Insights

Aggregate Borrow-Cost Budgeting

Alphanume Team · February 27, 2026

Treating borrow as a portfolio-level expense.

Borrow cost is the largest single expense category for a short-side strategy. Treating it as a portfolio-level budget rather than a per-position consideration produces meaningfully better aggregate outcomes — it forces explicit prioritization of which positions earn their borrow expense and which don't, and it caps the total drag on portfolio returns.

Why budget at the portfolio level

Per-position borrow-cost analysis is necessary but not sufficient. The portfolio-level view captures:

Total annualized borrow expense across all positions.
Marginal contribution of each position to total borrow.
Trade-offs between holding higher-borrow names and lower-borrow names.
Operational implications for total capital deployment.

The annual borrow budget

A reasonable framework:

Set a target annual borrow expense as a percentage of AUM. Typical: 3-7%.
Track running annualized borrow cost across the portfolio.
Refuse new positions whose addition would push the total above target.
Re-evaluate existing positions whose contributions exceed reasonable thresholds.

Computing the running cost

For each open position:

Annualized borrow cost = Position value × Annualized borrow rate

Sum across positions to get portfolio-level annualized borrow expense. Divide by AUM to get the percentage drag.

Track this metric daily; flag when it approaches target ceiling.

The per-position prioritization

When borrow budget is constrained, prioritize positions by expected-return-to-borrow ratio:

Position A: 8% expected return, 30% annualized borrow over expected 60-day hold = 5% net.
Position B: 12% expected return, 100% annualized borrow over expected 60-day hold = -4.6% net.
Position C: 5% expected return, 5% annualized borrow over expected 60-day hold = 4.2% net.

Position A is the best risk-adjusted use of borrow budget. Position C is acceptable. Position B is uneconomic.

Why expected return matters more than nominal borrow rate

A 100% annualized borrow rate sounds prohibitive but isn't always. If the expected return is high enough and the holding period is short enough, the trade can still net positive. Conversely, a 10% borrow rate sounds modest but can be uneconomic if the expected return is also small.

The relevant ratio is expected-return-per-day vs borrow-cost-per-day, not the absolute borrow rate.

The capacity dimension

Borrow budget interacts with capacity:

If borrow budget allows 5% drag at $100M AUM, the strategy can spend $5M annually on borrow.
A typical $1M average position with 20% borrow rate generates $200K annual cost.
The budget supports 25 such average positions running concurrently.
If positions average 60-day holds, total annual position-events = 25 × (365/60) ≈ 150 trades.

The math defines the operational scale.

The opportunity cost

Borrow cost is the explicit cost. Opportunity cost is implicit: the trades not taken because the borrow budget is exhausted. Tracking both:

Refused trades: maintain a log of positions screened out due to borrow-budget constraints.
Post-hoc analysis: which refused trades would have been profitable? The pattern informs whether the budget is set correctly.

Sleeve-level allocation

Borrow budget can be allocated across sleeves:

High-borrow sleeves (structured financing): smaller borrow allocation.
Moderate-borrow sleeves (discrete offerings, lock-ups): standard allocation.
Low-borrow sleeves (mid-cap ATMs): larger allocation.

The allocation can be expressed in basis points of AUM per sleeve.

Operational mechanics

For systematic implementation:

Daily refresh of per-position borrow rates from broker locate data.
Recompute portfolio-level annualized borrow expense.
Compare to budget ceiling.
Apply budget gate to new-position entries.
Flag existing positions whose rates have moved materially since entry.

The discipline test

The hardest part of borrow budgeting is refusing high-conviction trades because the budget is exhausted. The discipline:

The budget is the binding constraint, not the conviction.
If budget is exhausted, either reduce existing positions or refuse new entries.
Track refused-vs-realized trade outcomes to validate the budget level.