A Funny Thing Happened to the KPI

Tax departments are unusually difficult to measure because reasonable people disagree about what success means — and because measurement is not neutral. It changes behavior. A department judged mainly on cash tax reduction will find ways to reduce cash taxes. A department judged on filing timeliness will optimize the calendar. A department judged on “no audit adjustments” may become so conservative that it leaves value on the table, or stops documenting judgment because documentation creates visible disagreement. Choose the metric carelessly and you will get exactly what you asked for. That is the problem.

Enterprise Tax Decisions & Outputs & KPIs & Objectives → Mandate → Controls → Outcomes → Learning ▲ │ └──── Feedback loop: metrics should change behavior, not merely describe history ────┘

Exhibit 1. A KPI is useful only when the organization has agreed what the tax function is trying to accomplish.

Executive summary

• Choose the mandate before the metrics. A KPI is useful only when the organization has agreed what the tax function is trying to accomplish. Define the mandate in four layers — execute, protect, advise, build — and measure each one on its own terms.
• Tax is an optimization problem, not a minimization problem. The objective is maximum after-tax enterprise or family outcomes subject to risk, liquidity, legal, reporting, and operational constraints. Sometimes the right answer is paying more tax.
• The KPI is now literally tradable — and somebody already hedged one. Cboe has filed to list binary options on corporate KPIs (23 companies, more than 100 metrics), and a third party has already placed $3.0 million of prediction-market trades that mirror a football coach’s bonus schedule. Goodhart’s law acquired a Bloomberg terminal. A KPI is a proxy for a good state of the world; it is not the good state itself.
• Never hedge the cost of success (Principle 10). If the tax team creates $20 million of durable value and the bonus pool rises by $500,000, the $500,000 is the invoice for getting what management asked for — 2.5 cents on the dollar. Score the net outcome; do not neutralize the win.
• A balanced scorecard beats a headline number. Five dimensions — reporting and execution, risk and control, planning and value, business partnership, capability and resilience — with leading indicators alongside lagging outcomes. No single dimension is “the KPI.”
• Treat the familiar metrics with suspicion. Effective tax rate, cash taxes paid, “savings generated,” returns filed on time, audit-adjustment counts, department cost: each is easy to like and easy to game. Every one needs a companion metric.
• Keep the dashboard small. Eight to twelve executive measures with trend, threshold, and commentary. Twenty-five KPIs become wallpaper.

01 One Function, Several Businesses

Alternative asset managers and family offices sit at a difficult intersection. They are businesses, investors, fiduciaries, employers, counterparties, and taxpayers — and, in many cases, reporting hubs for other taxpayers. A single economic event can touch the management company, multiple funds, portfolio companies, founders, trusts, co-invest vehicles, foreign investors, and a dozen jurisdictions before lunch.

That creates three overlapping tax systems inside one organization. First, the tax of the operating enterprise: management-company income, compensation, state and international footprint, employment taxes, financial reporting. Second, product and investor tax: allocations, withholding, UBTI and ECI, blockers, K-1 and K-3 reporting, after-tax investor outcomes. Third, owner and family tax: carry, basis, estimated payments, estate and gift planning, residency, trusts, philanthropy. The reporting burden is itself a strategic fact — Schedules K-2 and K-3 alone extend partnership reporting into partner-specific international detail that feeds other people’s returns — so the modern tax function depends on data lineage and workflow design as much as on statutory interpretation. Tax authorities are moving the same way; the OECD’s tax-control-framework work expects documented strategy, assigned responsibility, testing, and assurance, and complexity that once lived in spreadsheets is increasingly visible to administrations through structured data.

Exhibit 2. Same tax function, different businesses. The scorecard has to know which one it is measuring.

02 The Responsibility Map: What Tax Actually Owns

A good KPI system begins with a responsibility map. If the organization expects tax to “own” every tax-related outcome while tax has no authority over source data, legal structures, investment timing, treasury, or investor communications, the metrics will be unfair and unhelpful. The reverse failure is just as common: tax cannot claim to be strategic if it enters important decisions after the economics have hardened. We define the mandate in four layers.

