The record, including where it went wrong
Each case states what the state did, what the operator did, and what resulted. Outcomes are described, not rated. Where a programme was ended or restructured, the reason is given — that reason is usually the most useful thing in the case.
Seventy years of state land allocation
Bar length is the operating period. The colour marks how each programme stands now — running as designed, restructured, or ended. Most of these have been running long enough to show what happens after the first generation, which is where several of them changed.
The problems appear in the second generation, not the first
Every programme here reported success in its opening decade. The difficulties that defined them — settler debt at replanting, aquifer depletion, monitoring that lapsed, repayment that never came — surfaced fifteen to thirty years in, well past the point at which the instrument could easily be changed.
A tenure decision taken now is being taken for a period longer than most of these programmes have existed.
A result elsewhere is not a forecast here
Precedent is the strongest material in this application and the easiest to over-read. These four rules govern how the cases are used, and they are the reason each case separates what transfers from what does not.
1 · Structures transfer. Agronomy does not
The Malaysian settlement scheme is the clearest illustration. Its crop was oil palm in a wet tropical climate — nothing about that yield, that rotation or that water regime has any application in the Kingdom. What does transfer is the financing structure: development cost treated as a recoverable loan, title issued on repayment, and the state left holding the lending role because settler land could not be pledged. That structure would behave the same way under any crop.
2 · Water regime is the first filter
| Programme | Water basis | Bearing on Saudi application |
|---|---|---|
| Saudi wheat programme | Non-renewable, declining | The only case that ran in the Kingdom, and water is why it ended |
| Egypt reclamation | Largely non-renewable | Comparable water position; the sequencing failure is the transferable lesson |
| Morocco land leasing | Rainfall and surface, under drought stress | Allocation machinery transfers; water assumptions do not |
| Australia pastoral | Rainfall, arid rangeland | Rangeland tenure and enforcement transfer well |
| Israel leasehold | Managed, substantially reused | Tenure doctrine transfers; supply position does not |
| Nigeria anchor scheme | Rainfall, humid | Offtake and input structure transfers; production system does not |
| Malaysia settlement | High rainfall, tropical | Financing structure transfers; nothing agronomic does |
3 · Read the failure, not the launch
Programmes are documented most thoroughly at announcement and most usefully at review. The figures that matter in these cases — repayment rates, settler debt, cultivated area against target, land condition — were published years afterwards, often by an auditor or a lender rather than by the implementing body. Where both exist, this module uses the later figure.
4 · Absence of precedent is not evidence against
Five sub-models in the catalogue have no located implementation. That records a gap in the record, not a verdict on the mechanism. Partnership structures in particular are ordinary instruments in other sectors; no agricultural land programme operating them has been found, which is a reason to examine them carefully rather than a reason to exclude them.
Every one of these programmes was changed by something outside the land instrument
Wheat cultivation ended through procurement quotas, not through land policy. Settler hardship arrived through replanting finance, not through the allocation. Rangeland condition degraded through monitoring that lapsed, not through the lease terms. Repayment failed through enrolment politics, not through the contract.
The mechanism chosen determines what the state can do later. It does not determine what will go wrong. On the evidence here, what goes wrong usually arrives from an adjacent system, and the useful question about any model is what leverage it leaves the state when that happens.