Ethiopia sells just a third of planned T-bills as yields fall and the one-year bill recovers

An analysis of the 1,004th T-bill auction — the government sold barely a third of its plan as yields fell across the curve and the one-year recovered.

Two weeks earlier, the one-year bill had been the worry: on May 13 the 364-day tenor drew fewer bids than the government offered and cleared at a punishing 19.00% cut-off. The 1004th auction settled that question. Bidders returned to the one-year in force, its cut-off eased to 15.65%, and the government did exactly what a reluctant borrower does when long money looks expensive: it trimmed the offer. It trimmed, in fact, far more than the one-year: the government put up just ETB 16.73 billion of paper against a published plan of ETB 44.81 billion, barely more than a third of what the calendar penciled in.¹

Within the framework of consistent T-bill operations, governments use the primary market not only to raise incremental credit but as a mechanism for active management of the existing debt stock. Under those conditions the operative question is rarely "how much new money," but "on what terms is maturing debt being refinanced." Ethiopia's domestic market fits that pattern, carrying a large stock of short-dated obligations that fall due in a near-continuous stream. With roughly ETB 91 billion of T-bills maturing over the coming 30 days, the 1004th auction is best read as a rollover operation, not new borrowing.

Against that maturing wall, the government chose to refinance only a fraction through this auction. It accepted ETB 16.73 billion, the entire amount it offered, equivalent to about 18% of what comes due over the next 30 days. ⓘThe ~ETB 91 billion maturing figure and the share-of-wall refinanced are our own computations from accepted-amount history; the government's actual near-term redemption profile may differ due to settlement timing, supplementary auctions, or funding raised outside the T-bill market. The balance must be met from elsewhere: drawn-down balances, foreign-currency proceeds, or direct central-bank financing. Investors, for their part, asked for far more paper than was on offer: bids reached ETB 58.37 billion, a bid-to-cover ratio of 3.49, the strongest cover of the current easing cycle.

Auction snapshot · May 28, 2026 (1004th auction)

The government accepted the full amount it offered at every tenor, so acceptance was 100% of the actual offer. Measured against the calendar, however, the auction was far smaller than planned: the ETB 16.73 billion offered was ETB 28.08 billion below the ETB 44.81 billion the Ministry of Finance had scheduled for May 28, a change from the auction plan that frames everything else in the result.

Offer composition: a smaller auction, tilted the same way

The May 28 offer kept the now-familiar shape, heaviest in the middle and at the long end, even as its total shrank. The 182-day (six-month) bill made up 35.0% of the amount offered, the 364-day 30.0%, the 91-day 25.0%, and the 28-day 10.0%. Relative to the May 13 auction the mix shifted modestly toward the three-month (91-day) tenor and away from the six-month, but the more consequential change was the level, not the composition: every tenor was cut.

An offer this far below plan usually signals the mirror image of an upsize. Where an above-plan offer points to an additional near-term cash need, an offer well below plan suggests the government met its funding needs from other sources, or judged that it did not need to roll the full maturing amount through the bill market this round. Determining the underlying catalyst requires deductive analysis; the auction itself cannot confirm it. What the auction does show is that the government faced a heavy maturing wall and chose to refinance only a sliver of it here, which is consistent with funding having been arranged elsewhere rather than with funding stress.

Read alongside demand, the small offer takes on a second purpose. With bids running at nearly three-and-a-half times the paper available, restricting supply was also a way to drive the clearing yield down. A borrower that wants cheaper money, and sees a crowded order book, has every reason to sell less and let competition do the work. That interpretation is consistent with the across-the-board fall in yields described below, though the auction alone cannot separate a deliberate cost-management choice from a simple absence of funding need.

Demand by tenor

Demand was firm across the curve, and firmest where it had been weakest. In absolute terms the 182-day tenor drew the most bids, ETB 22.53 billion against ETB 5.85 billion offered, a bid-to-cover ratio of 3.85, the highest cover of the auction. The 91-day followed at 3.71 (ETB 15.50 billion of bids against ETB 4.18 billion offered), and the 28-day at 3.32 (ETB 5.56 billion against ETB 1.67 billion). The 28-day also cleared at exactly the lowest yield bidders submitted, a "perfect-cover" reading in which every accepted bid came in at the floor.

The standout was the 364-day. Two weeks earlier it had been undersubscribed, with bids of ETB 11.99 billion falling short of the ETB 12.56 billion offered. This time it drew ETB 14.78 billion against an offer cut to ETB 5.02 billion, a cover of 2.95. Part of that swing is demand returning and part is supply withdrawn (bids rose by about a quarter while the offer was more than halved), but the direction is unambiguous: the one-year buyers came back, and they came back at a much lower yield.

How yields changed from the previous auction

Compared with the previous auction held on May 13, 2026, cut-off yields fell at every tenor, and the fall was steepest exactly where the May 13 result had been most strained.

