lambda, against submitted

The declared configuration of each arm, from data/arms.csv. submitted is this report’s reference.
switch	submitted	lambda	differs
tar_spec	submitted	customs	<– differs
rec_spec	sector5	secondary	<– differs
waste_spec	last_sector	treatment	<– differs
price_spec	coded	coded
eq_spec	coded	adjudicated	<– differs
wage_spec	numeraire	numeraire
armington_epsilon	0	0
wage_drift	0	0

lambda — Lambda (region-coded).

Run read from output/scenarios/arms/lambda/runs/baseline.RDS (cut 2026-09-29 06:52), reference from output/scenarios/arms/submitted/runs/baseline.RDS (cut 2026-09-23 09:50). Both ran all 100 periods.

What moves, and by how much

The fifteen widest gaps at t = 99, over the 29 aggregate variables these nine blocks cover. `relative` is the difference over the absolute reference value, so a sign change shows as a magnitude above 1.
variable	region	submitted	lambda	difference	relative
gdef	RoW	6.3961	15.5052	9.1091	1.4242
cab	RoW	-6.5366	-15.6482	-9.1116	-1.3939
cab	EU	6.5366	15.6482	9.1116	1.3939
gdef	EU	-6.6010	-15.7046	-9.1036	-1.3791
rec	RoW	4242.0097	1.6455	-4240.3641	-0.9996
rec	EU	439.1987	5.9859	-433.2128	-0.9864
mat	EU	647.1575	1090.2926	443.1352	0.6847
mat	RoW	7658.3368	11943.9482	4285.6114	0.5596
b_s	EU	1089.2191	751.1410	-338.0780	-0.3104
b_s	RoW	5334.8099	5739.2537	404.4438	0.0758
tb	EU	21.8614	20.8909	-0.9706	-0.0444
tb	RoW	-21.8614	-20.8909	0.9706	0.0444
ineq	EU	0.7348	0.7027	-0.0321	-0.0437
lf	EU	152.4000	147.5035	-4.8965	-0.0321
ydc	EU	250.8385	257.5847	6.7463	0.0269

Fit against the calibration targets

Fitness on the twelve-concept target set. `all_relative` is the old LEEDS objective (each deviation over its own target); `all_gdp` divides instead by the region’s value-added target, so a near-zero target cannot dominate; `real_relative` is the eight real quantities only. Lower is closer.
arm	all_relative	all_gdp	real_relative	worst_gdp_cell	worst_gdp_share
submitted	58.96345	0.019331	0.002010	Z2_b_s	55.44593
lambda	339.11772	0.028245	0.002161	Z2_b_s	55.31096

Economic

Macroeconomic

External

Social

Prices

Employment

Inequality

Debt.and.Wealth

Ecological

Provenance

This report is qmd/reports/arms/arm-baseline.qmd, rendered once per arm by R/render_arm_reports.R. It reads only cached baselines under output/scenarios/arms/<arm>/runs/baseline.RDS and runs no model.

The nine blocks are taken from ../LEEDS/policyreport 2/policy-report-aug13.Rmd. The two differences from that original: the two series are two arms rather than a baseline and a shock, so there is no shock line; and every block is drawn in levels, because the two arms start from different initial states and an index to each arm’s own t = 2 value compares those starting points rather than the paths.