A20 reader v04 numerical exhibits and calculation notes

All values below are historical source observations, arithmetic derived from retained displayed cells, or a labelled hypothetical. SOURCE_MAP.md resolves the source IDs to original files, hashes, URLs and exact locators. A basis point (bp) is 0.01 percentage point. Calculations use displayed precision, not hidden RBI unrounded data.

Exhibit 1 - fresh repo-linked loan spreads and fixed-weight envelope

Population: RBI Table IV.5, sector-specific fresh loans linked to the repo external benchmark; the three printed groups are public sector, private sector and domestic banks. Dates: April 2022 and August 2024. Source: S1, Chapter IV Table IV.5. All spread cells below are observed printed percentage points. Rate rises and envelope bounds are derived bp using the observed +250bp repo change. The envelope applies only if domestic is an exhaustive, comparable amount-weighted union of the two displayed bank groups at both dates.

Method: for group g, Δ rate_g (bp) = 250 + 100 × (spread_Aug - spread_Apr). If an unchanged public weight w lies between zero and one, Δ domestic = w Δ public + (1-w) Δ private, hence it must lie between the group changes. For education, 210 - max(170,157) = 40bp. Each two-decimal printed spread can differ from its unrounded value by at most 0.005 percentage point. The domestic change could be 1bp lower and the public change 1bp higher, leaving at least 38bp conditional violation. This is a rounding bound, not a confidence interval.

The education public-weight endpoints 77.52% and 40.52%, and a 42.92bp residual under one conditional decomposition, are algebraic implied quantities only if the exact pooling identity holds. They are not observed amounts, public-bank shares, lender switching, or an identified composition effect; the manuscript deliberately does not headline them. The August MSME group spread gap is only 0.05 percentage point, so its algebraic weight is weakly identified; two-decimal rounding permits 0-40% at that endpoint. April 2022 cells match S2 Table IV.5; S2 August 2023 and S3 March 2024 provide sparse intermediate education domestic spreads of 3.52 and 3.98 percentage points, both observed. None creates a monthly path.

Exhibit 2 - selected central-government auctions

Status: observed printed auction rows. Unit: percent for coupon/cutoff yield, ₹ crore face amount, years residual maturity. Population: named central-government dated-security reissues; cutoff yield is the individual auction cutoff, not accepted-bid weighted yield. Source locators: S5-S7 Statement 1 and PDF pages specified in source map.

Derived: the same-GS-2029 displayed cutoff-yield change is (7.4-7.09)×100 = 31bp, approximate because the February cell has only one decimal. The residual maturity fell 0.82 years. The August GS 2034 bond is a different instrument and cannot complete a same-bond three-date series. The retained DEA April-June 2022 report has an inconsistent Q1 weighted issuance-yield aggregate (6.95% in its narrative/Table 4.2, 7.23% in Statement 1); that ambiguous aggregate is not used here.

Exhibit 3 - gross Market Loans interest and matching book stocks

Population: central-government Market Loans accounting category, rather than the two selected securities. The interest values are observed actual gross Grant 39 item 2.01 for completed fiscal years (S8 PDF p152; S9 PDF p148). The stock values are observed book values in one 2025 Receipt Budget vintage, Statement 1(i), S10 PDF p60. Unit: ₹ crore, except derived ratios. Treasury bills, state debt and other liabilities are not in the book-stock denominator.

Derived: gross interest / ((opening stock + closing stock)/2) × 100; the ratio rises 0.037006 percentage point ≈ 3.7006bp. Gross interest increases ₹73,306.33 crore, 11.224207%. Two endpoint stocks do not observe the intra-year average. The ratio is not an auction yield, a coupon paid on the named bonds, cash interest, a time-weighted debt cost, or an August 2024 observation. Gross item 2.01 excludes the separate “Less accrued interest” adjustment and should not be confused with Grant 39 net total.

Exhibit 4 - explicitly hypothetical home-loan payment

Assumptions (hypothetical, not a surveyed borrower): ₹5,000,000 balance at a rate reset; 120 monthly payments remain; original nominal APR 7.5%, reset APR 10%; monthly rate is APR/12; reset before the next payment; no fee, prepayment or new borrowing. S11 Article 5.3(d)-(f) supplies the contract mechanism, not these values or a bank commitment. Local source calculation: session03-episode/AMORTISATION.json, SHA-256 7cc3210b13b8e76e2489a9e53f5c2122f1c8afe8efdc31e2445f6e80c91a83b1.

Formula for fully amortising monthly EMI: P r / (1 - (1+r)^-n) with principal P, monthly rate r, and n payments. At 7.5% APR and 120 months: ₹59,350.88. At 10% APR, retaining 120 months: ₹66,075.37, +₹6,724.48 or 11.33%, calculated before rounding the two EMI levels (subtraction of the displayed paise-rounded levels gives ₹6,724.49). Reader prose therefore says about ₹6,724. At 10% APR, retaining the first EMI: 145 full old-EMI payments and a smaller 146th final payment, thus 26 more calendar months than 120, if permitted. First-month reset interest is ₹5,000,000 × 0.10 / 12 = ₹41,666.67. Interest-only threshold for the old EMI is 12 × 59,350.8845679 / 5,000,000 = 14.2442% nominal APR. At equality the principal does not decline. Contractual term caps and bank discretion can override the purely mathematical tenor path.

Exhibit 5 - firm rates, distinct populations

Observed percent levels: S13 Annual Report 2022-23 Chapter III Tables III.6-III.7, and S14 Annual Report 2023-24 Tables III.7-III.8. Dates are end-March 2022/2023/2024; population is SCBs excluding RRBs; sector fresh and outstanding rupee-loan WALRs.

Separate observed chart changes: S1 Chart IV.29a, May 2022-August 2024, domestic banks, MSME fresh/outstanding +133/+95bp; large-industry fresh/outstanding +161/+71bp. S1 Table IV.5’s MSME +160bp is repo-linked fresh lending only. The annual levels, chart changes and Table IV.5 subgroup are not a mechanically joined series.

Exhibit 6 - bank lending and term-deposit rates

Unit: percent. Population: RBI press releases’ SCBs excluding RRBs and small-finance banks; fresh/outstanding rupee-loan WALR and fresh/outstanding rupee term-deposit WADTDR. March 2023 cells are observed in S15, 31 May 2023 release comparison bullets. August 2024 cells are observed in S16, 30 September 2024 release, with its merger-exclusion footnote. April 2022 cells are derived, not directly observed, by subtracting S1 Table IV.4 May 2022-August 2024 changes from S16 August levels.

S1 Table IV.4 gives August-minus-April changes of +190bp fresh loan, +119bp outstanding loan, +243bp fresh deposit, +190bp outstanding deposit; subtracting from 9.41/9.91/6.46/6.93 yields the derived April row. The March 2023 fresh/outstanding loan rates minus derived April levels are +181/+100bp, consistent with S13 Annual Report 2022-23 Table III.3’s May 2022-March 2023 memo row. This cross-check does not turn April into a directly observed press-release cell. S14 Annual Report discussion identifies savings deposits at roughly 30% and current accounts around 9% of total deposits at end-March 2024; S1 discusses declining CASA share and NIM pressure. The table cannot calculate NIM because the listed averages cover different flow/stock populations and omit other balance-sheet items.