Yes, you have read it right. This article is focused on impact of MLR changes, set by commercial banks, on the Bank of Thailand’s policy rate decisions, not vice versa.
During the first week of April, there was an event in banking sector which was surprising and somewhat confusing, at the same time. The event was the sudden cuts in Minimum Lending Rates (MLRs) by Thai commercial banks. In fact, all major banks had decreased its MLR by 25 basis points.
The surprising part was that the MLRs generally move after there is a change in the policy rate. Say, if policy rate decreases, then MLRs decrease. This time, however, the MLRs changed without any prior movement in the policy rate. As a matter of fact, this is not something unusually rare. To say it was a big surprise may be an overstatement.
Though the mild surprise this may be, more importantly, this does create confusion or doubt on the direction of the policy rate. Given weak development in economic recovery, there are constant speculations on further rate cut. This could be easily observed in capital market. One-year government bond yield, as determined by the market, has been lower than policy rate since early 2016. Currently, the yield is hovering around 1.4%, compared to the policy rate at 1.50% which has been maintained at current level since April 2015. In other words, even though market has been expecting the rate cut for sometimes now, there is no cut yet.
Therefore, the question is, given the new set of information on MLRs, ‘does the decrease in MLRs signal an imminent cut in policy rate?”
What do historical data suggest?
To entertain the question, we investigated historical data whether changes in MLRs could be used as a leading indicator of immediate change in the policy rate. Our data included the daily MLRs of the 5 biggest banks in Thailand and, obviously, the policy rate. The sample started from May 2000 when the Bank of Thailand (BOT) adopted inflation targeting and policy rate has been used as a primary monetary policy tool.
Among the 5 biggest banks, the MLRs have been adjusted, in total, for 202 times during the sample period, excluding the latest moves in April. Unfortunately, not all changes provided information relevant to our interest. In fact, changes in MLRs could be classified into 2 mutually exclusive groups – ‘influenced’ and ‘not influenced’ by prior change in the policy rate. Only the latter group, i.e. when changes in MLRs did not follow the change in policy rate, was relevant to our interest. As a reminder, we would like to answer if changes in MLRs could suggest an imminent change in policy rate, not the other way around.
Admittedly, to determine which changes in MLRs were ‘influenced’ by policy rate or not was somewhat a subjective exercise. Nevertheless, we identified that 123 out of 202 times of total changes in MLRs (or around 60%) were the result of prior changes in the policy rate. In other words, 79 times of changes in MLRs were not affected by prior changes in the policy rate.
Then, we could easily counted ‘how many times among these 79 times did we observe subsequent changes in policy rate?’ And the answer is 37 times – nearly half.
In other words, from historical perspective, there is roughly an equal chance that the policy rate will be adjusted in the next monetary policy meeting in synchronous with prior adjustment in the MLRs. Henceforth, we may conclude that changes in MLRs are not strong signals for immediate change in the policy rate!
How do behaviors change over time?
According to the result above, could we just flip a coin to determine the direction of the policy rate then? Not so fast! Since, our sample period ranged over 16 years, there could be some behavioral changes, such as changes in members of Monetary Policy Committee (MPC) or structure of Thai economy, which may affect the leading capability of the MLRs.
As a result, we looked exactly when changes in MLRs led to changes in the policy rate and when they were not on a timeline. Surprisingly, we observed a clearer pattern.
Before 2008, the results were mixed in a sense that changes in MLRs may or may not lead to imminent change in the policy rate. In fact, there were 37 changes in MLRs which led to changes in the policy rate, while there were 21 changes which did not.
However, the behaviors seems to drastically change since then. After 2008, there were 21 changes in MLRs which were not influenced by prior change in the policy rate, excluding the latest moves in April. Now, do you want to guess how many times were followed by change in the policy rate? The answer is……..zero. Clearly, something has changed.
Unfortunately, we do not have any concrete reason what caused the change in the behaviors. Also our analysis does not really imply causation.
However, our finding confirms that change in MLRs was a poor leading indicator of the policy rate in recent years.
All in all, our results suggested that the latest decrease in MLRs in April, per se, did not significantly amplify the probability of further rate cut. Please note that we didn’t say that it is impossible for the policy rate to change following the MLRs in the future. However, the decision on policy rate is more likely to depend on economic and financial market outlook at a time.
(Published in Bangkok Post on May 6, 2016)
TMB Analytics is the economic analysis unit of TMB Bank. Behind the Numbers is co-authored by Peerawat Samranchit and Naris Sathapholdeja. They can be reached at email@example.com