Executive answer
A defensible rate benchmark replicates credit-pricing conditions: date, currency, term, amount, rating, seniority, security, amortization, sector, geography, reference, options and fees. Searching for “dollar loans” and calculating a range does not sufficiently control comparability.
The process begins with loan delineation and term sheet. Then determine rating and implicit support, define search strategy, obtain contemporaneous data, filter, adjust and test. The conclusion should be reasonable for borrower and lender and reconcile with agreement and calculation.
A statistical range cannot cure weak data. Quality depends on comparables and adjustments. Another analyst should be able to reproduce the sample and understand rejection decisions.
Search specification
| Variable | Tested transaction | Search criterion | Possible adjustment |
|---|---|---|---|
| Date | signing/funding | contemporary window | curve/date |
| Currency | MXN, USD, EUR | same currency | swap/curve cautiously |
| Term | expected life | comparable band | interpolation |
| Rating | borrower/instrument | credit band | rating spread |
| Seniority | senior/subordinated | same priority | subordination premium |
| Security | yes/no/type | similar structure | guarantee value |
| Amortization | bullet/partial | similar profile | duration |
| Amount | principal | comparable scale | liquidity/size |
| Sector | industry | similar risk | qualitative analysis |
| Geography | market/country | similar exposure | country risk |
| Rate | fixed/floating | same structure | swap/fallback |
| Fees | total cost | include fees | all-in yield |
Complete before searching to reduce bias.
Market date
Price should reflect information reasonably available when agreed or modified. Use contract, commitment or funding date based on facts and explain. The signature may not be economic date if terms were committed earlier.
Set a window close enough to represent market and wide enough for observations. Under volatility, long windows distort. Moving comparables to a common date through curves may improve reliability but needs method and data.
Do not use later rates for convenience. If prepared late, reconstruct historic available information. Retain source files and timestamp.
Currency and reference
Nominal rates in different currencies contain different inflation and curves. Prefer same currency. Swap conversion may be possible but introduces assumptions and requires validation.
For floating rates identify reference, tenor, spread, floor, reset and fallback. One-month SOFR differs from three-month; TIIE and local references have separate markets. For fixed rates compare yield and duration.
Contract currency must fit delineation. A perfect USD benchmark does not cure debt an independent borrower would take in MXN.
Term, duration and amortization
Use expected life, not legal date alone. Prepayment and extension options matter. A five-year bullet concentrates risk; amortization reduces exposure. Duration compares cash-flow profiles more precisely.
Filter maturities around term and explain band. With sparse data, interpolate curves or adjust spreads with support. Do not mix revolvers, perpetual bonds and amortizing loans without control.
Grace periods and capitalized interest change profile. Address them qualitatively and quantitatively where material.
Borrower and instrument rating
Corporate rating estimates general risk; instrument rating may differ through seniority, security and recovery. Document method, ratios, sector, forecast and date. Separate stand-alone rating and implicit support.
Do not copy a notch uplift. Analyze strategic importance, integration, name, history and barriers. Implicit support is not an explicit guarantee.
Map agency and model scales. Distinguish issuer and issue ratings.
Comparable sources
Prioritize reliable internal comparables: borrower external debt, detailed bank offer or lender third-party credit. Check conditions. A secured facility does not compare directly with unsecured debt.
External sources include bonds, syndicated loans, credit databases and curves. Bonds provide observable pricing but differ in liquidity and issuance; loans may fit better but lack fields.
Describe source, coverage and data. Keep exports and criteria.
Filter strategy
Use objective criteria: instrument, date, currency, rating, term, seniority, sector, region, security, status and yield availability. Record initial, excluded and final populations. Do not delete observations solely to narrow range.
Review each comparable for convertibility, distress, related parties, restructuring and options. Automated filters miss issues.
If final sample is small, explain and triangulate with curves or offers. A large heterogeneous sample is not better.
If the benchmark does not preserve strategy, export and rejection reasons, it cannot be reproduced or defended even when its range looks reasonable.
Comparability adjustments
Adjustments should improve reliability and use data. Adjust date through curve movement, term through spreads, seniority or security through market evidence. Avoid subjective chains producing a target result.
Document formula and sensitivity. Show before and after. If adjustment exceeds differences across comparables, question the sample.
Avoid double-adjusting the same risk—for example, rating may already capture some country risk. Use a risk matrix.
Guarantees and security
A guarantee may raise instrument rating or reduce loss and spread. Define scope, guarantor, capacity, coverage, priority and enforceability. Do not treat a comfort letter as guarantee without analysis.
With an explicit guarantee, price the loan with it and analyze fee separately without double charge. Compare borrower cost with and without guarantee and incremental benefit.
Specific collateral also reduces risk. Document value, lien and recovery.
Fees and total cost
Coupon is not everything. Include arrangement, commitment, agency, guarantee, prepayment, unused-line and issue discount. Calculate all-in yield or effective cost over horizon.
Compare instruments on a common basis. A lower rate with a high upfront fee may be outside range. Identify the function compensated by each fee.
Do not add a margin when a comparable already includes the fee.
Range and point
Mexican law provides for ranges under its rules; financial analysis must still support placement. Being in range does not cure poor delineation. Explain why median or another point reflects terms.
Use consistent statistics and explain outliers. Do not trim toward a conclusion. Compare with borrower alternatives and lender return.
Run sensitivity for rating, term and date. Show exposure to a notch or curve change.
