In the Hebrew Bible, a Jubilee (Leviticus 25) is a periodic reset: every 50th year, debts are cancelled and ancestral land returns to its original family. Deuteronomy 15 describes a related, more frequent practice -- a sabbatical (7-year) release of debts. Both are collective, calendrical resets, not case-by-case mercy.
The most interesting historical wrinkle is Hillel's prosbul: a legal workaround that let a lender pre-declare a debt collectible through the courts, exempting it from the sabbatical release. Hillel introduced it, according to the Mishnah, because lenders had started refusing loans as the release year approached -- people in need couldn't get credit precisely because the relief mechanism existed. That anticipation problem (Deuteronomy 15:9 warns against it directly) is the central tension this lab is built to let you actually watch happen, instead of just read about.
The wealth dynamics underneath this lab aren't specific to Jubilee -- they're the standard agent-based credit model going back to Epstein & Axtell's Sugarscape (Chapter V), and to a body of econophysics work on wealth concentration. Random pairwise exchange alone (Drăgulescu & Yakovenko, 1999) produces an exponential wealth distribution, not the fat Pareto tail real economies show -- you need multiplicative dynamics (returns compounding on existing wealth) to get real concentration. Boghosian's Yard-Sale and Affine Wealth Models go further: they show wealth condensation -- one agent eventually owning everything -- is a mathematical inevitability under perfectly fair random trade, not evidence that the trades were unfair. Their "affine" model adds two parameters, χ (redistribution) and ζ (wealth-attained advantage), that fit real US/EU wealth distributions closely. A Jubilee, in that frame, is a periodic impulse added to the redistribution knob.
The other load-bearing idea here is Michael Hudson's argument that most ancient agrarian debt wasn't voluntary credit at all -- it was unpaid taxes, irrigation fees, and other arrears. That matters for the anticipation problem: Hillel's objection only makes sense if a lender is choosing whether to extend credit. If debt is arrears instead, there's no lender making that choice, and the anticipation critique collapses. This lab makes that a slider (arrears fraction) rather than a debating point.
One caution worth stating plainly, since it's easy to misread a chart as a discovery: that debt forgiveness reduces measured inequality for one period is trivially true and not worth simulating on its own. The non-obvious questions are what happens under repeated cycles -- does credit dry up as a reset approaches, does forgiveness just reset the clock on the same structural problem, and how does a king's collective mercy compare to individuals getting relief on their own?
Two independent relief mechanisms: a collective Jubilee (the king resets everyone's debt at once, on a periodic, random, inequality-triggered, or manual schedule) and an individual bankruptcy (any agent whose own debt burden gets bad enough for long enough gets their own discharge, with a credit lockout afterward -- no king involved). The Scenario dropdown on the lab page walks through the contrast directly: no relief at all, a working periodic jubilee, bankruptcy alone, and both together.
Full implementation notes -- the model's step-by-step mechanics, bugs found along the way, and why
the default parameters are what they are -- are in
design/jubilee-implementation.md in the repo.