Privatisation of Government Businesses, Services and Financing

1. Scope and definitions

What the word covers

Privatisation, at its broadest, is the transfer of ownership, financing or delivery of a government function into private hands. Four distinct transactions travel under that one word.

• An asset sale transfers ownership of a government business.

• Outsourcing keeps public funding and public responsibility but contracts delivery to someone else.

• Contestability makes a government provider compete against private providers rather than necessarily replacing it. Think of it as optional outsourcing.

• Private finance brings private capital into public infrastructure while ownership of the service stays public. These are sometimes called public-private partnerships (PPPs).

These blur at the edges. A 99-year lease of a port is an asset sale in everything but name. A franchise with the state setting fares and specifying timetables is closer to outsourcing than to a sale.

Each has a rough opposite. An asset sale reversed is nationalisation, outsourcing reversed is insourcing, and private finance reversed is direct public funding. Contestability transfers only if the public provider loses the test, so its opposite is intentional insourcing.

Some Commonwealth welfare services  aged care, child care and disability services, are outsourced today even though they did not arise through a discrete act of privatisation. They evolved over decades from grant-funded charitable provision into publicly funded delivery by non-government providers, whether under contract, individual entitlement or subsidy. The Commonwealth’s capacity to provide many of these welfare services directly is constitutionally constrained (Williams v Commonwealth). Insourcing is not available, and nor is a block grant to a service provider that no head of power supports. Delivery depends on payments to individuals, even where the money goes straight to the provider, or grants through the states.

What the word excludes

Privatisation often requires changes to administrative arrangements, regulation or funding. Those changes are necessary for the transfer but separate from it, and often can be effected without privatisation occurring. Getting these exclusions right matters, because arguments about privatisation often turn out to be arguments about funding models or regulation wearing a borrowed name.

Corporatisation is not privatisation. If the entity stays wholly government owned and staffed, nothing has moved, and corporatisation often precedes privatisation, which is why the confusion persists.

Regulatory separation, splitting a regulator from a provider, is governance reform, usually needed before contestability, not itself a transfer.

Funding mechanisms are not necessarily privatisation either: casemix funding, individualised budgets, social impact bonds and demand-side subsidies change who chooses and who pays, not who provides, except where the change moves provision, as contestable training funding did out of TAFE.

2. The case for and against

The case for privatisation

Supporters argue:

• Stick to the knitting. Government should not supply what the market would supply efficiently and effectively. Policy, binding decisions and regulation are core government work; catering, cleaning and warehousing are not. Public provision of the second kind displaces private firms and ties up capital and staff the state could use on work nobody else will do.

• Sharper incentives. Private owners lose profit and eventually the business when they waste money; a public agency just asks for a bigger appropriation.

• Contestability surfaces information. Market testing forces true costing, often for the first time. Defence found much of its 1990s savings came from the costing exercise itself, not from who won.

• Scale and expertise. A specialised provider running the same function for many clients gets better at it than a department running it only for itself.

• Capital. Private finance brings infrastructure forward instead of queuing behind other budget priorities, and shifts construction and availability risk to whoever manages it better.

• Clearer accountability. Separating purchaser from provider stops poor performance hiding inside one body; contracts make expectations explicit and measurable.

• Depoliticisation. Public ownership invites political use of the business: jobs preserved in marginal seats, plants sited for votes, prices held down before elections, investment deferred for budget optics. Sale removes the minister’s ability to do any of it.

• Competitive neutrality. A government business competing with private firms borrows more cheaply, pays less tax and faces no real dividend discipline, which distorts the market rather than disciplining it.

• The owner-regulator conflict. Where government owns the provider and regulates the sector, the regulator is compromised. Selling separates the two roles.

• Public pay above market rates. Public employers set pay by classification structure and centralised bargaining, not by what the work would fetch elsewhere. For lower-skilled roles the rate often sits well above what private firms pay for the same job. A tender exposes the gap, and closing it saves real money for taxpayers.

• Rules crowd out judgement. A public agency treats like cases alike because its decisions are reviewable, and the tailored decision is the one that gets overturned. Guidelines harden into rules and discretion disappears. Contracting buys back the discretion that accountability removed.

• Managerial and staffing freedom. Staffing caps, classification structures and procurement rules constrain what a public provider can do, and contracting is often a way around a constraint government imposed on itself.

