Funding a USD Loan Book with Swapped HKD

Option A: swapped HKD funding

• HKD funding rate: 4.10%.
• Cross-currency basis cost: 0.55%.
• All-in USD funding cost: 4.10% + 0.55% = 4.65%.
• NIM: 6.20% - 4.65% = 1.55%.
• Annual NIM on USD 200m: USD 200m x 1.55% = USD 3.10m.

Option B: USD deposits

• USD deposit rate: 5.05%.
• NIM: 6.20% - 5.05% = 1.15%.
• Annual NIM on USD 200m: USD 200m x 1.15% = USD 2.30m.

Difference: swapped HKD is 40bp cheaper, worth USD 800,000 per annum on the full USD 200m. On the headline cost, the relationship manager is wrong: swapped HKD is cheaper, not more expensive.

Break-even basis: the basis at which swapped HKD equals USD deposits is 5.05% - 4.10% = 95bp. The current basis is -55bp, so the basis would have to widen by 40bp before USD deposits became cheaper. Given the stated volatility of plus or minus 30bp, a 40bp move is plausible but not the base case.

The headline comparison is incomplete, and several constraints point against a full switch.

1. Capacity: the USD deposit book is only USD 80m, so the bank cannot fund USD 200m from USD deposits without buying more USD wholesale funding, which would itself be priced off the swap market or SOFR plus a spread.
2. Basis risk: the FX swap rolls every three months. If the basis widens from -55bp to -85bp, the swapped HKD cost becomes 4.95%, still slightly cheaper than 5.05%, but at -95bp it is equal and beyond that it is worse. The bank would be replacing a known deposit cost with a volatile basis exposure.
3. Liquidity and funding profile: the internal buffer requires 30% of wholesale funding to be longer than three months. A three-month FX swap does not satisfy that; the HKD deposit book is stable and large, which supports the swapped HKD route. Switching to USD deposits that are only 80m and short-dated would weaken the funding profile.
4. Rollover risk: a three-month swap requires refinancing four times a year. A basis spike or a USD liquidity squeeze at a roll date could force expensive funding or an asset sale.
5. Diversification: concentrating the entire book in one funding source, whether HKD or USD, is undesirable. The right answer is usually a mix, not a wholesale switch.

Conclusion: do not switch the full USD 200m. The 40bp saving is real, but it is not large enough to justify abandoning the deeper HKD deposit base or concentrating the funding in an 80m USD book.

I would recommend keeping the majority of the USD 200m funded by swapped HKD, using the USD deposit book selectively, and putting a formal basis trigger and tenor extension in place.

Recommendation:

1. Keep at least USD 120m in swapped HKD, preserving the 40bp saving (about USD 480,000 per annum on that tranche) and using the stable HKD deposit base.
2. Use up to USD 80m of USD deposits for diversification and to reduce reliance on the three-month swap roll. This gives a blended funding cost of about (120 x 4.65% + 80 x 5.05%) / 200 = 4.81%, and a blended NIM of about 1.39%, or USD 2.78m per annum.
3. Extend part of the swap tenor to 12 months to lock in the basis and to help meet the 30% longer-than-three-month funding requirement. A one-year cross-currency swap would fix the basis cost and reduce rollover risk.
4. Set a basis trigger at -90bp. If the three-month basis widens beyond that, the all-in swapped cost approaches or exceeds 5.05%, and the treasury should shift new funding to USD deposits or buy a basis hedge. The trigger should be reviewed monthly, not traded discretionarily.
5. Report both the headline cost and the basis sensitivity to the CFO, so the 40bp saving is understood as compensation for basis and rollover risk, not a free lunch.

Why: the decision is not just a cost comparison. It is a funding-mix decision that must respect capacity, liquidity-buffer rules and basis volatility. A blended approach captures most of the saving while keeping the bank within its liquidity and concentration limits.