> ## Content Index
> Fetch the complete content index at: https://www.incomestead.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# The Governance Score
- URL: https://www.incomestead.com/blog/governance-score/
- Published: 2026-07-28T07:15:24.000Z
- Updated: 2026-08-16T16:26:30.000Z
- Description: Most portfolio health checks grade what you own. This one grades how you run it — four components, a published formula, and a margin gate that caps the score.
- Author: Stefano Starkel
- Tags: Governance, #stage-2, #tool-governance-score

A **portfolio health check** usually means a look at what you own: holdings, returns, fees, maybe a risk rating. That check will not catch the thing most likely to hurt you. I run a leveraged, income-oriented book myself, and the bad weeks I have had were never caused by a bad holding. They were caused by a loan that crept up, a strategy that quietly doubled its weight, and a review I kept meaning to do. So the health check I actually run measures the operation rather than the inventory — four components, one hundred points, one number. We call it the **Governance Score**.

[Jump straight to the Governance Score ↓](#tool-governance-score)

TL;DR

- A real portfolio health check grades your *process*, not your picks.
- Four components, weights fixed: **Margin 35 · Allocation 30 · Cadence 20 · Policy 15**.
- Each component scores **0–100 on its own**, then is weighted. The result lands in one of five bands, from **Well-Governed** down to **Exposed**.
- **The margin gate:** a book in the Freeze zone is capped at 65 and a book in Forced is capped at 34 — which reads *Exposed* no matter how immaculate everything else is.
- The **band is the headline**; the number is rounded to the nearest 5\. A 73 and a 76 are the same conversation.
- Cadence is the one component you cannot self-assess honestly, so the by-hand version omits it and re-weights the other three. You can still run the whole thing in about twenty minutes.

On this page

1. [What is a portfolio health check?](#what-is-a-portfolio-health-check)
2. [The four things worth measuring](#the-four-things-worth-measuring)
3. [How to score each component by hand](#how-to-score-each-component-by-hand)
4. [Why margin caps everything else](#why-margin-caps-everything-else)
5. [What the number actually means](#what-the-number-actually-means)
6. [A worked example, scored line by line](#a-worked-example-scored-line-by-line)
7. [Run your own check this week](#run-your-own-check-this-week)
8. [Run it: Governance Score](#tool-governance-score)
9. [What the Score will not tell you](#what-the-score-will-not-tell-you)
10. [Frequently asked questions](#frequently-asked-questions)

## What is a portfolio health check?

A portfolio health check is a periodic review asking whether your portfolio is still in the condition you intended. The common version grades holdings — did they go up, are the fees reasonable, is the mix roughly right. That is a check on *results*, and results are the noisiest available signal about whether you are running the thing well. A good quarter can hide a book that is one bad month from a forced sale.

Morningstar's *Mind the Gap 2024* found that the average dollar in US mutual funds and ETFs earned 6.3% a year against the 7.3% those same funds returned over the ten years to 31 December 2023 — a shortfall Morningstar puts at about 1.1 percentage points before those figures are rounded, or, in its own words, investors **"lost out on about 15% of the return their funds generated"**, attributed to the timing of their own buying and selling ([Morningstar, *Mind the Gap 2024*](https://www.morningstar.com/content/cs-assets/v3/assets/blt9415ea4cc4157833/bltebc45c862e642793/6759e563cbd7d6cef415ac94/Mind%5Fthe%5FGap%5F2024.pdf?ref=incomestead.com) — the 2024 edition; Morningstar's landing page now hosts the 2025 report, which covers a different period). That study measures the timing of fund flows rather than governance, so do not stretch it further than it goes. But the direction is suggestive: the gap was not created by which funds people held. It was created by what they did, and when.

Behaviour is at least part of the variable, then — and unlike your holdings it is entirely yours to control. So a health check worth running has to measure behaviour, which means producing a number you can set beside last week's.

> A checkup that grades your holdings tells you how the market has been. A checkup that grades your process tells you how you have been. 

