Auction Radar
CBK primary issuance calendar with countdowns.
September 2026: what the government asked for, and what turned up
The Treasury offered Ksh 70.0B and took Ksh 58.8B — 84% of the amount advertised. Bids came to a median of 1.14x the amount offered, against 1.92x over the preceding 6 months.
Less money turned up this month than has been usual lately. Auctions that fall short tend to clear at higher rates, because the Treasury has to reach further to fill the book. That is good for a buyer, and it is also the market telling you why.
| Auction | Type | Bid |
|---|---|---|
| 2026-09-02FXD3/2019/015, SDB1/2011/030 | primary | 1.14x |
| 2026-09-07FXD4/2019/010 | switch | — |
2 auctions in September 2026, 1 with demand figures complete enough to measure.
What government paper actually paid, after tax and after prices
Across 312 auctions between 2010 and 2026, the rate the market accepted beat inflation in 301 of them. The typical auction left 5.78% once withholding tax and the inflation of the month it was sold in are both taken off.
Best on record
13.68%
IFB1/2023/017 on 2024-09-02 — cleared at 17.73% tax-free against 3.56% inflation.
Worst on record
-4.48%
FXD1/2011/020 on 2011-06-27 — cleared at 10.39% against 14.48% inflation.
The losses are not spread evenly. All 11 of them fall in 2011, 2012 and 2017 — years when the headline rate looked generous and inflation ran ahead of it. That is the thing to watch for rather than a reason to dismiss the record: a high nominal rate during a spike in prices is not the same offer as the same rate when prices are calm.
How contested the recent auctions were
Across the last 12 auctions, bids came to a median of 1.89x the amount offered, against 1.30x over all 184 auctions on record. 2 of 12 did not attract the money on offer.
| Auction | Offered | Bid | Taken |
|---|---|---|---|
| 2026-08-17IFB1/2019/016, IFB1/2021/018, IFB1/2021/021 | 150bn | 3.07x | 2.08x |
| 2026-07-27FXD1/2019/020, FXD1/2022/025 | 40bn | 2.15x | 1.58x |
| 2026-07-13FXD1/2021/020, FXD1/2022/010, FXD1/2026/030 | 70bn | 2.06x | 1.01x |
| 2026-06-08FXD1/2018/025, FXD1/2020/015 | 40bn | 0.86x | 0.86x |
| 2026-05-25FXD1/2021/020, FXD3/2019/015 | 50bn | 0.94x | 0.73x |
| 2026-05-11FXD1/2012/020, FXD1/2019/020, FXD1/2021/025 | 80bn | 1.33x | 1.18x |
| 2026-04-20FXD1/2026/030, SDB1/2011/030 | 20bn | 1.92x | 1.50x |
| 2026-04-06FXD1/2018/025, FXD1/2020/015 | 40bn | 1.87x | 1.25x |
| 2026-03-16FXD1/2019/020, FXD1/2021/025 | 60bn | 1.96x | 1.02x |
| 2026-02-16FXD1/2018/025, FXD3/2019/015 | 50bn | 4.27x | 2.01x |
| 2026-01-12FXD1/2019/020, FXD1/2022/025 | 60bn | 1.19x | 1.01x |
| 2025-12-08FXD1/2021/025, SDB1/2011/030 | 40bn | 1.33x | 1.18x |
When bids run well above the amount offered, aggressive bids get rejected and the clearing rate settles lower. A non-competitive bid takes whatever that turns out to be.
Bidders also disagree about price, and the auction says by how much. Across the last 24 sales the average rate asked across all bids ran 11 basis points above the average the Treasury actually accepted, against 10 over all 237 on record. That gap is the tail of bidders who asked for more than the government would pay and got nothing. A non-competitive bid sidesteps the question entirely — it takes the accepted average, whatever it turns out to be.
Follow CBK on WhatsApp
Auction announcements and results straight from the Central Bank, the moment they publish. This is the source — if we ever disagree with it, believe them.
Follow Pesa Smart KE on WhatsApp
Ours, not the Central Bank's: the same auctions worked through in plain language, with what the rate means after tax. Run with JiPange.