Borrower validation
Compare cost with expected use, capacity and offers. An independent borrower would not accept project-destroying conditions where alternatives exist, except under high risk. This does not cap rate at accounting return; it checks coherence.
Assess covenants and flexibility. A low rate with severe restrictions may not be better. Economic comparison includes terms.
Document treasury approval and alternatives.
Lender validation
Review funding cost, operation, capital and risk. Funding cost is not a maximum. A lender controlling credit risk differs from a conduit.
Confirm ability to lend and bear loss. Compare yield with investments of similar risk. Document monitoring.
Both parties should prefer the transaction over realistic alternatives.
Recommended file
- Term sheet and date.
- Capacity and delineation.
- Rating and support.
- Predefined strategy.
- Original export.
- Filters and rejections.
- Final comparables.
- Adjustments and formulas.
- Fees and total cost.
- Range and sensitivity.
- Bilateral validation.
- Agreement and calculation.
Illustration
Mexico borrows USD 10 million, senior unsecured, three-year bullet, BB rating. An initial USD bond set has 120 observations. Filter by date, rating, term and seniority; remove convertibles and distressed issues. Eighteen remain. Move to common date using curves and review fees.
A secured bank offer is excluded without adjustment but triangulates. Compare range with spread and capacity. Select near median for standard terms. A change to subordinated requires review.
The memo retains export and formulas. Updates do not overwrite the original.
Risk signals
- Later benchmark date without reconstruction.
- Mixed currencies.
- Group rating copied.
- Bonds and loans uncontrolled.
- Guarantee ignored.
- Rejections driven by outcome.
- Coupon only, no fees.
- Irreproducible adjustments.
- Range without selected-point rationale.
- No bilateral validation.
Governance and update
Update when agreed, materially modified or repriced. A fixed rate does not need annual repricing merely because markets move, but conduct should be monitored. A floating rate updates reference per agreement, not spread without an event.
Retain data under source licensing and permitted exports. Record analyst, reviewer and date. Independent review recalculates sample and rate.
Integrate benchmark with agreement, calculation, payments and reporting. The economic file should not stand alone.
Data quality control
Define required fields and units. Yield may be percentage, basis points or decimal; amount may be thousands or millions; term may be date or years. Normalize before calculation. Check duplicates, currency, negative yields, maturities and missing data.
Retain raw, working and output files. Never alter raw data. The working file records transformations and exclusions. Output shows final sample and statistics. Use stable identifiers for tracing.
A second reviewer should reproduce filters and sample adjustments. Test formulas, dates and percentiles. Record database version and access terms. If licensing prevents retaining data, preserve the permitted methodology and evidence.
From spread to contractual rate
Where comparables produce spreads, add them to a reference consistent in date and tenor. State whether spread is over spot, forward or average. For fixed rates combine curve and spread or use comparable yield without mixing concepts.
Calculate interest using contractual day count and calendar. A benchmark of 450 basis points over a reference does not prove the operating calculation uses the same reference. Reconcile every reset, floor and fallback.
Show an illustrative numerical example clearly separated from the actual conclusion. Readers should distinguish method from result.
Selected point and negotiation
Median may be practical when no other point is supported, but term-sheet facts can justify placement. Stronger protection, amortization and covenants may support a lower spread; subordination, concentration and flexibility may support higher. Do not apply a premium without evidence.
Retain offers and treasury comments. A negotiated in-range rate is stronger with contemporaneous alternatives. If agreed rate differs from study, explain a modification or error rather than silently changing the sample.
Fixed versus floating conversion
If operations require comparing fixed and floating instruments, document the swap curve and date, cash-flow schedule and conversion convention. Show sensitivity to reasonable curve movement. Do not treat today’s fixed equivalent as the one available at origination.
For variable loans, assess whether floor has value, especially near low rates. Caps, collars and reset lags can materially change economics. An instrument with embedded options may require exclusion where reliable adjustment is unavailable.
Credit spreads under sparse data
Where local-currency observations are scarce, consider a build-up approach using a base curve, rating spread and clearly separated adjustments. Triangulate against bank offers, group external debt or market indices. A build-up is not license to stack arbitrary premiums.
Document correlation among country, sector and rating. Avoid double counting. Use a sensitivity table and state which element drives conclusion. If uncertainty remains high, present a wider reasonable range and stronger monitoring rather than false precision.
Year-end review
Confirm principal, days, reference, spread, fees and interest agree across contract, calculation, bank, entry, withholding and returns. Investigate rounding, cut-off, FX or capitalization differences.
The benchmark may remain valid while terms remain unchanged; operating reconciliation is annual or more frequent. A material modification needs a new contemporaneous analysis with the original preserved.
Reviewer checklist
- Search criteria were defined before results.
- Date and currency match the transaction.
- Rating is documented and current at signing.
- Instrument seniority and security are controlled.
- Rejections have factual reasons.
- Adjustments improve rather than obscure comparability.
- Fees enter total cost once.
- Range and point are reproduced.
- Borrower and lender alternatives are tested.
- Contract implements the conclusion.
Record exceptions and their quantitative effect. Approval should identify preparer, reviewer and version.
Sources and cutoff
This article was verified as of August 2, 2026. Consult current Mexican Income Tax Law, the OECD Guidelines 2022, Chapter X, OECD financial transactions guidance and the OECD Mexico profile.
Zugzwang’s Loan Benchmark Memo documents rating, strategy, sample, adjustments, total cost, range and validation for a reproducible rate.