Reading Tea Leaves
Textual analysis reveals differences between the ECB and Fed
Tasseography is the art of interpreting messages found in tea leaves left at the bottom of a cup - though it can be applied, apparently, also to coffee grounds and wine sediments. Interpreting central bank policy can often feel a somewhat similar endeavour. In this short post, I attempt to divine meaning from documents reflecting discussion at the Fed and the ECB ahead of policy decisions. I think it is a useful approach - you may feel it is little better than shifting the tea leaves. Cheers!
The Fed’s Beige Book is a compendium of analysis compiled by all twelve Fed districts which provides policy-makers with an up-to-date assessment of the US economy. It is published about two weeks ahead of FOMC meetings. There is no clear equivalent for the ECB, but the European central bank does publish a discussion of policy meetings known as the Monetary Policy Account. Both the Beige Book and the Monetary Policy Account provide mostly qualitative assessments, which makes them attractive to commentators, and much less attractive to quantitative analysis - usually. However, textual analysis can help tease out features of both documents.
Among other great work by Len Kiefer is his suite of analysis tools looking at the Beige Book. Based on Len’s work let’s look at how the number count of the word ‘inflation’ has evolved in the Beige Book and the ECB’s MonPol Account since 2015 (the earliest date available for ECB MonPol Account). There are some interesting takeaways.
The size of the Beige Book is several times larger than the ECB’s MonPol Account, so occurrences have been ‘standardised’ for both documents, to aid comparison. I’ve also included the relevant official interest rate (Fed Funds and ECB Deposit Rate) to orient changes in policy against changes in text.
Strikingly, the ECB began to mention ‘inflation’ a lot more from mid-2021, about 9 months before the inflation alarm was raised by the Fed (March 2022). Yet despite obviously talking more about inflation, the ECB waited over a year before moving interest rates higher. In contrast, the rise in mentions of ‘inflation’ in the Beige Book preceded change to tighter Fed policy by just two weeks. Perhaps worth remembering.
In addition, the Beige Book appears to reflect a lot more concern about inflation from 2022 till early 2023 and also appears to have shown more concern than the ECB, whose standardised measure never reach the same high levels.
Then again, recent Beige Books have seen a very significant fall in mentions of ‘inflation’ - the last two publications both showed fewer mentions than appeared in the March 2021 report that preceded the first rate rise. This tallies with the FOMC’s ‘pause’ at the June meeting. In contrast, the latest available ECB MonPol Account (for meeting 14-15 June) shows no real decline in mentions of the word ‘inflation’, and so, I assume, no decline in concern for inflation by the ECB.
Len also created a handy ‘sentiment’ index using a custom lexicon (‘bing’) that assigns words either positive or negative value in a binary fashion. I’ve updated his work to include the latest Beige Books and and applied the same approach to the ECB’s MonPol Account over the same period. The sentiment index is intended to assess the overall economic situation, not inflation.
Both sentiment indices show a similar profile since 2015, but the Fed shows much greater shift to negative assessment of the economy at the beginning of the COVID19 pandemic.
More recent sentiment shows broadly balanced assessment, but with Beige Book reflecting consistently better outlook for the US economy than the ECB MonPol Account does for the Eurozone. The last Beige Book delivers a net positive score of +0.1, whereas the last ECB MonPol Account shows a net negative score of -0.08.