Multinational Operations and the Analysis of Financial Institutions

Phase 1 · Accounting: The Language of Business · Week 16 · ~18 focused hours · Mastery gate ≥ 85%

Bharat Precision Instruments owns one American subsidiary. In the year to 31 March 2026 that subsidiary sold $60.0 million of machine-tool spindles, spent $36.0 million making them and earned $7.5 million after tax. Not one of those figures is in dispute. Folded into the parent's rupee accounts, the same subsidiary contributed either ₹63.00 crore of profit and ₹391.50 crore of net assets, or ₹64.25 crore of profit and ₹312.75 crore of net assets.

Both answers are right. The net assets differ by ₹78.75 crore, which is 25.1799% of the smaller figure, and the whole gap traces to one sentence of management judgement in the notes about which currency the subsidiary really operates in.

Every ratio moves with it. Return on equity is 16.0920% on the first answer and 20.5436% on the second, 4.4516 percentage points apart. Debt to equity is 0.577778 times against 0.723261 times, higher by the same 25.1799%. Gross margin is 40.0000% against 41.1607%. Asset turnover is 0.724138 times against 0.816525 times. In Ohio nothing changed at all. One judgement changed in Pune. That's enough for a screening filter set at "debt to equity below 0.70" to shut out a company it used to let through.

Four thousand miles away, Chesapeake Union Bancorp publishes a liquidity coverage ratio of 134.7190% and a footnote nobody reads. The bank holds $7,000 million of agency mortgage-backed securities at amortised cost. The market says those bonds are worth $5,600 million. Take them out of the numerator, on the ground that selling them would crystallise a loss the bank has decided not to take. Now the same ratio is 92.3788%. The bank is above the regulatory floor on one reading and below it on the other, and both readings use its own published numbers.

These two problems look unrelated. They're the same problem. You already know how to read the economic facts. A rulebook restates them, and the restated version is the one the market quotes, screens on and lends against. Translation restates a foreign subsidiary. Basel restates a bank's balance sheet into two ratios. In both cases the published number comes out of a mechanical procedure with a few judgement inputs. If you can't run the procedure, you have no way of knowing which input the answer is sensitive to.

So you'll run it. By the end you can take one foreign subsidiary's statements and a rate history and produce both translations. You'll prove the translation adjustment two independent ways and say which of six ratios moves in which direction, and why. You can restate a subsidiary in a hyperinflating economy under two incompatible rulebooks and quantify the gap. You can take a bank's balance sheet and compute both Basel liquidity ratios end to end, score it on the six CAMELS letters, and name the letter it is weakest on with the number attached.

Currency work has a reputation for being fiddly. The reputation comes from people who memorise a grid of which line goes at which rate. You don't need the grid. You need one idea. Translation moves everything by the same factor, so it preserves every ratio built from one statement. Remeasurement freezes some lines and moves others, so it breaks all of them. Hold on to that and the grid rebuilds itself whenever you need it.

Learning objectives

By the end you can:

1. Apply the IAS 21, Ind AS 21 and ASC 830 indicators to determine a foreign operation's functional currency, and state which specific facts in a disclosure would flip the determination.
2. Translate a complete foreign-currency income statement and balance sheet under the current-rate method, and derive the cumulative translation adjustment both as a balancing figure and as a three-part decomposition that must agree.
3. Remeasure the same statements under the temporal method, identify every line carried at a historical rate, and derive the remeasurement gain or loss from the net monetary position and tie it to profit or loss.
4. Build the ratio grid for both methods under an appreciating and a depreciating functional currency, and predict the direction of each ratio's movement from which balance-sheet items carry the rate.
5. Prove the scaling identity that makes every mixed ratio under the current-rate method equal to its local-currency counterpart multiplied by the average rate over the closing rate.
6. Restate a subsidiary in a hyperinflationary economy under IAS 29 and Ind AS 29, compute the gain or loss on the net monetary position two ways, and quantify the difference against the US GAAP requirement to remeasure using the temporal method.
7. Account for a net investment hedge, place the effective portion correctly relative to the translation reserve, and compute the ineffective portion that reaches profit when the hedge is over-designated.
8. Separate transaction exposure from translation exposure on a foreign-currency borrowing, track the exchange difference across two year-ends and an interest settlement, and quantify the effect on profit before tax.
9. Reconstruct a constant-currency revenue figure from segment and geography disclosures, decompose the currency drag by segment, and state which convention management used.
10. Score a bank on the six CAMELS letters, compute the liquidity coverage ratio and the net stable funding ratio end to end from a stated balance sheet, and price the cost of closing a funding shortfall.
11. Read a general insurer's combined ratio, float, cost of float and reserve development, and identify where a currency translation distorts a ratio built from two statements at different rates.

Prerequisites & connections

Builds on. M1.09 gave you the functional-versus-presentation-currency distinction, the current-rate and temporal methods as a comparison table, and the cumulative translation adjustment derived as a plug on a subsidiary with no dividends and no historical-rate complications. Everything here starts from that derivation and extends it to a full statement, a dividend, a share-capital line at a historical rate, and a decomposition that has to close. M1.02 gave you other comprehensive income and the recycling question, which is where the translation reserve lives and how it eventually reaches profit. M1.03 gave you the equity block, and the translation reserve is one of its components. M1.08 gave you effective interest and the amortised-cost measurement that makes a held-to-maturity bond portfolio possible in the first place. M2.01 and M2.02 supply the six-ratio panel that the currency work below distorts, and M2.03 supplies return on invested capital. M5.01 gave you banks as a business, the twelve-KPI panel and the India-versus-US regulatory map. M5.03 gave you insurance, float and the combined ratio.

Feeds forward. M1.10 is the capstone filing read, where a multinational's translation reserve and a bank's liquidity disclosures both have to be found in a real document. M2.06 to M2.08 hunt the abuses, and constant-currency reporting is one of the most heavily managed numbers in any results release. M3.02 and M3.04 need consolidated equity before any multiple means anything, and consolidated equity is the thing the method choice moves. M7.06 owns foreign-exchange markets as markets, including forwards, parity conditions and how the rate you translate at came to be that rate. AA1.05 looks at a foreign subsidiary's net assets from the other side, as an allocation with a currency exposure that somebody has to decide whether to hedge.

Five neighbours own material that isn't rebuilt here. M1.09 owns consolidation itself, the elimination entries, purchase price allocation and the first derivation of the translation adjustment, and the pages below cite it without repeating it. M5.01 owns bank analysis as a business: what a bank sells, the twelve KPIs, the credit cycle, the Indian and American regulatory maps and the valuation frame. What follows adds the organising structure those KPIs hang on and the two Basel liquidity ratios computed end to end, and it cross-links for everything else. M5.02 owns non-bank lenders. M5.03 owns insurance in depth, including embedded value, the life-insurance margin ladder and the underwriting cycle. What follows takes only the four general-insurance numbers an analyst needs and asks what a currency translation does to them. M7.06 owns the foreign-exchange market. Where a neighbour owns a concept, use it and move on.

Core teaching content

4.14 Where a tool helps here, and where it must not

The work in this area divides cleanly. Building a translation from a stated rate history is arithmetic with a reconciliation attached, and a machine does it faster and, checked properly, more reliably than a person at the end of a long day. Deciding whether a subsidiary described as a distribution arm is really an extension of the parent is judgement about facts in a filing, and no model has a view worth acting on.

The failure mode is specific and it repeats. Ask a language model for a translated balance sheet and it will produce one that balances, because balancing is the pattern it has learned. It gets there by adjusting a line, without ever deriving the translation adjustment, and the output looks just like the correct answer. The defence is the double proof: the adjustment as a plug and the adjustment as a decomposition, computed separately, agreeing to the last decimal. A translation where those two disagree is wrong, and a translation where they were never both computed hasn't been checked.

Four checks, and they take five minutes. Run them every time, and most of all when the output looks right.

1. Does the subsidiary's own net assets roll forward before any rate is applied, opening plus profit less dividend equals closing?
2. Does the translation adjustment as a plug equal the sum of the opening-net-assets piece, the profit piece and the dividend piece?
3. Under remeasurement, does the remeasured balance sheet roll forward from opening remeasured net assets, with no residual, and does the remeasurement gain or loss derived from the net monetary position match the one derived as the income-statement residual?
4. Does the gap between the two methods' net assets equal the sum of the non-monetary items times the difference between the closing rate and their historical rates?

{
  "skill": "Translating and remeasuring a foreign subsidiary from a stated rate history, and recomputing a bank's Basel liquidity ratios from its own disclosures",
  "use": "Have an assistant extract the subsidiary's statements and the rate history from a filing into structured rows, then write the code that produces both the current-rate and the temporal versions, derives the translation adjustment as a plug and as a decomposition, derives the remeasurement gain or loss from the net monetary position, and asserts that the two routes agree and that both balance sheets roll forward.",
  "tools": ["Chat assistants with code execution - Claude, ChatGPT, Gemini", "Python: pandas for the statements, plain arithmetic for the translation", "A document reader grounded on the filing you already hold", "A spreadsheet for the side-by-side ratio grid"],
  "prompt": "From this filing, extract into tables: (1) the foreign subsidiary's opening and closing balance sheets and its full income statement in its own currency; (2) every exchange rate disclosed or derivable - opening, closing, average, the dividend date rate, and the historical rates for share capital, each inventory layer and each class of fixed assets. Then write Python that (a) translates under the current-rate method and computes the cumulative translation adjustment BOTH as the balancing figure and as the sum of the opening-net-assets, profit and dividend pieces, asserting they agree to 1e-9; (b) remeasures under the temporal method with cost of goods sold built layer by layer and depreciation at the historical rate, computes the remeasurement gain or loss from the opening net monetary position plus flows, and asserts the remeasured balance sheet rolls forward with no residual; (c) prints revenue growth, gross margin, asset turnover, current ratio, debt to equity and return on equity under the local currency and under both methods; (d) asserts that every mixed ratio under the current-rate method equals its local counterpart times the average rate divided by the closing rate. State every rate you had to assume rather than read. List every figure you could NOT reconcile.",
  "verify": "Do not accept a balance sheet because it balances. A model that cannot derive the translation adjustment will produce one that balances by adjusting a line, and the output is indistinguishable from a correct answer until you run the decomposition separately. Recompute the adjustment by hand from the three pieces before accepting the plug. Check that the historical rates used for inventory and fixed assets came from the filing rather than from the model's assumption, because inventing a plausible historical rate is the single commonest fabrication here and it moves net assets by more than any other input. Confirm the assertions actually fire by feeding the script a deliberately wrong closing rate. For the bank work, check that every run-off and available-stable-funding factor was taken from the applicable national rule for the reporting date rather than from the Basel text, that the 40 percent Level 2 cap was tested rather than assumed not to bind, and that the inflow cap at 75 percent of outflows was applied. Cite the filing, the note and the page for every input, and record the reporting date and the currency unit on every panel.",
  "diy": "You must be able to translate a five-line subsidiary, remeasure it, and derive the adjustment both ways on paper; the worked examples are the drill and the gate tests exactly that. A tool cannot tell you that a subsidiary described as a sales office has quietly become a manufacturer, that a translation reserve fails to reconcile because a net investment hedge is running through it, or that a bank's Level 2A assets are securities it has publicly committed not to sell. Those readings are yours.",
  "market": "IN"
}

A machine is most useful on the sensitivity grid. Recomputing both methods across a range of closing rates, and reading off where a covenant or a screening threshold is breached, is arithmetic nobody should do by hand more than once. Do it by hand once, on the two rate paths above, so you know what the grid is showing you.

