A minimum wage is a good answer to one cause of poverty and no answer at all to the others, which is what settles the question.
The case for it is that it acts on low wages — one of the principal causes of poverty at 5.2.2, and the one most other policies neglect. Support for the unemployed does nothing for a household in poverty where someone works full time, and that describes a large proportion of poor households in most countries. A minimum wage raises their income immediately, without waiting for the years that education or economic growth require. It also costs the government almost nothing, since the payment is made by employers rather than from tax revenue, which matters greatly in a country with limited fiscal capacity. And it widens the gap between what a job pays and what benefits pay, so it strengthens the incentive to work rather than weakening it as more generous benefits might.
The case against it begins with employment. Setting a wage above the level the market would produce means some workers cost their employer more than the value of what they produce, so firms employ fewer of them — and those most likely to lose their jobs are the least skilled and least experienced, the very people the policy was intended to help. Being unemployed at a higher minimum wage is worse than being employed at a lower one. The policy also reaches only those in work: the unemployed, the long-term sick and the elderly without pensions are untouched by it, and they are among the poorest. Firms may pass the higher wage cost into prices, which raises the cost of living and erodes the real gain. And in a country with a large informal economy the law cannot be enforced, so the poorest workers — who are precisely those working informally — receive nothing.
My judgement is that a minimum wage is an effective policy against in-work poverty and cannot be the most effective policy against poverty as a whole, because most poverty is not in-work poverty. The syllabus lists six policies — promoting growth, improved education, improved healthcare, more generous benefits, progressive taxation and a minimum wage — and they address different causes. Asking which is most effective in the abstract is like asking which tool is the best; the answer depends on what is broken.
For a household poor because wages are too low, the minimum wage is the direct remedy. For one poor because of unemployment, benefits and policies to create jobs are needed. For one poor through illness, healthcare provision matters most. For one poor because of age, pensions do. A government that introduced a minimum wage and nothing else would leave most poor households exactly as they were.
The level at which it is set also determines whether the main objection applies. Set modestly, close to what the market already pays the lowest earners, it raises incomes with little effect on employment. Set far above it, the employment loss can outweigh the gain. So it is not a policy that is simply good or bad, but one whose value depends on its calibration.
On balance the most effective single policy is improved education, because it addresses the cause of low wages rather than the symptom and its effects are permanent and pass to the next generation. But it takes a decade, which is why a minimum wage belongs alongside it rather than instead of it: one raises incomes now and the other removes the reason they were low.