## The four things worth measuring

Four things determine whether a leveraged, multi-strategy book stays governed. Each is scored **0–100 on its own terms**, then weighted into the composite. The weights do not move — a score you can re-weight after seeing it is not a score, it is an opinion.

| Component             | Weight | The question it answers                                          |
| --------------------- | ------ | ---------------------------------------------------------------- |
| Margin Safety         | 35     | Is the loan at a dose you can survive a bad quarter at?          |
| Allocation Discipline | 30     | Is the book still shaped the way you decided it should be?       |
| Cadence & Consistency | 20     | Are you actually looking, on a schedule, or only when worried?   |
| Policy & Structure    | 15     | Are the rules written down, where you cannot quietly amend them? |

Margin carries the most weight because it is the only one of the four that can end the operation outright — the lending view of that is [how a Forced reading caps the Score](https://www.incomestead.com/blog/borrow-against-stocks-without-selling/). Allocation is next because drift is the slow leak that changes what you own without any decision being made. Cadence sits at 20 because a process nobody runs is not a process. Policy takes 15 — smallest, but it is the component the other three are measured *against*.

Split another way: Margin and Allocation are *governable state* — what your book looks like right now. Cadence and Policy are *behaviour* — what you actually do. Behaviour is 35 of the 100 points, which is the whole argument of this site in one number.

## How to score each component by hand

Each component is measured the way its data actually behaves, which is why the four rules do not look alike. One is a curve, one is a deduction sheet, one is a habit, one is a ladder of evidence. Forcing them into one shape would be tidier and less honest.

**Margin Safety (weight 35).** Work out your margin utilization first. Take the account value at the top of your broker's screen — Net Liquidation Value at Interactive Brokers, Account Value at Schwab, portfolio value at Robinhood; every one of them has already subtracted what you borrowed — then **add your margin loan back** to get the gross figure, and divide:

utilization % = loan ÷ (account value + loan) × 100

One caveat before you trust that figure, because it errs in your favour. The identity is exact only when you hold no spare cash and nothing short. Spare cash inflates the denominator, so it makes your utilization read *lower* than it truly is — flattering you on the single number that gates the whole Score. On a book holding 10% cash that is enough to show Harvest where the real reading is Freeze. Short option positions pull the two figures apart as well. Either way, the denominator Incomestead uses is gross securities value — holdings only, cash excluded. If you hold spare cash or shorts, take that figure straight off your statement and divide the loan by it.

That utilization is not scored in steps — it runs down a **continuous line**, so 30.1% and 34.9% are not the same reading. Find your zone, then interpolate between its two endpoints:

| Zone    | Utilization     | Sub-score runs       |
| ------- | --------------- | -------------------- |
| Clear   | 0 to 30%        | 100 → 85             |
| Harvest | 30 to under 35% | 85 → 65              |
| Freeze  | 35 to under 40% | 65 → 35              |
| Forced  | 40% and above   | 35 → 0 (zero at 50%) |

So 38% sits three-fifths of the way through Freeze, and three-fifths of the way from 65 down to 35 is **47**. No margin at all scores 100\. The reasoning behind the zone boundaries, and the action each one asks of you, live in the [weekly margin management guide](https://www.incomestead.com/blog/using-margin-safely/).

**Allocation Discipline (weight 30).** This one starts at 100 and you subtract. A **Layer** is one strategy grouping — the job a slice of the book is doing, not an individual holding — and a **tolerance band** is how far its weight may wander from target before you act, in percentage points. Three deductions, taken in order:

| Deduction     | Rate         | Measured on                                                                     |
| ------------- | ------------ | ------------------------------------------------------------------------------- |
| Drift depth   | −2 per pp    | How far your *worst* Layer sits beyond its band, in percentage points           |
| Drift breadth | −6 per Layer | How many Layers are outside their band at all                                   |
| Concentration | −15 or −35   | Largest single position above 20% of the book (−15), or above 30% (−35 instead) |

The two drift deductions are deliberately different questions: depth asks how bad the worst one is, breadth asks how many are loose. A book can fail either way. Concentration is a universal hazard threshold rather than a personal preference: it scores everyone the same way, whatever you meant to hold. And however many of the three deductions land at once, the component floors at zero rather than going negative. **If you have not written targets down at all, this component cannot be computed** — you cannot be outside a band you never drew — so it folds into Policy below as "no written policy", which is a large and honest penalty. It does *not* quietly score zero and it does not score full marks. The full set of Layers is [the Incomestead Stack](https://www.incomestead.com/blog/strategy-allocation-incomestead-stack/), and the measurement itself is [allocation drift](https://www.incomestead.com/blog/allocation-drift/) — current weight minus target weight.

