Stunning policy experts around the world, the Central Bank of Turkey cut its overnight interest rate by 25 basis points, reducing the rate from 10.75% to 10.5%. NASDAQ called this a “surprise move.” Bloomberg said it was an “unexpected cut” with “only 3 of 19 economists in a Bloomberg survey predicting the move.”
In contrast to policy experts, Prattle’s algorithm determined that Turkey had been trending dovish since the end of February–making easing policy no surprise to us. Prattle scored a number of recent, important communications as dovish, including February 23’s meeting minutes and two speeches by the Deputy Governor Turalay Kenc.
In the minutes from their February meeting, the bank explicitly said that rates “would be reduced should… [evidence show] the effective use of the policy instruments laid out in the road map published in August.”
In his speech in Shanghai on February 26, Mr. Kenc cited the August policies as being a success, allowing the Central Bank of Turkey “to fight inflation by taking liquidity stabilizing measures for the foreign currency market and ensuring the stability of the financial system.”
Finally, it was Mr. Kenc’s speech in London on March 3 that most strongly hinted that a rate cut to take place later in the month. In a straightforward concluding statement, Kenc asserted that the “measures undertaken since August 2015 have helped Turkey outperform” expectations. In short, Kenc communicated that everything required for the bank to reduce rates had come to fruition.
Prattle’s algorithm detected this trend and scored every relevant communication from the Bank of Turkey since February 26 as increasingly dovish. As the graph below illustrates, there was a particularly precipitous drop in sentiment leading up to the rate cut decision, indicating a policy change may be imminent.
While the dovish trend was clear to our algorithms, it wasn’t as obvious to outside experts.
This is not due to any lack of diligence by industry analysts…but by the inherent difficulty of the task. Drawing patterns out of dense masses of central bank communications is incredibly challenging, requiring a level of objectivity and comprehensive evaluation that is near impossible for human analysts to achieve.
This kind of analysis, however, is exactly what Prattle’s algorithm was designed to do. Prattle’s expertly trained, domain-specific textual analysis program evaluates the entire scope of communications for 20 central banks around the world, producing comprehensive, unbiased, and quantitative assessments that identify the mood of each institution.
Prattle’s technology is precisely what the central bank watching world needs to predict the kind of policy decisions like what we saw from Turkey this week.
Interested in Prattle’s data? Reach out, and we can set you up with a free, no obligation trial of our Central Bank Sentiment Indexes.
The Prattle Team