Common mistakes & how experts think differently

1. Treating the functional-currency determination as a formality. The note is short, it rarely changes and it decides whether currency movements reach earnings. A novice reads past it. An expert reads it first, checks it against what the business actually does, and treats a stale determination on a subsidiary that has changed shape as a live analytical question. The ₹78.75 crore gap above is that sentence, priced.

1. Memorising the rate grid instead of the principle. Which line goes at which rate is recoverable in ten seconds from one question: would a functional-currency entity have recorded this item at a fixed past amount? If yes, historical rate. If it is a fixed number of foreign-currency units, closing rate. Everything else is the average. People who memorise the grid get inventory and cost of goods sold wrong under pressure. Those two are the only lines where the layer carries the rate, and the item doesn't.

1. Computing the translation adjustment as a plug and stopping. The plug is a number. It's right whenever the arithmetic is right, and wrong without warning whenever it isn't. Deriving it a second way, from the opening net assets, the profit and the dividend, is the only check that fires when a line has been translated at the wrong rate. Both exemplary practice and the gate require both routes.

1. Reporting revenue growth without saying what the currency did. An Indian parent reporting 16.6667% growth on a subsidiary that grew 11.1111% hasn't misled anyone, because the translation is correct. An analyst who repeats the 16.6667% into a growth-rate comparison against a domestic peer has. Compute the constant-currency figure yourself from the segment note before making the comparison.

1. Assuming the temporal method flatters or hurts. It does both, on different lines, in the same year. Here it raised operating profit by 9.9702% and lowered equity by 20.1149% on a falling rupee, and on a rising rupee it lowered gross margin while still raising return on equity. The mechanism is the stable part. Monetary items carry the rate and non-monetary items don't, so leverage and asset-based ratios move first and hardest.

1. Reading a net investment hedge as risk management rather than as presentation. The hedge is real and the exposure it offsets is real. What it changes in the accounts is where a currency movement is reported, and it changes that by ₹15.00 crore of profit in the example above. An expert checks the designated amount against the net investment, computes the ineffective portion if the designation exceeds it, and reconciles the translation reserve knowing the hedge is sitting inside it.

1. Netting transaction exposure against translation exposure. They are different exposures with different accounting and different cash consequences. A dollar borrowing in a rupee-functional entity produces a profit-and-loss hit every period whether or not the group also owns dollar net assets. Only a formal designation moves the two into the same reserve, and only up to the size of the net investment.

1. Taking a liquidity coverage ratio at face value. The published number is the rule's output. The analytically useful number asks what the bank could actually monetise. Securities it has told the market it will hold to maturity, at a price that would crystallise a loss, are liquidity in the regulation's arithmetic. In a crisis that's far less obvious. Chesapeake's ratio is 134.7190% on the first reading and 92.3788% on the second.

1. Scoring a bank on capital and asset quality and calling it a diagnosis. Those are the two letters that were already comfortable at every bank that failed in 2023. Management and sensitivity to market risk are where the answer was, and both can be quantified: the loan-to-deposit growth gap, the direction of provisioning, unrealised securities losses against common equity tier 1, and the duration of the securities book.

1. Comparing combined ratios across jurisdictions without fixing the denominator. The Indian and trade-basis convention divides expenses by net written premium and the US GAAP presentation divides by net earned premium. On Vindhya that is 1.6023 percentage points, which is larger than the difference most people are trying to detect. Put both insurers on net earned premium and then compare.

1. Reading a foreign insurer's yield on float without adjusting for the rate seam. Investment income comes at the average rate and float comes at the closing rate, so the translated yield is the local yield times the average over the closing rate. It moved by 0.168116 percentage points on the UK subsidiary above, which is small, and it wouldn't be small in a year the currency moved twenty percent.

1. Forgetting that the translation reserve eventually reaches profit. It recycles on disposal of the operation, in full, under both rulebooks. A group with two decades of accumulated translation losses will report a disposal gain or loss that has nothing to do with the price it got, and a reader who hasn't looked at the equity note will have no idea where it came from.

Worked examples

Every input is stated, so you can work along with a pencil before you read each step. Every answer is checked by a second independent route. Where a reconciliation exists, it's worked in full. All entities are synthetic and all rates are illustrative.

Worked example 1: The same subsidiary in a strengthening rupee, both methods (India, ₹)

BPI Americas' dollar statements are unchanged: revenue $60.0 million, cost of goods sold $36.0 million, selling costs $9.0 million, depreciation $3.0 million, interest $2.0 million, tax $2.5 million, net income $7.5 million, dividend $2.0 million. Opening net assets $39.5 million and closing $45.0 million, with the balance sheets stated earlier. Prior-year revenue $54.0 million.

This time the rupee strengthens: opening ₹82.00, average ₹78.00, closing ₹77.00, dividend at ₹79.00. Share capital remains at ₹70.00, the plant at ₹70.00, the opening inventory layer at ₹81.00 and the closing layer at ₹77.50. Prior-year average ₹80.00.

Solution. Step 1, translate under the current-rate method. Income at ₹78.00 and the balance sheet at ₹77.00.

Step 2, the translation adjustment as a plug. Opening net assets as first translated were 39.5 × 82.00 = ₹323.90 crore. Expected closing: 323.90 + 58.50 − 15.80 = ₹366.60 crore. Actual: ₹346.50 crore. The adjustment is −₹20.10 crore, a debit to other comprehensive income.

Step 3, the same figure decomposed. Opening net assets repriced, 39.5 × (77.00 − 82.00) = −₹19.75 crore. Profit repriced from average to closing, 7.5 × (77.00 − 78.00) = −₹0.75 crore. Dividend, −2.0 × (77.00 − 79.00) = +₹0.40 crore. Sum: −₹20.10 crore, agreeing with the plug.

Step 4, remeasure under the temporal method. Cost of goods sold is 12.0 × 81.00 + 39.0 × 78.00 − 15.0 × 77.50 = ₹972.0m + ₹3,042.0m − ₹1,162.5m = ₹285.15 crore, which is now higher than the translated figure, because the inventory in cost of sales was bought when the rupee was weaker. Depreciation is 3.0 × 70.00 = ₹21.00 crore.

Step 5, the remeasurement gain from the net monetary position. Expected: (−20.5 × 82.00) + (7.5 × 78.00) − (2.0 × 79.00) = −₹1,681.0m + ₹585.0m − ₹158.0m = −₹125.40 crore. Actual: −15.0 × 77.00 = −₹115.50 crore. The difference is +₹9.90 crore, a gain, because the entity owed dollars and dollars became cheaper. Decomposed: −20.5 × (77.00 − 82.00) = +₹10.25 crore; 7.5 × (77.00 − 78.00) = −₹0.75 crore; −2.0 × (77.00 − 79.00) = +₹0.40 crore. Sum: +₹9.90 crore.

Step 6, verify three ways. Opening remeasured net assets are unchanged at ₹265.10 crore, because the opening rates didn't change. Roll them: 265.10 + 66.45 − 15.80 = ₹315.75 crore, which the balance sheet confirms. The net-asset gap is 346.50 − 315.75 = ₹30.75 crore, being PP&E at 45.0 × (77.00 − 70.00) = +₹31.50 crore and inventory at 15.0 × (77.00 − 77.50) = −₹0.75 crore. The profit gap is 66.45 − 58.50 = ₹7.95 crore, being the cost-of-goods-sold difference of −₹4.35 crore plus the depreciation difference of +₹2.40 crore plus the remeasurement of +₹9.90 crore.

Read it. Every sign flipped except one. The translation adjustment turned negative, the remeasurement turned into a gain, gross margin under remeasurement fell below the local figure instead of rising above it, and the inventory piece of the net-asset gap changed direction. The one thing that didn't flip is that remeasured net assets remain below translated net assets, by ₹30.75 crore, because the plant was bought at ₹70.00 and both closing rates are above it. That reconciliation is a fact about when the plant was bought. It says nothing about this year. So a company that built its foreign plant a decade ago and one that built it last year will show unequal method effects on identical operations.

Worked example 2: The ratio grid, the identity behind it, and the group effect (India, ₹)

Take both rate paths for BPI Americas and build the six-ratio panel, then push the answer up to the consolidated statements. Bharat Precision's parent-only profit is ₹520.00 crore and its parent-only net worth is ₹2,850.00 crore.

Solution. Step 1, the grid.

Step 2, prove the scaling identity. Under the current-rate method every mixed ratio equals the local ratio times the average rate divided by the closing rate. On the falling rupee, 84.00 ÷ 87.00 = 0.965517, and 16.6667% × 0.965517 = 16.0920%. Asset turnover: 0.750000 × 0.965517 = 0.724138. On the rising rupee, 78.00 ÷ 77.00 = 1.012987, and 16.6667% × 1.012987 = 16.8831%, with turnover at 0.750000 × 1.012987 = 0.759740. All four agree with the grid to twelve decimal places.

