PF - Debt Sizing Query

Hi all,

I've been experimenting with debt sculpting methodologies and wanted to sanity-check my approach.

Historically, my model solved directly to a target maturity input (e.g. 25-years). Maximising debt sizing while forcing the debt balance to amortise exactly by the target tenor.

However, I noticed an issue in certain scenarios, particularly when:

• early-period CFADS is relatively weak,
• debt size is high,
• thus interest consumes most of the allowable debt service in those early periods.

In those cases, the model could still technically solve to the target tenor, but the principal profile became very small or nil in some early periods, as almost all the debt service was being absorbed by interest. 

The tenor solve was effectively forcing a solution that looked mathematically valid, because it only looked at closing balance at 0 at maturity, but wasn't particularly robust from a debt-service perspective in particular runs (depending on CFADs strength etc).

To address this, I added an additional constraint, in a usual debt-sheet, for a minimum-debt service buffer: = Permissible DS - Interest - Principal.

• If this is negative, then that goes into my master-delta section, as part of overall convergence criteria (with maturity delta, funding delta ... etc).

The resulting macro-methodology (call it mode #1) is now:

1. Solve natural debt size / gearing from DSCR sculpting.
2. Determine natural maturity.
3. If natural maturity is shorter than target
   1. elongate repayment profile
   2. adjusting DSCR (this can be iterated, based on timing of when this flag occurs)
   3. continue solving subject to a minimum debt service buffer test.

Overall, this is goalseeking based on DSCRs, and then it saves the gearing that is permissible, and then goalseeks the DSCR adjustment factor to elongate principal, but the gearing is the same.

This avoids situations where the model maximises debt at the expense of creating very weak principal repayment periods.

------------------------------------------

Now, I am trialing mode 2: maximises gearing directly subject to:

• target maturity,
• minimum DSCR,
• minimum debt service buffer.

This follows on from mode 1 above, but pretty much applies a bi-section search (high and low gearing /2 and converges iteratively) from the above. e.g: so gearing from mode 1 solved, then binary search from there.

My expectation was that Mode 2 might produce higher gearing than Mode 1, but in many cases the result is surprisingly similar.

My theory is that the same early-period interest constraint becomes binding, meaning the model reaches the same debt capacity regardless of whether I approach it through:

• natural tenor → elongation (Mode 1), or
• direct target-tenor optimisation (Mode 2).

Has anyone seen this behaviour before?

More specifically:

1. In lender-grade PF models, would you generally constrain debt sizing using a minimum debt-service buffer test if early-period interest is absorbing most debt service? Some advisors say some missing principal periods can be OK.
2. Would you treat this as a DSCR issue, a tenor issue, or simply evidence that debt capacity has already been reached?
3. Have people seen a meaningful increase in debt capacity from a direct target-tenor optimisation versus a natural-sculpting approach once interest-heavy early periods are present?
4. The macro can look convoluted, so from a lender-audit pov, would they prefer bi-section syntax throughout?