The Economic Machine

Phase 7 · Week 57 · ~20 hours

Where it sits. Almost every rupee in your bank account was created by a commercial bank at the moment it made somebody a loan. Not by the RBI. Not by a printing press.

For fifty-two weeks you have worked bottom-up: reading a single company's statements (Phase 1), judging whether to trust them (Phase 2), pricing what they say (Phase 3), grading the moat behind the numbers (Phase 4), learning each sector's dialect (Phase 5), and installing the philosophy and temperament to act (Phase 6). You can now answer the five spine questions about almost any business in isolation. But no business lives in isolation. Every company you will ever value is a small machine bolted inside a vastly larger one: an economy whose credit conditions set the demand for its products, whose interest rates set the discount rate on its cash flows, whose currency sets the value of its exports, and whose debt cycle decides whether the next three years feel like a tailwind or a trap.

That larger machine gets built here from first principles, exactly the way Ray Dalio builds it in How the Economic Machine Works: an economy is nothing but a pile of transactions, and once you see the transaction clearly, money, credit, banking, interest rates, saving, productivity, and the great debt cycles all assemble out of it. You will not learn to forecast the economy. Nobody reliably can, and a running theme of this phase is the humility that follows from that. You will learn to understand it mechanically, so that when the news says "the RBI hiked" or "the yield curve inverted" or "the US current-account deficit widened," you know precisely which gears just turned and what they touch.

This is the first module of the macro phase, and the mechanical foundation the rest of Phase 7 is built on. M7.02 puts the central bank on top of this machine; M7.03 turns the interest-rate lever into the "gravity" on all your valuations; M7.04 runs the debt arithmetic; M7.05 walks the machine through its historical breakdowns; M7.06 connects it across borders; M7.07 makes you build a live dashboard and certifies competency C8. The gate here is a mechanism exam ≥85%.

Suggested hour budget. Core content: 9h · worked examples (two-person credit economy; money creation two ways; S−I=CA; cycle-stage read): 3h · practice set: 4h · applied mini-project (an economy's "machine snapshot" from primary data): 2h · flashcards + mastery check: 2h.

App link. These mechanisms seed the Macro & Cycles branch's first lesson set and supply the mechanism cards ("loans create deposits," "spending = money + credit," "S − I = the current account") to the case engine. They are the conceptual backbone of the app's scenario-events system: every macro event the simulator injects later (a rate hike, a credit crunch, a deleveraging) is one of the gears below moving. They also begin the macro dashboard artifact you will finish and maintain in M7.07.

Learning objectives

By the end you can:

1. Model the economy as transactions: state that total spending = money + credit, that one person's spending is always another person's income, and that price = total spending ÷ quantity, and use these three facts to explain why total spending (not the "money supply" alone) drives short-run activity.
2. Define money precisely and separate it from credit: give money's three functions (medium of exchange, unit of account, store of value), explain why most modern money is bank deposits rather than cash, and articulate exactly how credit differs from money. Credit is a claim on future spending that must be created, serviced, and extinguished.
3. Explain credit and debt as claims on future spending: show why credit accelerates spending today (spending can exceed income) and constrains it tomorrow (spending must fall below income to repay), and distinguish self-liquidating "good" credit from consumption "bad" credit using the same ROIC-versus-cost logic you learned firm-by-firm.
4. Explain how banks create money: both models, fairly. Work the textbook money-multiplier (reserves → deposits, multiplier = 1 ÷ reserve ratio) and the modern endogenous-money correction ("loans create deposits"; Bank of England, 2014), state which the profession now favors and why, and reconcile the two on the same numeric example.
5. Explain interest rates as the master price: describe how the rate clears the supply of saving against the demand for borrowing, why it sits inside every valuation you build, and trace how a change in the rate ripples through borrowing, spending, and asset prices.
6. Derive and interpret the saving–investment identity: obtain S = I in a closed economy from the national-income identity, extend it to S − I = the current-account balance in an open economy, and read what a deficit or surplus actually means (a country importing or exporting saving). That reading is the bridge to M7.06.
7. Explain productivity as the only long-run driver: give the Solow intuition (diminishing returns to capital → long-run growth per worker must come from technology) and the endogenous-growth correction (ideas and human capital are produced inside the economy and need not run out), and locate productivity as the slow trend line under the cycles.
8. Assemble Dalio's three forces: layer the productivity trend, the short-term debt cycle (~5–8 years, run by credit availability and the central bank), and the long-term debt cycle (~50–75 years, debt building until deleveraging), name the four levers of a deleveraging and what makes one "beautiful," and describe (awareness level) the reserve-currency Big Cycle.

Prerequisites & connections

Builds on. M0.03 (the financial-literacy bootcamp: simple vs compound interest, real vs nominal, compounding, and the assumption throughout is that you can move fluently between a rate and its effect over time). M3.01 (time value of money: "a claim on future spending" is a future cash flow, and its present value is what an interest rate discounts; the whole logic of credit rests on the TVM you already own). M3.02–M3.03 (risk, return, and the cost of capital: the "master price" below is the risk-free rate that anchors every WACC you have built). M1.01 (double-entry: the money-creation section is nothing but bank-balance-sheet bookkeeping, and you will read those T-accounts here without blinking). M5.01–M5.02 (banks and NBFCs: you already know a lender's balance sheet; here you see what the whole banking system does to the money supply).

Feeds forward. M7.02 (central banks) puts a controller on the master price and the reserve supply described here. M7.03 (inflation, rates, the yield curve) turns "rates as the master price" into "rates as the gravity of asset prices." M7.04 (fiscal policy and sovereign debt) runs the r vs g arithmetic that governs the long-term debt cycle introduced here. M7.05 (cycles, bubbles, crises) is the short- and long-term debt cycles breaking down in real history (Minsky, 1929, 2008, Japan). M7.06 (FX, flows, the dollar) begins from the S − I = current account identity you derive here and from the reserve-currency Big Cycle. M7.07 (India macro + the dashboard) has you read all of this as live, rate-of-change data. Beyond Phase 7, this machine explains the demand line on every company model you will ever build: the credit cycle is a bank's whole world, the rate cycle is a capital-goods firm's, and the currency is an exporter's.

The one-sentence version of this module. An economy is a pile of transactions in which total spending equals money plus credit and each person's spending is another's income; credit (a claim on future spending, mostly created by banks the moment they lend) lets spending outrun income for a while and so produces cycles around a slow trend line that only productivity can raise, and layering the productivity trend with a ~5–8-year short-term debt cycle and a ~50–75-year long-term debt cycle (Dalio's three forces) is the whole mechanical story of how the world economy works.

Core teaching content

1. The economy is a pile of transactions

Start with the smallest possible unit and refuse to look away from it, because everything else is built by stacking it up. A transaction is one buyer giving money or credit to one seller in exchange for goods, services, or a financial asset. That is the atom of the entire economy. A cup of coffee bought with a ₹50 note is a transaction. A ₹40,000 crore acquisition paid for with cash and stock is a transaction. A government paying salaries, a bank buying a bond, a household taking a car loan: all transactions. There is nothing else. When people say the economy is complicated, they mean there are a great many transactions and they interlock; they do not mean the parts are exotic. Dalio's central pedagogical move, and it is the right one, is to insist that "the economy is simply the sum of the transactions that make it up," and that if you understand the transaction you can understand the whole by aggregation.

Now aggregate. Add up every transaction for a good, or a market, or a country, over a period, and you have total spending. Add up the goods and services side and you have total output (this is what GDP measures: the market value of final goods and services produced). Two facts fall straight out of the atom and they will do an astonishing amount of work for you:

Fact 1: Total spending = money + credit. A buyer can pay in exactly two ways: with money (settling the transaction on the spot, with finality) or with credit (a promise to pay later). There is no third way. So in any market, over any period:

Total spending = (total money spent) + (total credit spent)

This is why, in Dalio's phrase, "credit is the most important part of the economy and probably the least understood." Most people, watching the news, track money: the RBI's balance sheet, the Fed "printing." But in a credit-based economy the swings in credit dwarf the swings in money. When banks and households are eager to lend and borrow, credit surges and total spending surges with it, even if the quantity of money barely moves. When they pull back, credit collapses and spending collapses, again with money roughly unchanged. If you want to understand why the economy sped up or slowed down over the last year or two, look first at credit, not at money.

Fact 2: One person's spending is another person's income. The ₹50 you spend on coffee is ₹50 of the café's revenue. Its revenue pays a barista's wage, which the barista spends, becoming someone else's income, and on around the loop. Spending and income are the same rupees seen from the two ends of a transaction. This sounds trivial and is not, because it means spending is self-reinforcing in both directions: when I spend more, your income rises, so you spend more, so someone else's income rises. That is the multiplier. Run it in reverse: when I cut spending, your income falls, so you cut spending, so a third person's income falls, which is the vicious circle of a recession. An economy does not have a "spending problem" and a separate "income problem"; it has one problem seen twice.

Fact 3: Price = total spending ÷ quantity. For any good, the price is just the money-plus-credit spent on it divided by the quantity sold. Price = spending / quantity. This is the seed of inflation. If spending on a fixed quantity of goods rises, because credit surged, say, prices rise. If the quantity (output) rises to meet the spending, prices need not. So inflation is, mechanically, spending growing faster than the economy's ability to produce. (You will meet the classic shorthand for this, the quantity theory M·V = P·Q, in M7.03; note already that it hides credit inside the "velocity" term V, which is exactly why a pure-money view of inflation keeps getting surprised.)