Step 3, the headline gaps between methods, falling rupee. Net assets ₹391.50 crore against ₹312.75 crore, a gap of ₹78.75 crore, being 25.1799% of the smaller and 20.1149% of the larger. Profit ₹63.00 crore against ₹64.25 crore, a gap of ₹1.25 crore or 1.9841%. Return on equity 4.4516 percentage points apart. Debt to equity 0.145484 times apart, which is 25.1799% higher, the same percentage as the equity gap because the numerator is identical under both methods. Gross margin 1.1607 percentage points apart.

Step 4, the same gaps on a rising rupee. Return on equity 4.1620 percentage points apart. Gross margin −0.9295 percentage points, in the opposite direction. Net assets ₹30.75 crore apart and profit ₹7.95 crore apart, which is more than six times the profit gap on the falling rupee.

Step 5, push it to the group, falling rupee.

Read it. The profit gap almost vanishes at group level, at 0.2144%, and the equity gap doesn't, at 2.4294%. That asymmetry is why the method choice survives consolidation. Profit differences are small and offsetting. Balance-sheet differences pile up year after year in the same direction for as long as the currency trends. Total comprehensive income differs by ₹19.95 crore because under translation the currency movement is real income routed around profit, and under remeasurement it has already passed through profit. None of this means one method reports the business better. Both describe the same $7.5 million of American earnings, and neither is closer to the truth than the other.

Worked example 3: A net investment hedge, designated and over-designated (India, ₹)

Bharat Precision borrows $30 million and designates it as a hedge of its net investment in BPI Americas. Opening net assets are $39.5 million and closing are $45.0 million. The rupee runs ₹82.00 to ₹87.00. The translation adjustment for the year is ₹21.20 crore and the opening translation reserve was ₹25.50 crore.

Solution. Step 1, the loan. Drawn at 30.0 × 82.00 = ₹246.00 crore, retranslated to 30.0 × 87.00 = ₹261.00 crore. The exchange loss is ₹15.00 crore.

Step 2, test the designation. $30 million against an opening net investment of $39.5 million is 75.9494%, and against the closing $45.0 million it is 66.6667%. Both are below 100%, so the whole loss can be effective.

Step 3, place the loss. Without a designation the ₹15.00 crore is a foreign-exchange charge in profit. With one, it goes to other comprehensive income and into the translation reserve. Profit is ₹15.00 crore higher under the designation, and there is no cash difference at all.

Step 4, the reserve. Without the designation the reserve closes at 25.50 + 21.20 = ₹46.70 crore. With it, 25.50 + 21.20 − 15.00 = ₹31.70 crore, and the movement for the year falls from ₹21.20 crore to ₹6.20 crore.

Step 5, over-designate. Designate $50 million instead. The total loss is 50.0 × (87.00 − 82.00) = ₹25.00 crore. The effective portion is capped at the net investment: 39.5 × (87.00 − 82.00) = ₹19.75 crore. The remaining ₹5.25 crore goes to profit.

Read it. The hedge produces ₹15.00 crore of reported profit that the business didn't earn and removes ₹15.00 crore of reserve movement that the business did experience. It doesn't change a single rupee of cash or a single dollar of exposure. Rebuild the translation reserve from the subsidiary's net assets and you'll be out by the hedge amount. The answer is in the hedge note, and more arithmetic won't find it.

Worked example 4: Hyperinflation, under both rulebooks (Argentina, ARS, and US, $)

Pampas Riego SA's year, in millions of pesos: revenue 4,000.0, cost of goods sold 2,400.0, other costs 500.0, depreciation 150.0, profit 950.0, and a dividend of 800.0 paid on the closing date. Inventory moved 300.0 to 700.0. Gross plant 1,500.0, accumulated depreciation 300.0 opening and 450.0 closing. Monetary assets 500.0 to 900.0 and monetary liabilities 900.0 to 1,400.0.

The general price index ran 100.0 opening, 150.0 average and 220.0 closing, with the plant dating from 50.0, opening inventory from 100.0 and closing inventory from 190.0. Exchange rates were 90.0 pesos to the dollar opening, 130.0 average and 180.0 closing, with 45.0 for the plant, 85.0 for opening inventory and 160.0 for closing inventory.

Solution. Step 1, check the local statements. Equity moves 1,100.0 + 950.0 − 800.0 = 1,250.0. The net monetary position moves −400.0 + 700.0 − 800.0 = −500.0, where the flow of 700.0 is revenue less purchases of 2,800.0 less other costs of 500.0. Both roll, so the restatement can proceed.

Step 2, the monetary gain, route one. Restate the opening position and the flows into closing-date units: (−400.0 × 2.2) + (700.0 × 220 ÷ 150) − 800.0 = −880.000000 + 1,026.666667 − 800.000000 = −653.333333. The actual closing position is −500.0. The gain is +153.333333.

Step 3, the monetary gain, route two. Holding a net monetary liability of 400.0 through 120.0000% inflation gains 400.0 × 1.2 = +480.000000. Holding a net inflow of 700.0 from an index of 150 to one of 220 loses 700.0 × (220 ÷ 150 − 1) = −326.666667. The dividend, paid at the closing index, contributes 0.000000. Sum: +153.333333, agreeing with route one.

Step 4, the restated statements. Revenue 4,000.0 × 220 ÷ 150 = 5,866.666667. Cost of goods sold, layer by layer, 300.0 × 2.2 + 2,800.0 × 220 ÷ 150 − 700.0 × 220 ÷ 190 = 3,956.140351. Other costs 733.333333. Depreciation 150.0 × 220 ÷ 50 = 660.000000. Operating profit 517.192982, and with the monetary gain, net profit 670.526316. Inventory 810.526316, net plant 4,620.000000, total assets 6,330.526316, equity 4,930.526316.

Step 5, prove the restated balance sheet. Restated opening equity is (−400.0 × 2.2) + (300.0 × 2.2) + (1,200.0 × 4.4) = 5,060.000000. Roll it: 5,060.000000 + 670.526316 − 800.000000 = 4,930.526316, matching.

Step 6, translate, then do it the American way instead. At 180.0 the restated figures become equity $27.391813 million, profit $3.725146 million, total assets $35.169591 million, liabilities $7.777778 million and the monetary gain $0.851852 million. Under ASC 830 the same subsidiary is remeasured with no restatement at all: revenue $30.769231 million, cost of goods sold $20.692873 million, gross profit $10.076357 million, other costs $3.846154 million, depreciation $3.333333 million, operating profit $2.896870 million, a remeasurement gain of $0.726496 million and net profit $3.623366 million. Inventory $4.375000 million, plant $23.333333 million, total assets $32.708333 million, net assets $24.930556 million. Roll it: 25.751634 + 3.623366 − 4.444444 = $24.930556 million, matching.

Step 7, the gap, decomposed. Equity differs by $2.461257 million, which is 9.8725% of the American figure, and it is plant $2.333333 million plus inventory $0.127924 million. Profit differs by only $0.101780 million.

Read it. The two rulebooks disagree by a tenth of the subsidiary's equity and by almost nothing on profit, and the disagreement lives entirely in long-lived assets. The numbers don't say which is right. IAS 29 claims the plant is worth 4,620.0 million pesos of today's purchasing power. ASC 830 claims it cost $23.333333 million and always did. Both statements are defensible and they can't both drive a comparison, so a peer set mixing the two frameworks has to be adjusted before any price-to-book number means anything.

Worked example 5: A dollar borrowing across two Indian year-ends (India, ₹)

Konkan Chemicals draws $40.0 million at 6% on 1 October 2024, interest payable each 30 September. Rates: ₹83.50 on drawdown, ₹85.20 at 31 March 2025, ₹87.60 on 30 September 2025, ₹89.40 at 31 March 2026. Average rates ₹84.30 for the second half of FY25 and ₹88.10 for the second half of FY26. Profit before tax and before any exchange effect would have been ₹210.000 crore in FY26.

Solution. Step 1, FY25 principal. Recorded at 40.0 × 83.50 = ₹334.000 crore, retranslated to 40.0 × 85.20 = ₹340.800 crore. Exchange loss ₹6.800 crore.

Step 2, FY25 interest. Six months accrues $1.2 million, expensed at ₹84.30 for ₹10.116 crore, and the liability is retranslated at ₹85.20 to ₹10.224 crore. Exchange loss ₹0.108 crore. FY25 total exchange loss: ₹6.908 crore.

Step 3, the FY26 settlement. The $2.4 million coupon is paid on 30 September 2025 at ₹87.60, costing ₹21.024 crore. The half accrued at ₹85.20 settles at ₹87.60, an exchange loss of 1.2 × (87.60 − 85.20) = ₹0.288 crore. The half earned between April and September is recognised at ₹87.60 for ₹10.512 crore of interest expense.

Step 4, the FY26 second-half accrual. $1.2 million at ₹88.10 gives ₹10.572 crore of interest expense, retranslated at ₹89.40 to ₹10.728 crore, an exchange loss of ₹0.156 crore.

Step 5, the FY26 principal. 40.0 × 89.40 = ₹357.600 crore against ₹340.800 crore, an exchange loss of ₹16.800 crore.

Step 6, total and place it. FY26 exchange loss 16.800 + 0.288 + 0.156 = ₹17.244 crore. FY26 interest expense 10.512 + 10.572 = ₹21.084 crore. Cumulative exchange loss across both years ₹24.152 crore, on a principal still sitting at $40.0 million. Profit before tax falls from ₹210.000 crore to ₹192.756 crore, a decline of 8.2114%.

Step 7, designate part of it. With $24 million of the principal designated against a net investment, 24 ÷ 40 of the FY26 principal loss goes to other comprehensive income: ₹10.080 crore. What stays in profit is 17.244 − 10.080 = ₹7.164 crore, profit before tax is ₹202.836 crore, and the decline is 3.4114% instead of 8.2114%.

Read it. Two years of a stable dollar obligation produced ₹24.152 crore of charges, none of them cash, all of them reversible if the rupee recovers, and every one of them a real change in what the company will eventually have to pay. Don't take any of this as licence to treat the exchange loss as non-recurring. It recurs every period the currency moves, and over the life of the loan it will sum to the difference between the drawdown rate and the repayment rate on $40.0 million, which nobody yet knows.

