Map the exposures.
Identify the asset, direction, size and venue behind each position. Similar tickers are not enough to establish an offset.
FIG. 01 — A PORTFOLIO WITHOUT WALLS
Build with
less locked up.
A single view of exposure. A precise place for every pledge. Better architecture for capital across venues.
Explore the architecture ↗EXPOSURE ≠ COLLATERAL OWNERSHIP
ISOLATED ALLOCATION$5.0M
↓ILLUSTRATIVE PORTFOLIO ALLOCATION$3.2M
CAPITAL RELEASED IN THIS EXAMPLE$1.8M36% less collateral allocated
Hypothetical comparison for design illustration. Actual savings require a solved portfolio model and enforceable collateral controls.
One portfolio.
Two very different views.
A long at one venue and a short at another may offset economically. Separate margin systems can still demand capital from both sides.
MarginMesh proposes a shared exposure view before making an allocation. The hedge is useful only to the extent it can be trusted under stress.
ILLUSTRATIVE MATCHED EXPOSURE$800,000
VENUE A · LONG $1MVENUE B · SHORT $800K
The matched amount is a potential offset, not automatically available collateral.
Make the hedge visible.
Keep the pledges separate.
Stress the relationship.
Ask what happens when prices gap, basis widens or one venue liquidates before the other. Give the hedge less credit when the connection is fragile.
Allocate distinct collateral.
Assign each collateral lot to exactly one venue. Risk may be viewed across the portfolio; ownership and pledge records must remain precise.
Recheck when things change.
A smaller hedge, a new position or a changed risk limit should trigger a new allocation proposal. Capital efficiency needs continuous discipline.
What could an offset
change in practice?
In our separate teaching example, a $1M long and an $800K short each face a 20% stress assumption. Recognizing half the potential hedge changes the allocation.
COUNTED SEPARATELY$360K
PARTIAL HEDGE RECOGNIZED$200K
ILLUSTRATIVE CAPITAL RELEASED$160K
COLLATERAL LOT A / VENUE A$111,111
COLLATERAL LOT B / VENUE B$88,889
Rounded allocation of the $200,000 example requirement. Each lot is separate. This teaching rule is not the proposed CVaR optimizer or an enforceable cross-venue margin agreement.
Try the simple simulation ↗A hedge is only as good
as its weakest assumption.
Basis risk
Two exposures can drift apart. Model the relationship rather than assuming identical prices.
Liquidation timing
One venue may close a position before the other. A portfolio hedge does not ensure coordinated execution.
Venue failure
An offset that cannot be accessed may not protect the remaining position.
Pledge exclusivity
One asset cannot support two claims at once. Allocation requires distinct, enforceable pledge records.
Is this a live cross-margin service?
No. The site presents the research direction and an interactive two-venue teaching example. There are no connected venues, collateral transfers or executable margin agreements.
Can the model remove all collateral?
The simplified demo can show zero for a perfectly matched, fully recognized hedge. Real systems need additional assumptions and buffers; that result is not a usable margin policy.
Where is the mathematical model?
The article explains the simple example and includes the proposed CVaR optimization model for readers who want the formal constraints.
Read the article →