Hold these three facts. The rest of the module is mostly their consequences.

3. Credit and debt: claims on future spending

Credit is a lender's willingness to hand over purchasing power now in return for a promise of more later. Debt is the same arrangement seen from the borrower's side: the promise. They are two names for one thing, an asset to the lender and a liability to the borrower, and (recall M1.01) they are created together in a single double-entry stroke: the loan appears as the bank's asset and the deposit as the borrower's, in the same instant. Two willing parties can create credit out of nothing but agreement. No prior saving need be moved from a vault. This is the feature that makes credit powerful and dangerous in equal measure.

Why credit accelerates spending today. Without credit, your spending is capped at your income: spending ≤ income. Credit lifts the cap. With borrowing, spending = income + net new borrowing, and net new borrowing can be positive, so spending can exceed income. An economy where everyone's credit is expanding is an economy where everyone is spending more than they earn, and because one person's spending is another's income (Fact 2), everyone's income is also rising. It feels like prosperity, and in the moment it is indistinguishable from prosperity: sales up, wages up, asset prices up, confidence up, which makes everyone more creditworthy, which supports more borrowing. Credit feeding income feeding creditworthiness feeding credit is the self-reinforcing upswing of the cycle. Nothing about real productive capacity has necessarily changed; spending has simply been pulled forward from the future.

Why credit constrains spending tomorrow. Every rupee borrowed must be serviced (interest) and eventually repaid (principal). When the debt comes due, the arithmetic reverses: spending = income − debt service, and debt service is positive, so spending must fall below income to make room for repayment. The same self-reinforcement now runs downhill: less spending means less income means less ability to service debt means distress and default and still less spending. This is why a credit-driven boom is followed by a bust and not a plateau: the borrowing that made spending exceed income guarantees a later period in which spending falls short of it. Credit does not create wealth; it redistributes spending across time. You will prove this to yourself with numbers in Worked Example 1.

Good credit and bad credit: the ROIC test, at the scale of a nation. None of this makes credit bad. It makes credit conditional. The condition is exactly the one you learned company by company: does the borrowed money earn a return above its cost? Borrow ₹10 lakh at 10% to buy a machine that raises your earning power by ₹1.5 lakh a year, and the credit is self-liquidating. The investment generates the income that services and repays the debt, and you end up permanently richer. This is good, productive credit: debt used for investment whose return exceeds the interest, ROIC > cost of debt, the macro echo of the ROIC > WACC master test that has run through this whole program. Borrow the same ₹10 lakh to fund a wedding or a consumption binge, and there is no new income to service it. You have simply pulled future consumption into the present and must repay out of a future income that is no larger. This is bad, unproductive credit. A country whose credit growth funds factories, power, roads, and skills is building the capacity to service that debt; a country whose credit growth funds consumption and asset speculation is building fragility.

When you assess an economy's credit boom in §8, the first question is never "how much?" but "into what?"

5. Interest rates: the master price

An interest rate is the price of money-and-credit across time: what a borrower pays, and a saver earns, for moving purchasing power between the present and the future. Call it the master price because, more than any other single number, it clears the largest market in the economy and reaches into every valuation you will ever make.

It clears saving against borrowing. Picture two crowds. On one side, savers (households and firms with income they would rather lend than spend now) supply funds; the higher the rate, the more they are willing to save and lend (saving slopes up with the rate). On the other, borrowers (households, firms, governments wanting to spend more than their current income) demand funds; the higher the rate, the less they want to borrow (borrowing slopes down with the rate). The interest rate is the price at which the two crowds balance: the supply of saving meets the demand for borrowing. This "loanable funds" picture is the long-run intuition. In the short run, as §4 showed and M7.02 will detail, the central bank overrides the market and sets the short-term rate directly by controlling the price of reserves, which is exactly how it steers the whole machine. Both are true at once: the market pulls the rate toward the saving-borrowing balance; the central bank shoves it around that anchor to manage the cycle.

It sits inside every valuation. You have known since M3.01 that the value of any asset is its future cash flows discounted at a rate: PV = CF / (1 + r)^t. The r in that formula is built on the interest rate. So when the interest rate moves, the present value of every future cash flow moves with it: bonds, stocks, houses, a whole company's DCF. Lower the rate and every future rupee is worth more today, so asset prices rise; raise it and they fall. This is why M7.03 calls interest rates the "gravity" of asset prices: change gravity and everything with weight re-levels. It is also why the longest-duration assets (high-growth stocks whose cash flows are far in the future, long-dated bonds) move most when rates move, a fact that decides winners and losers in every rate cycle.

An analyst who values a business without knowing the rate regime is weighing objects without knowing the gravity.

It is the central bank's main lever on the cycle. Trace the chain, because you will trace it again in the practice set. The central bank cuts the rate → borrowing gets cheaper and saving less rewarding → households and firms borrow and spend more, and asset prices rise (lower discount rate) → the wealth and the spending feed incomes (Fact 2) → the economy accelerates. The central bank raises the rate → the reverse: borrowing dearer, credit slows, asset prices sag, spending and incomes cool. Because credit responds so strongly to its price, the interest rate is the throttle on the short-term debt cycle. Hold this chain; it is the spine of §8's short cycle and of all of M7.02.

Real versus nominal: never confuse them (from M0.03/M3.01). The rate you are quoted is nominal; what matters for behaviour is the real rate, roughly real ≈ nominal − expected inflation (the Fisher relation). A 12% loan in 6% inflation costs a real 6%; a 12% loan in 11% inflation costs a real 1% and is nearly free. This is not pedantry. It is why India's higher-inflation environment demands real-rate thinking, why negative real rates (nominal below inflation) quietly transfer wealth from savers to borrowers and fuel asset bubbles, and why "rates are high" is a meaningless statement until you subtract inflation.

9. Why this matters to an analyst

You will not forecast this machine, and Phase 7 will keep saying so. So why learn it in such mechanical detail? Four concrete uses, each of which touches the bottom-up work you already do:

• The rate regime is gravity on every valuation you build. Your DCFs, your comps, your reverse-DCFs all run through a discount rate anchored on the interest rate (§5). Know the rate regime and you know which way gravity is pulling every asset price, and which businesses (long-duration growth, leveraged, rate-sensitive lenders and developers) feel it most.
• The credit-cycle stage flatters or starves demand and margins. A credit-fuelled boom (§8) lifts sales, pricing, and margins across the economy, and makes cyclical earnings look like trend earnings, the exact trap behind the cyclical-P/E error of M5.09. Reading the credit cycle tells you whether the earnings in front of you are durable or borrowed from the future.
• A specific business's macro sensitivity is now legible. A bank or NBFC is the credit cycle (it creates money by lending, §4; it lives and dies by loan growth and credit costs). A capital-goods or real-estate firm is the rate cycle. An exporter or importer is the current account and the currency (§6). You can now name each business's macro exposure precisely, instead of waving at "the economy."
• Tail risk is a function of the long-term cycle. The difference between an ordinary recession you can underwrite and a deleveraging that impairs capital permanently is where the economy sits in its long-term debt cycle, the single most important thing to know before sizing risk in a late-cycle world (Phase 9's whole concern).

And the governing humility, which is the thesis of this phase and a direct application of the Marks–Soros–Buffett synthesis from Phase 6: the machine is understandable mechanically but not precisely forecastable. The gears are knowable; their timing is not. So the analyst's stance is to be robust, not predictive: to understand the machine well enough to know what could break, where the fragilities are, and how a given business would fare across regimes, rather than to bet the portfolio on calling the next turn. You will meet this principle again in every remaining module of Phase 7, and it is why M7.07 grades your dashboard on reading the present accurately, never on predicting the future.

Common mistakes & how experts think differently

"The central bank prints all the money." The single most common macro misconception, and §4 dismantled it: in a modern economy the great majority of money is bank deposits created by commercial banks when they lend ("loans create deposits"). The central bank creates the monetary base (M0) and sets the price of credit; the banking system creates the broad money (M3). The novice watches the central bank's balance sheet; the expert watches credit growth, because that is what actually drives the money supply and spending.

Getting the multiplier's causation backwards. Even people who know the 1/r arithmetic often believe reserves cause loans: that a bank must receive a deposit before it can lend. The reality is the reverse: lending creates the deposit, and reserves are found afterward at the price the central bank sets. The tell that exposes the error is post-2008 QE: trillions of reserves created, no proportional surge in lending or money, because lending was capital- and demand-constrained, not reserve-constrained. Hold the multiplier as a ceiling-arithmetic device; never as a causal engine.

Treating a current-account deficit as automatically bad. "The country is living beyond its means" is a moralisation of an identity (S − I = CA, §6). A deficit means only that a country invests more than it saves and imports the difference. Whether that is healthy depends on what the capital funds (productive investment vs consumption) and on whether the financing is sticky (FDI) or hot (FPI). These are exactly the ROIC > cost and fragility questions you ask of a firm's debt. Experts interrogate the composition and stability of the flow, not its sign.

Confusing money, credit, and wealth. Printing money does not create wealth (real goods and productive capacity do); credit is not income (it is borrowed against future income); and a rising money supply is not automatically inflation (it is inflation only if spending outruns output, Fact 3). The beginner treats these as interchangeable levers of "more money = richer." The expert keeps the stocks and flows rigorously separate: money settles, credit obligates, productivity enriches.