Worked example 6: Rebuilding a constant-currency claim (US, $)

Calder Instruments reports FY26 revenue of $2,373.714286 million, up 8.7086%, and states that revenue grew 11.0% in constant currency. Its geography note gives local-currency revenue and the average rates used.

Solution. Step 1, reported dollars. Americas $1,200.000000m and $1,320.000000m. Europe 640.0 × 1.0850 = $694.400000m and 700.0 × 1.0400 = $728.000000m. India 24,000.0 ÷ 83.00 = $289.156627m and 28,500.0 ÷ 87.50 = $325.714286m. Totals $2,183.556627 million and $2,373.714286 million, so reported growth is 8.7086%, which ties to the company's figure.

Step 2, constant currency at prior-year rates. Restate FY26 local revenue at FY25 rates: $1,320.000000m, 700.0 × 1.0850 = $759.500000m, and 28,500.0 ÷ 83.00 = $343.373494m, totalling $2,422.873494 million. Growth is 10.9600%, and the currency drag is −2.251337 percentage points. In dollars the currency cost $49.159208 million.

Step 3, the other convention. Restate FY25 at FY26 rates: $1,200.000000m, 640.0 × 1.0400 = $665.600000m, 24,000.0 ÷ 87.50 = $274.285714m, totalling $2,139.885714 million. Growth on that base is 10.9272%. The two conventions differ by 0.032805 percentage points.

Step 4, decompose the drag.

Step 5, check. The contributions sum to −2.251337 percentage points, matching the group drag computed in step 2. That agreement is the test that the rebuild used the right rates and the right base.

Read it. The company's 11.0% holds up. It falls between the two conventions' 10.9600% and 10.9272%, and the rebuild confirms it. The decomposition adds something the disclosure doesn't. India had the worse currency move, at 6.107143 points against Europe's 4.536290, and Europe still supplied nearly twice the group drag because it is 31.8013% of the base against India's 13.2425%. A reader who heard only about the rupee would have the ranking backwards. None of this tells you whether 10.9600% is good. A constant-currency figure removes the currency and leaves every other question untouched.

Worked example 7: A liquidity coverage ratio, line by line (US, $)

Chesapeake Union Bancorp (synthetic) is a $54 billion American regional bank. Its balance sheet, in millions:

Gross loans are $36,000 million against an allowance of $500 million. The agency mortgage-backed securities have a fair value of $5,600 million and qualify as Level 2A. The corporate bonds qualify as Level 2B. Of the short-term borrowings, $1,500 million matures within thirty days. The bank has undrawn commitments of $3,000 million to retail customers and $7,000 million of corporate facilities. Contractual inflows in the window are $700 million of retail loan payments, $900 million from non-financial corporates and $800 million from financial institutions.

Solution. Step 1, build the numerator in tiers. Level 1 is cash plus Treasuries: $7,000 million, no haircut. Level 2A is the mortgage-backed securities at market value with a 15% haircut: 5,600 × 0.85 = $4,760 million. Level 2B is the corporate bonds at a 50% haircut: 2,000 × 0.50 = $1,000 million. Uncapped total: $12,760 million.

Step 2, test the 40% cap. Level 2 totals $5,760 million. Forty percent of the uncapped total is $5,104 million, so the cap binds. Solve it from the other side: Level 1 must be at least 60% of the total, so high-quality liquid assets are 7,000 ÷ 0.60 = $11,666.666667 million, allowable Level 2 is $4,666.666667 million, and $1,093.333333 million of eligible securities are excluded. The 15% Level 2B sub-cap allows $1,750.000000 million against the $1,000 million held, so it doesn't bind.

Step 3, the outflows.

Step 4, the inflows and the cap. Retail loan inflows 700 × 0.50 = $350.00 million; non-financial wholesale 900 × 0.50 = $450.00 million; financial-institution inflows 800 × 1.00 = $800.00 million. Total $1,600.00 million. The cap is 75% of outflows, or $7,695.00 million, so the full amount counts. Net outflows: 10,260.00 − 1,600.00 = $8,660.00 million.

Step 5, the ratio. 11,666.666667 ÷ 8,660.00 = 134.7190%.

Step 6, stress the classification. The Level 2A credit comes entirely from bonds the bank holds at amortised cost, carries at $7,000 million and would have to sell at $5,600 million. Take them out and high-quality liquid assets fall to 7,000 + 1,000 = $8,000 million. The cap no longer binds, since $1,000 million of Level 2 is well under 40% of $8,000 million, or $3,200 million. The ratio becomes 8,000 ÷ 8,660.00 = 92.3788%.

Read it. The bank is 34.72 percentage points above the floor on the published basis and 7.62 points below it on the basis of what it has told investors it won't sell. Neither figure is a forecast. The first is the rule's answer and the second is a stress with one assumption in it, and the distance between them, 42.34 percentage points, is the size of the classification decision. The calculation can't tell you whether the bank would in fact sell in a crisis. That's a question about the board, and the balance sheet can't answer it.

Worked example 8: The net stable funding ratio, and the price of closing the gap (US, $)

Same balance sheet, one-year horizon.

Solution. Step 1, available stable funding.

Step 2, required stable funding.

Step 3, the ratio. 31,400.00 ÷ 32,425.00 = 96.8389%, against a 100% requirement. The shortfall is $1,025.00 million.

Step 4, price the first remedy. Term out $1,025.00 million of the short-term borrowing book past one year. That funding moves from a 0% factor to 100%, so available stable funding rises to $32,425.00 million and the ratio reaches 100.0000%. At a term premium of 75 basis points the annual cost is 1,025.00 × 0.0075 = $7.6875 million. After tax at 22% that's $5.996250 million, which is 1.3728% of the bank's net income of $436.80 million and 0.8836% of its pre-provision operating profit of $870.00 million.

Step 5, price the second remedy. Shrink instead. Commercial and industrial loans carry an 85% factor, so selling 1,025.00 ÷ 0.85 = $1,205.882353 million of them releases $1,025.00 million of required stable funding. That is 3.3497% of the bank's gross loan book, and it removes the spread those loans earn.

Read it. The shortfall is 3.16% of required stable funding and costs less than one and a half percent of net income to fix, so it's a nuisance and nowhere near a crisis. It's worth computing for the structure it exposes. $11,500 million of the bank's funding, being financial-institution money and short-dated borrowing, carries a zero available-stable-funding factor, while $16,500 million of commercial loans carry an 85% requirement. That mismatch sums up the business model in two numbers. The ratio doesn't predict a run. It measures a structure, and a bank can pass it on the last day of the quarter and fail it a week later.

Worked example 9: Scoring one bank on CAMELS (US, $)

Chesapeake Union's income statement for the year: net interest income $1,690 million, fee income $470 million, operating expenses $1,290 million, provisions $310 million. Tax is 22%. Average assets were $53,000 million and average equity $4,850 million. Risk-weighted assets are $34,000 million, tier 2 capital is $700 million. Non-performing loans are $612 million and net charge-offs for the year were $268 million. Prior-year gross loans were $31,600 million and prior-year deposits $32,400 million. The agency mortgage-backed securities book has a modified duration of 5.4 years.

Solution. C, capital adequacy. Common equity tier 1 is 4,400 ÷ 34,000 = 12.9412%. Tier 1 is 5,000 ÷ 34,000 = 14.7059%. Total capital is 5,700 ÷ 34,000 = 16.7647%. The leverage ratio on average assets is 5,000 ÷ 53,000 = 9.4340%. Comfortable on every measure.

A, asset quality. Non-performing loans are 612 ÷ 36,000 = 1.7000% of gross loans. The allowance of $500 million covers 500 ÷ 612 = 81.6993% of them and is 500 ÷ 36,000 = 1.3889% of the book. Net charge-offs are 268 ÷ 36,000 = 0.7444%. Adequate, with coverage below the 100% an American peer group would usually show.

M, management. The efficiency ratio is 1,290 ÷ (1,690 + 470) = 59.7222%, at the weak end of acceptable. Loans grew 36,000 ÷ 31,600 − 1 = 13.9241% while deposits grew 33,000 ÷ 32,400 − 1 = 1.8519%, a gap of 12.0722 percentage points funded by wholesale money. The bank is buying growth with borrowed funding and running a cost base that leaves little room.

E, earnings. Pre-tax profit is 1,690 + 470 − 1,290 − 310 = $560.00 million, tax $123.20 million, net income $436.80 million. Pre-provision operating profit is $870.00 million. Return on assets is 436.80 ÷ 53,000 = 0.8242% and return on equity is 436.80 ÷ 4,850 = 9.0062%. Net interest margin on earning assets of $52,000 million is 1,690 ÷ 52,000 = 3.2500%, and pre-provision profit is 1.6415% of average assets. Earnings are below the 1% return-on-assets bar and short of any plausible cost of equity.

L, liquidity. The liquidity coverage ratio is 134.7190% and the net stable funding ratio is 96.8389%, which is a breach. Loans are 36,000 ÷ 33,000 = 109.0909% of deposits. Wholesale reliance, meaning non-retail and non-operational funding over total assets, is 23,000 ÷ 54,000 = 42.5926%.

S, sensitivity to market risk. The unrealised loss on the held-to-maturity book is 7,000 − 5,600 = $1,400 million, which is 31.8182% of common equity tier 1. Marked, common equity tier 1 falls to $3,000 million, or 8.8235% of risk-weighted assets against the reported 12.9412%. A further 100 basis points, at a modified duration of 5.4 years applied to the $5,600 million market value and not the carrying amount, costs 5,600 × 5.4 ÷ 100 = $302.40 million, taking the marked ratio to 7.9341%, and a second step takes it to 7.0447%. Against an effective floor near 7.00% the bank has 93.41 basis points of headroom after one step and 4.47 basis points after two.

The verdict. Strong on capital as reported, adequate on asset quality, weak on management and earnings, breaching on one liquidity measure, and weakest on sensitivity to market risk, where two 100 basis point moves leave 4.47 basis points of marked capital headroom.