Exhibit 3. The four-layer mandate. Each layer gets its own metrics; none substitutes for another.

The distinction that makes this map useful: activity is not impact. Completing 1,000 returns is workload. Filing them accurately and on time is an execution outcome. Detecting a structural error before it reaches 1,000 investor statements is a control outcome. Redesigning the structure so the error cannot recur is capability building — and it is the layer most scorecards never see.

03 Optimization, Not Minimization

The conventional phrase “tax efficiency” is useful but incomplete. For an alternative manager or family office, the right objective is to maximize after-tax enterprise or family outcomes subject to risk, liquidity, legal, regulatory, reporting, reputational, and operational constraints. Sometimes that means paying more tax.

Run the numbers on a live example. A product structure saves 40 basis points of annual tax leakage. On a $2.0 billion vehicle, that is $8.0 million a year — real money. Now price the other side: the structure delays investor reporting by eight weeks, adds (say) $1.2 million of recurring blocker administration, creates foreign filing obligations, and makes redemptions harder to model. If the all-in friction is $1.5 million, the trade still clears — but it is a $6.5 million decision, not an $8.0 million victory, and whether even that holds depends on the product, the investor base, the expected holding period, and the operating capacity to run it. The tax department’s job is to quantify the trade, not to declare victory because a nominal rate fell. The family-office version is identical in structure: an estate-planning transaction can be technically elegant and still be a poor decision if it creates governance friction, demands impractical liquidity, or leaves the next generation unable to administer what it inherited. Stewardship is the objective; tax minimization is one input.

THE FIVE-NUMBER RULE

Every major tax recommendation should arrive with five numbers: expected benefit, downside exposure, timing, recurring operating cost, and sensitivity to the key assumption. It should also name the decision owner and the deadline. A technically correct answer without an owner is an expensive observation.

04 Goodhart Gets a Ticker

Goodhart’s law used to be an abstract warning from the social sciences: once a measure becomes a target, people start optimizing the measure. In 2026 the warning acquired a Bloomberg terminal. Cboe has filed with the SEC to list “binary KPI options” covering 23 companies and more than 100 metrics (SpaceX revenue, Nvidia data-center sales, Apple iPhone shipments, JPMorgan’s credit-loss provisions). Each contract settles at 0 or 1 depending on whether the reported number clears its threshold. “Prediction markets are demanding real-time signals, and we are bringing that activity to time-tested venues,” the exchange said. Read that as a management story rather than a market-structure story: the proxy a board uses to measure the business can now be bought and sold on its own, and the conceptual problem becomes impossible to ignore.

Matt Levine’s August 31 Money Stuff column supplies the comedy, via reporting from InGame. A third party that assists sports teams with hedging coach-bonus risk placed five block trades on Kalshi, worth a combined $3.0 million, on LSU’s football season: make the College Football Playoff, reach the quarterfinals, the semifinals, the title game, win it all. Individual blocks ran $300,000 to $900,000 — and the $3.0 million total payout if LSU wins the championship exactly equals coach Lane Kiffin’s championship bonus, with each intermediate payout sitting close to his round-by-round schedule. The schools did not place the trades; the exchange’s own rules would bar team employees from betting on their own games. Someone engineered this carefully.

Levine walks the logic through a car company: promise the CEO a $20 million bonus for delivering a million cars, then buy a $20 million contract on your own deliveries, and congratulations — you are now financially indifferent to hitting your own target. The reason the bonus exists is that the company expects to be better off in the state of the world where the cars get sold, even after paying it. “The executive’s bonus is, itself, the hedge.” It already costs more when things go well and less when they do not. Paying a prediction market to reverse that is a technically neat hedge of one cash flow and a rather odd hedge of the enterprise. (The column also asks whether a team trading its own results through an intermediary is just insider trading with extra steps — a genuinely interesting question we will happily leave to the prediction-market bar.)