**Cadence & Consistency (weight 20) — and why you should skip it.** This scores how consistently the book is *actually* governed: run it weekly and the sub-score is 100, near-weekly 85, sporadically 40, only-when-anxious 20, never 0\. Here is the honest part, and it is the reason this section exists at all: **you cannot self-assess this and neither can any calculator you run once.** A weekly habit is not visible from a single sitting, and a self-report is noise — which is why it is worth being precise about [how often to run the weekly reading](https://www.incomestead.com/blog/the-weeks-you-skip/). So the by-hand version — and the free tool — *omit* Cadence rather than let you flatter yourself, and redistribute its weight proportionally across the other three:

Provisional weights — Margin 43.75 · Allocation 37.5 · Policy 18.75

That is what you get without Cadence, and it is called a **Provisional Score** for exactly that reason. It is not a lesser number, it is a narrower one: three of four components, honestly weighted, and it says so.

Cadence is the component only a system that watches every week can measure — which is the one genuine thing the app does that a page of arithmetic cannot. I would rather say that plainly than sell you a self-graded habit score. Every other component here you can compute yourself, and [the other free tools](https://www.incomestead.com/tools/) will do that arithmetic in your browser.

**Policy & Structure (weight 15).** Not a checklist — a ladder of evidence, and where you sit on it is the whole score:

| Evidence                                       | Sub-score |
| ---------------------------------------------- | --------- |
| No target of any kind                          | 0         |
| "I have a target" — claimed, not written down  | 55        |
| A written charter: banded, dated, sums to 100% | 100       |

The gap between the middle rung and the top is 45 points on this component, and it costs an afternoon rather than money. The [one-page written charter](https://www.incomestead.com/blog/personal-investment-policy-statement/) is where those lines live. And note what writing one actually does: it lifts Policy from claimed to verified *and* it gives Allocation something to measure against, so it unlocks two components at once. Your Score does not rise because you used a tool — it rises because a written, banded target is the thing being measured, and before you wrote one there was nothing to measure against. Writing it is only half the job; the other half is [proving you operated the policy, not just wrote one](https://www.incomestead.com/blog/how-to-document-investment-decisions/).

## Why margin caps everything else

Here is the rule that makes this a governance instrument rather than a weighted average: **margin does not just weigh on the Score, it caps it.** A book in the Freeze zone cannot score above **65** however immaculate the other components are. A book in Forced is capped at **34** — the top of the worst band — so it reads *Exposed*, full stop. Existential risk is not something you average away against good behaviour elsewhere.

If that is where you have landed, the Score has already done its only useful job and the rest of this page is secondary. Whether you keep to any of it is [the structural question underneath a governed book](https://www.incomestead.com/blog/whole-life-vs-investing/). The action is to bring the loan down, and the zone-by-zone playbook for doing it — what to sell, what to leave, and in what order — is the [weekly margin management guide](https://www.incomestead.com/blog/using-margin-safely/). Read that before you read the rest of this.

This is the same precedence I built into my own governance engine — the software behind Incomestead — for when two rules fire at once: margin first, allocation second. A tidy allocation is worth nothing on a book that gets liquidated. It is also [why the margin zone gates the score rather than merely feeding it](https://www.incomestead.com/blog/portfolio-dashboard-vs-governance-tool/)

> You cannot earn your way out of a stretched loan with good housekeeping. The gate is there so the number cannot flatter you. 

One clarification, because this is where people misread our zones. Our **Forced** zone begins at 40% utilization, and that is a deliberately conservative house ceiling — not a regulatory edge. Under [FINRA Rule 4210](https://www.finra.org/rules-guidance/rulebooks/finra-rules/4210?ref=incomestead.com), the minimum *maintenance* requirement on long margin securities is 25% equity, which arithmetically means a loan can sit at roughly three-quarters of market value before that rule bites.

Read that carefully, because it is a floor, not an allowance. Under the Federal Reserve's Regulation T, initial borrowing against marginable equities in a standard margin account is capped at 50% of purchase price — so a standard account only arrives at three-quarters by falling a long way with the loan unchanged. Two honest exceptions: portfolio-margin accounts are risk-based and can permit more, and exempt securities such as Treasuries are treated differently. Brokers also impose house maintenance requirements above the 25% floor — commonly 30% to 40% — and can raise them without much notice. The gap between our 40% and any mechanical trigger is the entire point: it is the room in which you get to act on your own terms rather than the broker's.