Read it. Four of the six letters are the ones a screen would look at and all four are survivable. The two that aren't on the screen are where the bank is fragile, and they interact. The securities losses that hurt the S letter are unrealised because the bank won't sell. It won't sell because selling would crystallise them, and that same refusal is what props up the liquidity coverage ratio. The scorecard doesn't predict failure. It says where the bank breaks if it breaks, which is a different and more useful claim.

Worked example 10: A general insurer, and what translation does to it (India, ₹)

Vindhya General Insurance wrote ₹8,400.00 crore of gross premium and ceded ₹1,900.00 crore. The unearned premium reserve rose ₹350.00 crore. Claims incurred were ₹4,489.50 crore, commission ₹780.00 crore and operating expenses ₹1,050.00 crore. Outstanding claims reserves are ₹9,800.00 crore, unearned premium ₹3,150.00 crore and reinsurance recoverable ₹1,850.00 crore. Investment income was ₹780.00 crore. The opening claims reserve of ₹9,200.00 crore has been re-estimated at ₹9,540.00 crore.

Its UK subsidiary wrote £200 million of net premium in FY25 and £220 million in FY26, at average rates of ₹104.00 and ₹112.00 and closing rates of ₹106.00 and ₹115.00. The subsidiary's float is £900 million and its investment income £58 million, and its own combined ratio is 96.0000%. Indian net written premium in FY25 was ₹6,100.00 crore.

Solution. Step 1, the premium ladder. Net written premium is 8,400.00 − 1,900.00 = ₹6,500.00 crore, a retention of 77.3810%. Net earned premium is 6,500.00 − 350.00 = ₹6,150.00 crore.

Step 2, the ratios, both conventions. Loss ratio 4,489.50 ÷ 6,150.00 = 73.0000%. Expenses of ₹1,830.00 crore give an expense ratio of 29.7561% on net earned premium and 28.1538% on net written premium. Combined ratio: 102.7561% on the US GAAP basis and 101.1538% on the mixed Indian basis, a gap of 1.6023 percentage points. Underwriting result: 6,150.00 − 4,489.50 − 1,830.00 = −₹169.50 crore.

Step 3, float and its cost. Float is 9,800.00 + 3,150.00 − 1,850.00 = ₹11,100.00 crore. Investment income of ₹780.00 crore is a yield of 7.0270%. The cost of float is 169.50 ÷ 11,100.00 = 1.5270%. Profit before tax is −169.50 + 780.00 = ₹610.50 crore.

Step 4, the reserve check. Adverse development is 9,540.00 − 9,200.00 = ₹340.00 crore, being 3.6957% of the opening reserve and 55.6921% of profit before tax. Excluding it, the current-accident-year loss ratio is (4,489.50 − 340.00) ÷ 6,150.00 = 67.4715% and the combined ratio is 97.2276%.

Step 5, the subsidiary in rupees. Net written premium translates to 200 × 104.00 = ₹2,080.00 crore and 220 × 112.00 = ₹2,464.00 crore, which is 18.4615% growth against 10.0000% locally, so currency supplied 8.4615 percentage points.

Step 6, the ratios that survive and the one that doesn't. The subsidiary's combined ratio is 96.0000% in either currency, because both its parts are income-statement lines at the same average rate. Its float translates at the closing rate to 900 × 115.00 = ₹10,350.00 crore while its investment income translates at the average to 58 × 112.00 = ₹649.60 crore. The local yield of 6.4444% becomes 6.2763%, a distortion of −0.168116 percentage points, and the scaling identity confirms it: 112.00 ÷ 115.00 = 0.973913, and 6.4444% × 0.973913 = 6.2763%.

Step 7, the group. Net written premium was 6,100.00 + 2,080.00 = ₹8,180.00 crore in FY25 and 6,500.00 + 2,464.00 = ₹8,964.00 crore in FY26, growth of 9.5844%. At constant currency FY26 would be 6,500.00 + 220 × 104.00 ÷ 10 = ₹8,788.00 crore, growth of 7.4328%. Currency added 2.1516 percentage points.

Read it. Three separate distortions sit in one insurer's numbers and only one of them is about currency. The convention gap of 1.6023 percentage points decides whether the combined ratio prints above or below 101%. The reserve development of ₹340.00 crore decides whether the underwriting year was a 102.7561% loss or a 97.2276% profit, and it is 55.6921% of the profit the company reported. Currency adds 2.1516 percentage points to group premium growth and shaves 0.168116 percentage points off the translated yield on float. None of it settles whether the reserves are now adequate. Only the run-off triangle speaks to that.

Verify block

Every headline figure, at the precision the computation produced. Check your own working against it once you've done the work yourself.

• BPI Americas, local. Revenue $60.0m, gross profit $24.0m, operating profit $12.0m, net income $7.5m, dividend $2.0m; local net margin 12.5000%, gross margin 40.0000%, revenue growth 11.1111%. Net assets $39.5m to $45.0m; roll 39.5 + 7.5 − 2.0 = 45.0. Net monetary position −$20.5m to −$15.0m; purchases $39.0m; net monetary flow $7.5m.
• Falling rupee, current-rate (₹ crore). Revenue 504.00, cost of sales 302.40, gross profit 201.60, selling 75.60, depreciation 25.20, operating profit 100.80, interest 16.80, tax 21.00, net income 63.00, dividend 16.60. Cash 69.60, receivables 104.40, inventory 130.50, plant 391.50, total assets 696.00, liabilities 304.50, net assets 391.50, share capital 140.00. Opening net assets translated 323.90. Translation adjustment 21.20 as a plug and 19.75 + 2.25 − 0.80 = 21.20 decomposed. Opening retained 158.40 and reserve 25.50; closing retained 204.80 and reserve 46.70; equity totals 391.50.
• Falling rupee, temporal (₹ crore). Cost of sales 296.55, depreciation 21.00, operating profit 110.85, remeasurement −8.80, profit before tax 85.25, net income 64.25. Inventory 128.25, plant 315.00, total assets 617.25, net assets 312.75, retained 172.75. Remeasurement expected −121.70 against actual −130.50; pieces −10.25 + 2.25 − 0.80. Opening remeasured net assets 265.10 and opening retained 125.10; roll 265.10 + 64.25 − 16.60 = 312.75. Net-asset gap 78.75 = plant 76.50 + inventory 2.25. Profit gap 1.25 = cost of sales 5.85 + depreciation 4.20 + remeasurement −8.80.
• Rising rupee (₹ crore). Current-rate: revenue 468.00, cost of sales 280.80, depreciation 23.40, net income 58.50, dividend 15.80, net assets 346.50, total assets 616.00. Translation adjustment −20.10, decomposed as −19.75 − 0.75 + 0.40. Temporal: cost of sales 285.15, depreciation 21.00, remeasurement +9.90 (+10.25 − 0.75 + 0.40), net income 66.45, inventory 116.25, plant 315.00, total assets 585.25, net assets 315.75; roll 265.10 + 66.45 − 15.80 = 315.75. Net-asset gap 30.75, profit gap 7.95.
• Ratio grid. Local: growth 11.1111%, gross margin 40.0000%, turnover 0.750000x, current 2.333333x, debt to equity 0.577778x, liabilities to equity 0.777778x, return on equity 16.6667%. Falling rupee current-rate: 16.6667%, 40.0000%, 0.724138x, 2.333333x, 0.577778x, 0.777778x, 16.0920%. Falling rupee temporal: 16.6667%, 41.1607%, 0.816525x, 2.316092x, 0.723261x, 0.973621x, 20.5436%. Rising rupee current-rate: 8.3333%, 40.0000%, 0.759740x, 2.333333x, 0.577778x, 0.777778x, 16.8831%. Rising rupee temporal: 8.3333%, 39.0705%, 0.799658x, 2.339827x, 0.634046x, 0.853523x, 21.0451%. Scaling factors 0.965517 and 1.012987.
• Method gaps. Net assets 391.50 against 312.75, gap 25.1799% of the smaller and 20.1149% of the larger; profit gap 1.9841%; return on equity +4.4516 pp; debt to equity +0.145484x, being 25.1799% higher; gross margin +1.1607 pp. Mirror: return on equity +4.1620 pp, gross margin −0.9295 pp, net assets 30.75, profit 7.95. Operating profit 100.80 against 110.85, +9.9702%.
• Group. Parent profit 520.00 and net worth 2,850.00. Consolidated profit 583.00 against 584.25, gap 0.2144%; equity 3,241.50 against 3,162.75, gap 2.4294%; return on equity 17.9855% against 18.4728%, +0.4873 pp; other comprehensive income 21.20 against 0.00; total comprehensive income 604.20 against 584.25, gap 19.95.
• Net investment hedge. $30m loan 246.00 to 261.00, loss 15.00; hedge ratio 75.9494% of the opening net investment and 66.6667% of the closing. Reserve 46.70 without designation and 31.70 with; movement 6.20. Over-designation at $50m: total loss 25.00, effective cap 19.75, to profit 5.25.
• Pampas Riego. Historical profit 950.00, dividend 800.00, equity 1,100.00 to 1,250.00, net monetary position −400.00 to −500.00, purchases 2,800.00, flow 700.00, inflation 120.0000%. Expected closing position −653.333333; monetary gain 153.333333 = 480.000000 − 326.666667 + 0.000000. Restated revenue 5,866.666667, cost of sales 3,956.140351, gross profit 1,910.526316, other costs 733.333333, depreciation 660.000000, operating profit 517.192982, net profit 670.526316. Inventory 810.526316, plant 4,620.000000, total assets 6,330.526316, equity 4,930.526316; opening equity 5,060.000000; roll closes. Translated at 180.0: equity $27.391813m, profit $3.725146m, assets $35.169591m, liabilities $7.777778m, monetary gain $0.851852m. US GAAP: revenue $30.769231m, cost of sales $20.692873m, gross profit $10.076357m, other $3.846154m, depreciation $3.333333m, operating $2.896870m, remeasurement gain $0.726496m, net profit $3.623366m, inventory $4.375000m, plant $23.333333m, assets $32.708333m, net assets $24.930556m; roll $25.751634m + $3.623366m − $4.444444m. Equity gap $2.461257m, being 9.8725%, of which plant $2.333333m and inventory $0.127924m; profit gap $0.101780m.
• Konkan Chemicals. Drawdown 334.000, FY25 close 340.800, loss 6.800; FY25 interest 10.116 restated 10.224, loss 0.108; FY25 total 6.908. Coupon paid 21.024, settlement loss 0.288; first-half interest 10.512; second-half accrual 10.572 restated 10.728, loss 0.156; principal loss 16.800; FY26 total 17.244; FY26 interest 21.084; cumulative 24.152. Profit before tax 210.000 to 192.756, a fall of 8.2114%. With $24m designated: 10.080 to other comprehensive income, 7.164 to profit, profit before tax 202.836, a fall of 3.4114%.
• Calder Instruments. Americas $1,200.000000m to $1,320.000000m; Europe $694.400000m to $728.000000m, drag −4.536290 pp; India $289.156627m to $325.714286m, drag −6.107143 pp. Totals $2,183.556627m and $2,373.714286m; constant-currency FY26 $2,422.873494m; FY25 at FY26 rates $2,139.885714m. Reported growth 8.7086%; constant currency 10.9600% and 10.9272%; drag −2.251337 pp; convention gap 0.032805 pp; currency effect −$49.159208m. Contributions 0.000000, −1.442601 and −0.808736 pp, summing to the total. Weights 54.9562%, 31.8013%, 13.2425%.
• Chesapeake Union, liquidity. Level 1 7,000.00, Level 2A 4,760.00, Level 2B 1,000.00, uncapped 12,760.00; cap binds; high-quality liquid assets 11,666.666667, allowable Level 2 4,666.666667, excluded 1,093.333333, Level 2B sub-cap 1,750.000000. Outflows 360.00 + 500.00 + 2,250.00 + 2,800.00 + 2,000.00 + 1,500.00 + 150.00 + 700.00 = 10,260.00; inflows 350.00 + 450.00 + 800.00 = 1,600.00; cap 7,695.00; net outflows 8,660.00; ratio 134.7190%. Excluding the held-to-maturity book: 8,000.00 and 92.3788%.
• Chesapeake Union, funding. Available stable funding 5,000.00 + 2,500.00 + 11,400.00 + 4,500.00 + 4,500.00 + 3,500.00 = 31,400.00. Required 0.00 + 200.00 + 1,050.00 + 1,000.00 + 9,750.00 + 14,025.00 + 3,400.00 + 2,500.00 + 500.00 = 32,425.00. Ratio 96.8389%; shortfall 1,025.00; term-out cost $7.6875m, after tax $5.996250m, being 1.3728% of net income and 0.8836% of pre-provision profit; loan sale alternative $1,205.882353m.
• Chesapeake Union, CAMELS. Common equity tier 1 12.9412%, tier 1 14.7059%, total capital 16.7647%, leverage 9.4340%. Non-performing loans 1.7000%, coverage 81.6993%, allowance to loans 1.3889%, charge-offs 0.7444%. Efficiency 59.7222%; loan growth 13.9241%, deposit growth 1.8519%, gap +12.0722 pp. Pre-tax 560.00, tax 123.20, net income 436.80, pre-provision profit 870.00; return on assets 0.8242%, return on equity 9.0062%, net interest margin 3.2500%, pre-provision profit over assets 1.6415%. Loan to deposit 109.0909%, wholesale reliance 42.5926%. Unrealised loss 1,400.00, being 31.8182% of common equity tier 1; marked ratio 8.8235% against 12.9412%; a 100 basis point step on the 5,600.00 market value at a modified duration of 5.4 costs 302.40, taking it to 7.9341% then 7.0447%; headroom 93.41 bp then 4.47 bp.
• Vindhya General. Gross 8,400.00, ceded 1,900.00, net written 6,500.00, net earned 6,150.00, retention 77.3810%. Loss ratio 73.0000%; expense ratio 29.7561% and 28.1538%; combined 102.7561% and 101.1538%, gap 1.6023 pp; underwriting result −169.50. Float 11,100.00, investment income 780.00, yield 7.0270%, cost of float 1.5270%, profit before tax 610.50. Adverse development 340.00, being 3.6957% and 55.6921%; accident-year loss ratio 67.4715% and combined 97.2276%. UK premium 2,080.00 to 2,464.00, growth 18.4615% against 10.0000%, currency +8.4615 pp; float 10,350.00, income 649.60, yields 6.4444% and 6.2763%, distortion −0.168116 pp, factor 0.973913; combined ratio 96.0000% either way. Group 8,180.00 to 8,964.00, growth 9.5844%; constant currency 8,788.00 and 7.4328%; currency +2.1516 pp.