Now the tax version. Suppose your tax function creates $20 million of durable, risk-adjusted value — a sustainable rate reduction, positions that hold up on audit, credits that survive exam — and the bonus pool rises by $500,000. Here is the whole picture, by state of the world:

Exhibit 4A. One outcome, two states of the world.

$20.0M ($0.5M) $19.5M ██████ ██████ ██████ ▂▂▂▂▂▂ ██████ ██████ ██████ ██████ The bonus is 2.5% of the value ██████ ██████ created — the invoice for winning, ██████ ██████ not a loss to hedge ██████ Durable, Success-linked Net to the risk-adjusted compensation enterprise tax value

Exhibit 4B. The same three numbers, drawn to scale. The gold sliver is the hedge candidate.

Which column would you rather live in? The $500,000 is the invoice for getting exactly what management asked for. If the CFO’s first instinct is to hedge it, either the KPI is wrong or finance has discovered a very sophisticated way to be disappointed by success. The lesson runs past compensation: every KPI embeds a theory of what the organization wants, and a tax department can hit the number and harm the business, or miss the number and improve it. A KPI is a claim about the objective. It is not the objective.

05 Ten Principles for Choosing Tax KPIs

1. Measure outcomes and drivers. Filing accuracy is an outcome; source-data completeness and review exceptions are drivers. Both belong on the dashboard, because the drivers explain next quarter before it happens.
2. Normalize for complexity. A team supporting 300 simple domestic entities should not be compared mechanically with one supporting 80 entities full of tiered partnerships, cross-border investors, and bespoke allocations.
3. Separate controllable from shared outcomes. K-1 timing may depend on portfolio-company K-1s. Measure tax’s cycle time after complete data arrives, as well as total investor delivery time; the gap between them is a data-supply problem, not a tax problem.
4. Weight severity over frequency. Ten immaterial amended returns can matter less than one unnoticed withholding failure affecting a major investor or jurisdiction.
5. Reward early involvement. If tax is first invited after signing, a “planning miss” is an enterprise-process failure. Track when tax entered the decision, not merely what it said once it arrived.
6. Pair savings with durability and risk. A saving that reverses next year, creates a large uncertain position, or consumes the department’s capacity should not score like durable, low-friction value.
7. Let good risk metrics get worse first. A new issue register or control-testing program will surface more exceptions. That is better visibility, not worse performance; the KPI should track whether severe issues are found earlier, aged properly, and prevented from recurring.
8. Name the baseline. Every “savings” claim embeds a counterfactual. State it. An aggressive baseline can manufacture value the same way a tax-naive benchmark manufactures alpha.
9. Keep the dashboard small. Eight to twelve executive measures with trend, threshold, and commentary. Twenty-five KPIs become wallpaper; nobody reads wallpaper.
10. Never confuse the cost of success with a risk to be eliminated. If compensation, vendor fees, implementation costs, or reporting burden rise because the tax function produced a genuinely better enterprise outcome, score the net outcome. The unit of analysis is the whole state of the world, not the one line item that got more expensive because things went well. This is the same arithmetic as risk-adjusted tax value (expected benefit, minus recurring cost, minus probability-weighted downside); the Levine example just gives it a memorable corollary.

THE OPERATING RULE

Before hedging any cost, ask which state of the world produces it. A cost that arrives only when the organization wins is a feature of winning. Budget for it; celebrate it; do not neutralize it. And do not hedge yourself back to indifference between winning and losing.

06 The KPI Architecture: What Belongs on the CFO Dashboard

We would not begin with universal targets. We would begin with the architecture below, then set thresholds from the firm’s complexity, risk appetite, investor promises, staffing model, and maturity.

Exhibit 5A. Five dimensions, one scorecard. No single dimension is “the KPI” — strong performance is the portfolio of outcomes.

