Analysis: CBN Holdco Draft collides with BOFIA on capital and control

This is very open to discussion and should/can be challenged. The CBN’s draft guidelines for financial holding companies, released on June 10 with comments due July 9, have rightly generated conversation about capital surcharges and ownership floors.

But the more consequential story is not what the guidelines say. It is what happens when they sit beside BOFIA 2020 and the 2018 dividend circular.

Start with the elegant part. Section 19 of BOFIA caps a bank’s holdings in foreign subsidiaries at 10% of shareholders’ funds unimpaired by losses. I have written before about the friction this creates: the banks are directed to reduce stakes in subsidiaries, the same regulator approved transaction by transaction, the devaluation effect that inflates a breach without a single business decision being taken.

The new guidelines, by directing that foreign subsidiaries be held by the holdco rather than the Nigerian bank, dissolve the Section 19 problem in one structural move. Lift the offshore network out of the bank and up to the parent, and the bank’s Section 19 arithmetic resets to zero.

This is a coherent policy and deserves to be acknowledged as such. But notice what it implies. A mono-line Nigerian bank cannot lawfully hold a meaningful African network under Section 19. Only a holdco can.

The guidelines never say this aloud, but the combined effect of statute and draft is unmistakable: pan-African banking from Nigeria is now a holdco-only sport. Any board still romanticising the simplicity of the mono-line structure while nursing continental ambition is holding two ideas that the law no longer permits to coexist.

Now, the overlap doubles the burden. Section 16 of BOFIA prohibits any bank distribution until identifiable and accumulated losses are made good, statutory reserve transfers are completed, and provisions satisfy the CBN. The 2018 circular layers the NPL and composite risk rating gates on top, the framework I described as leaving empty plates at dinner for shareholders of constrained banks.

The draft guidelines now add a third filter at the holdco level. A holdco may not pay dividends until its preliminary and organisational expenses, share selling commissions, and capitalised expenses not represented by tangible assets are fully written off, adequate provisions exist for actual and contingent losses, and capital ratios are met.

A shareholder’s dividend now passes through 2 regulatory checkpoints in series. The bank must clear Section 16 and the 2018 circular to pay the holdco. The holdco must clear its own test to pay the shareholder. Each filter is individually defensible. Together, they create a transmission system in which value must survive two inspections before reaching the people who own it.

And here the guidelines set a trap for themselves. They confine the holdco’s permissible income to dividends from subsidiaries, approved shared services, interest on idle funds, and divestment profits. Dividends are, in substance, the entire diet. Yet that diet sits downstream of a bank dividend regime expressly designed to switch the tap off when a bank’s loan book misbehaves.

Consider the holdco whose principal bank carries NPLs above 10%. The bank pays nothing, lawfully and properly. The holdco above it remains an entity with board costs, compliance obligations, listed company expenses and now a capital requirement 20% above the aggregate of its subsidiaries, but with no lawful income of any substance.

The guidelines assume the dividend pipe flows. BOFIA and the 2018 circular exist precisely because sometimes it must not. I have seen no evidence that anyone modelled the 2 frameworks running simultaneously on a stressed group. The empty plates, it turns out, may soon be at the parent’s table too.

The first genuine clash is definitional, and it is the kind of thing that sounds academic until it costs someone a subsidiary. BOFIA and CAMA anchor the concept of control at more than 50% of issued equity or effective dominance through the familiar three-element power test.

The guidelines’ own definitions section faithfully reproduces this. Yet the operative rule demands that a holdco hold at least 51% of every subsidiary, registered as a person with significant control.

The space between 50% + 1 share and 51% is a sliver of the register, but the law now gives 2 different answers inside it. Hold 50.4% of a pension company, and you control it under the statute while breaching the guideline, with a 6-month divestment clock ticking once control is deemed lost.

A subsidiary instrument has set a stricter standard than the Act that authorises it, and where statute and circular disagree, the statute wins. The 51% figure is legally vulnerable, and any comment letter worth its postage should say so.

The 2nd clash is the one with the largest sums attached. BOFIA’s capital architecture, including Section 16 itself, recognises reserves and retained earnings as loss-absorbing capital. That is the entire logic of requiring statutory reserve transfers before distribution: retained profit is capital, and the Act treats it as such.

The draft guidelines erect a parallel capital regime at the holdco measured on paid-in capital only, paid-up share capital plus share premium, with retained earnings excluded entirely.

The same naira of retained profit, therefore, counts as capital in the bank under the statute and counts as nothing in the parent under the guideline. A decade of disciplined earnings retention, the very behaviour Section 16 was written to enforce, evaporates at the holdco door.

There is no statutory basis in BOFIA for a paid-in-only definition, and a guideline cannot amend an Act. The first holdco that this provision genuinely bites will have an ultra vires argument sitting in its lawyer’s drawer, and the CBN would be wise to fix the definition before someone is provoked into opening that drawer.

The 3rd tension is an old acquaintance in new clothing. When I wrote about Section 19, the uncomfortable question was whether regulatory approval attaches to the office or to the individual occupying it. The acquisitions now being questioned were reviewed, authorised, and stamped by the same institution that later identified the aggregate result as a breach. The regulator approved the journey, transaction by transaction, then objected to the destination.

The draft guidelines rebuild this exact pattern inside the holdco framework on day 1. Every existing sub-51% stake, every capital structure now falling short of the new aggregation test, was approved by the CBN at the time it was created. The guidelines give existing holdcos 6-months to notify their selected structure, but they are silent on grandfathering for positions that were perfectly lawful when blessed.