is not available from any registered source
  • VVIX is not available from any registered source
  • MOVE is not available from any registered source
SignalLevel1D5D20DPctlReadData
VIX minus realized 20d6.052.03 · 2026-09-03
SPY realized vol (20d, annualized)8.32%-1.04-5.6620good 2026-09-08
VIX (spot)15.72 · 2026-09-08
VIX (daily close history)14.32-0.88-0.19-0.8328good 2026-09-03

Growth Shadow · experimental

SignalLevel1D5D20DPctlReadData
Initial jobless claims206,0008,000-2,0002good 2026-08-29
Copper13542.824467.09 · 2026-07-01
Durable goods orders13.00-6.24 2026-07-01
Housing starts-13.48-14.48 2026-07-01
Industrial production1.08-0.30 2026-07-01
Nonfarm payrolls0.29%0.070bad 2026-08-01
Real GDP2.12-0.24 2026-04-01
Retail sales5.110.25 2026-07-01
Unemployment rate4.100.0020good 2026-08-01

Inflation Shadow · experimental

SignalLevel1D5D20DPctlReadData
10Y breakeven inflation2.350.000.040.1079bad 2026-09-04
Core CPI (index)2.47-0.28 2026-07-01
CPI YoY3.37-0.45 2026-07-01
Core PCE (index)3.340.01 2026-07-01
Average hourly earnings3.09-0.25 2026-08-01
WTI crude (momentum)91.487.5814.15 2026-09-01

Upcoming macro risk Shadow · experimental

Thu 10 Sepnothing scheduledlow
Fri 11 Sepnothing scheduledlow
Mon 14 Sepnothing scheduledlow
Tue 15 Sepnothing scheduledlow
Wed 16 Sep
FOMC rate decision + SEP dot plot
very high

Authored macro calendar (data/catalysts/_macro.json). No consensus data is registered from any authorised source, so no consensus or surprise column is shown. US market holidays are not modelled — sessions are weekdays. Sensitivity is an authored mapping, not a computed score.

Historical regime evidencePending validation

Conditional distributions are computed in the internal validation lab and publish only once the defining dimension's hypothesis passes its registered gates.

Portfolio interactionPending validation

Portfolio macro-exposure comparison requires a validated regime state to compare against.

Methodology, data sources and glossary

Known limitations

  • PMI/ISM has no authorised source (recorded factor-registry exclusion)
  • VIX term structure, VVIX and MOVE have no authorised source
  • ICE BofA OAS series carry a three-year window on FRED, so credit percentiles are shallow by construction
  • no market-implied FOMC path exists; policy-path readings are inferences
  • no macro consensus history is registered, so no surprise index is computed
  • US market holidays are not modelled in the event calendar

Data sources

  • Alpha Vantagemacro + prices (daily) · sole market-data vendor
  • FREDmacro series AV does not carry (daily) · documented source exception 2026-08-30; FRED® API, not endorsed by the St. Louis Fed
  • authoredFed stance, macro calendar (authored) · carried with its age

Validation status

Status SHADOW. Registered and awaiting evidence: H-CR-1, H-VOL-1, H-GR-1, H-IN-1, H-CR-2. Passed: none yet. Engine mrg-1.0.0, config mrg-1.0.0 / freg-1.0.0.

