It sounded like radical shock therapy. Last autumn, Donald Trump pledged to create a Department of Government Efficiency (Doge) to “dismantle government bureaucracy, slash excess regulations, cut wasteful expenditures and restructure federal agencies”. Elon Musk and Vivek Ramaswamy were announced as Doge’s leaders, with the Argentinian president Javier Milei’s “¡Afuera!” approach to government regulation and headcount the inspiration. The talk was of carving out trillions from US government spending over a decade.

Fast-forward to inauguration week and critics think Doge has been neutered before it’s even begun. First came the news that Ramaswamy wouldn’t be involved after all, prompting jokes that Doge’s first efficiency gain was getting rid of him. Then the executive order creating Doge unveiled a seemingly constrained mission: the “US Doge Service,” a rebrand of the existing US Digital Service, was primarily tasked with modernising federal technology and software. So much for taking a chainsaw to spending!

Whether this rebranding effort signals more modest aims or is just a Trojan horse remains unclear. Maybe the US tech bros behind Musk believe, like Tony Blair, that digital tech and modern AI will alone revolutionise government. Or perhaps Doge needs this official façade to sidestep the legal hurdles that plague federal advisory committees, while gaining the agency data it needs to make cuts. After all, in separate executive orders, Trump announced that Doge would be involved with a federal hiring freeze, while pausing new regulation and ditching internal diversity rules — moves more in line with the “major reform” Doge’s webpage promises.

But we should be clear about the pitfalls of the more limited interpretation: if Doge fixates on upgrading government IT, it won’t deliver on Trump’s original vision. Sure, it’s nice for users if agencies can process forms quickly and avoid glitch-ridden nightmares. Yet you can’t deliver genuinely “efficient” government by just improving departments’ software. In fact, a more user-friendly state may just boost demand for government, driving spending even higher and compounding the inefficiencies caused by overall government activity.

America’s federal government, like that of most developed countries, is already riddled with conflicting objectives that create inherent inefficiency. It supports renewables to cut carbon, yet taxes solar panel imports, raising the cost of decarbonisation. It subsidises infrastructure projects but enforces environmental, union and tariff rules that drive up their costs. It props up crops that are ingredients of unhealthy foods while funding anti-obesity campaigns. It subsidises flood insurance, driving up the costs of federal disaster relief funds it administers.

Slick code that helps bureaucrats, households and companies navigate all these programmes won’t eradicate these fundamental contradictions driving waste. They stem from a government juggling too many contradictory objectives at once. The only way to eliminate such inefficiencies is to limit government’s scope to begin with. Software improvements are small fry.

An IT revolution wouldn’t deliver the sweeping budget savings Musk once envisioned either. Consider a dramatic scenario: an AI programme replaces 75 per cent of federal jobs in delivering today’s output, saving a cool $302 billion. That’s less than a fifth of this year’s estimated US budget deficit and just 4 per cent of overall federal spending. Government’s real burden comes from the scope and extent of the programmes, pay-outs and regulations that legislation has facilitated, not the people and software administering them. Trimming the bureaucracy via better IT might streamline new rule-making and time costs, but it won’t meaningfully shrink the budget.

The same lesson applies to the UK, of course. For all the Labour government’s recent talk about bearing down on waste, if politicians refuse to reduce the state’s scope—ie, the subsidies, entitlements, and mandates that require the administrators to begin with — then we’re stuck with high and growing levels of spending. Tech tweaks alone make marginal differences. If they want real savings, both Labour and Doge must tackle what government does, not how it does it.