The scoreboard is public.
Asia is the largest economic opportunity in the world — and the largest squandered one. Across the region, the growth and innovation of billions of people is held back less by capability than by government: capital controls, opaque statistics, politicised institutions, policy that protects incumbents over builders.
We believe countries compete for capital the way companies compete for customers — and that this competition, made visible, is the most powerful force for reform.
The policies that work are not a mystery: open capital accounts, honest data, sound money, the rule of law. Pacific Alpha exists to keep the scoreboard public — grading governments against the world distribution, in dated, sourced judgments that cannot be quietly revised, so capital flows to the countries that earn it and the cost of bad policy becomes impossible to hide.
It has been done before. It was broken from the inside.
For nearly two decades the World Bank published a ranking called Doing Business. Governments competed on it — not rhetorically, but by rewriting law. India launched a national reform programme in 2014 aimed squarely at the areas it measured and rose from 142nd to 63rd in five years, on the back of a new insolvency code, tens of thousands of compliance requirements removed and thousands of provisions decriminalised. A number published in Washington was changing what happened in capital cities.
On 16 September 2021 the World Bank discontinued it. An independent investigation had found scores altered in favour of particular countries in two separate editions — China's in the 2018 report, and Saudi Arabia, the United Arab Emirates and Azerbaijan in the 2020 one. The pressure did not come from a rogue analyst. It came from the institution's own most senior leadership.
It did not fail because the idea was wrong. It failed because the instrument could not survive its own influence. The scoreboard became valuable enough that the people publishing it wanted to move it — and the method left enough discretion, applied quietly enough, that it could be moved without anyone outside noticing.
Every rule we operate under is an answer to that. Judgments are dated and append-only — a published grade is never moved, only superseded with the old one still visible beneath it. The scale is the world distribution, not a regional curve. Where a government's own statistics are contested we use independent sources and say that we are doing it. And we score ourselves in public, misses first. The job Doing Business did is vacant. We are rebuilding it so that its own success cannot corrupt it.
Nothing — AI is great at this. We use it heavily.
But ask it what it predicted about Malaysia last June. Every judgment here is dated, scored, and cannot be quietly revised. AI gives you a smart opinion today; it cannot give you an opinion it committed to a year ago and was willing to be wrong about.
That record is the product — and it cannot be back-filled, by us or anyone.
see the dated record →Calibrated probability — working shown, scored in public.
Credit risk has rated authorities — every bond, every covenant, every mandate cites them. Political-economy risk — the variable that decides whether a decade of returns happens at all — has no equivalent, anywhere. Pacific Alpha is building it, starting in the Asia Pacific. Authority of that kind is not claimed; it is accumulated, one dated, scored judgment at a time.
Predicting the path of a political system is a category error; the honest alternative is calibration. The slow variables — institutions, policy execution, monetary regime — are measurable. The phase transitions — wholesale leadership change — are identifiable. And calibration is testable: every event estimate starts from the base rate for its event class, moves only as far as named, dated evidence justifies, and publishes the result as prior → evidence → posterior, on the page. A vivid headline does not move a number unless the base rate says it should.
Headline odds are always decomposed — a 70% election times a 70% policy follow-through is a 49% real outcome, and we publish the multiplication. Once posted, estimates are immutable and publicly Brier-scored; each quarter's predictions are graded the following quarter, misses first. The discipline is mechanical, not aspirational: this site cannot be published if any estimate is missing its recorded prior.
The record is built to be cited — by people and by machines. The full judgment chain is machine-readable at corpus.json, because the next generation of research readers will be AI models answering questions like “how exposed is this portfolio to Taiwan?” — and models cite scored records the way they cite ratings agencies, rather than re-deriving the answer. Prose ages; a scored record compounds.
Brad Good
Pacific Alpha was built by someone who has worked inside Asia's decision-making structures — not just observed them. Brad Good has advised governments and corporations across Asia Pacific as a management consultant at Booz Allen & Hamilton and A.T. Kearney, carrying direct accountability for over $200 million in client P&L outcomes.
Fluent in Mandarin Chinese, Brad has spent two decades studying how Asia's political and economic systems actually function — below the headline level. He trained at Milton Friedman's institution (Chicago Booth) and holds a BA from UC Berkeley. He is also the author of four novels.
Pacific Alpha is the product of that experience: a framework for grading countries the way a credit analyst grades borrowers — systematically, publicly, and with accountability for every call.