Now the standing instrument. A macro dashboard is a small, fixed set of indicators, grouped by what they measure, that together locate the economy in its cycle. The point of a fixed dashboard is discipline: you look at the same things every month, so you see changes and surprises rather than chasing whatever is in the headlines. The full per-indicator reference is the reusable artifact artifacts/macro-dashboard.md, which you build alongside this module: what each indicator measures, the exact source (FRED code or RBI/MOSPI page), how to read it, what it signals, and the healthy/warning zones. What follows is the method: the grouping, the two lenses, and the mapping to the cycle.
Which indicators belong on a macro dashboard?
The ~18 indicators, grouped. (US/global core, with India add-ons; the artifact gives all the detail.)
| Group | Indicators | What the group tells you | Vintage: release date and revision status |
|---|---|---|---|
| Growth | Real GDP · PMIs (S&P Global / ISM manufacturing & services) · IIP (India) · retail sales | Is the economy speeding up or slowing? PMIs and retail sales are leading/timely; GDP is lagging confirmation. | Heavily revised: GDP revised twice plus annual benchmarks; PMIs and IIP revised, IIP's base rebased. Record the print date and whether it is first, second or final. |
| Labour | Unemployment rate / payrolls · jobless claims (US) · wage growth | How tight is the labour market? Claims lead, payrolls coincide, unemployment lags; wages drive services inflation. | Payrolls revised twice and benchmarked annually; claims revised weekly. Always note which vintage a three-month average is built from. |
| Inflation & policy | CPI · core PCE (US) / core CPI · the policy rate & its implied path · inflation expectations | Is inflation accelerating or cooling, and where is the central bank going? This group sets the gravity. | CPI and PCE revised with seasonal factors each year; index rebasing changes weights, not just levels. Record the base year. |
| Financial conditions | Yield curve (10y–2y and 10y–3m) · the 10-year yield · credit spreads (IG & HY) · financial-conditions / bank-lending-standards surveys | Are markets easy or stressed? The curve and spreads are the best leading recession signals in the whole dashboard. | Market data, so essentially unrevised, but the survey lines (lending standards) are quarterly and dated. |
| Global / FX | DXY & USD/INR · oil & gold · equities & the VIX | The global risk backdrop and the external channel; for India, the oil and rupee lines that light up the whole board (see the current account deficit). | Market data, unrevised; the reference rate is a fix at a stated time, so record which fix. |
| India add-ons | Current account (CAD) · FPI flows · FX reserves · GST collections · credit growth · fiscal deficit vs target | India's external balance, its financing, its formal-activity pulse, and its fiscal: the India macro story as six live numbers. | Balance-of-payments and fiscal lines revised for several quarters; GST and FPI flows near-final on release. |
Advertisement
How do you read a macro dashboard?
The two lenses, and always both. A dashboard read in levels is nearly useless. The market has already priced the level; what moves prices is new information, and new information comes in two forms (this is the payoff of M7.05's "assets are priced off surprises"):
- Rate-of-change: the direction and acceleration of each series. Not "CPI is 4.5%" but "CPI is 4.5% and falling from 5.2%, and the pace of the fall is slowing." The second derivative matters: an economy where growth is still positive but decelerating is a very different regime from one accelerating, and the turn in the rate-of-change usually leads the turn in the level. Read every indicator as a trend and an inflection, not a point.
- Versus-expectations (the surprise): the actual print against what the consensus expected. A CPI print of 4.5% is bullish for bonds if the market feared 5.0%, and bearish if it hoped for 4.0%. Same number, opposite market reaction, because prices already embed the expectation and only move on the gap. This is why economic-surprise indices (e.g., the Citi Economic Surprise Index) exist and why you must, for the market-moving indicators, know both the print and what was expected. The surprise is the signal; the level is old news.
How do you map the dashboard onto the cycle?
Mapping the dashboard onto the cycle: two frames. Once you have read the indicators as change-and-surprise, you place the reading on two maps that together define the "regime":
Frame 1: the growth–inflation 2×2 (the All-Weather / four-quadrants logic from M7.05). Every macro moment is, to first order, a combination of growth accelerating or decelerating and inflation accelerating or decelerating. Four quadrants, each with a characteristic asset winner:
| Inflation FALLING | Inflation RISING | |
|---|---|---|
| Growth RISING | Goldilocks / recovery → risk-on; equities (esp. growth/long-duration), credit, EM. The best quadrant for stocks. | Reflation / overheat → real assets, commodities, value/cyclicals, inflation-linked bonds; rates rising. |
| Growth FALLING | Slowdown / disinflation → recession → long-duration government bonds, quality/defensive equities, cash; the classic risk-off. | Stagflation → the hardest; commodities, gold, cash, real assets; punishes both nominal bonds and most equities. |
The dashboard's growth group + inflation group place you in a quadrant; the financial-conditions and global groups tell you how stressed the transition is; the surprise lens tells you which way the quadrant is about to shift. This 2×2 is the bridge from "what is the economy doing" to "what should I own", and it is exactly what the regime memo (see the regime memo section) has to resolve.
