What this site does, in plain English.
Money never sits still. Every day, capital moves between asset classes — out of tech and into utilities, out of bonds and into commodities, out of US stocks and into emerging markets. These shifts are called sector rotations, and they're the single biggest driver of which investments work and which don't in any given year.
Most people only notice rotations after they show up in headlines ("tech sells off," "banks under pressure," "gold breaks to new highs"). By then, the move is mostly over. Circadex tracks 79 global assets in real time so you can see the rotation happening across the whole map at once — not a single ticker in isolation.
You don't need a finance background to use it. The rest of this page explains the three big ideas that make the map readable, then walks through every signal type. Click Open the map whenever you're ready — it'll wait for you.
The core concept Circadex is built around
1 · What a market cycle is
The arc above: the four stages with typical sector leaders and example years from recent history.
A note on time horizons
Business cycle ≠ commodity supercycle
2 · Sector rotation
| Stage | Sectors that typically lead | Why |
|---|---|---|
| Early | Financials · Consumer Discretionary · Industrials | Rates low, credit expanding, consumers return. |
| Mid | Technology · Communication Services | Growth established, investors pay up for durable earnings. |
| Late | Energy · Materials | Demand peaks, inflation pressure rises, commodities benefit. |
| Recession | Staples · Utilities · Healthcare | People still need food, electricity, and medicine; discretionary spending falls. |
Grounded in
- Merrill Lynch / BofA Investment Clock (2004) — the canonical 2×2 of growth × inflation used across institutional asset allocation.
- Sam Stovall (1996) — Standard & Poor's Guide to Sector Investing. The practitioner codification of the US equity sector rotation sequence.
- Fidelity Investments— long-running “business cycle approach to sector investing” research, the most-cited retail-facing version of the model.
3 · How Circadex surfaces rotation
- Flow scoreCurrent position. Where a sector sits on the cycle right now — leading, gaining, neutral, weakening, or lagging.
- RegimeCurrent stage. The macro environment the classifier has detected — the stage the cycle is in.
- RatiosEarly warnings. Cross-asset relationships that historically move before a rotation is obvious (e.g. Gold/Copper rising = growth fears building).
- LeadersFirst movers. Assets that historically move first at a stage transition — e.g. KRE leads XLF by ~4 weeks at credit-contraction onset.
- CompassOne frame. The centroid trail shows capital's weighted centre of mass moving across the four quadrants — literally “the cycle turning”.
Honest caveat. No one sees cycles perfectly in advance. Rotation models describe tendencies, not laws — some cycles skip stages or compress them. Central-bank policy, geopolitical shocks, and technological shifts distort the classical sequence. Circadex surfaces the signals; the judgement is yours.
Sector rotation
Capital cycles between asset classes — usually in a recognisable order.
Regimes (the macro weather)
Markets behave differently depending on inflation, growth, and credit conditions.
Cross-asset ratios as early warnings
The ratio between two related assets often signals stress before the broader market reacts.
Prices stable or falling, growth healthy. The textbook 'Goldilocks' regime.
Often follows: Late credit contraction or fiscal-stimulus reset
Volatility crushed, credit spreads compressed, equities making new highs. Risk appetite stretched.
Often follows: Disinflationary growth running long
Prices rising broadly. Commodities, energy, and inflation-linked assets lead.
Often follows: Late euphoria meeting a supply or demand shock
Slow growth combined with high inflation. Most asset classes under pressure.
Often follows: Inflation persisting while growth fades
A geopolitical event (war, embargo) drives a supply shock that locks in stagflation.
Often follows: An external shock during inflation or stagflation
Lending tightens, spreads widen, banks pull back. Risk assets sell off.
Often follows: Stagflation forcing aggressive rate hikes
Risk assets sell off; capital rotates into defensives and cash. Not yet a recession.
Often follows: The first leg of credit contraction
Government spending props up growth; central banks lose primacy. Equities and commodities can hold up.
Often follows: Recovery from credit contraction with high debt
Capital rotates into sectors benefitting from government spending (defence, energy, infrastructure).
Often follows: Fiscal dominance maturing
The current regime is shown in the top-left of every page. Hover it for the plain-English description.
The Flow Map
An interactive map of 79 global assets — equities, bonds, commodities, currencies, and crypto. Each node shows whether capital is flowing in (green) or out (red). Nodes cluster by asset class so rotations are visible at a glance.
