The battle for the soul of the Conservative party post-election has already begun. This week, commentator and conservative activist Tim Montgomerie will officially launch his new agenda – “The Good Right” – a self-styled “compassionate conservative” call to arms for the Tories to become that bit less libertarian and that bit more willing to embrace state intervention.

Montgomerie believes that the key to achieving a smaller state with a lower tax burden is actually (in many areas) for the state to become more activist. Rather than simply cutting government spending directly (the supply of government), more targeted state action – such as state house-building, higher minimum wages, and greater lifestyle regulation – could reduce demand for it.

It’s an interesting thesis. Yet the historical evidence is not favourable to it. The state grew hugely in both size and scope through most of the twentieth century. Compared on a consistent basis (proportion of GDP at factor cost), state spending rose from 15.1 per cent of GDP in 1900 to 47.4 per cent in 2014, and is higher now than in any year between 1947 and 1979 – that supposed post-war consensus big-government age. Much of this rise in spending was heralded as being targeted towards certain limited aims, before that unhappy mix of electoral calculation, bureaucratic expansionism and lobbying of government led to inevitable mission creep.

Take the welfare state. Beveridge’s system was meant to provide a safety net, explicitly rejecting additional means-tested benefits. Yet over time, supplementary benefits or other means-tested income support became the most significant components of working age welfare.

The justification for state intervention in childcare in the 1990s was that targeted help could assist mothers with weak labour market attachment back into work, alleviating child poverty. Yet now we are moving towards a system of near-universal state funding of pre-primary care. The recent universal free school meals policy for 5-7 year olds was justified as being about giving every child “a good start in life”. There is already a lobby calling for all children aged 11 or under to receive subsidised dinners.

It is nearly impossible to think of a new government intervention that proved to be temporary, and which subsequently reduced spending. The evidence suggests overwhelmingly that state actions live up to Say’s Law: supply creates its own demand. Indeed, the lesson of welfare reform in the nineteenth century was that strong families and civil society institutions fill gaps when the state withdraws. There is no pain-free way of engineering them beforehand.

This does not mean we shouldn’t think carefully about what is driving state spending, of course. In many cases though, a thorough analysis of current problems would conclude that government should interfere less, not more.

The high housing benefit bill, for example, is certainly driven by a lack of supply of homes. But most evidence suggests that land-use planning regulations, particularly green belts around London, Cambridge and Oxford, are the most important factors in explaining the lack of new building in areas where people want to live. A liberalisation of these laws, not more state planning and state building, is what is required. Likewise, in addition to planning, there are a range of areas – from green energy subsidies and childcare regulations, to biofuels mandates and indirect taxes – where government policies substantially raise the cost of living, in turn fuelling demands for a “living wage” as the cost of essentials rises.

Rather than engage in yet more activism to correct existing state-induced problems, a more fruitful principle for governing may be “first do no harm”. Given the degree of state activity in our lives, in many areas this would require less government – or a so-called “libertarian” approach – and the opposite of what Montgomerie prescribes.

This article first appeared in City AM.