Practice set

Work each problem fully before you read its solution. It's tempting to peek, and you'll learn far less if you do. A calculator is assumed and no software is needed. Unless a problem says otherwise, answers within 0.5% of the stated figure count as correct; where a problem asks for a routing, a classification or a direction, the answer is exact. All entities are synthetic and all rates are illustrative. Assume dividends are paid on the date whose rate is given, and that flows other than dividends occur evenly and therefore carry the average rate.

P1 (guided). Decide the functional currency (India, ₹). An Indian pharmaceutical group owns a subsidiary registered in Ireland. The subsidiary holds the group's European marketing authorisations and employs eleven people in Dublin. It buys finished product from the Indian parent priced in rupees under an intercompany agreement reviewed annually in Pune, sells to European distributors in euros at prices set by a Pune committee against Indian cost, and holds no inventory it didn't receive from the parent. It is funded entirely by an intercompany loan denominated in rupees, and it remits its cash balance to Pune monthly. State the functional currency and name the three facts that decide it. Then state the single change that would flip the answer.

Solution. The functional currency is the rupee, so the euro-denominated books are remeasured with the temporal method and the difference goes to profit or loss. Three facts decide it. Sales prices are set in Pune against Indian cost. European competitive forces don't set them, so the primary sales-price indicator points to the rupee. Costs are overwhelmingly the intercompany purchase price, denominated in rupees. Financing is a rupee intercompany loan, and cash is remitted monthly instead of retained. So the operation has no autonomy, and its cash flows are directly and immediately available to the parent. Selling in euros to euro customers isn't decisive by itself. The standard asks what determines the price, and the invoice currency doesn't answer that. Here's the change that would flip it. If the subsidiary began sourcing and manufacturing in Europe, setting its own prices against European competitors, and funding itself from a European bank, the euro would become the functional currency and the same movements would move from profit to a translation reserve.

P2 (guided). The translation adjustment, both ways (India, ₹). An Indian parent's US subsidiary opens with net assets of $60.0 million, earns $12.0 million and pays a dividend of $4.0 million. Rates are ₹80.00 opening, ₹84.00 average, ₹88.00 closing, and ₹85.00 on the dividend date. Compute the translation adjustment as a plug and as a decomposition.

Solution. Closing net assets are 60.0 + 12.0 − 4.0 = $68.0 million, translated at ₹88.00 to ₹5,984.00 million. Expected: 60.0 × 80.00 + 12.0 × 84.00 − 4.0 × 85.00 = 4,800.00 + 1,008.00 − 340.00 = ₹5,468.00 million. The plug is ₹516.00 million, or ₹51.60 crore, credited to other comprehensive income. Decomposed: 60.0 × (88 − 80) = ₹480.00 million; 12.0 × (88 − 84) = ₹48.00 million; −4.0 × (88 − 85) = −₹12.00 million. Sum ₹516.00 million, agreeing.

P3 (guided). The remeasurement gain, from the position rather than the residual (India, ₹). The same subsidiary, now judged rupee-functional. Its net monetary position opens at −$18.0 million and closes at −$12.0 million, with a net monetary inflow from operations of $6.0 million and no dividend. Rates are ₹80.00 opening, ₹84.00 average and ₹88.00 closing. Compute the remeasurement gain or loss and decompose it.

Solution. Expected: (−18.0 × 80.00) + (6.0 × 84.00) = −1,440.00 + 504.00 = −₹936.00 million. Actual: −12.0 × 88.00 = −₹1,056.00 million. The difference is −₹120.00 million, or −₹12.00 crore. That's a loss through profit or loss, because the entity was net monetary short while the rupee weakened. Decomposed: −18.0 × (88 − 80) = −₹144.00 million on the opening position, and 6.0 × (88 − 84) = +₹24.00 million on the flows. Sum −₹120.00 million.

P4 (guided). Temporal cost of goods sold (US, $). A US parent remeasures a euro subsidiary. Opening inventory is €8.0 million bought at $1.1200, purchases are €46.0 million at an average of $1.0800, and closing inventory is €10.0 million bought at $1.0500. Revenue is €92.0 million. Compute cost of goods sold under the temporal method and under the current-rate method, and both gross margins.

Solution. Local cost of goods sold is 8.0 + 46.0 − 10.0 = €44.0 million. Temporal: 8.0 × 1.1200 + 46.0 × 1.0800 − 10.0 × 1.0500 = 8.96 + 49.68 − 10.50 = $48.1400 million. Current-rate: 44.0 × 1.0800 = $47.5200 million. The gap is $0.6200 million. Local gross margin is (92.0 − 44.0) ÷ 92.0 = 52.1739%, and it survives translation unchanged. Under remeasurement, revenue is 92.0 × 1.0800 = $99.36 million and gross margin is (99.36 − 48.14) ÷ 99.36 = 51.5499%, lower because the older inventory layers carried a stronger euro.

P5. The scaling identity, both directions (US, $). A subsidiary's local return on equity is 18.4000%. Compute the translated return on equity under the current-rate method when the average rate is $1.0800 and the closing rate $1.0500, and again when the average is $1.0500 and the closing is $1.0800.

Solution. The factor is the average rate divided by the closing rate. First case: 1.0800 ÷ 1.0500 = 1.028571, so 18.4000% × 1.028571 = 18.9257%. Second case: 1.0500 ÷ 1.0800 = 0.972222, so 18.4000% × 0.972222 = 17.8889%. Return on equity is a mixed ratio, taking profit at the average rate over equity at the closing rate, so it is the only one of the six panel ratios that moves under the current-rate method along with asset turnover. Nothing about the subsidiary changed in either case.

P6. The gain on the net monetary position (India, ₹). A subsidiary in a hyperinflationary economy opens with a net monetary position of −250.0 million local units and takes in 350.0 million of net monetary flows during the year, with no dividend. The general price index runs 100.0 opening, 160.0 average and 240.0 closing. Compute the monetary gain or loss two ways and state the year's inflation.

