A Shareholding Reconstruction of CHOAM, 10120–10191 AG: Dispute Records, Silent Partners and the Distribution of Spice Monopoly Rents

1. Introduction

Every Great House of the Landsraad measures its standing by its CHOAM directorships, and every Great House complains that it cannot tell what those directorships are worth. The Combine Honnete Ober Advancer Mercantiles, the universal development corporation through which melange and most other interstellar commerce pass, pays dividends to the Padishah Emperor and to the Great Houses. It also carries two partners who hold no visible seats at all: the Spacing Guild and the Bene Gesserit. The division of this income has never been set out in a single document available to any one holder.

Our question is how the melange monopoly rent was actually divided between 10120 and 10191 AG, the seven decades that preceded the Arrakis handover and the fall of the Corrino throne. We write from Kaitain in 10248 AG, when the Directorate first released parts of its pre-Jihad dispute series to accredited researchers.

Three arguments follow from the reconstruction. The first and central one is that the Combine's illegible shareholding worked as a stabiliser, whether or not anyone intended it to. Each holder could see its own gross dividend but not the shares of its rivals, and this kept grievance from hardening into coordinated action. Our second argument treats Imperial dividends as a plausible financing channel for the Sardaukar. The third reads the transfer of a fief contract as a mechanical reallocation of directorship weight, and applies this reading to the Arrakis handover of 10191 AG.

2. Sources and Method

The core source is the CHOAM Directorate's dividend dispute series, from which we recovered 47 proceedings dated 10120–10191 AG. Of these, 38 concern the core melange dividend and 9 concern sub-ventures whose share registers were published to their holders (CHOAM Directorate, 10120–10191 AG). A dispute record typically states the complainant's claimed entitlement, the Directorate's ruling and, in about half the cases, a partial figure for the total pool against which the claim was measured.

We combined these proceedings with financial disclosures lodged by 23 Great Houses during bankruptcy, marriage settlement or succession proceedings, most of them held in the Landsraad Public Archives. Nine of the 23 disclosures cover five or more consecutive fiscal years. For the Harkonnen and Atreides positions we also used household ledgers from Giedi Prime and Caladan (House Harkonnen, 10150–10191 AG; House Atreides, 10170–10191 AG). The governance context for these filings, in which property follows corporate rules while succession follows feudal ones, is set out in earlier work on the Great Houses.

The reconstruction pools records that were never held in one place. A House saw its own receipts and whatever pool figures surfaced in its own disputes. We see 47 disputes and 23 disclosures together, and each pool figure constrains every holder at once. Holders' shares were estimated by bounded least squares against the reported pool figures, and the uncertainty ranges come from resampling the disputes. With pooling, the 90% interval for a single Great House's share of total flow is typically about eight percentage points wide. Re-running the resampling on a single House's disputes and disclosure, we find that a House working from its own records alone could have narrowed that interval to no better than roughly twenty-five percentage points.

The two silent partners never appear as claimants, so their stakes are inferred indirectly. For the Guild we used fee schedules that the Guild disclosed to the Directorate (Spacing Guild, 10140–10191 AG). An ordinary transport or banking tariff is a cost the Combine pays for a service and is no part of dividend flow. The disclosed schedules, however, separate that tariff from a profit-participation line, paid as dividend-in-lieu and calculated against the same pool figures as the House dividends. Only that line was counted toward the Guild's share. The Bene Gesserit stake rests on marriage-settlement clauses that assign dividend streams to the Sisterhood and on residuals that the other holders cannot account for. It is the least certain figure we report.

Access came through the Economic Research Office, where the first author holds a post. Under the terms of release the Directorate reviewed the manuscript for the names of living directors and redacted eleven of them. It asked for no change to any estimate. Readers should nonetheless weigh the institutional interest of an author employed by the body under study.

3. The Directorship Lattice

The reconstructed distribution is given in the table at the end of this section. The central estimates sum to 100%; each range is that holder's 90% resampling interval.

Allocation of seats remains our own hypothesis. The dispute rulings refer to seats held by Imperial nomination and seats attached to Landsraad fief contracts, and they never mention a seat held by the Guild or the Sisterhood. We therefore model the silent partners as holders of dividend rights without votes. This squares with their absence from every claimant list.

Illegibility followed from the structure. A Great House knew its gross receipts to within a fiscal cycle, but the pool figure against which those receipts were measured changed from dispute to dispute. The House could not separate a fall in its share from a fall in the pool. The Imperial House was better placed, since it could see its own dividend and the pool figures quoted in its own rulings. Even so, the Guild and Bene Gesserit stakes were hidden from it as well. The cost of this blindness appears in the market for seats. A separate group of nine disclosures, not identical to the multi-year subset described in Section 2, records a sale of directorship rights. The prices realised lay between roughly 15% and 45% below the seller's own capitalised dividend receipts and showed no evident trend.