Reading "S = I" as behavioural advice or a guarantee. S = I is an ex-post accounting identity, not a promise that thrift causes investment or growth. In a slump, everyone trying to save more at once can lower spending, incomes, and thus actual saving: Keynes's paradox of thrift. The identity always holds after the fact; the behaviour of saving can be self-defeating in the short run. Experts distinguish the accounting identity from the behavioural dynamics.

Mistaking a credit-fuelled boom for trend growth. When credit is expanding, spending exceeds income and everything (sales, margins, asset prices, tax revenues) looks structurally strong, when much of it is cyclical and borrowed from the future (§3, §8). Valuing the peak of a credit cycle as if it were the trend is how analysts overpay at tops. The expert normalises to mid-cycle and asks how much of the current prosperity is income versus net new borrowing.

Believing GDP or productivity growth translates one-to-one into stock returns. It does not: share dilution, changes in valuation multiples, and the gap between the listed market and the whole economy all intervene; fast-growing economies have frequently delivered mediocre equity returns and vice versa. Productivity is the long-run ceiling on real earnings growth, not a delivery guarantee to shareholders. Experts hold "the economy" and "the market" as related but distinct.

Over-forecasting the machine. The deepest error is to mistake understanding the gears for the ability to time them. The gears are mechanical and knowable; their timing depends on politics, psychology, and shocks that are not forecastable. The expert uses macro to build robustness, knowing the fragilities and stress-testing a thesis across regimes, not to make confident calls on the next turn. Understanding without humility is more dangerous than ignorance, because it feels like knowledge.

Worked examples

Worked example 1: The two-person (then whole-economy) credit cycle

Objective: prove numerically that credit moves spending through time rather than creating lasting wealth, and see a boom-and-bust emerge from nothing but borrowing and repayment.

The micro version first. Priya earns ₹1,00,000 a year from her work and, with no credit available, spends all of it. Her ₹1,00,000 of spending is the local shopkeeper's ₹1,00,000 of income. The little economy's income and spending both equal ₹1,00,000, year after year, flat. The only way it grows is if Priya becomes more productive and earns more.

Now a bank appears and lends Priya ₹20,000 (she looks creditworthy: steady income, optimistic). In year 1 her spending jumps to income + net new borrowing = 1,00,000 + 20,000 = ₹1,20,000, a 20% boom. The shopkeeper's income rises to ₹1,20,000; feeling prosperous and creditworthy himself, he borrows and spends more, lifting someone else's income: credit → spending → income → creditworthiness → more credit, the self-reinforcing upswing. Nothing about the island's real productive capacity has changed; it merely feels 20% richer.

Come the reckoning. The ₹20,000 loan must be repaid next year with 10% interest, so ₹22,000. Assume Priya's earned income is back at its ₹1,00,000 base. Her spending now = income − debt service = 1,00,000 − 22,000 = ₹78,000, a 35% collapse from the ₹1,20,000 peak, and below her income. The shopkeeper's income falls to ₹78,000, so he spends less, and the downswing reinforces itself exactly as the upswing did. The credit did not make the island permanently richer. It pulled spending forward from year 2 into year 1, and year 2 paid it back with interest. (Had Priya instead borrowed the ₹20,000 to buy a tool that raised her earning power to ₹1,15,000, the loan would be self-liquidating and the growth real, which is the good-credit case of §3.)

The whole-economy version. Scale it up. Let the economy's productive income grow on a steady 2%/year trend (that is the productivity line, §7). Layer credit on top: net new borrowing adds to spending in the boom, and net repayment subtracts in the bust. Using Total spending = productive income + net new borrowing:

Read it. In years 2–4 spending runs above productive income. That is the credit-fuelled boom, and it feels wonderful. In years 5–6 spending runs below income as borrowers repay, which is the bust, and it feels like a recession, even though productive income (the trend line) never stopped its quiet 2% climb. Sum the net borrowing column across the cycle (+10 +12 +8 −12 −8 = +10, still being worked off) and you see the debt is being pulled back toward zero: the cycle redistributed spending across time and left the productivity trend as the only thing permanently standing. This six-row table is the entire short-term debt cycle of §8, and (stretched over decades with debt never fully repaid) the seed of the long-term one.

Worked example 2: Money creation, both models, one number

Objective: work the money multiplier and the endogenous "loans create deposits" view on the same injection, and reconcile them.

The central bank creates ₹1,000 of new reserves by buying a government bond from a non-bank (a pension fund, say), paying with newly created reserve money that lands in the fund's bank account. The reserve requirement is r = 10%. The seller matters: buying the same bond from a bank would be an asset swap that raises reserves by ₹1,000 and broad money by nothing, which is why M7.02's worked example 2 draws the distinction before the multiplier is applied.

Model A: the multiplier. The ₹1,000 lands as a deposit; the bank keeps ₹100 and lends ₹900; the ₹900 is redeposited, ₹90 kept, ₹810 lent; and so on. Total deposits = 1,000 / (1 − 0.9) = 1,000 / 0.10 = ₹10,000; total new loans = ₹9,000; reserves held = ₹1,000. Money multiplier = 1/r = 1/0.10 = 10. Causation in this story: reserves came first and were multiplied outward by successive lending. The central bank "pushed" ₹10,000 of money into being by supplying ₹1,000 of reserves.

Model B: endogenous ("loans create deposits"). Banks lend to whatever creditworthy demand exists, against their capital, creating deposits as they go. Suppose loan demand and bank capital support ₹9,000 of new lending. The banks simply make the loans, and each loan creates its own matching deposit in the same double-entry stroke:

Banking system
  Assets                     Liabilities
  + Loans      ₹9,000         + Deposits    ₹9,000

Broad money rises by ₹9,000 immediately (plus the ₹1,000 from the bond purchase = ₹10,000 of deposits), created by lending, not multiplied from reserves. Then the banks need reserves: to back ₹10,000 of deposits at a 10% requirement they must hold ₹1,000 of reserves, which they obtain from the central bank, which supplies them elastically, at its policy rate, because its job is to hit a rate, not ration a quantity. Causation in this story: loans came first and created deposits; reserves were found afterward at the central bank's price.

Reconcile. Both stories end at the identical balance sheet (₹10,000 of deposits, ₹9,000 of loans, ₹1,000 of reserves) and both respect 1/r as the ceiling. They differ entirely in causation and realism. Model A says reserves drive lending (central bank pushes); Model B says lending drives reserves (banks pull, central bank accommodates at a price). The decisive real-world test, QE creating trillions of reserves without a proportional surge in loans or money, sided with Model B. The 1/r arithmetic is a true ceiling; the causal engine is loan demand, bank capital, and the price of credit. Learn both; believe the second.

Worked example 3: S − I = CA for an open economy

Objective: compute national saving, investment, and the current account, and read what they mean.

An economy reports (₹ lakh crore, illustrative): output Y = 100, consumption C = 60, investment I = 25, government spending G = 20, net taxes T = 18.

• Private saving = Y − T − C = 100 − 18 − 60 = 22.
• Public saving = T − G = 18 − 20 = −2 (a government deficit).
• National saving S = private + public = 22 + (−2) = 20. (Check: S = Y − C − G = 100 − 60 − 20 = 20. ✓)
• The current account = S − I = 20 − 25 = −5, a deficit of 5.
• Cross-check via the spending identity Y = C + I + G + NX → NX = 100 − 60 − 25 − 20 = −5. ✓ Consistent.

Read it. This economy invests 25 but saves only 20, so it imports 5 of foreign saving: a current-account deficit of 5% of GDP, financed by a capital inflow. Two forces drive the external deficit: a robust investment appetite (I at 25% of GDP) and a government running a deficit (public saving −2) that eats into national saving. This is the "twin deficits" flavour (a fiscal deficit tends to widen the external one, all else equal; you will meet the mechanism in M7.04 and M7.06). Whether this is comfortable depends on §6's questions: is the imported 5 funding productive investment (self-liquidating, benign) or a consumption/fiscal binge (fragile)? And is the financing sticky FDI or hot FPI that could reverse in a sudden stop? This is precisely India's structural situation, and precisely why the current account is a headline number on the M7.07 dashboard.

Worked example 4: Reading the short-term-debt-cycle stage from indicators

Objective: practise the M7.07 skill of mapping an indicator panel onto a cycle stage.

An analyst is handed this snapshot of an economy:

Diagnosis: late-cycle expansion, approaching the top of the short-term debt cycle. The signature is unmistakable against §8: a credit-fuelled boom has pushed spending against the economy's capacity (low unemployment, high utilisation), so inflation has risen (Price = spending/quantity); the central bank is therefore raising the master price to cool credit and spending; the flattening/inverting yield curve is the bond market pricing in the slowdown those hikes will cause (M7.03); and spreads just starting to widen mark the very first cracks in credit confidence. What an analyst does with it: turn cautious on the most rate- and credit-sensitive exposures (long-duration growth stocks whose valuations are pure gravity; leveraged balance sheets; lenders facing a coming rise in credit costs; cyclicals whose peak earnings are being mistaken for trend), and remember that in the short-term cycle this downturn is likely curable by later rate cuts, unless the economy is also late in its long-term debt cycle, in which case the usual cure may fail (the question that separated 2007 from an ordinary slowdown). Contrast an early-cycle snapshot: high unemployment falling, credit growth reviving off a low base, rates being cut, curve steepening, spreads narrowing from wide levels. There the same logic points the other way, and you lean into the rate- and credit-sensitive exposures.