The case against privatisation

Opponents argue:

• Savings can be wage cuts, not productivity gains. Fewer staff doing the same work to the same standard is a real gain. Fewer staff doing less, with the shortfall absorbed as quietly degraded quality, is a wage cut wearing a productivity gain’s clothes, and the two are hard to tell apart because quality is harder to measure accurately. Paying less for the same work saves the budget without producing anything more efficiently.

• The state can’t transfer responsibility. When a provider fails, government still has to educate the children, house the asylum seekers, run the trains. The risk transfer that justified the deal is arguably partly fictional.

• Contracting destroys the option of doing it yourself. Once in-house capability is dismantled, the state can’t bid, benchmark or credibly threaten to take the work back, as Employment National’s and CRS Australia’s closures show.

• Accountability weakens. Commercial-in-confidence, contested FOI coverage and the distance between a minister and a subcontractor all reduce scrutiny.

• Transaction costs are large and invisible. The Commonwealth spent roughly $92 million contemplating visa-processing outsourcing, including $43.5 million to one consulting firm, and outsourced nothing; the replacement project was allocated $74.9 million and also failed.

• Selling income-producing assets is fiscal illusion. The budget looks better in the year of sale and worse forever after.

• Private finance defers the bill past the election. A government builds social infrastructure (eg schools and hospitals) now and takes the credit now, while the payments run for thirty years and fall on its successors. Private capital costs more than government bonds, around two to four points more, and part of the commitment never shows as debt.

• Some functions shouldn’t be delegated at all. Not because contractors do them badly, but because coercion exercised for profit is objectionable in itself, prisons and immigration detention are the usual examples.

Both sides are also argued by people with a stake. Unions oppose privatisation because their members’ jobs and pay are at risk. Consulting and contracting firms support it because they profit from the transactions and the contracts. Neither fact makes the argument wrong.

3. Why privatisation succeeds or fails

Most government functions could, in principle, be outsourced, sold or exposed to competition. Some are easy to manage that way. Others go off the rails, and a handful of recurring pressure points explain why: whether competition can exist at all, what the payment or regulatory design rewards, and whether the state retains enough information to know what is happening.

Each pressure point below is tagged with the transaction it applies to. Contestability inherits whatever outsourcing gets, so it is not listed separately.

3.1 Where competition is absent or cannot be sustained

A market need not be perfectly competitive. It needs several capable suppliers, the ability to compare and switch between them, and low costs of entry and exit. Without those, competition does not produce a benefit, it just changes who holds the power.

Natural monopoly and market power (asset sale)

A natural monopoly exists where one network serves a market more cheaply than two. Poles and wires, water pipes, rail track, ports serving a single hinterland. Duplicating the network wastes capital, so competition will not emerge no matter how the market is designed.

Selling a natural monopoly does not create a market. It converts a public monopoly into a private one and hands the buyer the power to extract economic rent. The buyer knows this and prices it into the bid, which is why network assets sell for large multiples. Governments then need a regulator to do what competition was supposed to do, hold prices near cost and keep the infrastructure maintained. English water and the Telstra copper network are the standing examples.

A government chasing proceeds can leave the monopoly element in, or contract the competition out where none exists, as NSW did by penalising container traffic at Newcastle to protect Port Botany. The price rises now and the problem arrives later.

Embedded regulatory functions (asset sale, outsourcing)

A regulatory power can be sold inside a business without anyone deciding to sell it. When the Australian Wheat Board became AWB Limited in 1999, what transferred alongside the marketing business was the statutory monopoly on bulk wheat exports, the single desk. That monopoly is what let AWB pay roughly $290 million in kickbacks to the Iraqi regime under the oil-for-food program undetected, a single mandated seller controls the contracting and nobody else sees the terms. The same risk sits in meat inspection, biosecurity inspection, private building certification and land titles concessions: wherever a firm both delivers a service and judges whether it meets a legal standard, it holds power over its own market. In each an accreditation and audit layer sits on top: the decision-maker is still paid by the party it judges, but registration, audits and deregistration sit with government. Identify the regulatory element before the transaction and build that layer, rather than discovering afterwards that nobody did.

