Gilt Edged Mini (GEM)

GEM is the use of ABMLP-X in order to run and consume - pro rata - real-world UK expenditure and interest rate time-series. A model run will typically consume time-series data that starts in either the year 1970 or, less typically, starting in the year 1695. A model run may end with either the latest values from the available time-series, or, if originating governing parameter settings, with values that represent some future date.

GEM represents the monetary system connections between the Government, a Central Bank and many, possibly thousands, of Producer and Household agents.

Development​

Supposition​

The generally accepted perception: Movements, especially large ones, in the price of a UK Gilt index are always the consequence of bond traders reacting to recent policy proposals. And to an extent, at the margins, near-term decisions show effects. I suspect, however, that the major price movements we see in Gilts result from longer-term system dynamics. The time lagged behaviours - velocity and acceleration effects - caused by an adjustment to one or more system governing parameters - adjustment(s) made months previously.

Analysis​

I use local Jupyter notebooks to wrangle and analyse both model (macro and agent-level) output as well as real-world UK economic data. I connect the output to a large language model (LLM) in order to further analyse and appreciate the datasets from a political-historical context.

Analysis includes, but is not limited to:

• When and why the UK spends:
  • Wars, pandemics and historical patterns.
• The evolving development of Household agent portfolio strategies:
  • In every model step, Household agents pursue wealth management strategies. These include, but are not limited to, the management of a Government money portfolio.

View a summary technical architecture.