Worked example 5: GDP by expenditure, and nominal versus real growth (India, ₹)

Objective: build GDP from the expenditure identity Y = C + I + G + NX, read its composition, and separate the nominal growth that fills headlines from the real growth that living standards actually track.

Aggregate every transaction over a year and sort the spending by who did it, and you have the expenditure measure of GDP: Y = C + I + G + NX, where C is household consumption, I is investment (real capital formation such as factories, machines, housing, inventories), G is government spending, and NX = exports − imports is net exports. It is the same identity that produced S − I = CA in §6, now read forwards. Take one economy for a year (₹ lakh crore, synthetic, India-shaped):

Step 1: assemble GDP. Y = 200 + 105 + 35 + (75 − 80) = 200 + 105 + 35 − 5 = ₹335 lakh crore. The shares (C ≈ 59.7%, I ≈ 31.3%, G ≈ 10.4%, NX ≈ −1.5%) are the fingerprint of a consumption-led but investment-hungry, mildly import-dependent economy: India's actual shape (check the current shares against MOSPI). Note I at ~31%: investment is a smaller slice than consumption but the volatile one, so it drives most of the swing across the cycle even though C is bigger.

Step 2: cross-check against §6. National saving S = Y − C − G = 335 − 200 − 35 = 100. Then S − I = 100 − 105 = −5, exactly NX. The current-account deficit of ₹5 lakh crore (≈ −1.5% of GDP) is not a separate fact. It is the same −5 seen from the other side: the economy invests 5 more than it saves and imports the difference. The expenditure identity and the saving identity are one identity read two ways.

Step 3: nominal versus real. Suppose next year nominal GDP is ₹368.5 lakh crore. Nominal growth = 368.5 ÷ 335 − 1 = 10.0%. But part of that is merely higher prices. If the GDP deflator, the economy-wide price index, rose 4.5%, then real growth ≈ nominal − inflation = 10.0 − 4.5 = 5.5% (the exact figure, 1.10 ÷ 1.045 − 1 = 5.3%, is close; the subtraction is the working approximation, the same Fisher move you used for real vs nominal rates in §5). Only the ~5.3–5.5% real figure reflects more actual goods and services; the other ~4.5 points is the rupee shrinking. Report a "10% GDP boom" without deflating and you have mistaken inflation for prosperity. That is the macro cousin of the real-vs-nominal trap that runs through the whole program, and the reason M7.07's dashboard tracks real growth.

Worked example 6: The debt-service ratio and the rate-reset squeeze (US, $)

Objective: quantify the debt-service ratio, and show how the master price (§5) becomes a spending squeeze through it, the microfoundation of the late-cycle downturn.

Section 3 said spending must fall below income when debts come due (spending = income − debt service). The debt-service ratio (DSR) puts a number on the squeeze: DSR = required debt payments (interest + principal) ÷ income. It is the single best gauge of how much room a borrower, a household or a whole economy, has left. Take a US household (synthetic):

• Disposable income = $80,000/yr; total debt = $200,000; the loans require $8,000/yr of principal repayment.
• Baseline, average rate 6%. Interest = 6% × 200,000 = $12,000; debt service = 12,000 + 8,000 = $20,000; DSR = 20,000 ÷ 80,000 = 25.0%. Income left for everything else = 80,000 − 20,000 = $60,000.

Now the central bank has spent a year raising the master price, and the household's floating-rate debt reprices from 6% to 9%:

• Interest = 9% × 200,000 = $18,000; debt service = 18,000 + 8,000 = $26,000; DSR = 26,000 ÷ 80,000 = 32.5%. Income left for other spending falls to 80,000 − 26,000 = $54,000.

Read it. A 3-percentage-point rate rise added exactly 3% × 200,000 = $6,000 of annual debt service, pushed the DSR up 7.5 points (25.0% → 32.5%), and cut the household's discretionary spending capacity by 6,000 ÷ 60,000 = 10%, with no change in income or in the debt principal. Multiply that across millions of borrowers, and because one person's spending is another's income (Fact 2), you have the mechanism by which rate hikes cool an economy: not magic, just arithmetic on the DSR. A rising aggregate DSR is one of the clearest late-short-cycle warning lights (the BIS publishes household and corporate DSRs by country), and a high starting DSR is exactly what makes an economy fragile heading into a long-term-cycle deleveraging (§8), because there is no slack left to absorb higher rates before defaults begin.

Practice set

Work each problem fully in writing before reading the solution. Numeric answers are graded to the stated tolerance.

Problem 1 (guided). In a market, buyers spend ₹800 crore of money and ₹400 crore of credit on 300 units of a good. (a) What is total spending? (b) What is the average price per unit? (c) Next year money spending is unchanged but credit spending doubles to ₹800 crore while quantity is unchanged; what happens to the price, and what is this called?

Solution. (a) Total spending = money + credit = 800 + 400 = ₹1,200 crore. (b) Price = spending / quantity = 1,200 / 300 = ₹4 crore per unit. (c) New spending = 800 + 800 = ₹1,600 crore; new price = 1,600 / 300 = ₹5.33 crore, up 33%. With quantity fixed, the surge in credit spending raised the price. This is inflation, and note it came from credit, not from any change in the money supply, which is exactly why a money-only view of inflation is incomplete (Fact 1, Fact 3).

Problem 2 (guided). Anil spends ₹500 at Meena's shop. Meena pays ₹300 of it in wages to Raju and spends ₹150 at Sara's stall. State each person's income created by this chain, and name the principle.

Solution. Anil's ₹500 spending is Meena's ₹500 income. Of it, ₹300 becomes Raju's income (wages) and ₹150 becomes Sara's income. The principle is one person's spending is another person's income (Fact 2). The same rupees are counted as spending at one end of each transaction and income at the other, which is why spending is self-reinforcing (the multiplier) in both directions.

Problem 3 (guided). Classify each as money, credit, or neither, and say why: (a) ₹2,000 of notes in your wallet; (b) the ₹50,000 balance in your savings account; (c) a ₹1,00,000 unused limit on your credit card; (d) a ₹10 lakh government bond you own.

Solution. (a) Money: currency, the most liquid medium of exchange (part of M0/M1). (b) Money: a bank deposit, which functions fully as money and is the bulk of broad money (M1–M3); most money is this, not (a). (c) Credit (potential): an unused limit is not money; it is a pre-approved claim on future spending that becomes debt (and creates a deposit) only when drawn. (d) Neither money nor credit-in-use for you: it is a financial asset, your holding of someone else's debt (the government's); it is a store of value but not a medium of exchange, so not money. The lesson: money settles; credit obligates; a bond is a stock of saved wealth, i.e. someone's debt held as your asset (§3, §6).

Problem 4. A banking system faces a reserve requirement of r = 4% (near India's CRR of recent years, check the current level). The central bank injects ₹500 crore of new reserves. Compute (a) the money multiplier and (b) the maximum deposits and (c) maximum new loans the system can support. (d) State the one-sentence reason the actual expansion is usually far smaller.

Solution. (a) Multiplier = 1/r = 1/0.04 = 25. (b) Max deposits = 500 × 25 = ₹12,500 crore. (c) Max loans = deposits − reserves = 12,500 − 500 = ₹12,000 crore (or 500 × (1−r)/r = 500 × 0.96/0.04 = 12,000). (d) In reality banks create deposits by lending to whatever creditworthy demand exists against their capital, so the actual expansion is limited by loan demand, bank capital, and the price of credit, not by the reserve ceiling; the 25× is a maximum almost never reached (the endogenous-money correction, and the QE evidence).

Problem 5. A bank makes a ₹40 lakh business loan to a creditworthy borrower. (a) Write the bank's double-entry at the moment of lending. (b) What happens to the economy's money supply, and by how much? (c) Two years later the borrower repays the ₹40 lakh principal; what happens to the money supply?

Solution. (a) At the instant of lending: Assets: + Loan ₹40 lakh; Liabilities: + Deposit ₹40 lakh (in the borrower's account). (b) The money supply rises by ₹40 lakh immediately, because the new deposit is new money, created by the act of lending ("loans create deposits"); no prior deposit was moved. (c) On repayment both entries are cancelled, loan asset and deposit both disappear, so ₹40 lakh of money is destroyed. Across a whole economy, net repayment (deleveraging) shrinks the money supply, which is why deleveragings are deflationary unless offset by money printing (§8).

Problem 6. The central bank cuts its policy rate sharply. Trace, in order, the effect on (a) the cost of borrowing, (b) household and business borrowing and spending, (c) asset prices, and (d) incomes, then name which of Dalio's three forces you are describing.

Solution. (a) Borrowing gets cheaper (the master price falls) and saving less rewarding. (b) Households and firms borrow and spend more (spending can exceed income via net new borrowing). (c) Asset prices rise, because a lower discount rate raises the present value of all future cash flows (PV = CF/(1+r)^t); rates are gravity. (d) The extra spending and the wealth effect raise incomes (one person's spending is another's income), which supports still more borrowing: the self-reinforcing upswing. This is the short-term debt cycle (~5–8 years), driven by credit availability under the central bank's hand.

Problem 7. A closed economy has Y = 250, C = 160, G = 50. There is no foreign trade. (a) Compute investment I. (b) Compute national saving S. (c) State the identity that links them and what it means.