Contract incompleteness and hold-up (outsourcing, private finance)

No contract anticipates everything, and the more specific the assets and the longer the term, the more must be renegotiated in a market of one. Refusing to renegotiate then isn’t about resolve, the alternative is often losing the service outright, which no government can absorb, so leverage genuinely changes hands at signing.

Step-in rights, automatic handover on default and non-recourse financing all put the risk on identifiable equity rather than on government. They work better than penalty clauses, because a contractor in enough trouble to trigger a penalty usually cannot pay it. The Cross City and Lane Cove tunnels in Sydney both went into receivership, equity was wiped, and both transferred to new owners with no state payment at the point of transfer. The Cross City Tunnel went through it twice, in 2006 and again in 2013. No state payment is not the same as no cost. NSW paid Connector Motorways $25 million in 2006 to delay the Epping Road surface changes by five months, which was the price of buying back road policy discretion the contract had already sold.

Government as monopsony buyer (outsourcing)

Market power runs in both directions. Where government outsources health and welfare services, it is usually the only buyer that matters, and a market with one purchaser behaves like a market with one seller: it sets a price, and providers accept it or leave.

A fee below the true cost of safe delivery does not produce a shortfall the provider absorbs, it produces quietly degraded service, since the alternative is exiting a market with one customer. The Aged Care Royal Commission traced chronic understaffing to funding levels set by the single payer, not operator behaviour alone; NSW’s Northern Beaches Hospital showed the same pattern under a fixed fee per public patient.

The adjustment lands on wages and staffing. A single payer setting a fixed price removes the way a shortage would normally lift pay, because the provider cannot pass the cost on. Shortages show up as vacancies, casualised work and turnover instead.

It also thins the market it buys from: providers who cannot get a better price elsewhere and cannot pass on costs either run margins too thin to invest, or exit, leaving whoever survived government pricing rather than whoever delivers the best care.

3.2 Where regulation replaces competition

The Averch-Johnson effect (asset sale)

Where a regulator allows a firm to earn a fixed rate of return on its asset base, the firm has an incentive to enlarge the asset base. Every additional dollar of capital earns the allowed margin. The rational response is to build more than the network needs, choose capital-intensive solutions over cheaper operating ones, and defend a generous valuation of existing assets.

Averch and Johnson identified this in 1962. It shows up in the Australian electricity debate as gold-plating, where network businesses over-invested in 2008-2013 because the regulated asset base determined their revenue. Whatever view is taken of that specific claim, the incentive is real and it is created by the regulatory design rather than by private ownership as such. Publicly owned networks subject to the same rules face the same incentive.

Under-investment and asset sweating (asset sale)

The opposite regime produces the opposite pathology. Where a regulator caps prices and lets the firm keep whatever it saves, the firm has an incentive to defer maintenance and renewal. Cutting capital expenditure raises returns immediately. The consequences appear years later, usually after the current owners have sold.

This is asset sweating. It is the standard criticism of the privatised English and Welsh water companies, which paid substantial dividends while deferring infrastructure renewal, and of Britain’s Railtrack before the crashes that led to its collapse in 2001.

These two failures are mirror images. Rate-of-return regulation invites over-investment, price caps invite under-investment. Modern incentive regulation combines them, adding investment allowances, service standards and periodic resets. That manages the trade-off rather than removing it, and shifts it onto the regulator’s judgement, formed from information it holds only imperfectly.

3.3 Where output cannot be fully observed

Quality shifting and corner-cutting (outsourcing)

Contracts specify what can be measured, and providers optimise what is specified; the gap between the two is where quality goes. Count placements and providers maximise placements, not job durability. Count occupancy and beds fill. Count response times and calls get answered while problems go unsolved. The dimensions of quality that are hard to observe, dignity, patience, thoroughness, honest advice, are the dimensions that degrade, because degrading them saves money and nobody can prove it happened. This is not dishonesty, it is the predictable result of paying for one thing and hoping for another, and it is worst where the user is vulnerable, cannot judge quality, and cannot switch: prisoners, detainees, aged care residents, people with disability, children in care.

