gen z·economics

Session 19· ·money supply

The Quiet Print: Crisis-Money Without a Crisis

In the year to 31 July 2026 India's money supply grew 14.7% while nominal GDP grew about 9% — and the RBI cut the CRR, bought ₹2 lakh crore of bonds and ran a $10bn swap to help it along, in a year with no crisis to justify any of it. The alibi for the inflation that may follow — Trump, oil, fertiliser — is being written in advance. This session teaches you to read the two RBI tables where the argument has to be settled.

By Harsh
32 min read
  • #money-supply
  • #rbi
  • #crr
  • #liquidity
  • #inflation
  • #india
  • #japan

Session 19 · The Quiet Print: Crisis-Money Without a Crisis

Locked

Sign in with your ISB email to watch this recording.

Macroeconomics with Prof. Tantri
The Quiet Print: Crisis-Money Without a Crisis — opening framing.
Worked example on the board, step by step.
Q&A: a common misconception, and how to reason past it.
browser read-aloud · no download

Here is the misconception this session exists to kill, and this time it is a misconception about the future: when Indian inflation shows up, you will be told it came from outside — Trump’s tariffs, oil, fertiliser, Iran, take your pick. That story will be plausible, loudly repeated, and mostly beside the point. Imported price shocks go up and come down; the professor’s words: “Oil is fine. Oil will go up and down.” What does not come down on its own is money. And in the twelve months to 31 July 2026, India’s money supply grew 14.7% — ₹41.4 lakh crore — against nominal GDP growth of about 9%. That gap alone is not unprecedented; this country ran wider ones through the 2000s boom. What is remarkable is that the central bank spent the year actively widening it: a CRR cut, ₹2 lakh crore of bond buying, a $10 billion swap — the full crisis toolkit, deployed in a year with no crisis to point at. The evidence is not hidden. It sits in two documents the RBI publishes on schedule, and the whole second half of this session was the professor reading them to us line by line, like a bank inspector. First, though, he closed the loop on Japan — because Japan is the cautionary tale of what happens when this class of adjustment is postponed for twenty years.

Japan: the missing adjustment, and a photographed to-do list

Last session built the machine: Japan funding a state-run carry trade with printed reserves at 0%, and the market never charging for the risk. This week’s recap added the one-line insight that should survive in your notes forever — the missing link was the risk premium on the currency itself:

How do you devalue a stock whose returns are fixed? You mark down its price. How do you devalue a money whose interest rate is pinned at zero? You mark down what it buys — inflation. For twenty-five years Japanese households refused to run that adjustment: they expected zero inflation, so they saved, did not demand higher prices, and the whole low-rate, stable-yen, borrow-at-home-invest-abroad machine kept running. That refusal is now ending. Japan has run ~2% inflation for four years — trivial by our standards, seismic for a country calibrated to zero (ask anyone who has eaten in America lately what a few percent a year does: the $25 meal is now $50). The Bank of Japan has hiked to 1.0%, the highest since 1995, with 1.25% likely next; the current government wants low rates and big deficits, which is exactly the mix that leaks out through a falling currency; and Japan — the largest foreign holder of US Treasuries — has discovered that threatening to sell them is negotiating leverage.

Which produced the spectacle of the year. In late July–early August, Japan and the US ran the first joint yen-buying intervention since 1998 — Tokyo deploying something like $70 billion by the class’s account — BoJ money-market data implies ~$59 billion, which is the figure to quote — and Washington a token slice. How do we know Washington’s number? Because Treasury Secretary Bessent brought a handwritten note to a Camp David cabinet meeting, a photographer’s zoom lens did the rest, and the world read his to-do list: “To Do: Buy Japanese Yen $5–10 bil.” (CNBC, The Japan Times, Axios — and here the professor’s reading, which you should carry as interpretation rather than settled fact: the US is not being charitable; it is buying yen so that Japan does not raise yen by dumping Treasuries into a fragile market. Contemporary reporting frames Washington’s motive more conventionally — heading off a competitive devaluation, backing an ally’s currency defence. The Treasury-market rationale was widely discussed and is plausible; it was never confirmed.) The yen moved from ~163 to ~158. Stabilised, for now. The underlying cause — Japanese households finally repricing their own money — has not gone anywhere.