The Context Engine
Explains why money is moving. Ratio signals flag when asset relationships hit historical extremes. Regime classification tells you the macro environment you're in. Geopolitical events show where disruption is concentrated.
Your Position
Track your portfolio alongside global flows. See whether your holdings are aligned with or against current momentum — before the move becomes obvious.
Flow score: price-implied capital momentum. +100 = maximum inflow, −100= maximum outflow. Based on 1-week returns (40%), 4-week returns (30%), and relative sector performance (30%), percentile-ranked against 5 years of the asset's own history.
How to read it
- +93US Semiconductors is leading its equity peer set — top percentile against 5 years of history.
- +10EU Industrials on the positive side but nowhere near extreme — worth watching where it moves next.
- −82US Energy one of the weakest assets in its peer group over the last ~month.
What it isn't
- Not literal ETF share-creation / redemption data (enterprise-priced and lagged).
- Not a prediction — describes what has already happened.
- Not a buy or sell signal.
Grounded in
- Jegadeesh & Titman (1993) — Returns to Buying Winners and Selling Losers, Journal of Finance. The foundational empirical paper on price momentum.
- Asness, Moskowitz, Pedersen (2013) — Value and Momentum Everywhere, Journal of Finance. Cross-sectional momentum across asset classes.
- Mansfield Relative Strength — classic practitioner overlay of price vs benchmark.
- IBD Relative Strength Rating— Investor's Business Daily's 1–99 price-performance ranking; the retail-familiar analogue of our −100 / +100 scale.
- Dorsey Wright RS Matrix — widely used institutional RS tool.
Limitations
- Supply shocks and forced selling can produce momentum without reflecting capital preference.
- Short-term reversal is a real risk near extremes (mean-reversion).
- The 5-year history window means regimes that predate it aren't in the percentile calibration.
A regime is an algorithmic classification of the current macro environment based on inflation, growth, credit conditions, and geopolitical signals. Circadex identifies 9 regime types:
- Disinflationary GrowthGrowth slowing, inflation falling — historically favours bonds and quality equities.
- InflationRising prices, supply or demand driven — commodities, energy, and inflation-linked assets tend to outperform.
- StagflationSlow growth + high inflation — most asset classes under pressure; commodities are partial hedges.
- Credit ContractionCredit tightening, spreads widening — risk assets under pressure; cash and short-duration bonds favoured.
- Fiscal DominanceGovernment spending driving growth — equities and commodities can hold up despite tight monetary policy.
- Late-Cycle EuphoriaLate expansion, compressed risk premiums, low volatility — equities leading but fragility building underneath.
- Correction / RotationRisk assets selling off, capital rotating into defensives and safe havens.
- Fiscal Dominance RotationFiscal stimulus-driven rotation — capital moving into sectors benefiting from government spending.
- Geopolitical Shock StagflationGeopolitical disruption causing supply shocks — safe havens and energy in focus; broad risk-off.
How the classification works
Each snapshot runs through a transparent rule-based scorer. Every regime has a list of signals (e.g. VIX > 25, HY spread > 5%, yield curve inverted) that add weighted points. The regime with the highest normalised score wins; confidence is high when the gap to runner-up is large, low when regimes are close. The full rule logic lives in the open at src/lib/pipeline/classifyRegime.ts.
Grounded in
- Dalio (Bridgewater) — “All Weather” framework & Merrill/BofA Investment Clock. The underlying 2×2 of growth × inflation environments that every regime maps into.
- Estrella & Mishkin (1998) — Predicting U.S. Recessions: Financial Variables as Leading Indicators. Yield-curve inversion as the most robust single recession predictor.
- Ang & Bekaert (2002) — Regime Switches in Interest Rates. VIX + rate dynamics as regime-distinguishing variables.
- Gilchrist & Zakrajšek (2012) — Credit Spreads and Business Cycle Fluctuations, American Economic Review. HY credit spread as a business-cycle / credit-stress signal.
- Sargent & Wallace (1981) / Leeper (1991) — foundational fiscal-dominance literature. Defines when fiscal policy drives the monetary/inflation regime.
- Minsky (1986) — Stabilizing an Unstable Economy; Kindleberger — Manias, Panics, and Crashes. Late-cycle euphoria & credit-fragility peaks.
- Blanchard & Galí (2007) — stagflation / supply-shock regime analysis.
