Data shines light on Tibor fragility

In Japan, yen Libor was killed off partly because it was too reliant on expert judgement, and the rates market moved largely onto transaction-based overnight rates. But two other interbank offered rates used domestically survive, despite having no transaction volume at all.

While that might look like a recipe for disaster in the post-Libor world, it’s not quite that straightforward.

The two benchmarks in question are the Japanese yen Tokyo Interbank Offered Rate, or D-Tibor, which reflects prevailing rates in the unsecured call market. According to a 2022 Japanese Bankers Association Tibor Administration (JBATA) survey, Japanese banks, securities companies and insurance companies have notional exposures of roughly $912 billion in loan assets tied to D-Tibor, and a further $1.3 trillion notional in derivatives linked to the benchmark.

The other benchmark is Euroyen Tibor (Z-Tibor), which is derived from prime bank transactions and quotes in Japan’s offshore yen rates market.

In August last year, the JBATA released a consultation paper on fallback issues for D-Tibor and Z-Tibor, the results of which are due to be published before March 31.

It has been known for some time that Z-Tibor’s days are numbered. Back in 2019, the JBATA consulted with the industry on a potential integration of the two benchmarks and indicated that the most likely scenario would be that of “retaining Japanese yen Tibor and discontinuing Euroyen Tibor”.

Two years later, the JBATA announced that the timing of Z-Tibor’s discontinuation – if adopted – would be the end of 2024. The administrator plans to launch a further consultation on whether to officially discontinue Z-Tibor. That seems all but inevitable, given that the offshore market on which it is based has been shrinking for some time.

While no such death sentence has been handed down for D-Tibor yet, the robustness of the benchmark nevertheless deserves some scrutiny.

The publication of a periodic review by JBATA in March 2022 revealed that between January 4, 2021 and December 30, 2021, 100% of six-month D-Tibor – the most used tenor – was based on indicative price quotes. The quotes are automatically included in the first level of the waterfall structure, but after actual transactions and committed quotes. The waterfall structure sets out different layers for banks’ submissions to the benchmark in order of priority to enable the rate to be published in all market circumstances.

In the methodology of Libor and the hybrid methodology of Euribor, indicative quotes are treated as expert judgement. So based on that classification, six-month D-Tibor would appear to be fully derived from precisely the thing regulators have been trying to get away from.

However, the Tibor administrator treats expert judgement separately from indicative quotes, placing the latter in level four of the waterfall. The report says the administrator never had to reach that level during the observation period.

It treats indicative quotes as better than expert judgement because of the methodology it uses. If a reference bank calculates submission rates using indicative quotes, then a change from the previous day in the mean rate of quotes published by brokers must be mechanically added to or deducted from the submission rate on the previous day. So, even though indicative quotes are not executed as actual transactions, the submission rates are nevertheless based on observable and objective data and a mechanical calculation process that should reflect actual market conditions.

Regardless of the definitional difference, it may seem worrying to some that a key benchmark is not connected to any actual transactional data, and underscored the importance of determining the right fallbacks for the two Tibor rates.

If D-Tibor is used as a fallback to Z-Tibor, users might be advised to begin thinking about a further fallback to TONA or TORF should the benchmark one day run into the same problems that precipitated Libor’s recent demise.