Frame 2: the debt-cycle stage (Dalio, from M7.01/M7.05). Overlay the quadrant onto where in the credit/debt cycle you are, because the same growth–inflation reading means different things early versus late in a cycle:
- Early cycle (recovery): growth turning up from a trough, inflation low, central bank easy, curve steep (bull steepener), spreads narrowing from wide → risk-on, add cyclicals and credit.
- Mid cycle (expansion): growth solid, inflation building, central bank neutral-to-tightening, curve flattening, spreads tight → stay invested, watch for the turn.
- Late cycle (boom/top): inflation high, central bank tightening hard, curve flat or inverting, spreads complacently tight, credit growth hot → de-risk; the inversion is the alarm.
- Recession / deleveraging: growth negative, central bank cutting/easing (QE), spreads blowing out, curve re-steepening (bull steepener) → defensive; long-duration government bonds; wait for the early-cycle turn.
The debt-cycle lens is what stops you from being fooled by a single quadrant read: "growth rising, inflation falling" is Goldilocks early in a cycle (buy) but can be the deceptive calm before the top late in one. Combining the 2×2 with the cycle stage is the difference between a mechanical read and an analyst's read.
A note on India versus the global board. For a US-centric read, the growth/labour/inflation/financial-conditions/global groups are the dashboard. For India you overlay the six India add-ons, and you weight the board differently: oil and the rupee carry more signal (the current account channel), the credit-growth and CAD lines are your external-and-financial-stability tells, GST collections are your best high-frequency formal-activity pulse, and the fiscal-deficit-vs-target print is your discipline check. You run both boards, because India does not live in isolation: a Fed tightening cycle (the global board) transmits to Indian yields, the rupee, and FPI flows (the India board) through the M7.06 channels.
Why do data revisions matter for a dashboard?
One column on the template does nothing until a number moves, and then it does more work than the rest of the board. Every figure on the dashboard arrives at least twice, and several arrive four times. A macro series is not a fact; it is an estimate that gets revised, and the revision is often larger than the signal the first print appeared to carry. A board that records only the latest value has discarded the one piece of information that decides whether last month's read still stands. Four hazards, in the order they bite.
Statistical revisions. US real GDP is published as an advance estimate, then a second and a third, then revised again at the annual and five-yearly benchmarks. US non-farm payrolls are revised in each of the two months after the first print and benchmarked once a year against unemployment-insurance records, and that benchmark has repeatedly moved by hundreds of thousands of jobs. A three-month payroll average built from first prints and the same average built from settled data are different numbers, and they can land on opposite sides of the break-even rate the whole read turns on. Practice item P19 works exactly that case.
Base changes and rebasing. India's consumer price index and index of industrial production are rebased periodically, and a rebasing changes the weights, not only the level. A basket whose food weight falls reports a different inflation rate from identical underlying prices. GDP base-year revisions do the same to every ratio with GDP in its denominator: the debt ratio, the fiscal deficit, the current-account balance. A five-year chart that spans a base change is two series drawn as one unless you say so.
Seasonal adjustment. The factors are re-estimated annually and applied backwards, so an adjusted history changes even when no raw datum did. A month you read as a slowdown can quietly become a month you would have read as noise.
Definition and coverage changes. Series get new sample frames, new deflators and new coverage. India's employment statistics and the treatment of imputed rent in the consumer price index are the standing examples; the level shifts and the pre-change history is not comparable without a bridge (check the current definitions in the source's own methodology note before you compare across a break).
The instrument is the vintage column added to the template above. For every indicator record three things: the release date of the print you are reading, which estimate it is (advance, second, final, benchmarked), and the base year or index vintage. Then one working rule. A conclusion drawn from a first print is provisional until the print settles, and the memo says so rather than implying a confidence the data cannot carry.
One consequence for the regime memo is worth stating plainly. The surprise lens survives revision better than the change lens, because the consensus a print is measured against was formed on the same vintage, whereas a revision to the previous month changes the computed change without any new information about the current one. When the two lenses disagree after a revision, name which one moved and why, and re-read the quadrant before you re-write the stance.