4. Imperial Dividends and the Sardaukar

Applied to the pool estimates for 10170–10191 AG, the Imperial share yields directorship income of 1.4–2.3 billion solaris per standard year. For the cost of the Sardaukar we combined provisioning and garrison accounts from Salusa Secundus with Imperial Treasury transfers to the training establishment (Salusa Secundus Garrison Command, 10160–10191 AG; Imperial Treasury, 10175–10191 AG). The resulting estimate is 1.1–1.9 billion solaris per standard year.

The ranges overlap between 1.4 and 1.9 billion; their ratio lies somewhere between about 0.7 and 2.1, so the income could have fallen short of the cost or roughly doubled it. We call the two commensurate and go no further. We cannot show that dividends were earmarked for the legions. What we can say is that the Throne's corporate income was of the right order to carry them without a levy on the Landsraad, and that no such levy appears in the ledgers we examined. Earlier work treats the prison planet's harsh conditions as a deliberately preserved input to Sardaukar production, whose costs belong on the same account.

If the hypothesis holds, an interruption in the harvest on Arrakis would then have reached the Sardaukar within one or two fiscal cycles, a link Landsraad debate seldom named.

5. Fief Transfer as Reallocation

Arrakis was held as a quasi-fief, meaning a planetary stewardship granted under a CHOAM company contract with the fief and its commercial concession bundled together. The Harkonnens held it on those terms for about eighty years, and in 10191 AG it passed to House Atreides by Imperial decree, granted this time as a siridar-fief but under the same CHOAM company contract. Our model attaches directorship weight to the contract, so a change of holder moves votes and dividend rights without altering the lattice.

Seven earlier fief transfers between 10120 and 10191 AG can be traced in the dispute series. In six of them, directorship votes recorded at the next Directorate sitting moved to the incoming House within one fiscal cycle. The seventh transfer was contested and the votes were held in escrow. The Arrakis contract was larger than any of these. Scaling Harkonnen ledger receipts from Arrakis against the pool figures for 10170–10191 AG, we estimate that it carried 3–6% of total dividend flow. That share falls within the Great Houses' aggregate of 51% and is comparable to the entire Bene Gesserit central estimate. The Harkonnen ledgers show the Baron's factors moving liquid holdings off Arrakis during the handover season, which is the behaviour the model predicts from a holder about to lose contract-attached rights.

Transfer also bears on the Imperial House. Work on Corrino succession law shows that Landsraad ratification made each accession a moment of bargaining between the Throne and the Great Houses. Our own inference is that contract-attached seats were one of the goods bargained over, and so a currency the Throne could issue and withdraw. On that reading, reassigning Arrakis moved some of that currency into Atreides hands.

6. Discussion

Illegibility offers one explanation for the Combine's durability over these decades. Had shares been common knowledge, a coalition of Great Houses could have seen which holder was over-rewarded and acted against it. Under the actual arrangement each House suspected the others separately and could prove nothing. The dispute series allows a partial test. In disputes over the opaque core dividend, 5 of 38 (13%) were brought as coordinated petitions by two or more Houses. The figure rises to 5 of 9 (56%) for disputes over sub-ventures with published registers (Fisher's exact test, two-sided, p = .013).

The sub-venture group is small, and legible and opaque ventures may differ in ways other than legibility, including the size of the stakes and the rank of the Houses involved. The broader claim, that opaque arrangements outlast transparent ones where claimants are many and small, therefore remains a hypothesis. It is compatible with the record without being established by it. A related argument was made earlier by one of us in the Directorate's working papers, which treat disclosure as a cost to be minimised (Iskadar, 10236 AG).

Opacity was not free. Houses could not price their holdings, directorships sold at the erratic discounts reported in Section 3, and capital was likely misallocated. We read the arrangement as a trade of allocative efficiency for political stability. The Throne's advantage in it was informational. Its share of dividend flow was smaller than that of the Great Houses in aggregate, but among the holders it could see furthest into the lattice while staying hidden from everyone else.

7. Limitations of the Record

Survival is the first problem. The 47 proceedings are those the Directorate chose to keep and later chose to release. Disputes settled privately, or recorded in series still withheld, are missing, and they may differ systematically from those that survive. The 23 disclosures come mostly from Houses under legal stress, and such Houses may not represent the Landsraad as a whole.

The silent-partner estimates depend on indirect evidence. The Guild range rests on fee schedules disclosed by the Guild itself, which had reasons of its own to present them selectively. Separating the profit-participation line from the ordinary tariff also depends on the Guild's own classification, which we could not audit; if part of that line was in substance a service charge, the Guild figure is overstated. The Bene Gesserit range rests largely on residuals, so any error elsewhere in the model, including this one, ends up in it. Solaris figures are also nominal across seven decades. We did not deflate them, and comparisons between the early and late parts of the window should be treated with care.

Finally, the Sardaukar comparison shows that the magnitudes match but not that the money moved. Treasury transfers name the training establishment without naming the source of funds. Only the release of the Privy Purse accounts of the late Corrino household, which remain withheld, could confirm or refute the financing hypothesis.