Modern central banks have a number of important responsibilities. Chief among them is setting monetary policy, but others include ensuring financial stability, regulating financial institutions and providing payment systems. A great deal of research is required to make sure the central bank is making the right judgments in exercising its responsibilities, which is why it typically employs researchers on staff.

Modern LLMs make it easy to parse the volumes of research that the major central banks have put out over decades. We can easily assess and compare, for example, the volume of research output as well as the topics that they cover.\n\nIn terms of volume of research, the major central banks — the Federal Reserve, European Central Bank, Bank of England and Bank of Japan — exhibit different patterns. Larger central banks, not surprisingly, produce more research.

The Bo E, like the Fed, stands out for having a big increase in output following the financial crisis. The FT reported recently that the Bo E’s research infrastructure is undergoing substantial change in the wake of the Bernanke Review, with the research hub being shut down and several senior researchers leaving. The trend towards fewer outputs may well therefore continue.

All together from more than 5,000 staff research papers since 1997 across our central banks, we asked the machine to read each abstract to analyse the topic of each paper. Monetary policy is unsurprisingly the number one in terms of the share of total output per central bank. The Federal Reserve has maintained a steady share on the topic at about 40 per cent since the late 1990s.

The Bank of Japan, on the other hand, steadily increased its research on monetary policy as time went on and the country remained mired in economic stagnation, which saw very little inflation. The ECB and the Bo E both saw declines in study of monetary policy after the financial crisis from relatively high levels, but this has partially rebounded since the 2021-22 inflationary surge.\n\nThe second major topic of study at central banks has been financial stability. The share of outputs on this topic rose substantially in the aftermath of the 2007-08 financial crisis.

Prior to the crash, there was essentially no research at the major central banks on this topic. That has since changed and the topic retains a substantial share of the total for each of the central banks. One point that emerges from the general patterns in the figures above is that research seems reactive to events rather than proactive.

Had there been more study of financial crises and other financial stability topics in the lead-up to 2007-08, central bank policymakers may have been better prepared to deal with events. But on the positive side, the quality of research has been increasing across the board, at least according to the LLM’s judgment of each paper based on the abstract alone. Quality improvements might not be so valuable if, as when the financial crisis took central banks by surprise, there is insufficient knowledge and expertise on relevant topics.