And the tell to watch is the one this class has repeated since INFS: not the Fed, the US 10-year. When the professor first said it, the 10-year was at ~3.9%. The Fed has cut since. The 10-year on the day of this class, 23 August 2026: ~4.7%. Policy rates are what central banks want; long yields are what the world’s savers charge. The cheap Japanese funding that used to flow out and compress everyone’s long rates is going home — uncovered parity collecting its debts, a decade late.

So much for the country that printed into dead inflation expectations. Now the uncomfortable part: what is our central bank doing?

Before the data: a policy win, and a bill nobody explains

Two things opened the class that belong in your notes even though neither is on the syllabus.

First, the professor took a small victory lap, carefully. He had criticised the FCNR window loudly and in public — the argument being that it raised some $50–60 billion of foreign-currency debt to pile into forex reserves that were already in good shape, while quietly exposing India to a sudden capital outflow two or three years out. Then, on 7 August, the RBI said it had no plan to discontinue the scheme; days later it pulled the deadline in from 30 September to 31 August. His framing was exact and worth copying: “I don’t want to claim causality, but at least there was an appreciation that this was not required.” Note the discipline in that sentence. He will not claim a causal effect he cannot identify — the same standard he applies to the Spanish credit register, applied to his own influence.

Second, a genuinely open puzzle, handed to the room. India’s general-government debt is about 82% of GDP — respectable, better than most rich countries, and the number every official comparison leads with. But India’s interest cost is among the highest in the world at ~5% of GDP, against Japan’s ~1% on debt of 130–140%. Put it in rupees and it stops being abstract: the centre pays about ₹15 lakh crore in interest against total tax revenue of ₹34 lakh crore — nearly one rupee in three — and the states add roughly ₹6 lakh crore more.

And he closed off the lazy answer before anyone could reach for it:

His hypothesis is the one Session 18 built the machinery for: the balance sheet is more stretched than the headline says. Entities classified as private in the national accounts are not private — when LIC borrows, when SBI borrows, when the petroleum companies borrow, that is sovereign borrowing in all but name, because none of them can be allowed to default. Consolidate them and the ratios stop looking flattering, and the interest bill starts to make sense.

Which is why the standing homework from last session now has a deadline and a destination. Lakshit has a first pass at the consolidated balance sheet of the Indian state; the professor’s offer, on the record: “I can set up a meeting with RBI or someone, because I don’t think anyone has that kind of data in a consolidated manner.” Followed by the line that explains why this room exists at all — that the output of these Sunday sessions should be people writing in newspapers and presenting to policymakers, and he will make the introductions. A reminder, if one is needed, that the assignments handed out here are not exercises.

Open the RBI’s own PDF

No leaks required here. Go to the RBI’s site → Statistics → Data Releases → the fortnightly “Scheduled banks’ statement of position in India” — the Section 42 return, a consolidated balance sheet of every bank in the country. The professor put it on screen and made us find the rows ourselves. Follow along; this is the useful knowledge part:

Item II, “Liabilities to others.” Demand liabilities: ₹28.76 lakh crore on 25 July 2025 → ₹34.31 lakh crore on 31 July 2026. Time liabilities: ~₹204 → ~₹235 lakh crore. And the sum that matters, item 2A — deposits other than from banks, all scheduled banks: ₹238 → ₹274 lakh crore. (A live catch from the back rows: the professor first read ₹233, the commercial-banks-only figure, against an all-scheduled-banks number — Ayush flagged it, the professor corrected on the spot. Compare like with like. The class runs on exactly this kind of pushback; bring it.)

Row 8, bank credit: ₹185 → ₹225 lakh crore — a ~21.6% jump that runs well ahead of the ~14–16% credit growth reported elsewhere for this period, so carry the pair as read-in-class until both legs are checked on the same return. Loans create deposits — the money side of the Keynesian plumbing — so both sides of the sheet swelled together.

