LLMs widely used in economics modeling by the end of 2026?
➕
Plus
15
Ṁ726
2027
43%
chance

It has been suggested that LLMs could be used to ingest large amounts of economic data (accounting books, economic facts about a community) to create an oracle capable of answering questions relevant to policy. For instance one could take 'a small village' (as Tyler Cowen put it) and train a LLM on a wide array of unstructured data to implicitly model the village's economy. An agency such as a central bank could then use the trained LLM to predict the effects of policy decisions, e.g. what would the effect on unemployment be if the interest rate rose by 0.25%?

Will at least one central bank or government of an OECD country publicly disclose having trained such a model with the intention to use it, by Dec 31, 2026?

Releasing a working paper or equivalent official communication with details of the training and intended application would count towards a YES resolution. Merely speculating on possible applications or declaring that such a project is in process would not. Researchers working on behalf of other institutions but sufficiently related to a central bank or government would still count as long as usage for policy planning is considered and the trained model is regarded as adequate for that purpose.

Other machine learning or simulation efforts that are not an LLM trained on text or other unstructured data will not count. Limited geographic scope, as long as at least a town or similarly sized community is covered, is not grounds for resolving NO.

Get
Ṁ1,000
and
S1.00
Sort by:

It's very likely that someone would train a model and release a paper about it. Actually use it in practice for making useful predictions - less so.