Rebalancing Threshold Design

Rebalancing is often described with a simple rule: “Rebalance at 5 percentage points.”

But behind that simple rule lies a more fundamental question:

What exactly are we trying to keep stable?

Thresholds determine:

• how often you trade
• how much your risk profile may drift
• how large structural deviations you tolerate

To understand threshold design, we need to define what is being measured.

---

1. What does a threshold measure?

Let:

• w_target = target allocation (e.g., 40%)
• w_actual = current weight
• Δ = w_actual − w_target

Rebalancing is triggered when the deviation becomes “too large”.

The question is how we define “too large”.

---

2. Absolute thresholds (percentage points)

Rule:

Rebalance if |Δ| > c

where c is a fixed number of percentage points.

Example:

• Target: 40%
• c = 5pp
• Rebalance below 35% or above 45%

Strength

• Directly controls structural drift
• Simple and transparent

Weakness

Works poorly for small allocations.

If target = 7.5% and c = 5:

• Band: 2.5% – 12.5%
• That is an enormous proportional deviation.

---

3. Relative thresholds (percent of target weight)

Rule:

Rebalance if |Δ| > r × w_target

Example:

• r = 25%
• Target 40% → threshold 10pp
• Target 8% → threshold 2pp

Strength

• Automatically scales with position size

Weakness

• Small allocations get very tight bands
• Large allocations get very wide bands

---

4. Hybrid thresholds

Example:

Rebalance if |Δ| > max(5pp, 25% × w_target)

This ensures:

• Large allocations always have a minimum structural band
• Small allocations scale proportionally
• Extreme distortions are avoided

---

5. Visual comparison

The chart below compares the absolute rule (±5pp), the relative rule (±25%), and the hybrid rule.

The hybrid rule applies the absolute threshold for small allocations and transitions to the relative rule beyond 20%.

---

6. Threshold size and turnover

Tighter bands:

• Reduce structural drift
• Increase trading frequency

Wider bands:

• Reduce turnover
• Allow larger temporary risk shifts

Threshold design is therefore a trade-off between:

• Risk control
• Transaction costs
• Tax frictions

---

Conclusion

The debate about “percent vs percentage points” is ultimately about:

What do we want to keep invariant?

Absolute rules stabilize structure.
Relative rules stabilize proportional deviation.
Hybrid rules balance both.