Then cash, from the Weekly Statistical Supplement: currency with the public is ₹42 lakh crore — 12% of GDP, back to pre-demonetisation territory. A year ago it was ~₹37 lakh crore. “We have no idea where this money is or what it is doing — but we are back to that level.”

Now do what the professor did — one addition and one division. Deposits ₹274 + cash ₹42 ≈ ₹316 lakh crore of money in the system, against nominal GDP of — his words on the tape — “around 325 lakh crore last year, now 345.” Checked: MoSPI rebased the national accounts to a 2022-23 base in February 2026, printing FY26 at ₹346.36 lakh crore and FY25 at ₹318.07. His FY26 figure was very nearly exact; his FY25 sits about seven lakh crore high. (Earlier versions of these notes quoted him at ₹350 for FY26. That was our rounding, not his — corrected against the recording.) Money added in one year, on the RBI’s own M3 for the same fortnight: ₹41.4 lakh crore — the professor’s ₹41 was right. The numerator is beating the denominator — by less than the class arithmetic suggests, but it is beating it:

Model №1 — one year on the WSS: the race

Deposits (2A)
Bank credit (row 8)
Currency with public
Nominal GDP

Sit with the toggle for a second. In rupee terms everything grew, which is normal and tells you nothing. In ratio terms money rose — but by how much depends entirely on a denominator the class rounded. On RBI’s own M3 over trailing-four-quarter GDP, the honest reading is 85.9% → 90.5%, a rise of about 4.6 points — less than the eight or nine claimed in class, more than the three that a first, over-eager correction of these notes reported. His 90% level is right; his seven-to-eight-point jump is not. And that level is one India first reached in the pandemic and has not really left since — a plateau revisited, not a breakout. The sharper fact is not the stock but the policy: a CRR cut, ₹2 lakh crore of bond buying and a $10bn swap, in a year with nothing to fix:

Editor's note · what survived the fact-check, and what did not — 24 August 2026

These notes were put through an outside review after publication, then checked against the RBI's own print. Both the professor and the first version of this page got things wrong, in opposite directions. The record, settled.

What the professor got right. For the fortnight ended 31 July 2026 the RBI reports M3 at ₹3,22,81,302 crore — up 14.7% on the year. That is ₹41.4 lakh crore of new money: his ₹41 was accurate. Money-to-GDP on trailing-four-quarter GDP is ~90.5%: his "90%" was accurate too. Read his numbers off the return and they hold.

What he overstated. The jump. On one consistent basis the ratio moved 85.9% → 90.5%, about 4.6 points — not the seven or eight claimed in class. And "unprecedented" does not survive: 14.7% is high, but India ran 17–21% through 2005–08 without an inflation spiral. Nor is the level a new frontier — broad money to GDP stepped up in the pandemic and never came back down. Roughly 90% in 2026 is a plateau revisited.

What this page got wrong, twice. The dial first read "90.3%, +8.8 points" — built by comparing a July-2026 stock on the class's deposits-plus-cash proxy against a March-2025 stock on RBI's M3, over GDP figures (₹350/₹325) that match no published vintage; MoSPI rebased to a 2022-23 base in February 2026, printing FY26 at ₹346.36 lakh crore and FY25 at ₹318.07. Corrected once, it over-shot the other way to "+2.9 points" by reading the wrong July fortnight. It now reads ~90.5% with its uncertainty band visible, because the denominator convention genuinely moves the answer: trailing GDP gives 90.5%, the new-series fiscal-year print gives 93.2% — above the pandemic peak — and the old series 89.7%.

The claim that survives is narrower and stronger. Not unprecedented money growth, but crisis-grade policy activism without a crisis: a CRR cut worth ₹2.5 lakh crore, ₹2 lakh crore of bond purchases and a $10 billion swap, in a 7%-growth year. Harder to dismiss precisely because it does not overreach.

