Musk and Milei belong to the same cult of disruption

The cult of "disruption" is quickly spreading from the private sector to the public. Javier Milei, the president of Argentina, has axed ten ministries, reduced public spending by a third in real terms, and slashed red tape. Donald Trump, the incoming U.S. president has tapped the world’s best known business disrupter, Elon Musk, to head a new department of government efficiency, or DOGE.

Even the British Labour government has leapt on the trend. Prime Minister Keir Starmer has accused "too many civil servants" of "being comfortable in the tepid bath of managed decline." The Cabinet Secretary Pat McFadden (whom many regard as the real deputy prime minister) has called for a new age of "start-ups" and "disruptors" in the public sector. Airbnb, WhatsApp and Spotify have all emerged from nowhere to disrupt their respective industries; why can’t we apply the same daring-do spirit to the public sector?

It is easy to see why this sentiment is so popular. Public-sector productivity lags far behind private-sector productivity in part because it is so hard to sack duds or reward superstars. Public-sector managers are addicted to adding layers of "supervisors": The ratio of managers to enlisted personnel in the U.S. military has doubled over the past two decades. They are also experts in resisting change. The greatest comedy about the civil service, "Yes Minister," features the dance between a minister, Jim Hacker, who chases after bright (or harebrained) ideas and his chief civil servant, the permanent head of the Department of Administrative Affairs, Sir Humphrey Appleby, who expertly blocks him. "Yes" always means "no," and "not quite now" always means "never."

Some of this inefficiency is due to "Baumol’s law." The economist William Baumol noted that boosting efficiency in labor-intensive organizations is hard (it always takes four people to play a string quartet) and most public-sector organizations are labor-intensive. But the AI revolution offers a chance to automate bureaucratic procedures now performed by paper shufflers. Bureaucracies were invented to deal with the flow of paper ("bureaucracy" means "rule of desks"). The AI revolution will inevitably produce fundamental changes in the nature of the state.

But is "disruption" the best way of dealing with our frustrations or seizing the opportunities offered by new technology? The public sector is fundamentally different from the private sector. Private companies can fail — indeed, the cult of disruption is all about increasing the rate of "creative destruction." The state must endure — not least because its core job is to pick up the pieces left behind by private-sector smashups. Milei recently told the Economist that "my contempt for the state is infinite." But it was the state that saved the capitalist system from collapse in 2008 when the finance sector’s enthusiasm for disruptive innovation went too far.

The public sector also enjoys powers over private citizens that the private sector does not (or at least has not since the demise of the East India Company in 1858). The state can deprive people of their liberty or, in some countries, their lives. Do we really want our legal systems disrupted by zealots or governed by untested algorithms? The great (conservative) political scientist James Q. Wilson pointed out that "it is not hyperbole to say that the constitutional order is animated by the desire to make the government "inefficient." The Founding Fathers constrained the government’s freedom of action in all sorts of ways to prevent it from riding roughshod over their peoples. The price of liberty is sometimes frustration.

The final problem with the doctrine of "disruption" is that it can be counterproductive: The mayhem it creates ends up discrediting the wider case for reform. Liz Truss’s plan to shake the British establishment out of its complacency with a "faster still" economic revolution flamed out after only 49 days. Elon Musk’s plans to give the U.S. government the Tesla-and-Twitter treatment look just as half-baked. Musk has promised to slash $2 trillion from the country’s expenditures without offering any details of what he would cut (Social Security payments?). His cohead of DOGE, Vivek Ramaswamy, has hinted that he will sack half of federal employees. Yet the real hiring splurge in the U.S. government has been in contractors, not direct employees, whose number had remained flat for decades. All this sounds like the traditional sloganeering to get rid of "fraud, waste and abuse" rather than a considered plan to modernize government.

AI only adds more honey to the disruption trap. Private companies are already beginning to scale back some of their expectations about "large language models" and the rest. AI is far from a "free lunch": introducing it properly involves a lot of upfront expenditure in computer power and personnel. AI also suffers from all sorts of bugs such as "hallucinations," which are a severe problem in the private sector and could prove fatal in the public one. Significant AI-driven improvements could be a decade in coming.

I hope this does not come across as a Sir Humphrey-style defense of the status quo. Too much of the public sector is badly broken (as anybody who has tried contact the UK tax authorities can confirm) and too much is simply tepid. But "disruption" should be reserved for extreme circumstances such as Argentina. Sometimes the best way to fix problems is through quite arduous work: the British Passport Office has been revolutionized in recent months without any resort to chain saws. Often it is though long-term investment: Singapore’s government machine is the envy of the world thanks to decades of patient capacity-building.

Disruption can certainly be a powerful tool, but only if it is treated as a controlled experiment rather than a big bang that could destroy essential public services. The U.S. Army is a master of controlled experiments. In 2018, it established a new command, the Army Futures Command, to experiment with new war-winning techniques, locating it in the high-tech hub of Austin, Texas, far from the Pentagon bureaucracy. It has also chosen three brigades to test and improve new equipment in a program called "transforming in contact." Pat McFadden wants to recruit "crack teams of problem solvers" from the private sector to work with established civil servants, encouraging them to think more creatively rather than putting them out of a job.

But there is a limit to how much we can gain from a combination of long-term reforms and controlled disruption. The deeper problem with the public sector is not the people who run it but the people who use it. The combination of an aging population and a stagnant economy means that a growing number of countries can no longer afford the largesse of the postwar era. And the only viable long-term solution to this problem (barring a productivity miracle) is to cut big entitlements rather than to pretend that we can force the public sector to produce miracles. What really needs to be disrupted is not so much the workings of government as the public’s expectations.

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners. Adrian Wooldridge is the global business columnist for Bloomberg Opinion.