Reducing union power is a standard labour market reform, and the case for it is a genuine economic argument rather than simple hostility to unions.
The case for it is that a union raises wages for its members by restricting the supply of labour or by threatening to withdraw it, and a wage held above the level that would clear the market means firms employ fewer workers than they otherwise would. Those excluded are the outsiders — the young, the inexperienced, the unemployed — who would have accepted the lower wage and are now not employed at all, so the union has raised the pay of those in work partly at the expense of those out of it. Unions can also resist changes in working practices, so firms cannot introduce new technology or reorganise production, which holds productivity down. Where wages are bargained nationally rather than plant by plant, a wage appropriate to a prosperous region becomes unaffordable in a poorer one, so firms do not open there and regional unemployment persists.
The case against it is that the assumption underneath the argument — that without unions wages would be set by a competitive market — is frequently false. In a town with one large employer the firm has substantial power over the wage, and a union is what restores the balance rather than what distorts it. Removing it in that situation lowers wages below the value of what workers produce and transfers income to the employer without raising employment. Unions also serve functions that flexibility arguments ignore: they enforce safety rules, provide training, and give workers a channel to raise problems, which reduces staff turnover and the cost of replacing people. And a workforce with no security may be a workforce that will not accept change, because every reorganisation now threatens a job with nothing to fall back on.
My judgement is that this should not be decided at the level of the economy at all, because the right answer differs between labour markets within it. Where many employers compete for workers, a union pushing wages above the competitive level does reduce employment and the case for limiting its power is sound. Where one employer dominates, the union is offsetting the employer's power, and weakening it lowers wages without creating a single job. A national law does the same thing in both cases, and so gets one of them wrong.
The second consideration is which powers are reduced. The power to strike without notice, to prevent the introduction of new equipment, or to compel membership is difficult to defend on economic grounds. The power to bargain collectively over pay, to represent a member facing dismissal, or to insist on safety standards is doing work that would otherwise have to be done by regulation — so removing it does not eliminate the cost, it moves it to the state.
On balance a government should target the specific practices that block productivity improvement rather than union strength as such, and should be most cautious where employers are concentrated. Treating the question as one of how much power unions should have in general guarantees a policy that is wrong somewhere.