Solution. The closing position is −250.0 + 350.0 = 100.0. Expected in closing units: (−250.0 × 240 ÷ 100) + (350.0 × 240 ÷ 160) = −600.0 + 525.0 = −75.0. The gain is 100.0 − (−75.0) = +175.0 million, and it goes through profit or loss. Decomposed: holding a net monetary liability of 250.0 through the year gains 250.0 × (240 ÷ 100 − 1) = +350.0; holding an inflow of 350.0 from an index of 160 to 240 loses 350.0 × (240 ÷ 160 − 1) = −175.0. Sum +175.0. Inflation for the year is 240 ÷ 100 − 1 = 140.0000%.

P7. An over-designated net investment hedge (India, ₹). An Indian parent has a net investment of $80.0 million in a US subsidiary and designates a $100.0 million dollar borrowing as a hedge of it. The rupee moves from ₹82.00 to ₹88.00. Compute the total loss on the borrowing, the effective portion and the amount that reaches profit.

Solution. The total loss is 100.0 × (88.00 − 82.00) = ₹600.00 million, or ₹60.00 crore. The effective portion is capped at the net investment: 80.0 × 6.00 = ₹480.00 million, or ₹48.00 crore, to other comprehensive income. The remaining ₹12.00 crore is ineffective and goes to profit or loss. Designating more than the net investment doesn't increase protection. It turns the excess into an unhedged exposure that reports through earnings, the opposite of what the designation was meant to do.

P8. A dollar loan across two year-ends (India, ₹). An Indian company draws $25.0 million at 5.5%, interest paid annually at each year end. The rupee closes at ₹82.40 on drawdown, ₹86.10 at the first year end and ₹89.90 at the second, with average rates of ₹84.00 and ₹88.00. Compute the exchange loss on the principal in each year, the cumulative loss, and the interest expense in each year.

Solution. Year 1: 25.0 × (86.10 − 82.40) = ₹92.50 million = ₹9.2500 crore of exchange loss. Year 2: 25.0 × (89.90 − 86.10) = ₹95.00 million = ₹9.5000 crore. Cumulative ₹18.7500 crore on a principal that never changed from $25.0 million. Interest is 25.0 × 0.055 = $1.375 million a year, expensed at the average rate: ₹11.5500 crore in year 1 and ₹12.1000 crore in year 2. The interest expense rose 4.76% in rupees on an unchanged dollar coupon. That's a second and separate currency effect, and companies routinely present it inside finance costs without separating it.

P9. Rebuild a constant-currency figure (US, $). A US group reports three geographies. The Americas earned $500.0 million then $540.0 million. The United Kingdom earned £300.0 million then £324.0 million, at $1.2600 and $1.2000. Japan earned ¥42,000 million then ¥46,200 million, at ¥145 and ¥158 to the dollar. Compute reported growth, constant-currency growth at prior-year rates, and the drag.

Solution. Reported: FY25 is 500.000000 + 378.000000 + 289.655172 = $1,167.655172 million and FY26 is 540.000000 + 388.800000 + 292.405063 = $1,221.205063 million, so growth is 4.5861%. Constant currency restates FY26 at FY25 rates: 540.000000 + 324.0 × 1.2600 + 46,200 ÷ 145 = 540.000000 + 408.240000 + 318.620690 = $1,266.860690 million, giving growth of 8.4961%. The drag is −3.910026 percentage points. By segment, local growth was 8.0000%, 8.0000% and 10.0000%, against reported growth of 8.0000%, 2.8571% and 0.9494%. Japan grew fastest in local terms and slowest in dollars. Read only the reported column and you'd get that pattern backwards.

P10. A liquidity coverage ratio with a binding cap (US, $). A bank holds $6,000 million of Level 1 assets, Level 2A securities with a market value of $5,000 million and Level 2B securities with a market value of $1,200 million. Stressed outflows are $9,200 million and stressed inflows $1,900 million. Compute high-quality liquid assets, net outflows and the ratio, and state what the ratio would be if the cap didn't exist.

Solution. After haircuts, Level 2A is 5,000 × 0.85 = $4,250 million and Level 2B is 1,200 × 0.50 = $600 million, so the uncapped total is $10,850 million. Level 2 of $4,850 million exceeds 40% of $10,850 million, which is $4,340 million, so the cap binds. High-quality liquid assets are 6,000 ÷ 0.60 = $10,000.0000 million, allowable Level 2 is $4,000.0000 million and $850.0000 million is excluded. The inflow cap is 75% of $9,200 million, or $6,900 million, so all $1,900 million of inflows count and net outflows are $7,300 million. The ratio is 136.9863%. Without the cap it would be 10,850 ÷ 7,300 = 148.6301%, so the cap costs 11.6438 percentage points.

P11. A funding shortfall, priced (US, $). A bank's available stable funding is $24,800 million and its required stable funding is $25,850 million. Compute the net stable funding ratio and the shortfall, then price two remedies: terming out short-dated wholesale funding at a 90 basis point premium, and selling loans that carry an 85% required-stable-funding factor.

Solution. The ratio is 24,800 ÷ 25,850 = 95.9381%, so the shortfall is $1,050 million. Terming out $1,050 million of funding that currently carries a zero factor to a maturity beyond one year raises available stable funding to $25,850 million and the ratio to 100%, at an annual cost of 1,050 × 0.0090 = $9.4500 million. Selling loans instead requires 1,050 ÷ 0.85 = $1,235.2941 million of disposals to release the same required stable funding. The first remedy costs an interest spread and keeps the assets. The second keeps the spread cost at zero and gives up the loans' margin, so the comparison is between the term premium and the net interest margin on $1,235.2941 million.

P12. Read an Indian bank's disclosed panel (India, ₹). A bank discloses gross advances of ₹3,20,000 crore, a gross non-performing asset ratio of 3.10% and provision coverage of 70.6%. Capital adequacy is 15.2% with common equity tier 1 at 13.2%. Net interest margin is 3.62%, cost to income 47.8%, current and savings account share 41.2% and credit to deposit 87.9%. Return on assets is 1.38% and return on equity 15.12%. The liquidity coverage ratio is 128.4%, the net stable funding ratio 117.1%, and the economic-value-of-equity sensitivity −18.3% to a 200 basis point rate rise. Compute the gross non-performing amount, the provisions held, the implied net non-performing ratio and the unprovided balance, then compute implied leverage and the net worth that a ₹4,20,000 crore balance sheet would imply. Assign each disclosed number to a CAMELS letter and name the weakest letter.

Solution. Gross non-performing assets are 3,20,000 × 0.031 = ₹9,920.00 crore, against provisions of 9,920.00 × 0.706 = ₹7,003.52 crore, leaving ₹2,916.48 crore unprovided. The net ratio is 2,916.48 ÷ (3,20,000 − 7,003.52) = 0.9318%, just inside the 1% comfort line. Implied leverage is return on equity divided by return on assets: 0.1512 ÷ 0.0138 = 10.956522 times, so a ₹4,20,000 crore balance sheet implies net worth of ₹38,333.33 crore. By letter, capital adequacy and common equity tier 1 go to C. The non-performing ratios and provision coverage go to A. Cost to income goes to M, and net interest margin, return on assets and return on equity to E. The current and savings share, credit to deposit, liquidity coverage and net stable funding all go to L, and the economic-value sensitivity to S. The weakest letter is S, at −18.3% of economic value for a 200 basis point move, which on the implied net worth of ₹38,333.33 crore is roughly ₹7,015 crore of value at risk from a rate move the bank doesn't control. Every other letter sits at or above its comfort threshold.

P13. A US property and casualty insurer, two conventions (US, $). An insurer writes $3,800 million of net premium and earns $3,600 million. Losses are $2,484 million, loss adjustment expenses $180 million and underwriting expenses $1,026 million. Compute the loss ratio, both expense ratios, both combined ratios and the underwriting result. Then recompute the combined ratio excluding $140 million of adverse prior-year development.

Solution. The loss and loss-adjustment ratio is (2,484 + 180) ÷ 3,600 = 74.0000%. The expense ratio is 1,026 ÷ 3,800 = 27.0000% on the trade basis and 1,026 ÷ 3,600 = 28.5000% on the GAAP basis, so the combined ratio is 101.0000% or 102.5000%, a gap of 1.5000 percentage points. The underwriting result is 3,600 − 2,484 − 180 − 1,026 = −$90.00 million. Excluding the adverse development, the loss ratio falls to (2,664 − 140) ÷ 3,600 = 70.1111% and the GAAP combined ratio to 98.6111%, an underwriting profit. Which figure describes the business depends on whether last year's reserves were an unlucky estimate or a policy, and only the run-off triangle across several years answers that.

P14. Currency and the insurance ratios (US, $). A US insurer's UK subsidiary earns £480 million of net premium, incurs £312 million of claims and £158.4 million of expenses, and holds float of £1,400 million producing £96 million of investment income. The average rate is $1.2400 and the closing rate $1.1900. Compute the combined ratio in both currencies and the yield on float in both, and explain the difference.

Solution. The combined ratio is (312.0 + 158.4) ÷ 480.0 = 98.0000%, and it is 98.0000% in dollars too, because all three inputs are income-statement lines translated at the same average rate. The local yield on float is 96.0 ÷ 1,400.0 = 6.8571%. Translated, investment income is 96.0 × 1.2400 = $119.04 million and float is 1,400.0 × 1.1900 = $1,666.00 million, giving 7.1453%, a distortion of +0.288115 percentage points. The reason is the scaling identity: 1.2400 ÷ 1.1900 = 1.042017, and 6.8571% × 1.042017 = 7.1453%. Any ratio crossing the income statement and the balance sheet carries that factor, and any ratio inside one statement doesn't.

Applied mini-project

One foreign subsidiary, translated and remeasured, and one bank scored.

There are two deliverables, and you build each one from primary filings. Every figure must be traced to the company's own document on its website, on BSE or NSE, or on EDGAR, with the year, the note number and the page recorded. Where a rate isn't disclosed, derive it and say how.