Cherry-picking and cream-skimming (outsourcing)

Where a provider is paid the same for every customer but customers differ in cost, it profits by attracting the cheap ones and avoiding the rest. In employment services this is creaming and parking: providers concentrate on job seekers closest to work, who trigger outcome payments with little help, while the hardest cases get minimum compliance servicing and are left alone, improving the provider’s return while defeating the policy’s actual objective. Charities whose mission was those clients creamed too. The incentives are just too strong. In transport it is the profitable urban route served and the thin rural one abandoned, which is why deregulated bus markets need tendered social routes bolted on. The fix, differential payments by client difficulty, minimum service requirements, universal obligations, is expensive and imperfect, since it requires the purchaser to know how costly each client or route really is, and the provider always knows better.

Community service obligations (asset sale, outsourcing, private finance)

Governments often want to guarantee a service standard nationally, so they impose a community service obligation, funded or unfunded. Where it is funded, Australia Post’s uniform letter rate or Telstra’s old payphone obligation, the provider uses its information advantage to inflate the price of meeting it. Where it is unfunded, the provider minimises it, and sometimes pays the penalty rather than comply. Either way the obligation erodes. The buyer lobbies to narrow it, technology makes it look obsolete, and governments eventually agree. A CSO is a political commitment whose content gets negotiated down, not a permanent guarantee bought at the point of sale.

3.4 Information and the loss of capability

Regulatory capture and information asymmetry (asset sale, outsourcing)

The regulator depends on the firm for information about the firm. It cannot verify costs, asset conditions or demand forecasts independently, because the operational knowledge sits inside the business. This asymmetry is structural and it does not go away with more diligent regulators.

Over time the relationship tends to tilt. The industry employs more specialists than the regulator, pays them better, and offers the regulator’s staff careers. Industry frames the technical debate because it supplies the technical material. Capture is usually not corruption. It is a slow convergence of worldview.

The same firms often occupy several seats at once. PwC sat inside a bidding consortium for a system the government was buying advice about. Consulting firms advise on the design of outsourcing, bid for the resulting work, and are engaged to review it afterwards.

Confidentiality and information leakage (outsourcing)

Contracting requires government to hand over information, and information cannot be handed back. To run a procurement the state discloses its cost structures, forward plans and weaknesses; to run a service the contractor gets the personal data of the people using it, welfare records, visa applications, while the state stays legally responsible without controlling the systems or staff handling it.

Disclosed information does not stay put. A PwC partner brought inside Treasury under confidentiality agreements to help design anti-avoidance law later monetised what he learned; such agreements are a promise, enforceable only after the damage. The flow is asymmetric too: government information moves outward through procurement, while contractor information moves inward slowly and is shielded from Parliament by commercial-in-confidence, so the contractor learns more about the department than Parliament learns about the contractor. And once a function is contracted, the state stops generating its own cost data: within a cycle or two nobody in government knows what the work should cost, only what it is being charged, and benchmarking depends on asking the market, which asks the incumbent.

4. Ownership is not the decisive variable

Public provision fails the same ways

Almost every failure mode above appears in directly provided government services, and nothing about public ownership guarantees that clients will be treated honestly, respectfully or responsively.

Robodebt is the clearest Australian case. It was designed, built and run inside government, with no contractor holding the decision. It raised hundreds of thousands of unlawful debts against people who owed nothing, defended the practice for years against internal legal advice, and caused documented harm including suicides. The Royal Commission reported in July 2023. Commissioner Holmes called it a crude and cruel mechanism, neither fair nor legal, and a costly failure of public administration in both human and economic terms.

The pattern holds more broadly. Government agencies optimise to measured targets exactly as contractors do, which is why waiting list definitions get managed and emergency department clocks get stopped. Public infrastructure is deferred and sweated when budgets tighten, on the same logic as a regulated network under a price cap. Public monopolies capture their departments through control of operational information, which is the same asymmetry that afflicts regulators. Royal commissions into institutional child abuse and into youth detention at Don Dale examined state-run institutions.

Public provision has one structural disadvantage that is rarely stated. The client usually cannot leave. A person dissatisfied with a government monopoly has no exit, only complaint, and complaint is handled by the organisation complained about. Where contracting creates genuine choice between providers, it gives the user an option that public monopoly does not.

It has a real advantage, though smaller than it looks. A minister can direct an agency to change tomorrow, where changing a contractor means renegotiating or waiting out the contract. But direction runs on ministerial attention, and attention moves on. Reform survives only if it is built into structure, funding or headcount first. Contract terms outlast ministers, which cuts both ways.