Solution. In a closed economy Y = C + I + G, so I = Y − C − G = 250 − 160 − 50 = 40. National saving S = Y − C − G = 250 − 160 − 50 = 40. Thus S = I = 40, the closed-economy identity: national saving equals investment, because real investment must be financed by real (deferred-consumption) saving. It is an accounting identity, true ex post, not a claim that thrift causes investment.

Problem 8. An open economy has national saving S = 30 and investment I = 45 (₹ lakh crore). (a) Compute the current-account balance. (b) Is it a surplus or deficit, and what does it mean in words? (c) Name one condition under which this is benign and one under which it is dangerous.

Solution. (a) CA = S − I = 30 − 45 = −15, a deficit of 15. (b) A current-account deficit: the economy invests more than it saves and imports the difference (15) as foreign saving via a capital inflow, so the rest of the world is lending to it / buying its assets. (c) Benign if the imported capital funds productive investment (self-liquidating, ROIC > cost) and the financing is sticky FDI; dangerous if it funds consumption or a fiscal binge and the financing is hot, reversible portfolio money that can stop suddenly (forcing the currency down and rates up). That is India's core external vulnerability, the bridge to M7.06.

Problem 9. Explain, using the Solow idea of diminishing returns to capital, (a) why a poor country can grow fast simply by accumulating capital, (b) why that growth eventually slows even if it keeps investing, and (c) what is required for output per worker to keep rising in the long run.

Solution. (a) A capital-scarce economy has very high returns to new capital (a worker's first machine adds enormous output), so building more capital yields rapid catch-up growth toward the frontier: the mechanics of the East Asian miracles and India/China's high-growth phases. (b) Diminishing returns: each additional unit of capital, with technology fixed, adds less than the last, so accumulation approaches a steady state where new investment only replaces depreciation and output per worker stops rising: the "middle-income slowdown" lens. (c) Long-run growth in output per worker must come from rising technology / total factor productivity A, and, per endogenous-growth theory, from the ideas and human capital (non-rival, potentially increasing-returns) produced by R&D, education, and good institutions.

Problem 10. Distinguish the short-term and long-term debt cycles. For each, state the approximate length, the driver, and whether a recession within it can be cured by the central bank cutting interest rates. Then explain why the long-term cycle eventually can't be cured that way.

Solution. Short-term debt cycle: ~5–8 years; driven by credit availability oscillating under the central bank, which raises rates to cool booms and cuts to revive slumps; recessions can be cured by cutting rates, because debts remain serviceable. Long-term debt cycle: ~50–75 years; driven by debt building up relative to income across many short cycles (each low leaving slightly more debt, rates ratcheting down) until the burden is unsustainable; recessions/deleveragings cannot be cured by rate cuts, because debt-service already consumes too much income and rates have reached ~0% (the central bank is "out of ammunition"). The economy must then deleverage, reducing debt/income via the four levers (austerity, defaults, printing, transfers), ideally balanced into a "beautiful deleveraging."

Problem 11 (timed, ~6 min). You are shown: unemployment high but falling; bank credit growth reviving off a low base; the central bank cutting rates; the yield curve steepening; credit spreads narrowing from wide levels; inflation low. (a) What stage of the short-term debt cycle is this? (b) Name two types of business exposure you would lean into and one you would be wary of, with reasons.

Solution. (a) Early-cycle recovery / early expansion, the mirror of Worked Example 4: the central bank has cut the master price, credit and spending are reviving, and the bond market (steepening curve, narrowing spreads) is pricing recovery and easing risk. (b) Lean into: (i) rate-sensitive, long-duration exposures (quality growth / long-cash-flow businesses) that gain most as the discount rate falls; (ii) credit-cycle-levered lenders (a well-capitalised bank or NBFC) whose loan growth accelerates and credit costs fall as the cycle turns up. Wary of: businesses whose trailing earnings reflect the just-passed downturn and might be mistaken for trend, and highly defensive names that lag in an upswing. But the sharpest caution is reserved for reading an early-cycle within a late long-term debt cycle, where the recovery may be weak and debt-constrained. (Reasoning over a single "right" pick is what earns the marks.)

Problem 12 (timed, ~8 min, synthesis). In no more than eight sentences, trace a full credit expansion and its reversal, from a central-bank rate cut through to the eventual constraint, naming at each step the relevant principle from this module.

Solution (model answer). A central-bank rate cut lowers the master price of credit, so borrowing and spending rise and spending exceeds income via net new borrowing (§3, §5). Because one person's spending is another's income (Fact 2), incomes rise too, which raises creditworthiness and supports still more borrowing: the self-reinforcing upswing of the short-term debt cycle (§8). Banks accommodate by creating deposits as they lend ("loans create deposits," §4), so the money supply and asset prices climb (lower discount rate, §5). Eventually spending presses against the economy's capacity and, since Price = spending/quantity, inflation rises (Fact 3). The central bank raises the master price to cool it; borrowing and spending slow, spending falls below income as debts are serviced, and the downswing reinforces itself into recession (§3, §8). In an ordinary short-term cycle the central bank can cut rates again to cure it, because debts remain serviceable. But if this has recurred for decades with debt ratcheting up (the long-term debt cycle), debt-service and near-zero rates leave the central bank unable to stimulate, forcing a deleveraging, the reduction of debt relative to income via austerity, default, printing, and transfers (§8). Throughout, the productivity trend (§7) is the only line that permanently rose; the rest was spending moved through time.

Problem 13 (guided). New credit-funded spending of ₹100 crore enters an economy where households spend 60% of any extra income (marginal propensity to consume, MPC = 0.6) and save or pay tax on the rest. (a) Trace the first three rounds of induced spending. (b) What is the spending multiplier, and by how much does total spending ultimately rise? (c) If instead MPC = 0.8, what are the multiplier and the total?

Solution. (a) Round 1: the ₹100 cr is spent and becomes ₹100 cr of someone's income; round 2: those recipients spend 0.6 × 100 = ₹60 cr; round 3: 0.6 × 60 = ₹36 cr; and so on, each round 60% of the last. (b) The rounds form a geometric series, so the spending multiplier = 1 ÷ (1 − MPC) = 1 ÷ (1 − 0.6) = 2.5, and total spending rises by 100 × 2.5 = ₹250 crore. (c) With MPC = 0.8, multiplier = 1 ÷ (1 − 0.8) = 5, total = 100 × 5 = ₹500 crore. This is Fact 2 (one person's spending is another's income) turned into arithmetic, with the identical geometric-series structure as the money multiplier of §4, and it is why a credit impulse amplifies rather than merely adds. A higher MPC means less leaks out to saving each round, so the multiplier is larger.

Problem 14 (guided). A US-scale open economy reports (US$ trillion, synthetic): GDP Y = 28, consumption C = 19, government G = 5, net exports NX = −0.9. (a) Use the expenditure identity to find investment I. (b) Compute national saving S. (c) Show that S − I equals NX. (d) State the current-account balance as a percentage of GDP.

Solution. (a) Y = C + I + G + NX ⇒ I = Y − C − G − NX = 28 − 19 − 5 − (−0.9) = 28 − 19 − 5 + 0.9 = $4.9tn. The trap is the sign of NX: because net exports are negative, subtracting them adds 0.9; mishandle the double negative and you get $3.1tn. (b) S = Y − C − G = 28 − 19 − 5 = $4.0tn. (c) S − I = 4.0 − 4.9 = −$0.9tn = NX ✓, so the open-economy identity of §6 holds exactly. (d) CA ÷ Y = −0.9 ÷ 28 = −3.2%, a current-account deficit of about 3.2% of GDP: the US importing saving from the rest of the world, financed by the world's appetite for dollars and dollar assets (§6, §8).

Problem 15. An Indian household (synthetic) earns ₹15,00,000/yr and carries a ₹40,00,000 loan at 9% interest, repaying ₹1,40,000 of principal a year. (a) Compute its debt-service ratio (DSR). (b) The loan is floating; the RBI's hikes push the rate to 11%, so recompute the DSR. (c) If inflation is running at 6%, what is the approximate real interest rate on the loan at the 9% nominal rate? (d) In one sentence, connect the DSR move to spending.

Solution. (a) Debt service = interest + principal = 9% × 40,00,000 + 1,40,000 = 3,60,000 + 1,40,000 = ₹5,00,000; DSR = 5,00,000 ÷ 15,00,000 = 33.3%. (b) At 11%: interest = 11% × 40,00,000 = ₹4,40,000, debt service = ₹5,80,000, DSR = 5,80,000 ÷ 15,00,000 = 38.7%, up 5.4 points from a 2-point rate rise. (c) Real rate ≈ nominal − inflation = 9 − 6 = 3% (Fisher, §5): the "high" 9% nominal rate is a modest 3% in real terms, which is exactly why real-rate thinking matters in India's higher-inflation setting. (d) Because spending = income − debt service, the higher DSR directly shrinks the household's room to spend on everything else: the rate channel of the short-term debt cycle, working one balance sheet at a time.

Problem 16. US nominal GDP rises from $27.0 trillion to $28.35 trillion over a year, while the GDP deflator increases 2.4% (synthetic, US-shaped). (a) Compute nominal GDP growth. (b) Compute approximate real GDP growth. (c) A commentator says "the economy grew 5% last year." What is wrong with that as a statement about output, and what is the correct read?