Deliverable one: the subsidiary

Find a filing that discloses enough about one foreign subsidiary to rebuild it. Indian annual reports carry a statement of subsidiary financials, often called AOC-1, giving each subsidiary's share capital, reserves, total assets, total liabilities, turnover and profit. Many US filers give the same detail in a guarantor or significant-subsidiary footnote. Pick the largest foreign subsidiary you can rebuild, and state what you had to assume.

The first artefact is the rate history. Assemble the opening, closing and average rates for the year, the rate on any dividend date, the historical rate for share capital, and your best estimate of historical rates for inventory and fixed assets, each with its source and its basis. Test 1: the average rate you use must reproduce the group's disclosed translated revenue for that subsidiary to within 1%, or the difference must be explained.

The second is the two translations. Produce the subsidiary's income statement and balance sheet under both the current-rate and the temporal methods. Test 2: the translation adjustment must be computed both as the balancing figure and as the three-part decomposition, and the two must agree to the smallest unit you are printing. Test 3: the remeasured balance sheet must roll forward from opening remeasured net assets with no residual, and the remeasurement gain or loss derived from the net monetary position must equal the one derived as the income-statement residual.

The third is the ratio grid. Revenue growth, gross margin, asset turnover, current ratio, debt to equity and return on equity, in the subsidiary's own currency and under both methods. Test 4: every mixed ratio under the current-rate method must equal its local counterpart multiplied by the average rate over the closing rate, and you must show the multiplication. Don't just assert it.

The fourth is the reconciliation and the note. State the net-asset gap between the two methods and show it equals the sum of each non-monetary item multiplied by the difference between the closing rate and its historical rate. State the profit gap and show it equals the cost-of-sales difference plus the depreciation difference plus the remeasurement. Then write one page: which method the company uses and on what stated grounds, what the choice does to reported growth and to consolidated equity, whether a net investment hedge is running through the translation reserve, and what an analyst should adjust before comparing this company with a domestic peer.

Deliverable two: the bank

Take one bank, in either country, that publishes enough to rebuild its liquidity ratios. The Pillar 3 disclosure, the liquidity coverage ratio disclosure template and the regulatory call report or the Indian equivalent are the sources.

The fifth artefact is the CAMELS scorecard, six letters, every one carrying at least two computed numbers with the formula shown. Use the twelve-KPI panel you already have for capital, asset quality, earnings and liquidity, and build management and sensitivity from the growth gap, the efficiency ratio, the provisioning direction, the unrealised securities position against common equity tier 1, and any disclosed duration or economic-value sensitivity. Test 5: each letter must carry a number and a stated threshold, and the weakest letter must be named with the figure that makes it weakest.

The sixth is the two Basel ratios, recomputed. Build the liquidity coverage ratio from the disclosed high-quality liquid asset composition and outflow categories, testing the 40% Level 2 cap and the 75% inflow cap explicitly, and build the net stable funding ratio from the disclosed available and required stable funding tables. Test 6: your recomputed figures must land within 5 percentage points of the bank's disclosed figures, or every difference must be attributed to a named category and factor.

The seventh is the stress. Recompute the liquidity coverage ratio excluding any securities the bank holds at amortised cost and has stated it intends to hold to maturity, and recompute common equity tier 1 marking those securities to their disclosed fair value. Then apply a further 100 basis point rate rise using the disclosed duration. Test 7: state the headroom to the bank's own effective capital requirement, in basis points, after each step.

Deliverable. Two pages of numbers and two pages of argument. The numbers pages carry the rate history, both translations, the ratio grid, both reconciliations, the CAMELS scorecard, both Basel ratios and the stress, with the seven test residuals shown in full. The argument pages answer two questions in figures. First, how much of this group's reported growth and consolidated equity is a currency artefact, and what would change if the functional-currency judgement were revisited. Second, which CAMELS letter would fail first at this bank and what specific event would trigger it. Close each with the single assumption whose movement would change your answer, and quantify what it would take.

The rubric runs to 21 points. You pass at 17, with no zero on any row.

Any artefact whose test fails scores zero for that row, however well the note reads. A translation adjustment that agrees with itself only because it was computed once isn't evidence of anything. The second route is the entire point, and a submission that skips it has produced a number nobody has checked, including its author.

Reading & resources

• IAS 21 and Ind AS 21, read against ASC 830. The functional-currency indicators are two pages in each and they are the whole judgement. Read the IAS 21 indicators, then the six ASC 830 economic factors, then ask which of them a filing you already hold actually addresses. [Free] [Intermediate]
• IAS 29 and Ind AS 29, with the ASC 830 highly-inflationary paragraphs beside them. The shortest way to see two standard-setters solve one problem in opposite directions. Read IAS 29's restatement mechanics first, then the American requirement to use the reporting currency as the functional currency, and the reason for the divergence becomes obvious. [Free] [Advanced]
• IFRS 9 and Ind AS 109 on hedge accounting, specifically the net investment hedge paragraphs, and ASC 815-35. Short, and they settle where the effective portion goes, how effectiveness is measured against the net investment, and what happens on disposal. [Free] [Advanced]
• The Basel Committee's liquidity coverage ratio and net stable funding ratio standards, and the disclosure templates that accompany them. The factor tables are the reference you will keep returning to. Read them once end to end so you know what the categories mean, then use the national rule for anything you are actually computing. [Free] [Intermediate]
• The Reserve Bank of India's master directions on the liquidity coverage ratio and the net stable funding ratio, and the corresponding US agencies' liquidity rule. National implementations differ on run-off factors, on what counts as a stable retail deposit and on the treatment of derivatives. The differences are large enough to change a conclusion. [Free] [Intermediate]
• The Uniform Financial Institutions Rating System, the source document for CAMELS. It describes what a supervisor looks at under each letter, which is a better checklist than any secondary summary, and it is short. [Free] [Intermediate]
• The FFIEC's Uniform Bank Performance Report for any American bank. Free peer-group comparison on every ratio in the panel, computed consistently, with the bank's percentile against its peers. Nothing in Indian disclosure matches it. [Free] [Beginner to Intermediate]
• A bank's Pillar 3 disclosure alongside its annual report. The Pillar 3 document carries the risk-weighted asset build, the liquidity templates and the interest-rate sensitivity that the annual report summarises. Reading them together is how the sensitivity letter becomes computable. [Free] [Intermediate]
• The 2023 US regional bank failures, in the regulators' own reviews and not the commentary. The supervisory post-mortems set out which letters were rated what, and when. They're the best available case study in a bank that looked adequate on capital and asset quality while failing on sensitivity and liquidity. [Free] [Intermediate to Advanced]
• Any Indian filer's AOC-1 statement of subsidiary financials, and any US filer's guarantor footnote. These are the disclosures that make the mini-project possible, and knowing what they contain changes what you can rebuild. Open three and note what each does and doesn't give you. [Free] [Beginner]
• The quarterly results release of a large Indian IT services company, and of a large American consumer multinational, read for the constant-currency disclosure. Both publish it every quarter, both define it slightly differently, and comparing the two definitions in a single sitting teaches more than any description. [Free] [Beginner to Intermediate]
• A general insurer's run-off triangle disclosure, in either country. Several years of reserve development in one table, and the only place you can answer the reserving question instead of just raising it. [Free] [Advanced]
• Kieso, Weygandt and Warfield, Intermediate Accounting, on foreign currency, and Choi and Meek, International Accounting, on translation and its analysis. The first drills the mechanics; the second is unusually good on why the translation methods produce the distortions they do. [Paid] [Intermediate]

One exercise is worth more than rereading any of these. Take a company you already follow that has a large foreign subsidiary, rebuild its translation from the subsidiary statement, and try to reconcile your translation reserve movement to the one in its equity statement. Whatever fails to reconcile is a hedge, an acquisition, a disposal or a rate you guessed wrong, and finding out which is an hour that teaches more about the equity note than any chapter.

Teach it back & journal

Feynman prompt. Find someone who can read a balance sheet and has never thought about currency. Put two sets of figures for BPI Americas in front of them, one showing net assets of ₹391.50 crore and a return on equity of 16.0920%, the other showing ₹312.75 crore and 20.5436%. Tell them both are correct, that the subsidiary earned $7.5 million either way, and ask which company they'd rather own. Let them commit before you explain anything.

Then make them find the mechanism themselves. This is the part people want to skip, and it's the part that does the teaching. Show them that the difference is ₹78.75 crore, that ₹76.50 crore of it is one line, and that the line is a factory bought when the rupee was at ₹70.00. Don't tell them the rule. Ask why a factory's rupee value should depend on the year it was bought, and let them argue both sides, because both sides have a case and the standard setters have not settled it either.

Finish with the part that catches everyone. Ask them to predict what happens to the two answers if the rupee strengthens instead. Almost everyone says the gap reverses. Show them that remeasured net assets stay lower, at ₹315.75 crore against ₹346.50 crore, because both closing rates are still above ₹70.00. If they then ask, unprompted, when the factory was built before commenting on any multinational's balance sheet, the idea has installed.

For the bank half, one exchange is enough. Give them Chesapeake Union's liquidity coverage ratio of 134.7190% and ask whether the bank is liquid. Then hand them the fair value of the held-to-maturity book and ask again. The interesting moment is when they realise nobody lied.

Journal prompt. For two weeks, every time you open a results release, find the currency disclosure before you look at any growth number. Write down four things: the functional currency of the largest foreign operation, the movement in the translation reserve, whether a net investment hedge is running through it, and the constant-currency figure with the convention that produced it. Do this for at least six companies, split between Indian and American filers, including one with a subsidiary in a hyperinflating economy if you can find one.

Then answer in writing. How many of the six could you reconcile the translation reserve movement for, and what was left over in the ones you couldn't? Note which company's reported growth changed most when you stripped the currency out, and whether any commentary you read mentioned it. For the American filers, did any hold securities at amortised cost at a fair value materially below carrying value, and what did that do to their capital ratio when you marked it?

Close with a paragraph on the habit this area demands. Currency and regulation both take facts you can already read and route them through a rulebook, and every routing decision is disclosed somewhere. The usual failure is accepting a published ratio without asking which rule produced it. Which of the four checks are you most likely to skip when a deadline is close? Write down what goes wrong when you skip it, and what you will change this month so that skipping costs more than doing.