Exhibit 5B. Two executive measures per dimension; everything else is drill-down.

07 Metrics We Would Treat with Suspicion

Every metric below is popular for a reason, and every one can mislead for a reason. The effective tax rate deserves special caution in private capital: it is often the wrong primary metric for pass-through structures, and even for management companies or blockers it moves for reasons unrelated to the tax team’s quality. Academic work treats tax risk as distinct from the level of tax avoidance itself; volatility and governance matter independently of the headline rate. The board’s question is not “how low is the rate?” It is “how durable, explainable, and governable is the result?”

Exhibit 6. Seven familiar metrics and one brand-new green light — the one that should make you nervous.

08 Managers: Follow the Investor and Deal Lifecycle

For an alternative asset manager, the tax department is part of the product. Investors experience tax through withholding, tax distributions, K-1 and K-3 timing, state schedules, UBTI and ECI, blocker leakage, and the character of returns. A manager can post excellent pre-tax performance and still create avoidable after-tax friction. Four manager-specific KPI families close that gap.

• Investor reporting reliability: committed-versus-actual delivery dates, material correction rate, K-3 request fulfillment, state-schedule completeness, and repeat investor inquiries by issue type.
• Structural tax leakage: basis points of recurring taxes and unrecoverable withholding inside blockers, feeders, and portfolio structures — shown by strategy and investor archetype, never as one blended number.
• Deal readiness: percentage of material acquisitions, financings, restructurings, continuation vehicles, and exits where tax diligence and modeling occurred before investment committee or signing; open tax conditions tracked to close.
• Allocation and waterfall assurance: reconciliation of tax allocations to governing documents and economics, model change controls, exception testing, and the count of material post-close allocation surprises.

For private funds we also like an investor tax friction index: a small composite of reporting delay, leakage, withholding complexity, state burden, and recurring investor exceptions. Its purpose is not to rank products mechanically. It exists to force the product team to admit that two economically similar wrappers can create very different investor experiences.

09 Family Offices: Stewardship, Continuity, Decision Quality

The family office needs a different center of gravity. It files fewer investor statements but faces more concentrated, personal, path-dependent decisions; a missed election, an unreconciled basis schedule, or a trust-administration failure can affect one family for decades. Four additional dimensions carry most of the weight.

• Family tax liquidity forecast: variance between forecast and actual federal, state, foreign, trust, and entity-level tax cash needs, with explanations for material differences.
• Basis and attribute integrity: percentage of material entities and assets with current, supportable basis, holding period, suspended-loss, credit, and carryforward records that reconcile across entities and advisers.
• Estate and succession readiness: percentage of priority planning decisions with quantified tax outcomes, governance implications, valuation support, liquidity analysis, implementation owners, and annual review dates.
• Adviser orchestration: aged open items across accountants, lawyers, trustees, investment managers, and administrators. Duplicate work and unresolved responsibility are family-office risks in their own right.

One mistake deserves its own sentence: complexity is not sophistication. A structure that requires continual heroic coordination among six advisers may be technically impressive and operationally weak. The right KPI question is not “how many structures do we have?” It is “can the family understand, govern, fund, and defend the structures it has?”

10 Implementation Without a Second Compliance System

A KPI program can become its own bureaucracy. Refuse. Most metrics should come from systems the team already needs — filing calendars, issue trackers, provision models, investor-service logs, workflow tools, transaction pipelines, control testing. If a metric requires a monthly manual scavenger hunt, it will not survive its second quarter. Three layers: a CFO dashboard of eight to twelve measures with trend, threshold, and commentary; a deeper operating dashboard for the tax leader; detailed diagnostics for process owners that never masquerade as executive KPIs.

Exhibit 7. Four cadences, four audiences, four questions. The annual one matters most — tax risk appetite should not be invented by the tax department in isolation.

11 A Practical Starting Scorecard

Below is an illustrative 100-point framework. The weights should move with the organization — a fund complex in heavy K-1 season weights reporting more; a family office entering a generational transition shifts weight toward planning, liquidity, and governance. Debating the weights is half the value, because the debate exposes what management actually expects from tax.