## What the number actually means

Five bands. Two things to hold on to before you read them: the **band is the headline and the number is secondary**, and the number you display is rounded to the nearest 5\. The band is always read from the real, unrounded score — which is what makes a Forced book capped at 34 read *Exposed* rather than rounding itself up into a friendlier word.

| Band                  | Score  | What it says about the operation                               |
| --------------------- | ------ | -------------------------------------------------------------- |
| Well-Governed         | 90–100 | On-target, deep in Clear, weekly, documented. Keep reading it. |
| Governed              | 75–89  | Solid, with minor gaps you already know about.                 |
| Loosely Governed      | 55–74  | Real gaps — drift, thin cadence, or a Harvest-zone loan.       |
| Ungoverned            | 35–54  | Largely running on memory. Exposed, and fixable.               |
| Exposed / At the Edge | 0–34   | Ungoverned and/or margin-critical. The Forced gate lands here. |

Keep two words apart in your head, because they look alike and are not: a *zone* is your margin utilization — Clear, Harvest, Freeze, Forced — while a *band* is where your Score lands. A book can be in the Clear zone and still land in Ungoverned, if nothing is written down and everything has drifted.

One low reading is also not an instruction. A single week can knock a book down a band and back up by Friday. In the app, a slide only warrants action once it *persists across two consecutive weekly readings* — the same two-week persistence rule the margin zones use. A by-hand score, taken once, has no history to apply that to, which is exactly why the second reading matters more than the first.

## A worked example, scored line by line

A drifting operator. Margin utilization **38%**. Their worst Layer sits **8 percentage points** beyond its tolerance band, **3 Layers** are outside their bands altogether, and their largest single position is **25%** of the book. They say they have a target but it is not written down anywhere.

| Component         | Working                                                | Sub-score  |
| ----------------- | ------------------------------------------------------ | ---------- |
| Margin            | 38% — three-fifths through Freeze, so 65 − (3/5 × 30)  | 47         |
| Allocation        | 100 − (2×8 depth) − (6×3 breadth) − (15 concentration) | 51         |
| Cadence           | Omitted — cannot be measured from one sitting          | —          |
| Policy            | Claimed but unwritten — the middle rung                | 55         |
| Provisional Score | (0.4375 × 47) + (0.375 × 51) + (0.1875 × 55)           | 50         |
| Band              | 50 falls in 35–54                                      | Ungoverned |

Nothing here is a catastrophe on its own. A 38% loan is uncomfortable, not fatal; three loose Layers is a Saturday's work; the target exists, it is just in their head. Together they read **Ungoverned**, and the itemisation tells you where to start: the single largest recoverable block is Policy, because writing the target down lifts Policy from 55 toward 100 *and* makes Allocation measurable against something real.

Now watch the gate do its job. Take a second operator whose allocation is perfect and whose charter is written — 100 on both — but whose utilization is **43%**. Margin scores 24.5\. Their weighted total comes to about **67**, which would read Loosely Governed. But 43% is the Forced zone, so the cap applies and the Score is **34 — Exposed**. Two of the three scored components perfect — more than half the weight of the whole instrument — and one number overrules them. That is not a bug in the arithmetic; it is the arithmetic agreeing with reality about which problem is load-bearing.

## Run your own check this week

Twenty minutes and your own brokerage screen, once your targets are written down — if they are not, budget an afternoon for that first, and score Policy honestly in the meantime. Copy this card and fill it in.

The Provisional Governance Score card

1. **Margin \_\_\_ /100.** utilization = loan ÷ (account value + loan). Then interpolate: Clear 0–30% runs 100→85 · Harvest 30 to under 35% runs 85→65 · Freeze 35 to under 40% runs 65→35 · Forced 40% and above runs 35→0, hitting zero at 50%. No margin = 100.
2. **Allocation \_\_\_ /100.** Start at 100\. Subtract 2 per percentage point your *worst* Layer sits beyond its band; 6 per Layer out of band; 15 if any single position tops 20% of the book, or 35 instead if it tops 30%. Floor of zero. No written targets = not computable; fold it into Policy.
3. **Cadence — skip it.** You cannot measure a weekly habit in one sitting, and guessing corrupts the number.
4. **Policy \_\_\_ /100.** No target at all = 0 · claimed but unwritten = 55 · written, banded, dated and summing to 100% = 100.
5. **Combine:** (0.4375 × Margin) + (0.375 × Allocation) + (0.1875 × Policy).
6. **Apply the gate:** if you are in Freeze, your Score is the lower of that total and 65\. If you are in Forced, it is the lower of that total and 34.
7. **Read the band, not the number:** 90+ Well-Governed · 75–89 Governed · 55–74 Loosely Governed · 35–54 Ungoverned · 34 and below Exposed.

