Today we're looking at quasi-equity instruments: hybrids that sit between debt and equity and require judgment on both sides of that line to value properly. There are several types worth covering, convertible loans, SAFEs, loans with warrants attached, but we'll start with one: a loan with warrants attached.
What is a loan with warrants attached?
The structure is common wherever a lender is taking on real credit risk without the collateral or covenants a traditional bank loan would carry, venture debt being the clearest example. The lender extends a loan on terms that look thinner than the risk would otherwise justify, and gets warrants, the right to buy equity later at a fixed price, attached as a kicker. The loan and the warrant are usually detachable: two separate instruments bundled into one transaction, priced together at origination but capable of being valued, and later sold or exercised, independently.
How do you price it with no market price to check against?
Pricing a loan-plus-warrants package from scratch means estimating credit spread, equity volatility, and the warrant's expected term, all without a market price for either the loan or the underlying equity to check the answer against. That's a lot of assumptions stacked on top of each other, and calibration is what keeps the model honest: anchor the combined value to what was actually paid for the package at transaction date, and back out whichever inputs are needed to make the model agree with that price, an approach AICPA describes. Suppose a fund pays $200m for a package of debt and warrants together. The warrants are valued directly, using equity volatility and their expected term, at $15m. Debt gets the residual, $185m. The implied market yield on that $185m is then checked for reasonableness. The spread isn't assumed, it's solved for.
It's worth being explicit about what the warrant is actually doing in that structure. It isn't just an upside sweetener bolted on for good measure, part of the lender's expected return is coming through the warrant rather than the coupon. Warrants exist to compensate the lender for the risk being taken on, and for providing capital that few other sources would extend on these terms, so a model that prices the loan on spread alone and treats the warrant as an afterthought is missing where a meaningful share of the return is actually built in.
What changes at each measurement date?
At each later measurement date, the same three pieces get repriced separately. The underlying equity is revalued first, ideally itself calibrated to a recent transaction. The warrant is then repriced off that updated equity value, using current volatility and remaining term. The loan is repriced using its contractual cash flows discounted at a current market yield, reflecting whatever's happened to credit quality and market rates since origination. Sum the three, and if the underlying equity, volatility and yields haven’t moved much, the total should land close to the original price paid, that's the sanity check that the model is still behaving consistently, not drifting from what calibration established at the start.
How the worked example splits at origination and is re-checked later. The spread is the output of the calibration, not an input to it.
This simple addition approach rests on one assumption worth being explicit about: that the loan and the warrant are genuinely separable. That holds here because the warrant is a detachable, stand-alone instrument rather than a feature embedded inside the debt itself.
What to watch for: change-of-control provisions. Many of these structures include terms that force early repayment of the debt, or truncate the warrant's remaining life, if the company is sold. Both the warrant's expected term and the loan's expected repayment date need to reflect those triggers as of the measurement date, not just run mechanically to stated maturity, otherwise the model is pricing an exit path the instrument doesn't actually have.
None of this removes judgment, it just disciplines it. Every input, volatility, spread, expected term, is an estimate, and the only real check available is whether the combined value stays consistent with what was actually paid and with how the underlying business has genuinely changed since, with each of the assumptions being backed by a reliable source. That consistency has to be maintained at every measurement date, not assumed, which is what fair value measurement for these instruments ultimately comes down to.
Two of the three inputs in that loop are market data rather than judgement calls, and that is where a documented source beats a defensible-sounding estimate. The smartZebra interest rate engine supplies the credit-spread leg - spreads calculated for maturities from 1 to 30 years across 15 sectors, plus a synthetic rating tool for borrowers with no observable debt - and the beta and volatility data supplies the equity leg, each figure carrying the source timestamp and calculation path an auditor asks for when the question turns to where the assumption came from.
Related pages
- Credit Spreads & Interest Rates - the spread and market-yield leg of the build-up, by tenor, sector and rating, plus synthetic ratings for unrated borrowers
- Beta Factors & Volatility - the equity-volatility input the warrant is priced from
- Introduction to Loan Valuation and the Private Debt Market - how credit risk gets priced where there is no market quote to check against
- Valuation Approaches for Loans and Borrowings - the loan leg on its own, before a warrant is attached to it
- Calibration in Cost of Capital - Friday Digest - the same anchoring discipline applied to discount rates
- Incremental Borrowing Rate (IFRS 16) - Friday Digest - base rate plus credit spread, fixed at a date and revised only on triggers
