Interest Rate Benchmarking for Intercompany Loans: CUP Approaches and Yield Construction

01The Most Appropriate Method for Loans

For pricing the interest rate on a loan that has been accurately delineated as debt, the comparable uncontrolled price (CUP) method is generally the most appropriate. A loan is a relatively standardized instrument, and the market produces an abundance of observable prices for debt: bond yields, syndicated loan rates, and quoted lending terms. This is why CUP is the workhorse method for financial transactions in a way it is not for many other controlled transactions.

Two forms of CUP are available, and the analysis prefers the more reliable one where both exist:

• Internal CUP. Where one of the parties has entered into comparable uncontrolled loans with independent third parties, those transactions are internal comparables. This is often the most reliable benchmark, because the borrower’s actual third-party borrowing, or the lender’s actual third-party lending, reflects the real terms available to that party. A recent comparable bank loan to the borrower is powerful evidence of its arm’s length cost of funds.
• External CUP. Where no suitable internal comparable exists, the rate is built from external market data: yields on bonds issued by independent companies of similar credit quality, or rates on comparable third-party loans. This is the more common situation, and it requires constructing a yield rather than reading a single price, because no external bond will match the tested loan on every characteristic.

Cost-based approaches, in which the lender’s cost of funds is marked up, are sometimes seen but are generally weaker, because they price the lender’s position rather than the borrower’s arm’s length cost of borrowing, which is what the arm’s length principle asks about.

02Constructing the Yield

When an external CUP is used, the arm’s length rate is constructed as a risk-free base plus a credit spread. Each component is set by the loan’s economically relevant characteristics.

2.1 The Risk-Free Base

The risk-free base is the return on a near-riskless instrument in the same currency and for the same tenor as the tested loan, usually a highly-rated government bond yield. Two characteristics of the loan fix this base:

• Currency determines which risk-free curve applies. A loan is priced off its own currency’s risk-free rate, not the lender’s or borrower’s home-currency rate. Currency also embeds inflation and country expectations, which is why the same borrower faces different rates on loans in different currencies.
• Tenor determines the point on the yield curve. Because the curve is generally not flat, a five-year loan carries a different base than a one-year or ten-year loan. The base is read at the tenor matching the loan.

2.2 The Credit Spread

The credit spread is the premium over the risk-free base that compensates the lender for the borrower’s default risk. It is set by the borrower’s credit rating, derived as described in the credit rating article, and built from comparable instruments: bonds or loans issued by independent borrowers of the same rating, in the same currency, at a similar tenor.

Because no comparable matches the tested loan exactly, the spread is adjusted for differences in the characteristics that affect risk:

The all-in arm’s length rate is the risk-free base plus the adjusted credit spread. This constructed yield is what the intercompany rate is tested against.

03Selecting and Screening Loan Comparables

Building a credit spread from external data requires a set of comparable instruments, assembled with the same comparability discipline used throughout benchmarking. The comparables must match the tested loan on the characteristics that drive pricing, and differences that cannot be matched must be adjusted for or the set narrowed. The screening logic mirrors a benchmarking search:

• Starting population: drawn from a bond or loan database.
• Rating filter: limited to the borrower’s rating band.
• Currency filter: limited to the loan’s currency.
• Tenor filter: a range around the tested loan’s maturity.
• Structure screen: instruments differing materially on seniority, security, or structure are removed or adjusted.

One screen matters more here than in many searches:

• Pricing date. Interest rates move with the market, so the comparables must reflect conditions at the time the loan was priced, normally the date the loan was entered into. A rate benchmarked to data from a different rate environment is not comparable, however well matched on credit characteristics.

04A Worked Yield Construction

Consider Solvang Marine, a fictional shipping subsidiary, borrowing five-year funds in US dollars from its group treasury. The borrower has been rated BBB after the implicit-support analysis.

Step one, the risk-free base. The five-year US government bond yield at the pricing date is 4.05 percent. This is the base, set by the loan’s dollar currency and five-year tenor.

Step two, the credit spread. A screen of the bond database for BBB-rated issuers, in US dollars, at tenors around five years, with senior unsecured ranking comparable to the tested loan, yields a set of seven comparable spreads over the risk-free rate, in basis points: 178, 188, 196, 205, 215, 222, and 240.

Step three, aggregate to a range. Applying the OECD interpolation method to this set:

Reading the result:

• The arm’s length spread runs from 192 to 218.5 basis points, giving an all-in interquartile range of 5.97 percent to 6.24 percent, with a median of 6.10 percent.
• If the intercompany rate on the Solvang facility falls within this range, it is treated as arm’s length.
• If it falls outside, the median at 6.10 percent is the usual point of adjustment.

The interquartile range and interpolation method here follow the same conventions applied in any benchmarking study; the mechanics are common to profitability and interest-rate benchmarking alike.

05Common Pitfalls in Rate Benchmarking

Several recurring errors weaken interest-rate studies, and each traces back to a mismatch on an economically relevant characteristic:

• Currency mismatch. Pricing a loan off the wrong currency’s risk-free base. The base must match the loan’s currency, not the parties’ home currencies.
• Tenor mismatch. Reading the risk-free base or the comparable spreads at the wrong point on the curve. Both must match the loan’s tenor.
• Stale market data. Using comparables from a different rate environment. They must reflect the pricing date, even if well matched on credit.
• Ignoring the rating. Building a spread from comparables of a different rating band prices a different risk. The comparable set must match the borrower’s rating, which is why the rating determination precedes and feeds pricing.
• Overlooking security and ranking. A senior secured comparable does not price an unsecured subordinated loan. Differences in ranking and collateral must be adjusted for.

The practical takeaway is that rate benchmarking is comparability analysis applied to debt: the reliability of the rate depends on how closely the comparables match the tested loan on the characteristics that drive price, and on adjusting for the differences that remain.

06Pricing in the Loan Workflow

Pricing is the final step, and it depends on everything before it:

• Delineation established that the arrangement is debt and identified its characteristics.
• Capacity fixed the arm’s length principal.
• The rating fixed the borrower’s creditworthiness and therefore the spread band to search.

Only with those settled does the yield construction produce a defensible rate. A rate built without them, off an unexamined principal, an underived rating, or mismatched comparables, is exposed at exactly the points a reviewer will test. The flagship guide sets out the full sequence as a workflow the reader can run end to end.

Comp-Press · Transfer Pricing Practitioner’s Guidance. General best-practice reference, not legal or tax advice.