President-elect Donald Trump is said to be interested in the privatization of the U.S. Postal Service, a prospect that also appeals to his DOGE project and its allies in Congress. Yet debating this issue — and privatization more generally — is almost pointless without first defining terms. “Privatization” can be good, bad or uncertain.

The best kind of privatization is when private suppliers can replace state provision outright. Consider Poland in the 1980s, when most of the economy was state-owned and state-run. After the fall of communism in 1991, many of those state-owned businesses were replaced with private corporations. Since then, Poland has gone from being a poor country to one with living standards close to those of Western Europe.

Privatization is not the only reason, of course, but it was an essential part of the story. When it comes to food and consumer products, for instance, it is much better to let the market operate.

But Poles did not privatize everything. They generally left water companies and electricity providers in the public sector, for example. This is the second category of privatizations: those that are uncertain in their impact.

Water and electricity are two essential services where there is no easy way to get privatization exactly right. It is simply impractical to have many firms selling the product to a single group of households — not in the same way that, say, many cow farmers can produce and sell cheese. It costs too much to lay the basic piping or wires.

One option is to have a private entity with monopoly privileges but regulated prices. Another is to have a set of “common carrier” wires and allow multiple producers to use the network on regulated terms of access. A third is just to have the government own and run the company.

Involving the private sector may give better incentives for cost reduction as well as innovation, since profit maximization is a strong impetus for those kinds of improvements. The efficiency of the private company, however, is also a source of problems. A private company may be efficient at lobbying the government for cronyist privileges. That may lead to higher prices, overly generous reimbursement for cost increases, tougher barriers to entry, or entrenched technologies that favor the incumbent.

In other words: If embedded in an imperfect system, corporate efficiency is not always a pure virtue.

In the U.S., privately owned and publicly owned water utilities show, on average, roughly equal performance. Perhaps that is a disappointing result, but it is consistent with the “public choice” theories favored by many free-market economists.

A third kind of privatization is when business adds a layer of activity to a preexisting government function. For instance, some states have “privatized” their Medicaid services by outsourcing Medicaid provision to private health insurers. The Medicaid program has not gone away or been turned over to the private sector — rather, companies have a role in administering the system.

This kind of “layered” privatization, like the second kind of privatization, can work out either for the better or for the worse. One recent study shows this privatization increased the costs of Medicaid significantly without providing offsetting benefits. The private companies have done a good job — for themselves — of extracting more revenue from the system. Yet Medicare Cost Advantage, which creates a private layer of service on top of Medicare, run by insurance companies, does offer significant benefits to those who opt for it.

The lesson here is that talk of “privatization” per se is meaningless without elucidating which kind of privatization is under consideration. One class of privatizations, such as in Poland, is close to an unalloyed good, and economists are right to tout it. But in a large number of other cases, which are frequent in modern Western democracies, privatization works only when judiciously applied.

Given those categories, what about the USPS? Under a 2018 plan, a privatized postal service would save money by delivering the mail fewer days per week and sending it to more centralized locations, rather than to people’s doorsteps. Not everyone, including many members of Congress, finds that desirable. I have questions of a different sort: Namely, how good would a newly involved private company be at manipulating the federal government to get extra delivery subsidies, entry barriers and price privileges?

Maybe postal privatization can work, but it is unlikely to succeed in an environment of crony capitalism. And at least so far, that seems to be what is on tap from the Trump administration.

Tyler Cowen is a Bloomberg Opinion columnist, a professor of economics at George Mason University and host of the Marginal Revolution blog. This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.