 Hi rimau1
Yes the IRR is an annual rate of return.
I would say the likely range is 16-18% p.a. shareholder IRR through to YE 2028 IF Gravis earn any Performance fee, which I think is probable. With an outside chance of up to 21% p.a.
My starting point is that Gravis know the portfolio better than I ever will, so it is of interest to see what our shareholder returns would be if they achieve the Performance Fee they signed up to.
You could argue Gravis are not that competent to have ended up in a wind-down scenario, and they have lost key staff, so may not achieve any Performance Fees. I recognise that’s possible. My money is on them earning Performance Fees, but nothing is certain, and others may take a different view.
Obv. if Gravis don’t earn any Performance Fee, then shareholder IRR would be lower than 16%, but I don’t have any special insight on a probable worst case – it would need e.g. modelling of worst case losses on Problem and Watch-list loans, or assessing the impact if there is a high interest rate environment in 2028 to try and sell the large (51% of Face Value) rump into, none of which I have looked at.
Equally if Gravis can pull loans forward materially, or an early sale of the complete remaining portfolio can be achieved, we might get an IRR higher than my range above, over a shorter period. I’d be happy with that, but again my analysis based on Gravis Performance Fees probably doesn’t add anything to that discussion. |
Some smart investors here have exited GABI higher up. I didn’t follow suit as I believe Gravis knows the portfolio best and have signed up to a Performance Fee that, if achieved, gives respectable shareholder returns from here.
I have finally done a “proper” spreadsheet to test this. My guess (central estimate) from the current share (bid) price is approx 17% shareholder IRR through end 2028, assuming Gravis achieve mid-range of their Performance Fee. Or approx 16% if they just reach the minimum threshold for Performance Fees.
The key to over-achieving on that, is if Gravis can materially pull forward the realisation of loans, vs what I assumed. The major risk, obviously, is if one or more large loans go bad, in excess of the haircuts I've assumed.
Gravis also have an 8.8% shareholding, which should better align them with shareholders overall.
So for me, any drift in GABI shareprice, during the potentially long fallow periods for redemptions/distributions, would be an opportunity to top up, rather than a reason to exit now.
But E&OE, DYOR, NAI, YMMV, etc.
Perspective
The threshold for Gravis to earn Performance Fees on the wind-down is 12% IRR on an “adjusted portfolio value” of £321m as at 31 Dec 2023. The 12% is measured on cash receipts after transaction costs so OPEX etc not deducted.
Tailwinds for them, in achieving this 12% IRR, include:
- Portfolio weighted average yield of 8.7% p.a. (per 2023 AR)
- The £321m starting-value is an 11% discount off the portfolio NAV at YE 2023, and much more off portfolio face value
- That portfolio NAV has been calculated using a weighted average discount rate of 10.5% p.a. (per 2023 AR, though I do find this hard to credit)
- Lower market interest rates from here may help (e.g. in getting loans re-financed elsewhere, or getting an acceptable price for a portfolio sale of, say, 2028+ loans)
Workings
I profiled cashflows (loan redemptions + interest) that result in Gravis achieving 12% IRR on the “adjusted portfolio value” of £321m. I started from the loan-redemptions reported so far in 2024, then used the Gravis loans spreadsheet to model a reasonable time-profile for future redemptions of the remaining loans. I assumed all 2028+ loans are wrapped up in 2028, (pulling realisation of all 2028+ loans into 2028). I massaged the resulting cashflow-profile (via haircuts off face value) to arrive at a Gravis 12% IRR from the notional investment of £321m at Y/E 2023. It needed 5% discount in 2025, 10% in 2026, 15% in 2027, and 26% in 2028 (with 2028+ all pulled into 2028) to massage the Gravis IRR down to 12%.
Lots of assumptions and calculation short cuts in there, but all to try and get a reasonable time-profile for the redemptions+yield cashflow that Gravis is being measured on. The resulting cash profile is constrained by the need to model a 12% IRR. So any changes in the time-profile or estimated loan face value, rental yield etc, will principally affect the levels of discount off face that I had use to get to the 12%, more than changing the resulting cashflows that give a 12% IRR.
The purpose of all the above was to see what shareholder IRR (from here) resulted from those gross cashflows, after deducting estimated operating costs including Gravis fees.
When I plug in existing-shareholders’ starting investment (266m shares x 68.4p Bid = £182m market cap at the Bid Price as of June 28th) and use the same estimated future cashflows from here, as above (minus estimated costs/expenses), I get a shareholder IRR of 15.9% from here. This relationship (shareholder IRR vs Gravis 12% IRR) should be somewhat robust to the exact cashflow-timing and haircuts assumed (given these “have to fit into” the Gravis 12% IRR target, for my purposes).
Gravis need to reach the 12% IRR at Portfolio level to earn any Performance fees. Their Performance fee is capped at £14.7m, which would require £73.7m “excess return over the 12% IRR” (before deducting the Performance Fee). But this is slippery to model. It may be achieved more by pulling redemptions forward, than by collecting larger amounts (i.e. reducing discounts/haircuts). My understanding is that it is calculated on a loan by loan basis, once 12% IRR is achieved on the overall portfolio. So even with just 12% IRR at portfolio level, significant Performance Fees could be earned, with outperforming loans paying out for Gravis, but balanced for shareholders by others that underperform.
My guess is that if Gravis earn half the max Performance fee, shareholder IRR increases by 1 to 2% over the 12% Gravis IRR scenario, so approx 17% shareholder IRR is my central case.
It’s a very rough calculation, but encouraging fwiw.
Would be interested if anyone else has done a similar calculation, or has comments on methodology.
Assumptions:
For redemptions I used the 2024 announced redemptions to date, then timings based on the Gravis loans spreadsheet going forward, applied to my estimate of current face value outstanding.
I assumed increasing haircuts off face value from 2025 on, and that all 2028+ loans are wrapped up in 2028 for a 26% haircut in 2028 overall.
I grouped redemptions into quarterly buckets, except 2028+ all in a single (very large: 51% of total) bucket.
I didn't explicitly model bad-loan (and interest) write-offs, so those were only covered by the overall haircuts off face-value I assumed.
For Expenses/costs: I used £1.5m pa fixed, + Gravis management fee of 0.7% of portfolio face value (not 0.75% of NAV, as per Circular, as I didn’t have qtr by qtr NAV numbers). I didn’t add liquidation costs, hoping this is offset by using face value not NAV for Gravis periodic fees. Could be improved! I deducted £4.4m of Performance fees (30% of the max, all in 2028) even with Gravis IRR of only the 12% portfolio-level threshold, for the reasons given above, and deducted £7.4m Performance Fee for my mid-range “central estimate” scenario.
Please treat this post as if written by an anonymous BB poster you’ve never met and whose judgement, and capacity for spreadsheet errors, is unknown ! :-) |
Register for real-time alerts, custom portfolio, and market movers
Hot Features
Premium
