Hyper Multi-Objective Optimisation

Public Policy policy-design state-capacity

Policies and institutions fail when they are laden with several objectives, resulting in a system that fulfils none of them. Just like an engineering system design returns a null solution when strict conflicting objectives are applied at the same time, public policies trying to optimise for several objectives end up failing.

Origin

The framework was introduced in Anticipating the Unintended as an application of multi-objective optimisation from engineering to public policy. In hardware design, a chip must balance speed, size, temperature range, and cost; add too many strict constraints and the design becomes impossible. The same logic applies to policymaking.

What it says

Even before a policy is drafted, it faces five ex-ante design constraints: equity, efficiency, cost, political feasibility, and ease of implementation. The danger arises when governments pile on additional objectives. Because objectives often conflict, trying to optimise for all of them simultaneously produces a null solution — a policy that fails on every dimension.

Two approaches can resolve the trap:

• Augmentation: Create separate agencies or policies, each optimised for fewer objectives. One instrument per target.
• Withdrawal: Accept that some objectives are best handled by markets or society, not by the state. This requires the humility to say what government cannot do.

Applied

India’s tax policy is a textbook case. Dr M. Govinda Rao noted that in addition to revenue and investment incentives, Indian tax policy is loaded with objectives such as backward-area industrialisation, infrastructure encouragement, small-scale promotion, employment generation, charitable activity, scientific research, and SEZ development. The result is an enormously complicated structure that opens avenues for evasion, avoidance, and rent-seeking.

MGNREGA began as a wage-employment safety net. Over time it was asked to also build durable rural assets, deepen democratic participation, generate climate-resilient infrastructure, and serve as a macro stabiliser. The instrument is now optimising for five things at once — and struggling to deliver on its original purpose.

Traffic police are asked to enforce rules and manage congestion. The two objectives often conflict: strict enforcement slows traffic; easing flow requires overlooking violations. Neither objective is met.

When it falls short

Real politics demands multi-objective optimisation. No government can announce that it will ignore equity, or cost, or political feasibility. Withdrawal is harder than augmentation because every objective has a constituency. The framework tells you what to aim for; it does not tell you how to get there politically.

Related frameworks

• One Instrument, One Target — the formal rule that hyper multi-objective optimisation violates.
• Wicked Problems — what happens when objectives are not just numerous but fundamentally contested.
• All Things Governments Do — the menu of tools, many of which should be left unused.

Further reading

• Tinbergen, J. (1952). On the Theory of Economic Policy.

Originally explored in A Framework a Week: Hyper Multi-Objective Optimisation on Anticipating the Unintended.