Solution. (a) Nominal growth = 28.35 ÷ 27.0 − 1 = 5.0%. (b) Real growth ≈ nominal − deflator inflation = 5.0 − 2.4 = 2.6% (exact: 1.05 ÷ 1.024 − 1 = 2.5%). (c) The 5% is nominal: it blends more output with higher prices; only the ~2.5–2.6% real figure measures extra goods and services, i.e. the change in living standards. Quoting the nominal number as "growth" overstates real progress by the full 2.4-point deflator, the same nominal-vs-real error flagged for interest rates (§5), now applied to output. Welfare tracks the real line, not the headline.

Problem 17. A borrower takes ₹10,00,000 at 10% interest (synthetic). (a) In productive use it buys equipment that raises annual earning power by ₹1,80,000, so compute the return on the borrowed money and state whether the credit is self-liquidating. (b) In consumption use it funds a wedding that generates no new income, so what services the debt now? (c) State the general condition, in the ROIC-versus-cost language used above, that separates good credit from bad.

Solution. (a) Return on the borrowed money = 1,80,000 ÷ 10,00,000 = 18%, against a 10% cost of debt; interest is 10% × 10,00,000 = ₹1,00,000/yr, so the investment throws off 1,80,000 − 1,00,000 = ₹80,000/yr after servicing interest. The debt is self-liquidating and the borrower ends up permanently richer. This is good, productive credit. (b) Nothing new services it: the ₹1,00,000/yr of interest (and eventual principal) must come out of an unchanged income, so the borrowing simply pulls future consumption into the present: bad, unproductive credit. (c) Good credit funds an investment whose return exceeds its interest cost (ROIC > cost of debt), so it generates the income to repay itself; bad credit does not. It is the macro echo of the ROIC > WACC master test, and, scaled to a nation, the difference between a credit boom that builds capacity and one that builds fragility (§3, §8).

Problem 18 (timed, ~10 min, synthesis). An open economy reports (₹ lakh crore, synthetic): Y = 300, C = 180, I = 95, G = 40; aggregate debt = 450 at an average interest rate of 7%; nominal GDP is growing 11%. Indicators: bank credit growth high and accelerating, capacity utilisation high, inflation above target and rising, the central bank hiking. (i) Compute net exports NX. (ii) Compute national saving S and confirm S − I = NX; state the current account as a % of GDP. (iii) Compute debt-to-GDP and annual interest as a % of GDP. (iv) With nominal growth g = 11% and the debt's interest rate r = 7%, is debt/GDP tending to rise or fall, all else equal? (v) Name the short-term-debt-cycle stage.

Solution. (i) NX = Y − C − I − G = 300 − 180 − 95 − 40 = −15. (ii) S = Y − C − G = 300 − 180 − 40 = 80; S − I = 80 − 95 = −15 = NX ✓; current account = −15 ÷ 300 = −5.0% of GDP, a deficit (invests 95, saves 80, imports the 15 difference, §6). (iii) Debt/GDP = 450 ÷ 300 = 150%; interest = 7% × 450 = 31.5, so interest/GDP = 31.5 ÷ 300 = 10.5%. (iv) Because g (11%) > r (7%), income (the denominator) grows faster than the debt compounds, so debt/GDP tends to fall even without net repayment: the benign r < g case that M7.04's debt arithmetic formalises (the dangerous regime is r > g). (v) Late-cycle expansion, approaching the top of the short-term debt cycle, since accelerating credit, high utilisation, rising above-target inflation, and a hiking central bank are the Worked-Example-4 signature. Scoring: all five parts correct in time = fluent; ≤3 = redo Worked Examples 5–6 and Problems 13–14 before the mastery check.

Applied mini-project: an economy's "machine snapshot"

Task. Build a one-page "Economic Machine snapshot" of one economy (choose India or the United States) entirely from primary data sources, and use it to locate the economy on Dalio's three forces. This rehearses, in miniature, the M7.07 dashboard capstone and forces you to touch the real data behind every mechanism above.

Data to pull (name the source for each figure, no secondary summaries).

• India: real GDP growth and its trend (MOSPI/NSO, mospi.gov.in); bank credit growth and the money-supply aggregates M0–M3 (RBI DBIE database, dbie.rbi.org.in); the policy repo rate and CRR/SLR (RBI website); CPI inflation vs the 4%±2% target (MOSPI/RBI); the current-account balance as % of GDP and the gross saving and investment rates (RBI, and the Ministry of Finance Economic Survey, which tabulates saving/investment).
• United States: real GDP growth (BEA, via FRED, fred.stlouisfed.org); bank credit and money (M2) growth (FRED); the fed funds target rate (Fed/FRED); CPI and core PCE inflation vs the 2% goal (BLS/BEA via FRED); the current-account balance as % of GDP and the national saving and investment rates (BEA via FRED).

Required contents.

1. The transaction view (spending): state whether credit growth is accelerating or decelerating and what that implies for total spending right now (Fact 1). One line.
2. The master price: the current policy rate, its recent direction, the real rate (policy rate minus inflation), and one sentence on what that gravity is doing to asset prices (§5).
3. The external balance: the current-account balance and the saving–investment gap (S − I), with one sentence on whether the country is importing or exporting saving and how the deficit/surplus is financed (§6).
4. The productivity trend: one line locating the long-run growth trend (frontier-innovation vs catch-up/accumulation) and what that says about durable growth (§7).
5. The three-forces read: name the short-term-debt-cycle stage (early / mid / late expansion or contraction) from the indicator panel (Worked Example 4), a one-line view on where the economy sits in its long-term debt cycle (low-debt-and-room vs high-debt-and-fragile), and one sentence on the Big-Cycle backdrop (reserve-currency status, awareness level).
6. The "so what?" for an analyst: two sentences on what this snapshot implies for a rate-sensitive business and for a credit-cycle-levered business (e.g., a lender) you could name.

Scoring rubric (self-score honestly; pass ≥11/14).

Case Lab: run case XL-7.1 (Reading one economy's machine snapshot) in the app's Cases tab for this module: a guided pass over a single economy's GDP composition, credit growth, real policy rate, and S − I position, ending in a three-forces read that mirrors this mini-project step by step.

Reading & resources

• Ray Dalio, How the Economic Machine Works: the ~30-minute animated video (free on YouTube and economicprinciples.org) and the companion essay "Economic Principles." This is the anchor resource for the module; watch it before you start and again after. The transaction view, "credit is the most important and least understood part," and the three-forces template all come from here. [Free] [Beginner]
• Bank of England, McLeay, Radia & Thomas, Money creation in the modern economy (Quarterly Bulletin 2014 Q1), with its companion Money in the modern economy: an introduction. The free, authoritative, plain-English primary source for endogenous money ("loans create deposits"). Non-negotiable reading for §4. [Free] [Intermediate]
• Ray Dalio, Principles for Navigating Big Debt Crises (2018): free PDF from Bridgewater/economicprinciples.org; the full anatomy of the short- and long-term debt cycles, the four levers, and the "beautiful deleveraging," with 48 historical cases. Read Part 1 ("The Archetypal Big Debt Cycle") now; the case studies feed M7.05. [Free] [Intermediate]
• Ray Dalio, Principles for Dealing with the Changing World Order (2021): the Big Cycle of empires and reserve currencies; the animated video is free, the book is paid. Awareness-level for this module; developed in M7.06. [Video Free / Book Paid] [Beginner→Intermediate]
• Marginal Revolution University (Tyler Cowen & Alex Tabarrok), mru.org: free, excellent short videos; work the Solow model series (for §7) and the money and the banking system series (for §2, §4). The single best free structured macro course. [Free] [Beginner]
• N. Gregory Mankiw, Principles of Economics (or Macroeconomics): the chapters on The Monetary System, Money Growth and Inflation, Saving, Investment, and the Financial System, and Production and Growth (the Solow intuition). The cleanest textbook treatment; note it teaches the multiplier as the base case, so pair it with the BoE paper. [Paid] [Beginner]
• Paul Krugman & Robin Wells, Macroeconomics: strong on the open economy and the S − I = CA / balance-of-payments chapters (§6). [Paid] [Beginner→Intermediate]
• Reserve Bank of India: the "Functions and Working of the Reserve Bank of India" handbook (free PDF) and the DBIE statistical database (dbie.rbi.org.in) for M0–M3, credit growth, the current account, and rates. Your primary India data source here and in M7.07. [Free] [Intermediate]
• FRED (Federal Reserve Bank of St. Louis), fred.stlouisfed.org: free, the primary US data source for GDP, money (M2), credit, the fed funds rate, inflation, and the current account; learn to build and save a chart here now. [Free] [Beginner]
• Robert Solow (1956), "A Contribution to the Theory of Economic Growth" and Paul Romer (1990), "Endogenous Technological Change": the primary papers behind §7; skim for the ideas (diminishing returns to capital; non-rival ideas), not the algebra. [Free] [Advanced]
• Zerodha Varsity — Macroeconomics module (zerodha.com/varsity): free, India-focused primer on money, RBI, inflation, and the external sector; good for grounding the mechanisms in Indian institutions. [Free] [Beginner]
• Niall Ferguson, The Ascent of Money: the reading-spine book for Phase 7; the early chapters on the rise of banking and credit make §2–§4 vivid and historical. [Paid] [Beginner]
• Going deeper: Bank for International Settlements, "Credit-to-GDP gaps" and "Debt service ratios" statistics (bis.org → Statistics) [Free] [Intermediate]. The primary cross-country data behind Worked Example 6 and §8: watch how the household/corporate DSR and the credit-to-GDP gap lead cyclical turning points. Pull the latest India and US readings yourself and verify against the BIS tables.
• Going deeper: Carmen Reinhart & Kenneth Rogoff, This Time Is Different: Eight Centuries of Financial Folly [Paid] [Intermediate→Advanced]. The long-run empirical companion to Dalio's long-term debt cycle (§8): centuries of credit booms, sovereign defaults, and the r-versus-g debt dynamics you met in Problem 18, developed further in M7.04–M7.05.

