Originators tend to inherit a delivery structure rather than choose one. Whatever was set up when the shelf was built keeps running, and the question of whether it still fits the business gets asked only when something breaks: a warehouse line gets tight, a product line grows faster than expected, or a quarter-end lands badly.
It is worth asking on purpose. The three structures solve genuinely different problems, and the right answer depends far more on how predictable your production is than on what the loans look like.
Flow: sell as you close
Loans are sold individually or in small, frequent batches, typically against a standing arrangement with agreed parameters and pricing mechanics.
What it is good at
- Balance-sheet velocity. Capital comes back quickly and the warehouse line stays available, which is usually the whole point.
- Small, steady operational load. A little work continuously, rather than a lot of work at once.
- Fast feedback. If something in your underwriting or file assembly is causing friction, you find out this week rather than after two hundred loans have been written the same way.
What it costs
- Less certainty at lock than a forward commitment gives you, since pricing follows the market as you deliver.
- Parameters have to be agreed in advance, which means loans outside them fall out of the flow and need handling separately.
Flow tends to fit originators with steady, reasonably homogeneous production who care most about keeping capital turning.
Bulk: sell an assembled pool
A defined set of loans is packaged, priced and traded as one transaction.
What it is good at
- Heterogeneous production. A pool can carry variety that no flow parameter set would accept, because it is priced as the specific collection of assets it is.
- Deliberate balance-sheet moves. Exiting a product, reducing a geographic concentration, monetising a seasoned book: all are pool-shaped decisions.
- Seasoned collateral. Payment history is evidence, and bulk is where seasoned loans get looked at as seasoned loans.
What it costs
- Concentrated operational effort: tape construction, a data room, a diligence cycle and a settlement, all in one push.
- Loans sit on the line while the trade is assembled and executed, so warehouse cost is real and worth counting.
Bulk fits lumpy or varied production, and any situation where the objective is a specific change to the balance sheet rather than routine turnover.
Forward: commit ahead of origination
A commitment is made for loans that have not been written yet, against agreed characteristics and a delivery window.
What it is good at
- Certainty at lock. You know your exit before the loan exists, which is what makes aggressive front-end pricing defensible rather than hopeful.
- Predictable production. If you can forecast what you will write, a forward converts that forecast into an economic advantage.
What it costs
- Delivery discipline. A commitment you cannot fill is a problem, and one you overfill needs somewhere else to go. This is the real constraint, and it is operational rather than financial.
- Less flexibility if your product mix shifts inside the window.
Forwards fit originators with a reliable forecast and enough production discipline to hit a window. They are a poor fit for a shelf whose mix is still moving.
The question that actually decides it
Set the collateral aside for a moment and ask two things:
- How predictable is your production, three months out? High predictability makes forwards available. Low predictability makes them a liability.
- What is the binding constraint right now: capital, certainty, or operational capacity? Flow solves capital velocity. Forwards solve certainty. Bulk solves a balance-sheet composition problem, and it consumes operational capacity to do it.
Most originators of any size end up running more than one structure at once, and that is usually correct rather than a sign of drift: flow for the core product where the parameters are known, bulk for the seasoned or off-parameter production that flow will not take, and a forward where the forecast is genuinely solid.
How this looks from the buy side
We bid all three, and the underlying analysis does not change between them. Loans are scored individually on their own attributes whether they arrive one at a time or a thousand at once, and a trader reviews the output before any bid goes out. What changes is the packaging and the timeline around it.
The one thing worth saying to a seller weighing the choice: do not pick a structure to make the loans easier to buy. Any buyer worth trading with can handle varied collateral. Pick the structure that fits how your business actually produces, and let the pricing process deal with the variety.
