Separate the Goals from the Optimisation
Let experts choose the route, never the destination
- Difficulty
- Moderate
- Time to result
- ~months to results
- Steps
- 6
- Confidence
- 68%
Appelbaum's answer to 'why listen to economists at all if they cannot forecast?' is to split the job in two. Economics, he argues, is an excellent system for disciplining choice-making: it forces you to model factors, weigh costs against benefits and think systematically about trade-offs. That is real, transferable capability. What it does not confer is any special standing on what the goals of policy should be. The failure of the last fifty years was collapsing the two — treating people who were good at optimisation as though they were also authorities on the destination. The corrective is procedural: society (or the leadership team) decides what it wants; the specialists are then genuinely able to help get there. It also inoculates against a second failure — the tendency, per Galbraith, for what gets called sound analysis to mirror the needs of whoever funded it.
Origin
Extracted from Masters in Business, where Binyamin Appelbaum sets out the argument behind his book The Economists' Hour about the rise and limits of economists in American public policy.
Core principles
- 01Technical skill at reaching an objective says nothing about which objective is right.
- 02Experts earn deference on means, never on ends.
- 03Forecasting ability and analytical ability are different capabilities.
- 04Whoever bears the consequences should set the goal.
- 05A model disciplines choice; it does not choose.
How to run it
- 1
Write the goal down first
Fix the objective before any analysis is commissioned, so the analysis is scoped to serve it rather than to define it.
Pro tip Circulate the goal in writing; verbal goals drift toward whatever is easiest to measure.
Watch out If the goal only appears after the model runs, the model chose it.
- 2
Name who owns the goal
Be explicit that goal-setting sits with the people who bear the consequences, not with the analysts. Appelbaum argues there is no special reason to give economists a greater hearing on that subject than anyone else.
Watch out Prestige and technical fluency quietly convert into unearned authority over ends.
- 3
Commission means, not ends
Ask specialists to lay out routes, costs, benefits and trade-offs against the stated goal. This is where their systematic modelling is genuinely superior to intuition.
Pro tip Ask for at least two viable routes so the choice stays yours.
- 4
Discount the forecasts
Treat predictions about the future as the weakest output of the process. Appelbaum notes that every generation produces prominent economists announcing that recessions have been solved.
Pro tip Ask what the same method predicted five and ten years ago.
Watch out Confidence in a forecast is uncorrelated with its accuracy.
- 5
Audit whose interests the analysis serves
Check funding, incentives and who benefits if the conclusion is accepted. Appelbaum cites a celebrated paper praising the corporation whose footnote discloses corporate funding.
Pro tip Look at the whole pyramid — ask what the analysis implies for the middle and bottom, not only the top.
Watch out Analysts may sincerely believe conclusions that happen to serve their funders; sincerity is not independence.
- 6
Re-open the goal on a schedule
Revisit the objective at set intervals independent of whether the optimisation is going well, so a well-executed wrong goal gets caught.
Watch out Momentum makes an unchallenged goal look like consensus.
In the wild
Under Volcker and Greenspan the Fed's mission narrowed to one objective — control inflation — on the theory that everything else would follow. Economists optimised hard and successfully against that number while financial regulation was treated as unnecessary. The optimisation worked; the goal was too narrow. After 2008 the institution reversed course, accepting responsibility for financial stability and employment as well, and appointed the first non-economist chair in half a century.
→ A broader mandate replaced the single-number goal after the crisis exposed what the goal had excluded.
A leadership team asks its analytics group whether to prioritise margin or retention. The analysts model both and, because margin is easier to measure and improve, the margin case looks stronger — so it becomes the goal by default. Applying the split, the executives instead decide first that retention is the objective the business is judged on, then hand analytics the narrower question of the cheapest route to it. The same team, the same models, a different destination.
→ Analytical capability is preserved while goal-setting stays with the people accountable for it.
Common mistakes
Treating optimisation skill as wisdom
Being rigorous about trade-offs does not make someone right about which trade-off is worth making.
Buying the forecast with the analysis
The systematic cost-benefit work can be sound while the prediction attached to it is worthless; they are sold together and should be separated.
Ignoring the funder
As Galbraith's line has it, what is called sound analysis often mirrors the needs of the affluent. Skipping the funding check imports someone else's goals as if they were neutral.
Is it for you?
Best for
Leaders and citizens who rely on specialists — economists, analysts, consultants — and need their input without surrendering direction.
Not ideal for
Purely technical problems where the objective is genuinely uncontested and measurable.
From the transcript
“Economics is a wonderful system for disciplining choice making. It really helps you to think carefully about costs and benefits.”
“Where I think we go wrong is in privileging the views of economists about what the goals of policy should be. There's no special reason…”
From the episode
Binyamin Appelbaum Discusses Monetary Policy
Binyamin Appelbaum