IMAGE VIA GOOGLE
ON FEBRUARY 13th 2024 the Nepal Rastra Bank introduced a standing deposit facility, or SDF. The aim was to keep the weighted average interbank rate from falling below the lower bound of the central bank’s interest-rate corridor. For the first year, not much changed. Then, between February 13th 2025 and May 26th 2026, something curious happened. Over 314 trading days, the interbank rate dropped below the floor on only 24 occasions. On 22 of those days, the facility was not available.
The SDF is not a standing facility in the usual sense. The Nepal Rastra Bank offers it only three days a week: Sunday, Tuesday and Thursday, excluding holidays. Banks must apply in person at the SDF desk during fixed hours. The deposits mature on the next SDF day. This is a far cry from the automated, overnight facilities run by the European Central Bank or the Reserve Bank of India. The Federal Reserve takes a different route, paying interest on all reserve balances held at its regional banks. In Nepal the manual and intermittent design creates natural gaps.
The data, compiled by Victor Kumar Sapkota, an assistant director at the central bank, show what happens in those gaps. On days when the SDF was available, the interbank rate fell below the floor just twice in 168 observations, or 1.2% of the time. On days without the facility, the rate dropped below the floor 22 times in 146 observations, or 15.1% of the time. The probability of a breach was 12.7 times higher on non-SDF days. A Fisher’s exact test, suitable for small expected frequencies, returned a one-sided p-value of 0.00000101. The association is not subtle.
Both failures on SDF days merit a closer look. One occurred on a Friday that was declared a working day after a Thursday holiday. The day was a half-day for banks. The reduced operating window may have prevented some institutions from placing their surplus liquidity with the central bank. In other words, even when the facility is nominally available, narrow hours can defeat its purpose. The other failure is not explained in the paper, but the asymmetry remains overwhelming.
The paper excludes the first year of operation to allow for institutional adjustment. Mishkin’s rule of thumb–that policy effectiveness tends to appear only after a year or longer–is cited as justification. Over the subsequent 15 months, SDF days reduced the probability of a floor breach by 92.1%. The relative risk reduction is sharp enough that the author recommends a simple fix: make the facility available on all days with interbank transactions.
That recommendation, written into a working paper published this month, had already been acted upon. On May 27th 2026, one day after the data series ended, the Nepal Rastra Bank revised its open-market operations procedure for the eighth time. The SDF is now offered daily, excluding public holidays. The same revision notes that an online application system is under development. Automation would remove the other friction–the in-person visit to the desk.
The paper’s author acknowledges that the analysis assumes independence across trading days. Liquidity conditions may well persist from one day to the next, and a more sophisticated time-series treatment could refine the estimates. But the core finding is robust enough to survive that caveat. The floor held when the door was open. It was breached when the door was shut. Central banks designing corridor systems might note that a facility’s availability matters as much as its rate. ■