The Modern Analyst's Addendum

Everything above teaches this skill from first principles, by hand. That is how you learn it, and the mastery check still tests it that way. This addendum shows how a working analyst amplifies the same skill today. It adds; it never replaces. (R1/R10)

AI-Augment this skill

{
  "skill": "The transaction view of the economy, money versus credit, the two models of bank money creation, the interest rate as the master price, the S = I and S − I = CA identities, productivity as the only long-run source of growth, and Dalio's three forces",
  "use": "Macro is the phase where a language model is simultaneously most fluent and most dangerous, and the split is clean: it is good at MECHANISM and terrible at LEVEL. The mechanisms in this module are stable, published and heavily written about — the Bank of England's 2014 paper, the Dalio essay, the national-accounts identities — so a model has read them and can trace a chain or write a double entry competently. What it cannot do is tell you where the economy is right now. Ask it for the repo rate, credit growth, the current-account balance or CPI and it will return a fluent, confident, undated number that looks exactly like the one you needed and may be a year stale. That is the worst failure available in Phase 7, because it is invisible. So: (1) CHAIN-TRACING — give it a shock ('the RBI hikes 50 bp', 'the government widens the deficit by 1% of GDP') and ask for the full transmission chain with the SIGN at each step and the identity that constrains it, then check the chain against §5 and §6 yourself. (2) BALANCE-SHEET DRAFTING — 'write the double entry, both sides, for a bank making a loan / a borrower repaying / the central bank buying a bond from a non-bank', which is §4's whole argument in bookkeeping form and the fastest way to see that reserves and deposits are different things. (3) SECTORAL-BALANCE BOOKKEEPING — hand it C, I, G, NX and T and ask it to close the three balances to zero and name which sector must absorb a change; the arithmetic is exact, so an error is detectable. (4) SOURCE NAVIGATION — 'which RBI Weekly Statistical Supplement table carries scheduled-commercial-bank credit', 'which FRED series is the monetary base'. It is good at where a number lives and unreliable about what the number is.",
  "tools": ["Chat assistants — Claude, ChatGPT for transmission-chain tracing, double-entry drafting and identity bookkeeping", "Coding assistants — for the sectoral-balance engine and the credit-impulse panel below", "Document/PDF AI — for RBI Bulletin and Monetary Policy Report chapters, the Economic Survey's saving-and-investment appendix, and the BoE money-creation paper"],
  "prompt": "Using only the figures I supply, close the three sectoral balances for this economy and show that they sum to zero: Y=<>, C=<>, I=<>, G=<>, T=<>. Report the private balance (S_private − I), the public balance (T − G), the foreign balance (−CA), and the sum. Then state which single sector's balance must change, and by exactly how much, if the public balance improves by <X> and the current account is unchanged. Do NOT supply, estimate or recall any current value for any of these variables, any policy rate, or any inflation rate — if you think a figure is missing, name the RBI or FRED table I should pull it from and stop.",
  "verify": "The identities are the verification, and they are exact, so this is a hard test rather than a plausibility read. Any decomposition a tool hands you must satisfy C + I + G + NX = Y to the last decimal, and (S_private − I) + (T − G) + (−CA) = 0 to the last decimal. If it does not close, the numbers are either invented or stitched together from different vintages — both fatal, and both invisible until you run the check. Separately, and this is the discipline the phase is built on: replace every LEVEL in the output with one you pulled yourself, and write the release date next to it. A macro figure without a date attached is not a figure. The mechanism a model gives you can be checked against §4–§6 in ten minutes; a stale level cannot be checked at all without going to the source, which is why you go to the source first.",
  "diy": "The gate is unaided: you assemble GDP from Y = C + I + G + NX, derive S − I = CA from it on paper, write the bank's two-line double entry at the moment of lending and again at repayment, run the 1/r multiplier arithmetic in both directions, compute a debt-service ratio, and read a short-term-debt-cycle stage from an indicator panel. Nothing in this callout is admissible in that room."
}

Modern Data Analysis

By hand first. You built the six-row credit table and watched spending run above income and then below it; you worked the ₹1,000 injection through both money-creation models to the same balance sheet; you derived S = I and S − I = CA from the spending identity; you computed a debt-service ratio at 6% and at 9%. Those stay in your hands, and the derivation of S − I = CA especially. It is four lines of substitution, and an analyst who can reproduce it on demand will never again treat a current-account deficit as a moral fact.

Today's workflow. Three things change, and none of them is speed. First, the identities become assertions. Written as code, C + I + G + NX = Y and the three-balance sum are assert statements the panel cannot be built without, which turns "I think these numbers are consistent" into "these numbers are consistent or the cell does not run", and it is the single cheapest catch for the commonest real error in macro work, which is mixing vintages (a GDP number from one release with a consumption number from a revised one). Second, the credit read moves from the level to the impulse. Fact 1 says spending tracks money plus credit; §8 says the short cycle is credit availability oscillating. The quantity that matches those claims is not the stock of credit and not even its growth rate but the change in the flow of new credit relative to GDP, the credit impulse, because that is the term that enters the spending identity. A credit stock that keeps growing at a decelerating rate is a drag on spending while still looking like an expansion. Third, the panel is transformed identically every quarter, so that a move in the number is a move in the economy and not a move in your method: same seasonal adjustment, same real-versus-nominal choice, same base year, written once.

Tools & sources (IN + US). pandas for the panel and the identity assertions, numpy for the arithmetic, matplotlib for the three-forces chart; India: the RBI Weekly Statistical Supplement (Table 1 for the RBI's own balance sheet, Table 5 for scheduled-commercial-bank aggregate deposits and bank credit, the money-stock tables for M0/M1/M3), RBI DBIE (dbie.rbi.org.in) for the same series as downloadable history, the RBI Handbook of Statistics on the Indian Economy for long annual runs, MOSPI/NSO for the quarterly and annual national accounts with the expenditure-side split (private final consumption expenditure, gross fixed capital formation, government final consumption expenditure, exports and imports) and for the CPI release, the Economic Survey's statistical appendix for the gross saving and gross investment rates that make S − I computable, and the BIS credit-to-GDP-gap and debt-service-ratio tables, which publish India's private-non-financial-sector DSR on a comparable basis; US: FRED by series id: GDP (nominal), GDPC1 (real), GDPDEF (the deflator that separates the two), M2SL (broad money), BOGMBASE (the monetary base, since the M0/M3 gap of §2 is these two side by side), TOTBKCR (bank credit at all commercial banks, the credit-impulse input), WALCL (the Fed's balance sheet, for the QE test of §4), FEDFUNDS and SOFR (the master price, policy and market), CPIAUCSL and PCEPILFE (headline and core), UNRATE and TCU (the capacity side of Fact 3), GSAVE and GPDI (saving and investment, for S − I), NETEXP, and TDSP (household debt service as a percent of disposable income, which is Worked Example 6 measured rather than synthetic), with BEA NIPA Table 1.1.5 as the primary for the expenditure split and BEA's International Transactions tables for the current account. State plainly which side is free. Every US series above is free, machine-readable and versioned, with a documented vintage. On the India side, RBI, MOSPI, the Economic Survey and BIS are free, but they are PDFs and portal exports rather than a versioned API, and the long, consistent, cross-checked panels (state-level aggregates, harmonised monthly series, anything needing a single reconciled history) are behind CMIE Economic Outlook or CEIC, both paid. That asymmetry is not a footnote; it means an Indian macro panel costs you assembly time that a US one does not, and it is the reason the M7.07 dashboard is specified around a small number of series you can genuinely maintain.

