Every Signal grades the previous week's call before making a new one, including the weeks it missed.
-
A Cheaper Dollar Is Not Relief, 7 Nov
That a weaker dollar only reaches you if the channel between it and your currency is open, and that Nigeria's was blocked by domestic liquidity rather than anything happening in Washington. The named test was specific: if the DMO's next auction cleared above 16 percent on the 364-day, the debt office was not declining to pay up so much as briefly able to avoid it.
Right. The 364-day cleared at 15.89 percent on 11 November, up nine basis points from 15.80 but still below the line we drew. The debt office allotted ₦713.30 billion of an ₦800 billion offer against ₦1.12 trillion of bids, so it is still taking less than it is offered. The dollar-index leg of that edition's test we cannot settle here and are not claiming.
-
Nobody Went First, 1 Nov
That lenders, central banks and one very large company had all been asked the same question and almost all of them answered it the same way: not yet. We named two things that would weaken it first. A Bank of England hike on 5 November. And a debt office forced to take materially more than it wanted at rates back above 16 percent.
Half right. The Bank of England hiked four days later, 8-1, so the first falsifier triggered on schedule. The second did not: the DMO took ₦509.41 billion of an ₦800 billion offer against ₦2.045 trillion of bids, holding stop rates between 15.40 and 15.80 percent. The deferral broke where we said it would hold and held where we said it would break.
-
Lenders Started Choosing, 24 Oct
That lenders had begun pricing governments individually rather than as a class, and that the gap between what lenders want and what borrowers will pay was widening.
Half right. In Europe it held. France still pays about 153 basis points over Germany, down from the week before but far above September's 117, and Moody's declining to downgrade is what moved the price rather than anything Paris did. In Nigeria it inverted. The debt office was offered ₦1.91 trillion, took ₦71.16 billion and pushed its own one-year rate down to 15.80 percent. The lender did not set that price; the borrower did, because it did not need the money. We called the direction and missed that a borrower sitting on enough liquidity can refuse.
-
One Shock, Four Answers, 18 Oct
That one energy shock was being read four different ways, and that Britain's hike consensus would hold. We named the falsifier: if UK September CPI came in at or below 3.1 percent on 21 October, the November consensus would unwind as fast as it formed.
Right. UK CPI came in at 3.5 percent, not 3.1, the highest since October 2025, with core at 2.8 and services at 3.7, all above forecast. The falsifier did not trigger and the consensus hardened rather than unwinding: markets now price two rises before year end. The energy read held too, with fuel prices up 29.8 percent year on year against 23.0 in August.
-
Nobody Is Being Paid to Wait, 11 Oct
That nobody is being paid to wait, because the price of committing money over time had broken in four markets at once, and that in Nairobi the money would keep refusing duration.
Right. Kenya's 15 October auction put the argument beyond doubt. The 91-day drew 622.46 percent of what was offered while the 364-day drew 66.43 percent, the only tenor to come in under its target. The yields barely moved: 8.75 percent for three months against 9.0383 for a year. The falsifier we named also held. Airtel Money's unconditional open on 14 October did not take it meaningfully above the £1.96 offer.
-
The Exit Was Already Written, 4 Oct
That Airtel Money would have a soft debut, because it priced below its own indicated range and the sellers were partly discharging an obligation rather than meeting demand.
Right so far. Conditional dealings opened on 9 October and the shares traded below the £1.96 offer. Reuters called the debut muted, and the sale raised around £582 million rather than the $800 million first targeted. Unconditional dealings on 14 October are the full test, so this is not settled.
-
Read the Corridor, Not the Cut, 27 Sept
That the 350 basis point cut repaired the price of money without expanding the supply of it, so naira fixed income and the government's refinancing cost would reprice first while bank borrowers waited.
Too early. The cash reserve requirements that cap how much of each naira a bank can lend have not moved, so nothing yet contradicts it. That edition named December credit data as the test. A week of silence is not a result.
-
One Pipeline. Four Currencies, 20 Sept
That one pipeline attack had already repriced four currencies in nine days, and that Nigeria’s bond market had priced an easing the central bank had not agreed to, after the DMO cleared a reopened June 2038 bond at 16.85 percent, 94 basis points below August, on thinner bids.
Half right. The transmission read held. The Nigerian call did not. The monetary policy committee cut 350 basis points two days after we published, so the bond market had been reading the central bank correctly rather than running ahead of it. That was the edition’s own stated falsifier, and it triggered.
-
Money Got More Expensive Everywhere. Except in Nigeria, 13 Sept
That a Fed hike on 16 September would narrow Nigeria's yield advantage without closing it, moving the marginal dollar rather than all of them. We put the odds at anywhere from 50 to 85 percent depending on where you looked.
Half right. The Fed hiked 25 basis points to 3.75 to 4 percent in a unanimous vote. The spread behaved exactly as described, with Nigerian ten year paper clearing 16.79 percent two days earlier. But we quoted a range wide enough to be right whichever way it went. The mechanism was the call. The probability was not.