Editor’s note: This is general educational material on how Kenyan government debt operations work. It is not investment advice and it is not a recommendation about any security. It is based on the Central Bank of Kenya auction documents and National Treasury strategy papers listed at the end.
Kenya’s debt office periodically asks investors to give up a bond before it matures and take a longer one instead. No cash is repaid on the principal. The holder surrenders one Treasury bond, receives a different Treasury bond with a later redemption date, and the government’s repayment calendar changes shape. The operation is called a switch auction, and the Central Bank of Kenya runs it as fiscal agent for the Republic of Kenya. The July 2026 exercise, which moved holders out of FXD1/2021/005 and into FXD1/2012/020, shows exactly how the mechanism is built and what it does for the issuer.
What the July 2026 switch put on the table
The prospectus named two securities. The source bond was FXD1/2021/005, ISIN KE7000006655, a ten-year issue with 0.3 years left to run and a maturity date of November 9, 2026. The destination bond was FXD1/2012/020, ISIN KE4000003949, a twenty-year issue with 6.3 years to maturity, redeeming on November 1, 2032. The source bond carried a coupon of 11.2770 and attracted withholding tax at 15%. The destination bond carried a coupon of 12.0000 and withholding tax of 10%.
The offer was for KES 10 billion. The sale period ran from June 26, 2026, to July 13, 2026, with bids closing at 10.00 am on the auction date and settlement two days later on July 15, 2026. Non-competitive bids were accepted between KES 50,000.00 and KES 50,000,000.00; competitive bids started at KES 2 million per CSD account per tenor. Only investors holding the source bond free of encumbrance on the auction date were eligible, and anyone with an outstanding pledge had to cancel it five days before settlement to take part. Participation was voluntary, in part or in whole, and holders who did nothing kept the original bond on its original terms.
How the two bonds are converted into each other
The exchange works because both securities are turned into money before they are compared. The Central Bank fixed the source bond at a yield of 8.8322% and a dirty price of 102.7442 per hundred of face value. The destination bond was sold multi-priced on bids quoted in yield, exactly as in an ordinary bond auction, and the prospectus published a pricing table converting each quoted yield into a clean price. At a quoted yield of 12.0000% the clean price was 99.9599. Accrued interest on the destination bond was 2.1429 per KES 100, so the dirty price at that yield was 102.1028. Withholding tax is computed on clean prices.
From there the arithmetic is mechanical. The nominal amount of the source bond tendered, multiplied by its dirty price, gives the consideration, which is the value the holder is handing over. That consideration divided by the destination bond’s dirty price gives the face value of the new bond received. The two prices almost never divide into whole units, so a residue is left over; the Central Bank said any remaining cash below the minimum investment amount of KES 50,000.00 would be refunded on the settlement date. Everything else is a book entry. Allocations appeared under the Bids tab of the DhowCSD investor portal, and successful investors saw their portfolios updated with the allocated amounts.
Why the Treasury runs them at all
The 2026 Medium Term Debt Management Strategy sets out the reasoning. Its stated aims include reducing refinancing risk by cutting short-term domestic debt and lengthening the average time to maturity of the total portfolio, and it lists liability management operations among the tools for doing so. The numbers show why. At end-2025 the average time to maturity of the domestic portfolio was 6.4 years against 10.0 years for the external portfolio and 8.3 years for the total. Debt maturing within one year stood at 13.3 percent of the total. The strategy sources 84 percent of gross borrowing domestically over the medium term and 16 percent externally, so the domestic maturity profile is where most of the rollover pressure sits.
Kenya has used the same logic on its foreign currency debt. In February 2025 the National Treasury issued a USD 1.5 billion Eurobond maturing in 2036 and used part of the proceeds to repurchase USD 579 million of the USD 900 million Eurobond due in 2027, an operation the strategy describes as spreading repayment obligations over a longer horizon. A domestic switch does the same job without any new money, because the consideration never leaves the government’s books.
Analysis: what the July result says about how switches clear
The result notice is the more informative document. Against KES 10,000.00 million offered, bids worth KES 8,159.65 million at cost came in, a performance rate of 81.60 percent and a bid-to-cover ratio of 1.03. The Central Bank accepted KES 7,954.71 million, of which KES 7,944.29 million was competitive and KES 10.43 million non-competitive. The amount switched out of the source bond was KES 7,914.70 million. The market weighted average rate was 12.8118 percent and the weighted average rate of accepted bids 12.8076 percent, giving a price of 98.6714 per hundred at the average yield.
Three things follow from those figures. First, the issuer took almost everything it was offered. A bid-to-cover of 1.03 and an acceptance almost equal to the bids received means the accepted amount tracked the bids submitted rather than being cut back at the margin; on these figures the binding limit was the volume tendered, not the level of the bids. Second, the clearing yield of 12.8076 percent sat well above the 12.0000 coupon printed in the pricing table, so the destination bond changed hands below par, at 98.6714. A holder switching therefore received more face value than the consideration surrendered, which is the ordinary result when the destination line is priced under par. Third, the switched face value of KES 7,914.70 million is smaller than the accepted amount of KES 7,954.71 million, a reminder that the two legs are equal in value rather than in nominal.
What the documents do not establish is whether the operation shifted the aggregate maturity profile in any meaningful way. A single switch of roughly KES 8 billion out of a line maturing in November 2026 is small next to a domestic portfolio whose average time to maturity is measured in years, and the strategy paper reports the profile annually rather than after each auction. The August 2026 exercise suggests the debt office is widening the technique: it offered holders of Treasury bills 2685/091, 2646/182 and 2574/364, along with bond FXD1/2012/015, a move into FXD4/2019/010. Bringing bills into a switch attacks the short end directly, which is where the strategy says the refinancing problem is concentrated. A reader tracking this would compare the size of successive switches against the redemption due in each window, and watch whether the clearing yield in a switch sits above or below the yield on comparable new issuance in the same month.
