Family office AI prompt

Allocation versus selection

Return split into allocation and selection, with a scorecard for each manager.

For
COO, MFO/advisor
Topic
Performance
The prompt
Copy

Evaluate our portfolio’s performance over the past [three] years and work out how much of the return came from asset allocation versus manager selection. Calculate excess return against an appropriate benchmark, tracking error and information ratio for each manager, and tell me which managers added risk-adjusted value. State the benchmarks you used.

Example output · Claude with Masttro connected
Example answer from Claude connected to Masttro MCP: return split into allocation and selection, with a scorecard for each manager

Illustrative output based on Masttro's demo family office data.

What to look for
1
Allocation versus selection

How much of the return came from the asset mix and how much from picking managers.

2
A manager scorecard

Excess return, tracking error and information ratio for each manager.

3
The benchmarks stated

The answer names the benchmarks used, so you can challenge them.

Before you run it

Replace [three] with the number of years to evaluate.

Ask next

Follow-up prompts

Use [MSCI ACWI] and [Bloomberg Global Aggregate] as the benchmarks instead.

Copy

Show the attribution year by year.

Copy

Rank managers by information ratio only.

Copy
Why it's useful

Why split returns into allocation and selection?

Good results can come from being in the right asset classes or from choosing good managers inside them. Families pay very different fees for each, and it helps to know which one is working.

This prompt takes monthly returns by asset class and manager from Masttro and asks the AI to run the attribution and a risk-adjusted scorecard. Benchmarks come from the AI tool, since Masttro doesn't store them for this client, so the prompt asks it to state them.

01
Shows where returns came from

Separates the asset mix from manager choice.

02
Risk-adjusted

Tracking error and information ratio show who added value for the risk taken.

03
Transparent benchmarks

You can swap in your own benchmarks with a follow-up.

FAQ

Questions about this prompt

Where do the benchmarks come from?

From the AI tool. Masttro supplies monthly returns by asset class and manager, and the prompt asks the AI to state the benchmarks it chose.

Can I use my own benchmarks?

Yes. The first follow-up names specific indices to use instead.

What is an information ratio?

Excess return over the benchmark divided by tracking error. Higher means a manager added more return for each unit of active risk.

Can I see attribution year by year?

Yes. A follow-up breaks the result down by year.

Masttro MCP

Run this prompt on your own Masttro data

Connect Claude or ChatGPT to the data you consolidate in Masttro. Read-only, scoped to each person's permissions, and switched on by your Super Admin.

Claude or ChatGPT
You paste the prompt here
Your AI tool
Masttro MCP
Applies your Masttro permissions
Read-only
Your Masttro data
Holdings, transactions and performance across every entity