What Circadex adds on top
- The specific 9-regime taxonomy, including sub-divisions (geopolitical-shock stagflation, fiscal-dominance rotation, correction-rotation distinct from credit contraction). Standard frameworks stop at 4; we widen the catalogue for retail legibility.
- Numeric thresholds (e.g. “VIX > 25 scores 25 points for the geopolitical regime”). These are calibrated against historical ranges but are our judgement, not a published method.
- The rule-based weighted-scoring architecture itself. Most academic work uses Markov-switching models; we chose transparent rules so users can audit exactly why a regime was assigned. Source in the repo.
Limitations
- Rules are static. If macro structure changes in ways we didn't anticipate, the classifier may mis-label. Thresholds are reviewed periodically.
- Regimes may lag real-time conditions. The classifier reads the current snapshot; turning-points show up in the data only after they've moved variables past thresholds.
- Regime classifications are descriptive. They name a historical pattern the market is behaving like; they are not predictions of future returns.
A ratio edge tracks the price relationship between two assets — for example, Gold vs Copper, or High-Yield bonds vs Investment-Grade bonds. The ratio value is ranked against its own 5-year history to produce a percentile.
A ratio at the 90th percentile or above is flagged as an extreme high — historically unusual. A ratio at the 10th percentile or below is flagged as an extreme low. Extremes have historically preceded mean-reversion, but are not guaranteed to do so.
How to read a ratio extreme — three layers
- SHOWWhat the number means descriptively — which asset is outperforming which, and by how much vs its own 5-year range.
- CONTEXTWhat has historically tended to follow readings at this level — with honest caveats (sample size, window, hit rate). Not a forecast.
- ASK2–3 prompts the reader should ask themselves. Questions, not instructions.
Circadex surfaces patterns and questions. It doesn't tell you what to do — that's your call, and your responsibility.
A contagion chain models a sequence of historically correlated asset moves — for example, an oil shock that flows through energy costs, to food prices, to emerging market stress.
Stages are tracked as they activate. This is a pattern-recognition tool based on historical correlations, not a causal model. Not all chains complete in sequence, and historical patterns may not repeat.
Certain assets have historically moved before related assets — the leading asset tends to anticipate the direction of the lagging asset by a number of weeks. Circadex tracks 13 such pairs.
When a leader and lagger show strong divergence in flow scores, it is flagged as a signal worth watching. Historical lead times are approximate and vary across market cycles.
In commodity bull markets, capital has historically flowed through sectors in a roughly sequential order: Oil/Energy → Natural Gas → Fertilisers → Grains → Coal → Uranium → Industrial Metals → Silver → Gold. Earlier sectors tend to peak before later ones begin.
The rotation sequence bar on the Sector Rotation page shows where current capital flows sit in this historical pattern. Segments are coloured by each commodity's live flow score. This is a pattern-recognition model, not a prediction — sequences may skip, reverse, or stall.
Secular signals identify structural macro shifts that persist for years rather than months. Circadex tracks 5 such indicators: Real Assets vs Financial Assets, International vs US, Small vs Large Cap, Commodities vs Tech, and Dollar Weakness.
Each signal compares average flow scores between two groups of assets. When more secular signals are active, it suggests a structural regime shift (e.g., from financial-asset dominance to real-asset dominance) rather than a short-term cyclical rotation.
Circadex is an incubation project of Daliah Group B.V. — a market-intelligence tool that aggregates publicly available price and macroeconomic data to identify patterns in capital flows. It does not provide investment advice, recommendations, or signals of any kind. Full legal terms are in the Terms of Service.
Flow scores and regime classifications are algorithmic outputs derived from price data — they reflect observed historical patterns, not predictions of future performance. Past patterns do not guarantee future results.
Circadex is not a licensed financial advisor, broker, or investment manager. Nothing on this platform constitutes investment advice under MiFID II, the Financial Services and Markets Act 2000, or any other regulatory framework.
AI-generated content (weekly briefings) is clearly labelled. It is produced for informational purposes only and must not be construed as advice. This platform complies with EU AI Act Article 52 transparency requirements for AI-generated content.
Data sources: Yahoo Finance, FRED (Federal Reserve), CoinGecko, GDELT. Data may be delayed, incomplete, or inaccurate. Daliah Group B.V. accepts no liability for decisions made on the basis of information presented on this platform.
Always consult a qualified financial professional before making investment decisions.