Howard Marks makes two points in Mastering the Market Cycle that should anchor any attempt to invest by the cycle. We can never know where the cycle is going, but we can, and must, know roughly where we stand in it. And the economic cycle and the market cycle are related but not synchronised: the market is itself a leading indicator, usually turning six to nine months before the real economy, and swinging far beyond what fundamentals justify because of the psychology pendulum. Indicator dashboards tell you the phase; valuations and credit conditions tell you how much the market has already paid for it. Both must be read together.
This is the framework behind the Cycle Compass. It uses the standard four phases and a chain of indicators to infer which one we are in.
The four phases
| Phase | Economy | Policy and credit | Psychology |
|---|---|---|---|
| Early recovery | Activity turns up from trough; unemployment still high but peaking; inflation low or falling | Aggressively easy; credit thawing; steep curve | Disbelief. Best risk-reward of the cycle |
| Mid expansion | Broad, self-sustaining growth; earnings breadth widens; capex revives | Normalising to neutral; credit freely available | Growing confidence; corrections get bought |
| Late cycle | Growth positive but decelerating; capacity tight; inflation and wages rise; margins peak | Tightening; curve flattens or inverts; spreads start widening | Euphoria and FOMO; risk highest when it feels lowest |
| Contraction | Output and earnings fall; unemployment rises; defaults climb | Pivot to easing; credit rationed | Capitulation, from which early recovery is born |
No phase announces itself. You infer it from the balance of evidence.
The indicator chain
Leading indicators turn three to twelve months before the economy and are used to anticipate the next phase. Coincident indicators confirm the phase you are in. Lagging indicators confirm the phase you have just left; they are most useful for avoiding false signals (inflation and unemployment peak after the turn) and for timing policy reversals. The discipline: form the hypothesis with leads, confirm with coincidents, use lags only as a cross-check, never as a trigger.
The United States has the cleanest architecture. The Conference Board publishes leading, coincident and lagging composites every month, and the components (yield curve, ISM new orders, initial claims, building permits, credit spreads, core capital goods orders) are all on FRED.
India has no official composite leading index, so you assemble one. The OECD publishes a composite leading indicator for India; RBI research periodically builds coincident indices; and the practitioner’s dashboard that most India macro desks actually watch is:
- Leads: HSBC PMI new orders and export orders; the 10-year G-sec minus the 91-day T-bill; non-food bank credit growth; new project announcements (CMIE); two-wheeler, commercial-vehicle and tractor sales; RBI consumer confidence expectations; OBICUS capacity utilisation (75% sustained has historically triggered private capex); FPI flows and market breadth; non-oil non-gold imports.
- Coincidents: GST collections and e-way bills; IIP and the core sector index; power demand, rail freight, port cargo, FASTag tolls; passenger vehicles, air traffic, UPI; quarterly revenue growth ex-BFSI.
- Lags: CPI and WPI; unemployment; bank asset quality; operating margins and wage growth; the policy rate itself.
Two India-specific calibrations. India’s trend growth is high, so a “contraction” is usually a growth recession (GDP 4–5%, PMI in the low fifties), not negative GDP, and thresholds must be set accordingly. And India’s economic cycle is increasingly domestic (credit, capex, consumption), but its market cycle remains hostage to crude oil, the dollar and FPI flows. US late-cycle stress reaches Indian equities through flows and oil long before it reaches Indian GDP, so the US dashboard must always be read alongside the Indian one.
Sector rotation by phase
Once the phase is known the logic is mechanical: own what benefits from the next six to twelve months of the cycle, not the last six.
- Early recovery: banks and NBFCs, autos, real estate, consumer discretionary, industrials, small and mid-caps, metals. Underweight FMCG, pharma and utilities, which lag, and cash.
- Mid expansion: capital goods and infrastructure, banks (credit growth), IT and global cyclicals, manufacturing, power and energy capex, selective mid-caps. Underweight deep defensives and long-duration bonds.
- Late cycle: energy and commodities, FMCG and staples, pharma and healthcare, utilities, gold, low-volatility and quality factors, large over small. Underweight small and mid-caps, high-beta cyclicals, richly valued growth and leveraged financials.
- Contraction: FMCG, pharma, gold, liquid funds and short gilts, quality and low-vol; then, near the trough, begin accumulating early-recovery cyclicals while the headlines are still terrible.
All of these are available as NSE-listed rupee ETFs, and a core of a broad-market index ETF, gold and cash is held through every phase, sized up or down but never to zero.
The governing principle
Macro is a lean, never a veto. The phase reading scales the sector tilts through a confidence multiplier; a mixed reading means a smaller tilt and a bigger core, not a coin toss. And deployment is tranched: a base tranche now, a conditional tranche while the dashboard holds, and an event tranche kept for a genuine dislocation or a genuine resolution. The dashboard tells you the odds, never the outcome. The goal is to be positioned so that being roughly right is very profitable and being wrong is survivable.
This is an analytical framework for study, not investment advice. I am not a SEBI-registered investment adviser.