And the case that this is benign, which the argument is owed: money demand rises as a banking system deepens, so a drifting ratio can be the signature of formalisation rather than a warning light. The channels that turn money into inflation are visibly absent — housing up 3–4%, equities flat, and the RBI simultaneously draining ₹2.5 trillion at the short end through VRRR auctions, which is liquidity management, not deficit monetisation. The professor's Friedmanite lean is a testable prediction, and the test is the wage series he says nobody measures. Until it moves, the benign reading fits the data at least as well. Which is the entire reason the Wage Watch exists.

Where did it come from? Two levers, one of them squared

Lever one: the CRR. The Cash Reserve Ratio is the fraction of every deposit a bank must park at the RBI, dead. India ran at 4% for years; in June 2025 the RBI cut it to 3%, delivered in four tranches through September–November — about ₹2.5 lakh crore of reserves released. But the released cash is the smaller half of the story. The CRR sets the ceiling on the multiplier: with 4% reserves, ₹1 of reserve money can become at most ₹25 of deposits; at 3%, ₹33. Cut the CRR and you have not added a bucket of water — you have widened the pipe:

Model №2 — the multiplier machine

Round 1 lending
After 5 rounds
The limit

(Why does “life” run at ~5× rather than 33×? Banks hold buffers above the minimum, some money leaks into cash under mattresses, SLR locks a slice into G-secs. The ceiling is theory; the direction of the ceiling is policy — and in 2025 policy pointed it up. This is also why the professor said restoring money-supply discipline needs more absorption than the original injection once you have touched the CRR: you shrank the sponge and widened the pipe in the same move.)

Lever two: OMOs — the RBI buying bonds with new reserve money. Through late 2025 and January 2026 the RBI ran ₹50,000-crore purchase auctions four times over — ₹2 lakh crore of OMOs announced alongside a $10 billion three-year dollar-rupee swap, a ₹3-trillion package — on top of ~$25 billion of similar swaps through early 2025. In class the professor’s emphasis was the destination: when states borrow heavily and the RBI buys bonds to keep the market orderly, it is financing government borrowing with reserve money — and banks lend that reserve money on, creating loans and deposits (row 8 and item 2A, both of which you just watched jump).

Which brings us to the exchange this website exists for. Your correspondent asked: three months ago we measured ~8–9 lakh crore printed in the six months to April, but ~5–6 of it was absorbed — dollar swaps, due back in three years. If the delta is small, why worry? The professor’s answer, roughly: because you are looking at the gross of one operation, and the WSS shows you the net of everything — after all their injecting and absorbing, deposits still grew ₹36 lakh crore against a ₹20-lakh-crore GDP rise. And the CRR cut sits underneath the netting: shrink the reserve requirement and you must absorb even more just to stand still. Then he handed over the homework:

Accepted. The Money Printer, now running on the Bloomburger Terminal — live counters off the WSS anchors, a money-to-GDP dial, and a dated, sourced ledger of every tap (CRR cuts, OMOs, swaps) against every sponge (VRRRs, dollar sales, swap reversals), filterable by date range back to demonetisation. Play with it after this model, which is the same ledger in miniature:

Model №3 — the tap and the sponge, Jul 2025 → Jul 2026

CRR cut (4→3%)
OMO purchases
$10bn swap (3-yr)
Net primary liquidity
Implied deposit creation
What actually happened

This is an identity, not a mechanism — the most important sentence on this page. Sliding until implied meets actual tells you how much of the deposit growth the liquidity operations could arithmetically support. It does not show they caused it: deposits also grow on credit demand, government spending and capital inflows, none of which appear here. Two further cautions. The ₹5.38 L cr is durable-liquidity impact, and only the OMO leg is genuinely new central-bank money — the CRR cut freed reserves that already existed, so multiplying the whole sum overstates the base. And the swap's ₹0.88 L cr reverses in 2029: a sponge with a due date.