Exhibit 8A. A 100-point starting frame. A 92 should not automatically beat an 89 — thresholds exist to trigger questions, not to avoid them.

Reporting & execution ████████████████████████████████████████████████████ 25 Risk & control ████████████████████████████████████████████████████ 25 Planning & value ██████████████████████████████████████████ 20 Business partnership ███████████████████████████████ 15 Capability & resilience ███████████████████████████████ 15

Exhibit 8B. The weights, drawn. Half the scorecard is execution and control — the license to operate; the other half is why the function exists.

12 Measuring What Matters: “No Surprises,” Defined Correctly

CFOs often say they want no tax surprises. Sensible instinct; incomplete metric. A tax department cannot eliminate legislative change, market volatility, or audit risk. It can make surprises less surprising: identify material uncertainty early, quantify the range, assign the owner, document the evidence, and make the decision deadline visible. It tells the deal team when a structure changes after-tax return; it tells the founder when a liquidity assumption is wrong; it tells the CFO when the forecast is losing reliability. And it builds systems so the insight does not depend forever on one person remembering one spreadsheet.

That is a far more demanding standard than “the returns were filed,” and a far more useful one. The goal of tax KPIs is not to prove the department is busy, inexpensive, conservative, or aggressive. The goal is to show whether the tax function makes the enterprise more informed, more resilient, and better after tax.

And one final test for any scorecard: if hitting the KPI creates a state of the world management genuinely prefers, do not design the measurement system so cleverly that the organization becomes indifferent to getting there. The point of a tax scorecard is not to hedge the good outcome. It is to recognize it.

Selected sources and further reading

Matt Levine, “Money Stuff: Don’t Hedge the KPIs,” Bloomberg Opinion (Aug. 31, 2026), discussing Cboe’s proposed KPI binary options and reporting (via InGame) five Kalshi block trades totaling $3.0 million, placed by a third party assisting with coach-bonus hedging, that mirror LSU coach Lane Kiffin’s bonus schedule. The same issue’s link roundup includes “Why ‘Tax Alpha’ Is Silicon Valley’s New Obsession” — the subject of our August 29 client alert.

Cboe Global Markets, proposed Rule 4.80 (binary KPI options), File No. SR-CBOE-2026-061; Bloomberg, “Cboe Seeks to List Prediction Market Type Options on Earnings Metrics” (July 1, 2026).

Robert S. Kaplan & David P. Norton, “The Balanced Scorecard — Measures That Drive Performance,” Harvard Business Review (Jan.–Feb. 1992).

Internal Revenue Service, Partnership Instructions for Schedules K-2 and K-3 (Form 1065) (2025).

OECD, Co-operative Tax Compliance: Building Better Tax Control Frameworks (2016); OECD, Tax Administration 2025, ch. 6 (compliance management); OECD, Co-operative Compliance: A Framework (2013).

Deloitte, “Protecting Legacy — The Value of a Family Office: Management Considerations” (2024); PwC, Guide to Tax and Wealth Planning 2026 (family-office governance); EY, “The Key Questions Every Tax Leader Should Ask Themselves” (Sept. 2025).

Jihwan Choi & Hyungju Park, “Tax Avoidance, Tax Risk, and Corporate Governance: Evidence from Korea,” Sustainability 14:469 (2022), on tax risk as distinct from the level of tax avoidance.

Internal La Presa development materials reviewed for this piece include prior work on the modern alternative-asset tax director, tax operations and AI, after-tax investor outcomes, structural tax leakage, and tax-aware decision metrics.

Download the alert (PDF, 504 KB)

This alert is for general information only and is not tax, legal, investment, or accounting advice. Consult your own advisers before acting. Carlos A. Schmidt, MBT, MBA, CPA · Managing Member · carlos@lapresallc.com · (917) 558-6393.