Food pulled headline inflation down; reserves hold the key to rates

The headline fell; core did not

Annual CPI fell from 31.51% to 29.73% in September, according to TurkStat data released on 5 October. It is the first reading below 30% since November 2021. The market's first take was simple: disinflation is on track and a cut is coming at the Monetary Policy Committee (MPC) meeting on 22 October. Tradingpedia reported on 6 October that Commerzbank expects a 100 basis point cut this month. That reading is noise, because it ignores why the headline fell.

Monthly inflation was 1.84% in September, but the core C index, which excludes unprocessed food, energy, alcohol, tobacco and gold, rose 2.1%. Core ran above the headline. TurkStat data cited by Habertürk show the only large item pulling the headline down was food: food and non-alcoholic beverages fell 0.20% on the month. Citrus prices dropped 24.41% and other vegetables 17.62%. That is a seasonal and fragile decline. Strip out food and the price pressure is still there.

The broader core measure points the same way. TurkStat data show the special aggregate B index rose 2.01% on the month and 29.00% on the year in September. In services, the back-to-school effect is clear: higher education prices rose 25.79% on the month and university dormitories 19.99%. These items are seasonal but driven by demand and rents; unlike food, they do not reverse the following month. Twelve-month average inflation still stands at 31.49%.

Energy and housing are the sticky groups

Housing, water, electricity and gas rose 2.71% on the month in September and 39.99% on the year. Transport was the fastest-rising main group at 2.79% on the month, up 35.10% on the year. Both groups are running 5 to 10 points above the headline. What they share is energy prices and rents. Brent futures closed at $100.32 on 5 October; while the Hormuz disruption lasts, there is no concrete reason for that price to ease quickly.

In October, tax was added to this pressure. On 1 October the eşel mobil system, which offset fuel price rises with tax cuts, came to an end. According to Hürriyet, the excise on petrol will be 7.90 lira a litre in October, 11.36 lira in November and 14.8277 lira in December. The first step in October adds about 4.15 lira at the pump. November and December each bring a further tax step of roughly 3.46 lira. Transport's 2.79% rise in September is the picture before this schedule began.

The arithmetic of the year-end target

The official year-end forecast cited by Daily Sabah is 28.4%. TurkStat data put cumulative inflation from January to September at 24.32% already. For the forecast to hold, October, November and December combined must stay at about 3.3%. That means a monthly average of around 1.1%. September's 1.84% reading is about 0.7 points above that. With excise steps added in October and November, hitting this arithmetic depends on food prices falling every month.

The Commerzbank note on Tradingpedia spells out the same risk: inflation could re-accelerate towards year-end and keep the lira under pressure. The policy rate is 37% and the annual headline 29.73%, a gap of 7.27 points. The headline's decline widens that gap and opens room for a cut on paper. Annualised, however, core C's monthly pace of 2.1% points to a path well above 28%. The real rate buffer is not as wide as the headline suggests.

That distinction is decisive for the MPC. If the committee looks at the headline, it sees a 7.27-point buffer. If it looks at core and the tax schedule, it sees that buffer narrowing in October and November. Our reports of 2 and 4 October read the constraint through reserves and the external deficit; September CPI confirms the same constraint from the price side. When a food-driven decline meets a currency held steady by spending reserves, the safety margin for a cut gets thin.

The real constraint is reserves and the energy bill

The variable that will decide the cut is not inflation but the price at which the lira is being held steady. According to Dünya on 5 October, based on Matriks calculations, gross reserves fell by about $4 billion to $167.2 billion in the week to 2 October. The decline has run for six weeks; the loss since $188.4 billion on 21 August is $21.2 billion. That is an average of $3.5 billion a week. Official data are due on 8 October at 14:30.

Over the same period the currency barely moved. USD/TRY stood at 48.99 on 29 September and 49.18 on 6 October. Commerzbank says the lira is supported by central bank and state bank intervention. The price of that calm shows up in the risk premium. Türkiye's 5-year CDS was 220.50 basis points on 21 August and rose to 255.69 on 2 October. The day reserves peaked and the day CDS began to climb are almost the same.

The yield curve does not fully endorse the rate-cut story either. Bloomberg HT data put the 2-year benchmark at 39.78% on 6 October, 2.78 points above the 37% policy rate. At the two-year tenor, the market either does not believe in an easing path or is demanding an extra risk premium. In the Investing.com series, the 10-year yield jumped from 32.72% to 35.46% on 5 October. Whether the jump reflects selling pressure or a change of benchmark bond could not be verified.

The calendar to 22 October

Three data points stand before the MPC. The first is official reserve data on 8 October; if the $167.2 billion in the Matriks calculation is confirmed, the drain will have stretched into a seventh week. The second is reserves and FX deposits for the week of 15 October. The third is Brent: a price around $100 is the source of the pressure that kept transport and housing above the headline in September.

Our reading: the September data produced a justification for a cut but did not widen the room for one. A 100 basis point cut would bring the lira yield down to 36%. If weekly reserve losses continue at a pace of $3.5 billion, it would raise the cost of currency stability. If the CBRT does cut, it is more likely to pair the move with tight liquidity and state bank support. A large step before the reserve drain stops would put too much weight on a single month's fall in food prices.