# Sectoral balances and the credit ceiling — recomputed in-session (R3)
def balances(Y, C, I, G, T):
    NX   = Y - C - I - G                      # expenditure identity, closed by construction
    priv = (Y - T - C) - I                    # private balance:  S_private − I
    pub  = T - G                              # public balance:   T − G
    forn = -NX                                # foreign balance:  the rest of the world's
    assert abs(priv + pub + forn) < 1e-9      # the identity, as an assertion
    return NX, priv, pub, forn, priv + pub + forn
print(balances(100, 60, 25, 20, 18))          # WE3:  (-5, -3, -2, 5, 0)
print(balances(100, 60, 25, 20, 20))          # +2 of tax, NX held: (-5, -5, 0, 5, 0)
mult = lambda r, c=0.0: (1 + c) / (r + c)     # c = currency drain (cash held / deposits)
print(f"{mult(0.04):.2f}  {mult(0.04, 0.15):.2f}")        # 25.00  6.05
inc = [100 * 1.02**n for n in range(6)]                   # the productivity trend, §7
nb  = [0, 10, 12, 8, -12, -8]                             # net new borrowing, WE1
spend = [i + b for i, b in zip(inc, nb)]
print(f"{max(spend):.1f} -> {min(spend):.1f} = {min(spend)/max(spend)-1:+.1%}",
      f"| trend {inc[0]:.1f} -> {inc[-1]:.1f} = {inc[-1]/inc[0]-1:+.1%}")   # -17.1% | +10.4%

Verify. Prove the engine against the module's own worked figures before it touches a real economy. Worked Example 2: deposits 1,000 / 0.10 = ₹10,000, loans ₹9,000, reserves ₹1,000, multiplier 10.00×. Problem 4: multiplier 25, maximum deposits ₹12,500 crore, maximum loans ₹12,000 crore. Worked Example 3: private saving 22, public saving −2, national saving 20, S − I = −5, and NX = 100 − 60 − 25 − 20 = −5, the same −5 twice. Worked Example 5: Y = 335, S = 100, S − I = −5, nominal growth 368.5 / 335 − 1 = 10.00%, real growth exactly 1.100 / 1.045 − 1 = 5.26% against the 5.5% subtraction shortcut. Problem 14: I = 4.9, S = 4.0, S − I = −0.9 = NX, current account −3.21% of GDP. Problem 18: NX = −15, S = 80, debt/GDP 150.0%, interest/GDP 10.50%. Worked Example 6: DSR 25.0% → 32.5%, a 7.5-point rise, residual income 60,000 → 54,000, down 10.0%. Every one of those reproduced to the stated precision here. Then verify the thing the engine tells you back that you never typed: the DSR's sensitivity to the rate is debt ÷ income, so this household's 200,000 / 80,000 = 2.50 means every 100 bp of repricing costs 2.50 points of DSR, a one-line summary of how exposed a borrower is, computable for a household, a company, or a country.

Quantitative lens

Three things above are exact arithmetic rather than estimates, which makes them worth computing rather than asserting. Everything below is those same worked figures run through the engine above; the identities are properties of the accounts, and the illustrative paths are properties of the stated inputs, not measurements of any economy.

One sector's surplus is another's deficit, and the sum is zero: always, everywhere, to the last decimal. Take Worked Example 3's economy: Y = 100, C = 60, I = 25, G = 20, T = 18. The private sector's financial balance is S_private − I = 22 − 25 = −3; the government's is T − G = −2; the rest of the world's is −CA = +5. They sum to exactly zero. The same closes for Problem 18's economy, for the US-shaped Problem 14, and for the India-shaped Worked Example 5. Not approximately, not on average, but by construction, because every rupee borrowed by one sector was lent by another. This is S − I = CA from §6 with the private and public halves pulled apart, and it converts a vague argument into arithmetic: there is no configuration in which all three sectors run surpluses, and the phrase "the country should save more" is incomplete until it names which of the other two balances is going to move to accommodate it.

Which turns austerity into a subtraction you can do in your head. Hold Worked Example 3's economy and let the government balance its budget by raising net taxes from 18 to 20, with output, consumption, investment, government spending and net exports all unchanged. The public balance improves from −2 to 0. The foreign balance is unchanged at +5. So the private balance must go from −3 to −5. It deteriorates by exactly the 2 the government improved by, because the sum is fixed at zero and only two sectors were free to move. That is the arithmetic core of the austerity trap Dalio's first lever describes and M7.04 develops: unless the current account improves, a government deficit reduction is a private-sector deficit increase, one-for-one. Note precisely what the identity does and does not say. It says the three must sum to zero, so something absorbs the change; it does not say the private sector will absorb it while output stands still. In practice output and consumption move, which is the paradox of thrift of §6 arriving through the same accounts.

The money multiplier's ceiling collapses the moment you allow people to hold cash. The textbook 1/r assumes every rupee lent returns to the banking system as a deposit. Let a fraction c of deposits be held as currency instead and the multiplier becomes (1 + c) / (r + c). At India's r = 4%, the headline ceiling is 25.00×; at a currency drain of 15% of deposits it is 6.05×, a 75.8% reduction, and at 25% it is 4.31×. At the module's teaching r = 10%, 10.00× becomes 4.60× at the same 15% drain. There is a second, quieter point in the same arithmetic: even the pure 10× ceiling is a limit the re-lending chain approaches slowly. Each round retains 90% of the last, so reaching half the limit takes 7 rounds, 90% of it takes 22, and 95% takes 29. A model that presents the multiplier as a mechanism the central bank operates is describing something that would take dozens of sequential lending decisions to complete even if the causation were right. Refining the formula does not rescue it: (1 + c)/(r + c) is a better ceiling, and §4's point is that the ceiling was never the binding thing. The QE evidence settled the causation, not the constant.

And the cycle's whole visible drama lives in a term that nets to nearly nothing. Run the six-row path above. Spending peaks at 116.0 in year 3, when productive income is 104.0, so spending runs 11.5% above income; it troughs at 96.2 in year 5, when income is 108.2, running 11.1% below. Peak to trough, spending falls 17.1%. Over the same six years the productivity trend rises from 100.0 to 110.4, +10.4%, and never falls in any year. The entire recession is the credit term; the trend line does not participate in it. And the credit term's cumulative contribution across the six years is +10 +12 +8 −12 −8 = +10, still being worked off, which is the module's thesis stated as a number: credit produced a 17-point swing and left behind a residual, while productivity produced all 10.4 points of the durable gain.

Honest limits, and they are the point rather than a disclaimer. An identity constrains; it does not cause. The three balances summing to zero tells you that a change in one must be absorbed by the others, and says nothing about which one moves, in what order, or through what mechanism. The direction of adjustment is economics, and the arithmetic is silent on it. Reading "the government must run a deficit for the private sector to save" out of the identity is the same error as reading "thrift causes investment" out of S = I, and §6 already warns against it. The (1 + c)/(r + c) refinement is a better ceiling and still the wrong model, so do not let a more accurate constant make the multiplier feel more true. The six-row credit path is a hand-built illustration with borrowing numbers chosen to make a shape; it is not fitted to any economy's data and the −17.1% swing is a property of those chosen numbers, not an estimate of a business cycle's amplitude. Most importantly, the question an analyst actually wants answered here (how often does a credit boom of a given size end in a bust of a given size, and with what lead) is a base-rate question that needs the BIS credit-to-GDP-gap and DSR panel across countries and decades. That panel is free and it is named above; until you have pulled it and counted, the frequency is not yours to state, and this addendum does not state one.

Do it in code: build the panel with the two identities as assertions, so a vintage mismatch cannot pass silently; compute the credit impulse rather than the credit level; put the three sectoral balances on one chart with the zero line drawn, because the picture is the argument; and pull every level yourself, with its release date recorded next to it. See G1/MS1.03, G2/DA1.02 and G2/QD1.01.

Where this goes next: galaxy cross-links

• G1 · EC1.03 Economic Growth & Convergence takes the Solow sketch of §5 to its full form, with the production function, the convergence hypotheses and the growth accounting that decomposes a country's record into capital, labour and productivity. It also carries the cross-country evidence that GDP growth and equity returns correlate weakly over decade windows, which is the correction a reader needs beside the within-country, century-scale productivity claim made here.
• G1 · MS1.01 What Money Is: the three functions, commodity to fiat, and the deposits-are-money fact of §2 developed properly, including why a deposit is a claim on a bank rather than a thing you own.
• G1 · MS1.03 Central Bank Operations & Repo: how reserves are actually supplied at a price rather than rationed as a quantity: the corridor, the repo, and the operational proof of §4's causation claim.
• G2 · DA1.02 Pandas Data Wrangling: resampling, percent-change on the right axis, joining series with different frequencies and vintages, and the reindexing discipline that makes a macro panel reproducible rather than merely assembled.
• G2 · QD1.01 Data Pipelines & Vendors: pulling FRED and RBI series by id on a schedule, handling revisions, and knowing which vendor you are paying for what, which is the practical answer to the free-versus-paid asymmetry above.
• G3 · CN1.03 The So-What: turning a machine snapshot into the one sentence a decision-maker can act on, which is the difference between a dashboard and an argument.

Teach it back & journal

Feynman prompt. Write a one-page explainer for a smart 15-year-old titled "Where does money come from, and why do economies boom and bust?" You must convey, without jargon: that spending is money plus credit and that one person's spending is another's income; that most money is made by banks when they lend (a new loan creates a new deposit out of nothing but agreement); that borrowing lets you spend more than you earn for a while but forces you to spend less than you earn later, which is why economies rise and fall in cycles; and that the only thing that makes a country permanently richer is getting better at producing things (productivity), not borrowing. If your page needs the words "endogenous," "multiplier," or "current account," you don't own it yet. Say it in plain words a 15-year-old repeats back correctly.

Journal prompt. Three entries. (1) Before you started this, where did you think money came from: the government printing it, or somewhere else? Write your old mental model in one honest sentence, then the corrected one, and note which fact surprised you most (most people are startled that banks create money by lending). (2) Take one economy you follow (India or the US) and, in three sentences, place it on the three forces as best you can today: where in the short-term debt cycle, roughly where in the long-term one, and what the master price (real policy rate) is doing to asset prices. Do not worry about being right. Worry about using the mechanism correctly; you will grade the reasoning, not the call. (3) Name one company you have studied whose fortunes are especially tied to one gear of this machine (a lender to the credit cycle, an exporter to the current account and currency, a capital-goods firm to rates) and write the one macro variable you should now watch for it: a line you will carry into its next earnings read.

This module's flashcards and mastery quiz are wired into the app: see the node's Quiz and Reviews.