These failure modes come from the shape of the task rather than from who owns the provider. Hard-to-measure output, users who cannot switch and a provider that knows more than its overseer produce the same pathologies in both sectors. What differs is the remedy. Public failure can be directed, for as long as anyone is watching. Private failure has to be re-tendered, and only if the capability still exists.

Where it does not, the usual response is another layer of regulation. Quality commissions, safety commissioners, complaints bodies, inspectorates and worker screening regimes such as working with children and working with vulnerable people checks accumulate around contracted human services, each added after a failure the contract did not prevent. That is a real remedy, but it is slower and more expensive than the one it replaces, and it leaves the state paying twice: once for the service and again to watch it.

Governments and public sector unions both accept these mechanisms more readily around private provision than public, because of where the finding lands. Scrutiny of a contractor produces a finding about the contractor. The same scrutiny on an agency produces a finding about the minister, which neither side is eager to invite.

When it tends to work

Contracting reliably performs where the service is well defined and its quality visible, several capable suppliers exist and more can enter, assets are general purpose so a new supplier can take over, contracts are short enough to re-tender, and the government keeps enough expertise to specify the work, monitor it and take it back.

Refuse collection, catering, cleaning, vehicle maintenance, commodity information technology and construction meet most of these tests, which is why they were contracted first almost everywhere and why they generate least argument.

Prisons, detention, child protection, aged care and complex employment services meet almost none of them. When they are contracted out anyway, they require a stronger architecture around the contract to secure good outcomes.

5. Conclusion

Ownership matters less than four things around it: whether competition survives past the tender, whether quality can be seen and not just volume and cost, whether the user can go somewhere else, and whether the government still knows what the work costs.

Those four tests sort the transactions into a rough order, from settled to hardest to undo.

Selling a business into a competitive market is the easiest. The transaction ends, the market disciplines the buyer, and government walks away. CSL and Britain’s National Freight Corporation are the arguments here, and they are largely over.

User-funded infrastructure built with private finance works where the loss lands on equity. The Sydney tunnels failing was the system working: demand risk sat with the financiers, not government, and the assets found new owners without a bailout.

Outsourcing a well-defined service works and generates least argument. Refuse, catering, cleaning, vehicle maintenance, commodity information technology.

Outsourcing human services is the hardest ongoing problem in the set, because the output cannot be seen, the users often cannot leave and the arrangement never ends. It does not follow that keeping it in-house is safe. Robodebt was built and run entirely in government, and royal commissions into institutional child abuse and into youth detention at the Northern Territory’s Don Dale examined state-run institutions. The choice is between two hard options, not between a hard one and an easy one.

Private finance for infrastructure the government will be paying off is a different commitment. The state is the sole buyer for thirty years, the same hold-up problem outsourcing carries, dressed as a financing deal. Private finance handles construction and availability risk well, but the premium, two to four points above government borrowing, has to be earned back through bundling and whole-of-life design that almost nobody checks. Northern Beaches Hospital did not pass that test. Sydney Metro Northwest did.

Selling a natural monopoly is the one you can only undo at great cost. The buyer extracts rent and the regulator built to prevent it faces an information problem it cannot solve. English water is thirty six years of that argument. The government has now resolved to abolish the regulator. The largest company survives on terms negotiated with its creditors, with special administration held in reserve.

As a tendency rather than a rule: competitive asset sales, user-funded private finance and ordinary outsourcing are the arrangements that generally work best. Human-services outsourcing has to be judged case by case. Government-funded private finance has a high bar to clear. Natural-monopoly sales are the hardest to make work.

Contestability produces the clearest comparable evidence in the record. Where the same service was tendered repeatedly and in-house teams could bid, costs fell whoever won. Three practical consequences follow. Identify the natural monopoly elements and the embedded regulatory functions before the transaction and excise them, because separating a network later is never cheaper, and a regulatory power sold with a business hands the buyer authority over its own market. Keep enough in-house capability to specify, monitor and credibly re-tender. And count the transaction costs, which are large, recurring, and paid whether or not anything is delivered.

Human-services outsourcing is genuinely contested, and no design there is good, only less bad. Every model for delivering a complex service to people who cannot easily judge it, and cannot easily leave, will disappoint someone. The honest question is not which arrangement avoids failure. It is which failures a given design produces, who bears them, whether they can be minimised in number and seriousness, and whether, when they do occur, they can be seen and corrected.