Before portfolio choice
Portfolio theory begins with an investor and a set of available investments. An investment office also encounters the question that comes before that choice: what could become available if compatible investors acted together?
Differences that matter
A shared investment can depend on several commitments arriving on acceptable terms. One investor may need liquidity; another may face a concentration limit; a third may be able to participate only through a particular account. These differences need not require separate underlying investments. They do require an accurate account of who can do what, on which terms, with whose permission.
Independent mandates
Polycentric finance studies investment under independently governed mandates. The point of coordination can be a program rather than an entire portfolio. Investors may join a common investment while retaining different holdings, obligations and authorities around it. Funds, partnerships, separate accounts and conventional custom vehicles already solve parts of this problem. Their capabilities belong in the comparison.
When better alternatives change the group
The distinction between choosing investments and forming them has an unexpected consequence. Better options for an individual can weaken a shared opportunity. If his departure prevents an investment from reaching the scale it needs, the other investors can lose more than he gains. The flagship paper proves a precise version for fixed-term threshold programs with scale economies. It also shows where that logic fails or changes: congested access, changing terms and competing uses of capital require their own analysis.
Willingness, coordination and funding
Three questions help diagnose a failed program. Would the investors participate on the offered terms? Can they coordinate on doing so? Can the accepted obligations actually be funded? Transfers can improve an investor's terms. An assurance payment can change the incentive to subscribe when others may refuse. A funded backstop can cover a defined shortfall. Each has a cost and a party responsible for paying it.
A current book and clear authority
An investment office therefore needs more than a preference list. It needs a current book of claims, cash, obligations and evidence; a mandate that distinguishes preferences from hard restrictions and authority; and a record of accepted commitments that survives later preference changes. A promising program must fit those books jointly. Two attractive alternatives cannot both spend the same dollar.
What AI changes
AI changes the cost of this work on both sides. It can make separate service cheaper and can help shared implementation preserve individual circumstances. Persistent records can reduce repeated interpretation. Reusing common documents can reduce duplicated processing. The accompanying methods report measures those effects in a bounded experiment, including errors and service failures. It does not establish the full cost of a real investment office.
The office and the research
Argosy combines a personal investment office with shared programs. The research asks which arrangements create value, whose participation they require and who receives the gains. A program is worth forming only under terms its investors can accept, with authority and funding that support those terms.
References and scope
The paper’s models and the methods report’s processing results address different questions. Their assumptions and measurement limits are stated in the publications.
- John Komkov, The Organization of the Investment Frontier, 5 October 2026.
- John Komkov, Separate Records, Shared Processing, AI methods report, 5 October 2026.