Glossary

OAS
Option-adjusted spread — the extra yield a corporate bond pays over government debt of the same maturity, after stripping out the value of any early-repayment options. It is the market’s price for the risk that the borrower does not pay.
HY
High yield — bonds from companies rated below investment grade. They pay more because they default more often, so their spread widens fastest when credit conditions turn.
IG
Investment grade — bonds from companies the ratings agencies judge unlikely to default. The safer half of the corporate bond market.
CCC
The lowest tier of high-yield credit ratings, one notch above default. Because these borrowers are closest to trouble, their spreads move first and furthest when stress arrives.
NFCI
National Financial Conditions Index — a weekly measure published by the Federal Reserve Bank of Chicago that combines over a hundred indicators of how easy or hard it is to borrow. Zero is average; positive is tighter than average.
SLOOS
Senior Loan Officer Opinion Survey — the Federal Reserve’s quarterly survey asking banks whether they are making it harder or easier to borrow. It measures the supply of credit directly, rather than inferring it from prices.
HYG
An exchange-traded fund holding high-yield corporate bonds. Widely traded, so its price reacts within the day to shifts in appetite for credit risk.
LQD
An exchange-traded fund holding investment-grade corporate bonds — the safer counterpart to HYG.
FOMC
Federal Open Market Committee — the Federal Reserve body that sets US interest rates. It meets roughly every six weeks.
Treasury
Debt issued by the US government. Treasury yields are the base interest rate everything else in markets is priced against.
real yield
The interest rate after subtracting expected inflation — what a lender actually earns in purchasing power. It is the honest measure of whether interest rates are restrictive, because a high rate during high inflation may not be restrictive at all.
breakeven
The inflation rate at which an ordinary government bond and an inflation-protected one would pay the same. It is the bond market’s own forecast of inflation, readable every day.
TIPS
Treasury Inflation-Protected Securities — US government bonds whose payments rise with inflation. Comparing them with ordinary bonds is what produces a breakeven rate.
2s10s
The gap between the 2-year and 10-year Treasury yields. Normally the 10-year pays more; when it does not, the curve is inverted, which has historically preceded recessions.
3m10y
The gap between the 3-month and 10-year Treasury yields — the same idea as 2s10s, measured from the very front of the curve.
yield curve
The line joining the interest rates on government debt of different maturities. Its shape says what the market expects interest rates and growth to do.
fed funds
The overnight interest rate US banks charge one another, and the rate the Federal Reserve targets directly. Every other US interest rate is anchored to it.
basis point
One hundredth of a percentage point. A move from 4.00% to 4.25% is twenty-five basis points.
bp
Basis points — hundredths of a percentage point. Twenty-five bp is a quarter of one percent.
pp
Percentage points — the plain difference between two percentages. Three percent rising to five percent is a move of two percentage points.
CPI
Consumer Price Index — the best-known measure of US inflation, tracking what a representative basket of goods and services costs.
core CPI
Consumer price inflation excluding food and energy, whose prices swing for reasons unconnected to the wider economy. It is the more reliable read on the underlying trend.
PCE
Personal Consumption Expenditures price index — the inflation measure the Federal Reserve actually targets. It differs from CPI mainly in how it weights what people buy.
core PCE
The Federal Reserve’s preferred inflation measure: the PCE index excluding food and energy.
IP
Industrial production — a monthly index of what US factories, mines and utilities actually produced. A direct read on the goods economy.
GDP
Gross domestic product — the total value of everything a country produces. The broadest measure of economic activity, and the slowest to arrive.
PMI
Purchasing Managers’ Index — a survey of business purchasing managers, widely used as an early read on activity. This engine does not use one: no authorised source is available.
ISM
Institute for Supply Management — the body that publishes the best-known US purchasing managers’ surveys. Not available to this engine from any authorised source.
claims
Initial jobless claims — the number of Americans filing for unemployment benefits for the first time each week. The fastest honest read on whether the labour market is turning.
YoY
Year on year — comparing a figure with the same month a year earlier, which removes seasonal effects.
annualized
A short-period rate scaled up to what it would be over a full year, so a three-month change can be compared with a twelve-month one.
VIX
The market’s expectation of how much the S&P 500 will move over the next thirty days, derived from option prices. Often called the fear gauge, though it measures expected movement in either direction.
VVIX
A measure of how much the VIX itself is expected to move — volatility of volatility. Not available to this engine from any authorised source.
MOVE
A measure of expected volatility in the US government bond market, the rates equivalent of the VIX. Not available to this engine from any authorised source.
realized vol
How much a price actually moved over a past window, as against how much it was expected to move. Comparing the two says whether volatility was over- or under-priced.
implied vol
How much a price is expected to move in future, as inferred from what people pay for options on it.
variance premium
The gap between expected and actual volatility. When it is positive, sellers of options were paid more than the movement that followed turned out to be worth.
contango
When contracts for later delivery cost more than nearer ones — the normal shape for volatility, reflecting greater uncertainty further out.
backwardation
When nearer contracts cost more than later ones — usually a sign of immediate stress.
drawdown
The fall from a peak to the following trough. It measures the loss an investor would actually have lived through, rather than the return between two chosen dates.
SPX
The S&P 500 index — the five hundred largest listed US companies, and the usual shorthand for the US stock market.
S&P 500
The index of the five hundred largest listed US companies, weighted by market value.
NDX
The Nasdaq-100 index — the hundred largest non-financial companies listed on the Nasdaq, dominated by large technology names.
SPY
The oldest and most heavily traded exchange-traded fund tracking the S&P 500. Used here as the index’s price series.
RSP
An exchange-traded fund holding the same S&P 500 companies but weighting each equally. Comparing it with SPY shows whether gains are broad or concentrated in the largest few.
IWM
An exchange-traded fund tracking the Russell 2000 index of smaller US companies, which are more sensitive to borrowing conditions.
XLY
An exchange-traded fund holding consumer discretionary companies — the things people buy when they feel comfortable.
XLP
An exchange-traded fund holding consumer staples companies — the things people buy regardless. Comparing XLY with XLP is a long-standing read on risk appetite.
ETF
Exchange-traded fund — a fund whose shares trade on an exchange like a single stock.
breadth
How many individual shares are participating in a market move. An index can rise on a handful of large companies while most shares fall, which breadth reveals and the index alone hides.
DMA
Day moving average — the average closing price over a set number of past trading days, used to judge trend. A price above its 200-day average is conventionally an uptrend.
200-day
The average closing price over the last two hundred trading days — the most widely watched measure of a market’s longer trend.
percentile
Where a reading sits in its own history. A ninetieth-percentile value is higher than ninety percent of everything that came before it.
LEAPS
Long-dated options, conventionally those expiring more than a year out. They give exposure to a view over a long horizon for a fraction of the capital.
DTE
Days to expiry — how long an option contract has left to run.
hedge sleeve
The portion of a portfolio held specifically to offset losses elsewhere, sized as a multiple of its normal level.
FRED
Federal Reserve Economic Data — the St. Louis Federal Reserve’s public database of economic statistics.
ALFRED
The archival version of FRED, which stores each statistic as it was originally published, before later revisions. It is what makes an honest historical test possible.
point-in-time
Using data exactly as it stood on the day being studied, rather than as later corrected. Without it, a historical test quietly uses knowledge nobody had at the time.
shadow
Running a model alongside the live system, recording what it would have said, while it drives no decision. It is how a claim earns trust before it is given any.
DATA HOLD
The engine refusing to publish a fresh reading because a required input is missing or stale. Showing yesterday’s answer with today’s date would be worse than showing nothing.

Snapshot history

2026-08-31 · 2026-08-31 · 2026-08-31 · 2026-08-31 · 2026-08-31 · 2026-08-31 · 2026-09-01 · 2026-09-02 · 2026-09-05 · 2026-09-08 · 2026-09-09

This surface is produced by an unvalidated shadow engine and drives no portfolio decision. It is not trade advice. This report is produced under the MCH Regime Rule (Checklist Part A1). It is not trade advice.

MCH ADVISORY SERVICES · All research · Research, not personalised investment advicemchadvisoryservices.com ↗ · Research Standard v4 (decision-level) · Research OS ros-1.20.0 · US-listed