And the ratio’s path, so you can see how unusual the last fifteen months are:

Model №4 — money ÷ GDP: the 90% problem

Shape, not levels. The first five bars sit on older GDP and money vintages and are not strictly comparable with the last two — the same mixing error the outside review caught on the terminal's dial, left visible here rather than hidden. COVID's peak measures anywhere from ~90 to ~94% depending on how you date the GDP collapse against year-end money — the point survives any measurement: that level had a pandemic attached. This one doesn't. Dots, sources and a date filter on the Bloomburger Money Printer.

Fairness requires the caveat the professor attached: the motive is probably decent. Keep rates low, keep investment going — a legitimate aim, no conspiracy needed. And the minutes may yet be vindicated. But state the complaint in its narrow form, because the sweeping one is not true: the RBI’s bulletins and policy statements do carry the monetary aggregates. What is fair to say is that the published debate runs on rates and liquidity operations, and the quantity of money barely features in the reasoning. That is a criticism of emphasis, and it survives. “Nobody discusses it” does not — don’t take the sweeping version into an argument you intend to win.

The two arguments — and the one variable nobody measures

So ₹41 lakh crore of new money: what does it become? The class gave you both honest answers.

The Keynesian answer: if the economy has idle capacity, cheap money becomes quantity. Demand shows up, the unused factory runs a second shift, GDP grows into the money, ratios normalise, everyone declares success. This is the good ending — and note it is the same argument as the multiplier fights of Budget season.

The Friedman answer: if the constraint was never demand — if it is skills, land, logistics, the things a cheque cannot conjure (“even if you have demand, I cannot produce an iPhone tomorrow morning”) — then quantity responds briefly and prices respond durably. Expectations move, wages chase prices, and you are back to 6–7% inflation for two or three years with the RBI hiking into it — the Phillips-curve sessions told you that machinery cannot be exploited twice.

Model №5 — the Keynes–Friedman switchboard

→ becomes output
→ becomes prices

The professor’s stated lean — offered as a worry, not a verdict: “the evidence in the medium and long term is that the response is mostly through prices, and once prices start responding, they tend to be durable.” He was equally clear that he could be wrong and that this is not the insight for sale here. The insight for sale is cheaper and harder: know the quantum. Then judge arguments as they arrive.

Two live refinements from the Q&A, both worth keeping. Shorya, who raises debt for companies, confirmed the ground truth — banks are lending cheaper, secured and unsecured — and asked whether that at least helps the manufacturing push. Answer: it depends entirely on why credit got cheap. If intermediation genuinely improved — UPI, credit data, thinner spreads — that is a real reform and it sustains (the fintech-credit paper is the evidence). If it is printed reserves meeting sticky prices, the discount is a loan from the medium term, and Session 17 showed you — on twenty-three million Spanish loans — what banks quietly do to risk when money is artificially easy. Same lesson from Bharat’s suggestion that banking penetration explains the deposit jump: over ten years, absolutely; over twelve months, no. Levels versus changes — the oldest trap in this course.

What you can now do

Be the sensor — and ask the right question. The last fifteen minutes of the class were spent designing a measurement, and the design matters more than the enthusiasm. The variable that decides between the two endings is not “are wages rising.” It is the reservation wage: the raise below which your people start sending out CVs.

Two refinements he was insistent about. Adjust for productivity — if your revenue per employee is up because people learned to do more, their raise is not inflation; the same raise without the productivity is the problem. And ask crude, personal questions, because the sophisticated ones fail. His critique of the existing surveys is the sharpest thing in the session: IIM-Ahmedabad’s inflation-expectations tracker asks informed people what they forecast, and informed people just tell you the current print. “I’m not interested in his forecast.” The RBI, meanwhile, surveys people who often don’t follow the question at all. What nobody asks is the one thing that predicts the chain: at what number do you walk? Attrition at firms that skipped the hike is the same signal, observed from the outside.

