About
JR Barbieri, Founder of ParetoAlpha

JR BarbieriFounder

Eighteen years as CFO to multi-generational families. He built his own models because no platform could answer the questions they asked, then built ParetoAlpha so the next CFO does not have to.

18years as CFO to multi-generational families
$100M+in estate structures designed and signed off by counsel and CPAs
$1B+inter-bank dispute resolved on an audit

No family he has worked for is named on this site, and none will be. Yours would not be either.

What he built is on this site: the study (11,842 returns), the seal format, the verifier, and 32 controls checked by machine.

BA, San Francisco State University · MBA, UCAM (Murcia, Spain) · CFP® education program, UCLA Extension · Financial Engineering and Risk Management Specialization, Columbia · Finance & Quantitative Modeling for Analysts Specialization, Wharton

Why it exists

Every foundation and family office keeps a spreadsheet nobody admits to.

It is where the real questions get answered, by hand, and where the reasons go missing when the person who built it leaves. ParetoAlpha keeps the answer, the rows behind it and the decision, sealed, so the next person can check it. One ledger, two doors: a foundation’s filed 990-PF first; a family office’s book on the same ledger.

Three rules

What the product will not do, whoever asks.

Read-only

It sees. It moves nothing alone.

Connections cannot initiate a transfer. The one payment path needs two people.

Row-level

Your foundation, or your family, is the claim.

The database enforces it. A query without it returns nothing.

Human-approved

It stages. A person approves.

Every action staged, shown to a named person, logged.

The name

Pareto, and alpha.

Which few decide the rest. What is left after everything you can name. And a record that keeps both honest.

Read why it is called ParetoAlpha
Pareto, 1896

The few that decide the many.

Vilfredo Pareto, a railway engineer turned economist, found that income in every country he measured followed the same kind of curve, and that one number, α, set how steep it was. The lower α, the more the few decide. In a family’s book the few are a handful of positions, a handful of decisions and a handful of dates.

The system measures α on the family’s own rows, with its interval, and says nothing when there are too few rows to know.

0%0%25%25%50%50%75%75%100%100%Share of holdings, largest firstShare of the totalmedian reviewed office
At α 1.16, the top 20% hold 80% and the top 1% hold 53%. The lower α, the more the few decide; at α ≤ 1 the largest one dominates everything. The dot is real: in the latest quarter the median reviewed family office held 89% of its reported value in 10 of 16 listed names. A Pareto curve passes through it at α ≈ 1.33: one point, drawn, not an estimate.
Pareto, 1906

The frontier rules out. It never chooses.

Pareto’s second idea: an option is dominated when another is at least as good on every count and better on one. What is left is the frontier. It tells you what to stop considering. It cannot tell you which of the rest to pick, because that is a preference, and a preference belongs to the family.

That is the line the product holds. It describes and projects; it does not prescribe. For every decision on the record that moves money, it lays out the frontier across the five lenses: the path taken and the path not taken, which one is ruled out and on which lenses, and a blank wherever a lens has no figure. Among the paths left it ranks nothing. The choice is sealed as the family’s.

0%0%25%25%50%50%75%75%100%100%Cash available within 90 daysIncome a yearABCDEFGHIJ
5 of 10 options are on the frontier. Option H is ruled out: B is at least as good on every question asked and better on one. Click any dot. Illustration: made-up options.
Alpha, 1968

What is left after everything you can name.

In 1968 Michael Jensen defined alpha as what a portfolio earned beyond what its exposure to the market explains. It is a residual: everything you can name is taken out first. It only means something if every term is measured as it was known at the time. Graded with figures restated later, any decision can be made to look like genius or like negligence.

So every figure here carries two dates, the date it is about and the date it was known, and a decision is graded on what was in front of the people who made it.

100110120Q1Q2Q3Q4Q5Q6Q7Q8The quarter the mark is aboutknown at end of Q4
Solid: what the committee could see at the end of Q4, 4 marks. 2 of them have since been restated (ringed). Dashed: the same quarters as restated today, which is all a backtest run today can see. Grading the committee on the dashed line grades it on numbers it never had. Illustration: made-up marks.
For the quant: the definitions
Pareto tail
P(X > x) = (xm/x)α. The top fraction p holds p1−1/α of the total; 80/20 is α = log45 ≈ 1.161. At α ≤ 1 the mean is infinite and the largest observation dominates.
On the family’s book
Hill estimator on the top k = ⌈n/5⌉ positive outcomes: α̂ = k ÷ Σ ln(x(i)/x(k+1)), SE = α̂/√k, 95% interval reported. Refused below 30 outcomes, where the interval is wider than the answer. The reading follows the interval, not the point.
Dominance
a ≻ b ⇔ aj ≥ bj for every objective j, with strict inequality for at least one. The frontier is the set nothing dominates. Adding objectives can only grow it.
Jensen’s alpha
rp − rf = α + β(rm − rf) + ε. Measured honestly only with r(t | known at t), never r(t | known today): the bitemporal record stores both and answers either.
How we work

One foundation or family at a time.

Legal name
ParetoAlpha, Inc.
Product
Written once. Any rewrite shows.
Model
A free sealed proof for any foundation, no account. Published pricing, sold by order form.
Advice
Computes facts from your own data. Not investment, legal or tax advice.