Effective Independent Provider Count¶
Diagnostic estimator — instantiates Dependency Concentration Control
Collapses a weighted, correlation-adjusted dependency portfolio into a single honest number — how many genuinely independent providers you effectively have, which is usually far fewer than you can name.
Counting providers overstates your resilience twice: it ignores that reliance is uneven, and it ignores that "different" providers can be correlated. Effective Independent Provider Count fixes both in one figure. It weights each provider by its actual share of the exposure (so a vendor carrying 2% doesn't count the same as one carrying 60%) and then discounts for the common-mode correlations an audit has surfaced (so three names sharing one upstream collapse toward counting as one). The result is an effective number — you list eight providers but effectively have 2.3. Its defining move is producing a single, comparable scalar of independence: unlike a heatmap it doesn't show the shape, and unlike a top-share metric it doesn't stop at the biggest one — it rolls the whole distribution, correlations included, into one number you can track over time and set a threshold against.
Example¶
An asset manager holds client assets across what looks like a comfortably diversified set of six custodian banks. Effective Independent Provider Count weights each custodian by assets held — and immediately two of the six turn out to hold 80% between them, so the raw diversity is already thinner than it looks. Then it applies the common-mode discount: three of the six sub-custody through the same global network in one market, so they aren't independent for that market's assets. Rolled together, the effective count lands near ≈2.4 independent custodians, against six named.
That single number does the work a vendor list couldn't: it says the portfolio is roughly as concentrated as if it had two and a half custodians, which is what the risk committee actually needs to compare against its threshold and against last quarter. The estimator is the concentration analogue of the Herfindahl-Hirschman Index, whose reciprocal is read directly as an "effective number" of independent parties.[n1]
How it works¶
- Define the exposure unit. Fix what "reliance" means for this portfolio — assets held, transaction volume, spend, request share — because the count is only meaningful in a consistent unit.
- Weight by share, not by name. Compute each provider's fraction of total exposure, so dominance is captured rather than averaged away.
- Apply the correlation discount. Fold in the common-mode profile so providers that share a root are counted as fractions of one independent unit, not as separate ones.
- Collapse to one figure. Combine the weighted, discounted shares into a single effective-independent-count that is comparable across time and against a threshold.
Its distinguishing trait is that it is the one output that penalizes correlated redundancy numerically — adding a fourth provider that shares everyone else's upstream barely moves the number, which is exactly the truth a headcount hides.
Tuning parameters¶
- Exposure basis — the unit reliance is measured in. A count computed on spend can look very different from one on transaction volume; pick the basis that maps to the harm a failure causes.
- Correlation discount strength — how hard shared-root providers are collapsed toward one. Aggressive discounting tells the harsh truth but leans heavily on the audit's completeness; light discounting flatters the count.
- Weighting curve — how steeply large providers are penalized relative to small ones. A steeper curve makes a single dominant provider crater the number, matching the intuition that one 60% vendor is the whole risk.
- Alert threshold — the effective count below which the portfolio is flagged as too concentrated to accept without action.
When it helps, and when it misleads¶
Its strength is compression into an honest, trackable scalar: one figure that already contains both unevenness and correlation, easy to trend, easy to threshold, and far harder to fool than "we use six banks." It is the number a cap or a multi-sourcing rule can be written against.
Its failure modes are those of any single index. It inherits the audit's blind spots — a correlation the common-mode profile missed is a correlation the count silently treats as independence, so the number can look healthier than reality. The tidy figure also invites false precision over judgement calls (the exposure basis, the discount strength) baked into it, and the classic misuse is choosing the basis and discount that make the count clear the threshold rather than the ones that describe the risk. A scalar also throws away the shape a heatmap would show. The discipline that keeps it honest is to state the basis and discount openly, stress the number against a stricter discount, and read it alongside the map it summarizes rather than instead of it.
How it implements the components¶
Effective Independent Provider Count fills the measurement side of the archetype — it quantifies concentration, it does not detect, display, or govern it:
concentration_measure— it is the concentration statistic: a single effective-independent-count that folds unevenness and correlation into one comparable figure.
It consumes the per-provider provider_exposure_unit and weighted edges from the Weighted Dependency Graph rather than defining them. It does not itself trace the common-mode correlations it discounts by (that's Common-Mode Dependency Audit, which it consumes), render the exposure as a picture (that's Dependency Concentration Heatmap), or act on the number (Concentration Cap Policy, Multi-Sourcing Rule); the estimator supplies the scalar those mechanisms visualize and enforce against.
Related¶
- Instantiates: Dependency Concentration Control — the estimator supplies the single concentration figure the appraisal trends and thresholds.
- Consumes: Common-Mode Dependency Audit supplies the correlation profile that turns a raw count into an effective one.
- Sibling mechanisms: Common-Mode Dependency Audit · Dependency Concentration Heatmap · Top-K Exposure Share · Concentration Cap Policy · Multi-Sourcing Rule · Weighted Dependency Graph
Editorial Notes¶
Form Classification¶
Form family: Analysis, Modeling & Optimization
Rationale: Effective Independent Provider Count operates as a computation, comparison, model, or analytic representation used to infer, estimate, or choose because it collapses a weighted, correlation-adjusted dependency portfolio into a single honest number — how many genuinely independent providers you effectively have, which is usually far fewer than you can name.
Independent corroboration: The frozen evidence defines Effective Independent Provider Count as 'Collapses a weighted, correlation-adjusted dependency portfolio into a single honest number — how many genuinely independent providers you effectively have, which is usually far fewer than you can name', so its operative form is Analysis, Modeling & Optimization.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Economics & Finance
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Portfolio diversification and concentration economics supply the effective-number idea that converts unequal exposure shares into one independence measure.
Related originating lineages:
- Logistics & Supply Chain Management — Supply-chain resilience contributed the provider and common-upstream dependency problem.
- Statistics & Experimental Design — Correlation adjustment supplies the method for discounting providers exposed to the same failure mode.
Review resolution: The current reviewers agree that economics_finance is primary. For the reported differences (reported_ambiguity, alternate_origin_disagreement, encyclopedia_synthesis_disagreement), the evidence supports cross_disciplinary_synthesis, multi_domain, and logistics_supply_chain, statistics_experimental_design; these choices preserve materially formative origins without conflating later domain reach.
Attribution caveat: The exact provider-count artifact synthesizes financial concentration measures with supply-network dependence.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; medium confidence.
Notes¶
The number is only as trustworthy as the correlation input behind it, so it should never be reported without the discount assumption attached: "≈2.4 effective, under the audit's current common-mode profile." Detached from that caveat, a reassuring scalar can outlive the audit that justified it and quietly certify an independence the supply chain no longer has.
[n1] The Herfindahl-Hirschman Index sums the squared shares of participants in a market; its reciprocal, 1/HHI, is interpreted as the effective number of equally-sized participants — so an HHI implying "2.4 effective competitors" is the same maths this estimator applies to dependency shares. The correlation discount here is an extension beyond the classic index, which assumes participants are independent. ↩