Drashti Shah, already running this inside her own organisation, brought the complication: above ₹10 lakh, people judge a 5–7% hike against school fees rising 12% and call it a pay cut; below ₹10 lakh the same arithmetic bites harder. Inflation is not one number, it is a different number per household. The professor’s answer — build the median profile, or three or four bands with their own baskets, since a representative consumer is exactly what policy already assumes. And when she mentioned she had been interviewing women and found they think about prices differently, he stopped the class: if that holds up rigorously, it is “a new discovery” and a top-journal paper. (There is established work showing people import the inflation of the country they grew up in; whether gender does something similar appears to be open.)

So the Wage Watch now asks his question, not the obvious one: the walk-away threshold, whether the last hike beat productivity, and whether attrition followed where it didn’t. Two minutes, anonymous — no name, no email, no account id. His ambition for it, said out loud: “If you can create a group and launch something, this will be tracked by markets. Imagine if it starts influencing bond prices tomorrow. That will be amazing.”

Advise like you mean it. If money-supply growth converts to inflation, rates go up, not down. The professor’s advice to Shorya’s clients is above, on the record.

Make the monthly habit mechanical. Once a month: Section 42 for deposits and credit, the WSS for cash and reserves, DBIE when you want the series. Or open the Money Printer, where the counters, the dial, and the tap-and-sponge ledger now run continuously — that page is this session, automated.

Read the two China articles before next class — and read them as an argument, not a syllabus. With four minutes left he opened the next topic rather than trailing it, and set the reading as a pair, both in The Economist’s By Invitation slot. They sit thirteen days apart, and the second is a direct reply to the first.

28 July — Gita Gopinath, Pierre-Olivier Gourinchas and Hélène Rey. Two consecutive IMF chief economists and an LBS professor, arguing that the yuan is the wrong lever. They concede it is undervalued. Their claim is that an undervalued currency is what a particular configuration of domestic policy produces: suppress household consumption, let property investment collapse, and persistent surpluses and a weak currency come out the other end as the result, not the cause. Run thin private savings against large fiscal deficits and you get the mirror image. So the exchange rate is a symptom, and treating it as the disease fails twice over. Economically, revaluing by fiat with no change in the policy mix does little in the short run — Chinese export prices are set in dollars and adjust slowly — while cheaper imports in yuan terms mean China absorbs less of what the world produces, which is the opposite of what rebalancing requires. Politically, they read Plaza differently from the folk version: it worked because Japan co-operated and real macroeconomic adjustment followed, America’s fiscal stance included — not because five men signed something in a hotel. Their closing formulation is the one to carry: “When you push for growth-supporting reforms, you have a chance. When you push for an exchange-rate adjustment, you ask for conflict.”

Editor's note · a direction to get right

The class described the failure mode as a nominal appreciation producing a real appreciation that makes things worse. The article says the opposite, and the article is correct: a nominal appreciation unsupported by policy change morphs into a real depreciation, because deflation does the offsetting. The real rate is the nominal rate adjusted for relative prices — if Chinese prices are falling faster than its partners', the price channel eats the nominal move and can overshoot it. Same conclusion (the cure feeds the disease), opposite arrow. Worth fixing in your own notes before you repeat it.

10 August — V. Anantha Nageswaran, India’s Chief Economic Adviser, with P.S. Srinivas. Note who is speaking: India’s own CEA, arguing in a global forum about a third country’s currency. He grants the long-run point, then goes after the step doing the work — the conclusion only follows if savings and investment drive the currency and never the reverse. In China, he argues, the same state-led institutions that suppress consumption also stop the currency adjusting, so the exchange rate is not sitting outside the imbalance; it is one of the channels generating it. In the short run it is a lever in its own right — and the IMF’s own long habit of making devaluation standard structural-adjustment conditionality is the tacit admission that it is.

His empirical case is what to bring to class. China’s goods-trade surplus hit a record $1.2 trillion in 2025, up a fifth on the year, even as exports to America fell sharply — the shortfall more than made good by shipments elsewhere. A hyper-competitive currency forces trading partners to hold their own rates down, exporting the adjustment problem rather than solving it: China’s biggest export of late, in his phrase, has been its deflation. And the reform alternative he finds weaker than it looks — household consumption has sat near 40% of Chinese GDP for two decades, through all of which it was official policy to raise it. Hence the line that lands hardest: “A currency level is visible every day. A reform pledge can be obscured indefinitely.”

