I spend my working life inside global supply chains — tiers, tonnes, audits, ports. In 2027 I'm going back to first principles: the economics, the science and the politics of environmental governance, taught the way the Kennedy School teaches it. Then I'm handing the whole syllabus to whoever wants it.
ENRG is one of the policy areas at Harvard Kennedy School — the school that trains people to make public decisions rather than to study them from a distance. It is not an environmental science degree and not an activism programme. It is a training in how collective choices about scarce, shared things actually get made, financed, enforced and undone: emissions, energy, water, land, forests, minerals, and the institutions that arbitrate between them.
The area draws faculty from economics, earth science, law and politics, and connects to the research communities around it — the Belfer Center's environment and natural resources work, the Salata Institute for Climate and Sustainability, and the university-wide Center for the Environment. Students arrive as engineers, diplomats, regulators, operators; they leave able to build an argument that survives a budget committee.
Three habits define the teaching, and they are the three I'm stealing for 2027. Quantify honestly — every claim carries a number and an uncertainty. Name the counterfactual — compared to what, and who is worse off. Assume politics — a policy that cannot be implemented is not a policy, it's a preference.
Everything below is self-study, assembled from public Harvard material and the open standards my industry already runs on. No enrolment, no credential. The credential is that the work gets better.
“A supply chain is a governance system that nobody voted for. Fourteen thousand suppliers, and the only constitution is a purchase order.”
Why I'm studying this — the note I wrote on the first page
Foundations: why sensible people produce unsustainable outcomes.
Externalities, public goods, cap-and-trade versus carbon taxes, abatement cost curves, discounting the future. The single most useful mental model for anyone who negotiates with suppliers about cost.
Carbon budgets, feedbacks, the difference between a scenario and a forecast, adaptation versus mitigation. Read the Summary for Policymakers slowly, twice, and learn to spot the sentences that were negotiated.
The Kennedy School's core discipline: define the problem before the solution, state the counterfactual, present three real options with their losers named, and fit it on two pages.
Governance: how a rule travels from a chamber to a factory floor.
Why good policy dies: incumbent industries, distributional pain, weak enforcement capacity, and the honest politics of a just transition. Case-based, always with named actors.
Carbon border adjustment, leakage, competitiveness, and what happens to a developing-country exporter when the rules of a distant market change. The most consequential subject in my job and the least understood in my building.
Corporate sustainability reporting and due diligence regimes, forced-labour import controls, grievance mechanisms, and the difference between an audit and an actual remedy for a worker.
The chain itself: numbers I could defend under oath.
The GHG Protocol from the inside: category boundaries, spend-based versus primary data, allocation choices, double counting, and how a target becomes credible rather than decorative.
Beyond carbon — deforestation-free sourcing, water stress in basins we buy from, biodiversity and the commons problems that Ostrom's work explained long before anyone called it ESG.
Decarbonisation finance at tier two: green premiums, long-term offtake, blended finance, contract clauses that actually move capital, and the arithmetic of a supplier with a nine-percent margin.
Machines, judgement, and the part only a person can do.
Remote sensing for land-use change, document intelligence across customs and mill certificates, anomaly detection in supplier data — and a careful catalogue of where these models are wrong, confidently and expensively.
Risk frameworks, the EU AI Act's obligations and timelines, data provenance and consent, model documentation — plus the uncomfortable ledger: the energy, water and land cost of the intelligence I'm using to cut energy, water and land costs.
The Kennedy School's oldest lesson: distinguish the technical problem (someone knows the answer) from the adaptive one (people must change what they value). Then close the year by teaching all eleven modules to someone else.
Not exam questions. The ones that stay open — about machines, judgement, and what a capable human still has to be in 2030.
What is the smallest number I could defend, at the same table, to a regulator, a supplier and a scientist?
If a model can produce the answer, what exactly is the human still for?Answer it concretely, for one task, this week.
Which of my judgements are really pattern-matching — and which are value choices I must never delegate?
Who bears the cost when a prediction about my supply chain is wrong, and do they have any way to appeal it?
What does a factory floor know that no model has ever been trained on — and are we preserving that knowledge or quietly deleting it?
If AI drives the cost of measurement toward zero, what becomes worth doing that we once called immaterial?
What is the carbon, water and land footprint of the intelligence I am using to reduce footprints?
Where does trust actually live across fourteen thousand suppliers — in the data, in the contract, or in the relationship?
Which single incentive would I have to change for the right thing to happen when nobody is watching?
What am I optimising that I never chose?Every dashboard is an argument someone made before you arrived.
What must a 2030 colleague be able to do that no system will do for them — ask a better question, hold a room, say no?
If I am still doing this work in 2030, what will I wish I had started learning in 2027?