The one-year cut-off dropped from 19.00% to 15.65%, a decline of 335 basis points that reversed most of the May 13 spike in a single auction. The shorter tenors eased more gently: the 28-day cut-off fell from 10.25% to 9.50%, the 91-day from 11.85% to 10.81%, and the 182-day from 14.00% to 13.37%. Each of these is the lowest reading for its tenor in over a year: the 28-day since March 2024, the 91-day since August 2024, and the 182-day since November 2024. All sit well above the near-zero yields of the pre-reform era, when rates were administered rather than set at auction and so are not a fair comparison.

Two features of the May 13 result unwound here. The unusual gap between the one-year cut-off and its weighted average (19.00% against 15.43% two weeks earlier) narrowed sharply, to 15.65% against 14.72%, as the most aggressive bids that had pulled the May 13 cut-off up did not reappear. And the curve flattened: the spread between the one-year and the one-month tenor, which had stretched to 8.75 percentage points on May 13, closed to 6.15 points, unwinding the May 13 spike, though still wider than the 4.7-to-5.3-point range seen through April. The blended weighted-average yield across all tenors fell from 13.48% to 12.48%.

Six-auction runway: how the market got here

The 1004th auction's outcomes are the culmination of the previous auctions and their result trends. The six auctions that preceded it, ⓘThe six-auction window is chosen as a pragmatic span that captures near-term market movements; it is not a statistically derived lookback. from March 4 to May 13, 2026, trace a market in which yields fell steadily and demand strengthened, with the one-year tenor the recurring exception.

The six auctions before May 28 — plan, offer, demand, yields

Three threads run through the table. The government's blended cost of new T-bill debt fell from 16.71% on March 4 to 13.03% at this auction, a decline of more than 360 basis points in under three months. Demand, weak enough to leave the early-March auctions undersubscribed, firmed from April 1 onward and has stayed above 1.7× cover at every auction since. And supply did much of the work: where the government floated ETB 60–63 billion in late February and early March and could not fill it, it has since brought offers down toward, and now well below, the calendar plan, letting a crowded order book press yields lower.

The May 28 result is the sharpest expression of that supply discipline so far. By offering just over a third of the planned amount into the strongest demand of the run, the government pushed its cost of borrowing to the lowest of the cycle and pulled the one-year back from the May 13 spike. Every tenor extended a move already in motion; the one-year, May 13's single break, rejoined the trend.

What is driving these trends?

The broad decline in T-bill yields reflects a shifting economic backdrop, even as one indicator has turned.

Disinflation, with a recent setback. Inflation, above 20% only two years ago, has fallen steeply through the reform period. The Ethiopian Statistics Service reported headline inflation easing to a multi-year low of 9.4% in March 2026, sustaining the single-digit reading first reached late in 2025. The print released before this auction, however, showed inflation re-accelerating to 11.7% in April 2026, driven by higher food prices.² That uptick helps explain why the one-year, though much improved, did not fall as far as the shorter tenors: investors still demand extra compensation to lend for a full year while the price outlook is uncertain.

A tight monetary stance. The National Bank of Ethiopia (NBE) has held its benchmark policy rate at 15% since adopting an interest-rate-based monetary framework in August 2024, and the Monetary Policy Committee has signaled it will keep conditions restrictive until inflation is firmly contained. The July 2025 end of mandatory bank purchases of government bonds has freed bank balance sheets to absorb more T-bills,³ which helps explain the depth of demand at the shorter tenors (the 28- and 91-day).

A thin secondary market. Ethiopia's market for trading government securities after issue remains underdeveloped, which limits a holder's ability to sell a one-year bill before maturity. Investors demand extra compensation for that illiquidity, and it keeps the one-year yield elevated relative to the shorter tenors beyond what inflation expectations alone would justify.

What to watch in the next auction

The next auction is scheduled for June 10, 2026 under the Ministry of Finance's Q4 EFY 2018 calendar,⁴ which plans an offer of about ETB 37.5 billion, more than double the ETB 16.73 billion actually sold on May 28. The first thing to watch is whether the government returns to issuing near plan or extends the supply squeeze. If it keeps the offer small against strong demand, yields have room to fall further; if it normalizes supply to refinance the maturing wall, the extra paper could slow or halt the decline.

The one-year tenor remains the swing factor. Its cut-off has retraced from 19.00% to 15.65% in one step, and whether that easing continues depends on the next inflation reading and on how much one-year paper the government chooses to sell. A larger one-year offer met by steady demand would confirm the recovery; a renewed pullback in one-year bids would signal that April's inflation setback still weighs on long-dated appetite. The short end (the 28- and 91-day) is likely to stay comfortably oversubscribed. The NBE's next policy-rate decision and the next inflation movement will frame how the central bank balances inflation control against the government's borrowing needs.

Offer-composition shares, seasonality, the planned-versus-actual gaps, and the ~ETB 91 billion maturity estimate are own computations. This article explains and contextualizes the auction; it is not investment advice.

References