He also refuses the symmetry, and this is the argument with the most carry. An American fiscal deficit unwinds inside an electoral cycle, watched by an independent central bank. China’s savings-investment gap is rooted in local-government finance, state-bank balance sheets and the political weight of the tradable sector — slower and costlier to shift. Treating the two sides as equivalent lets the more entrenched one escape scrutiny. And the part that should matter most to this room: developing countries are squeezed from both directions — Chinese excess capacity crowds out their manufacturers while American deficits pull away the capital that would have funded them. His number to argue with: the IMF’s own 2026 External Sector Report puts yuan undervaluation at a midpoint of 21.3%.

Where they agree, which is easy to miss. Both want American fiscal adjustment. Both want Chinese demand-side reform. Both reject the fantasy that a revaluation on its own fixes anything — Nageswaran concedes explicitly that a sharp appreciation taken in isolation would intensify the deflation. The disagreement is about sequencing and leverage: whether the currency is the thing you ask for, or the thing you get once the reforms land. Come with a view on which.

Then, next: growth, finally, for a class or two. Four sessions running, the news has bumped it.

Editor's note · the numbers, audited

Class figures against the published record. Deposits: the professor initially read ₹233 lakh crore (scheduled commercial banks) against ₹274 (all scheduled banks); corrected in class to 238 → 274 — the like-for-like all-scheduled-banks pair. Demand liabilities 28.76 → 34.31; time ~204 → ~235; credit 185 → 225; currency with the public ~37 → 42; total money ≈ 316. GDP: class working numbers ~325 (FY25) → ~350 (FY26); MoSPI's FY25 print is ₹330.7 lakh crore — using it shifts the ratio by under a point, changing nothing. The 90.3% on the dial is the class's own arithmetic (316 ÷ 350). Forex reserves: ~$720bn, among the world's largest piles. The COVID-peak comparison is measurement-sensitive (~90–94% depending on dating); flagged wherever it appears.

Editor's note · dating the yen operation

For the eventual exam question: the yen touched ~163 in late July; on 31 July 2026, at Camp David, Treasury Secretary Bessent's notepad — "To Do: Buy Japanese Yen $5–10 bil" — was photographed at 11:33 EDT, and the joint intervention (the first since 1998) was confirmed in the days after ([CNBC](https://www.cnbc.com/2026/08/03/yen-intervention-us-japan-trump-bessent-katayama.html), [The Japan Times](https://www.japantimes.co.jp/business/2026/08/03/markets/japan-us-joint-yen-intervention/), [Al Jazeera](https://www.aljazeera.com/economy/2026/8/3/japan-and-us-confirm-rare-joint-intervention-to-prop-up-yen)). BoJ money-market data implies Tokyo's own buying may have run to ~$59bn in a single session — the class's "$70 billion" is the right order; the class's "$10–15bn" for the US brackets the note's $5–10. Washington's motive, per [Axios](https://www.axios.com/2026/08/03/yen-japan-treasury-bessent) and [Bloomberg](https://www.bloomberg.com/news/articles/2026-08-05/japanese-yen-intervention-why-bessent-wants-fed-to-expand-fima-backstop): keep Tokyo from raising dollars by selling its ~$1.2 trillion of Treasuries. Result so far: ~163 → ~158.

Editor's note · the standing homework, now a machine

The professor's assignment — track how much they bring in versus how much they absorb — is now the Money Printer on the Bloomburger Terminal: WSS-anchored counters ticking at the year's run-rate, the 90% dial, era presets (the Quiet Print, COVID QE, demonetisation — the one time the machine ran in reverse — and Powell's 2020 printer for scale), and the tap-and-sponge ledger, every entry dated and sourced. Two entries are marked ≈ pending reconciliation with the April deep-research note (the 8–9 printed / 5–6 absorbed decomposition, swaps due 2026–29) — being retrieved from the archives; the ledger updates the day it lands.

Sources