Write today's band down, and the number beside it rounded to the nearest 5\. Next week's reading, set beside this one, is what turns a number into a trend — and the trend is the only part that is actionable.

## What the Score will not tell you

A number invites over-reading, so let me be exact about its limits. The Governance Score does **not** grade your investments. A book of excellent holdings can score badly and a book of mediocre ones can score well, because it measures discipline, not judgment. It does not predict returns and it does not forecast markets — nothing here is a market view. It is not a risk model; it says nothing about volatility, correlation, or what your positions might do next. And it does not tell you whether your chosen risk level is right for you. That is your call, not the instrument's.

It is also deliberately coarse where it counts. The band is the headline and the displayed number is rounded to the nearest 5, because finer resolution would imply a precision the underlying inputs do not support — the difference between a 73 and a 76 is not information, whereas the difference between Governed and Loosely Governed is.

One more thing you are owed. The weights and the deduction rates are **my judgment, not a validated model**. Nobody has backtested 35/30/20/15 against outcomes, and the coefficients — 2 points per percentage point of drift, 6 per loose Layer — are a first calibration that will be tuned against a real distribution of scores before they are called final. I am not claiming a particular score predicts a particular result. The value is in reading the same four things the same way every week so that a change means something. What I will claim is that the formula is *published* and deterministic: same inputs, same score, every point itemised and traceable to a rule you can read. A scoring method you cannot audit is not one to trust with a leveraged balance sheet — including mine.

What it does tell you is whether the operation has slipped, in which of four places, and by how much. That is a narrow claim, and it is the honest one.

## Frequently asked questions

### What should a portfolio health check include?

At minimum: your leverage measured against a stated denominator, your current weights against written targets, evidence you are reviewing on a schedule, and confirmation that the rules exist in writing. Holdings and returns are the part most checks lead with and the part that tells you least about whether the portfolio is being run well.

### How often should I check my portfolio?

Weekly for the reading, rarely for the rules. A weekly cadence is short enough to catch a loan or a weight moving, and long enough that you are not reacting to noise — which is why a breach must persist across two consecutive weekly readings before it drives an action. The policy behind the reading should be reviewed annually or on a real life change, and never in the same week as a breach.

### Is there a free portfolio health check?

The method on this page is free and complete — the scoring card above is the whole instrument, and it runs on your own brokerage numbers with no tool required. That is deliberate: a scoring method you cannot audit is not one you should trust with a leveraged balance sheet. You can [compute it in a spreadsheet](https://www.incomestead.com/blog/portfolio-tracking-spreadsheet-vs-software/) and get the right answer, provided you pin the margin convention once and keep last week's reading — which is exactly where a hand-built version tends to drift.

### What is a good Governance Score?

Governed (75–89) is a realistic target and Well-Governed (90–100) means on-target, deep in Clear, weekly and documented. Two things are worth knowing before you judge your first reading. It will be a *Provisional* Score, because Cadence cannot be measured from one sitting, so it is drawn from three components rather than four. And the fastest points on the board are almost never in your holdings: moving Policy from claimed to written is worth 45 points on that component and it also unlocks Allocation, which is why an afternoon of writing usually moves a Score further than a month of trading.

### Why does margin carry the most weight?

Because it is the only component that can end the operation. Drift costs you shape, a skipped week costs you awareness, unwritten rules cost you enforceability — all recoverable. A loan that outruns the book in a falling market is not, which is why it does not merely weigh the heaviest at 35 points but caps the whole Score: Freeze at 65, Forced at 34.

Next

The card above is the whole method — and the **Governance Score** further up this page does the arithmetic for you. It reads whatever the other free tools have already measured in this browser, names every gap it has, and publishes the formula in full. Score yourself this week, then do it again next week: the second reading is where a number becomes a trend you can act on.

[Assemble your Score →](#tool-governance-score)

Governance is not a feeling about whether you are being careful. It is a number, read on a schedule, that disagrees with you when you are wrong. Work out yours this week — unlike most portfolio health checks, this one grades the part you can actually change. If you want the frame underneath it, start with [what portfolio governance actually means](https://www.incomestead.com/blog/portfolio-governance/).

Incomestead recommends. You decide.