It’s not so much that libertarianism is about unilaterally abolishing the government and letting companies and individuals do whatever they they so please, but rather it’s more like a conceptual framework for resolving disagreement and conflict when economic agents are either unable to choose not to avail themselves to the government or some other regulatory body. Such decentralized arrangements are motivated by incentives and economic and social self-interest.
A common example is a party of diners agreeing to split a bill. Without the government stipulating how it should be split, the diners are able to arrive at an outcome that everyone in the party agrees to. Maybe this means that diners who order an inexpensive meal such as pasta may have to subsidize someone who orders steak, but does so knowing that complaining may perceived by others are un-altruistic or stingy, so he or she is motivated by self-interest to agree.
The incentive of avoiding possible social ostracization is great enough to compel some individuals to act in ways that are not economical optimal yet in the context of the situation, socially optimal. Another example is Coase’s theorem, which is a conceptual framework for resolving deputes in regard to property rights and externalities in lieu of the government.
Although economists, in general, are not opposed to government, an area of interest to economists is to try to find solutions to problems without deferring to the government to wave a wand to fix it. Politicians and voters want wand-based solutions; economists avail themselves to the intellectual challenge of how competing agents acting on incentives can solve such problems, and in some instances the government is either unwilling or unable to do anything. So that means we have no option but to defer to non-government solutions.
Upon cursory glace, it’s evident the govt. has way more power than the private sector. The private sector can only merely restrict access to their services and is much more reactionary; the public sector has the authority to put people away forever and is proactive, restricting undesirable people from access to all services. Although companies have the power to seek civil redress, without the government to help enforce the judgement such as forcing the defendant to disclose finances and assets under oath, it is effectively useless as the defendant can just hide his money, lie, or just not show up to court.
This power imbalance may also explain why pundits and politicians, on either side of the aisle, seek govt. solutions to social problems, not private sector solutions. Just throw the book at the bad guys; problem solved. As bad as corporations can be, this is why public/govt. tyranny is worse than private sector tyranny. Yes, Facebook and Paypal can de-platform conservatives, but can not actually imprison them or easily stop them from using competing platforms. They can try, but it comes at a cost. This is why large companies aggressively spend billions to acquire competitors, sometimes at very large premiums.
Even well-intentioned govt. programs can cause great misery and suffering. Ironically, the notion of ‘human rights’ is often used as a pretext, intentionally or not, to deprive others of such rights. Anywhere from half a million to a million Iraqis died during the decades-long Iraq war under the guise of ‘human rights,’ in addition to thousands of US servicemen killed or permanently injured.
As an example of decentralized regulation, consider the dark market for 0-day exploits, as well as bug bounties. If a hacker finds a major vulnerability to a website and discloses it to the company, yet is rebuked or not compensated justly, he can sell it on the dark web anonymously , in which case the vulnerability is purchased by criminals and exploited for commercial gain, in the process infecting and harming thousands of users, and thus the ensuing bad press, consumer complaints, and loss of revenue compels the company to take action and offer better financial incentives for the confidential disclosure of such vulnerabilities, all without the government getting involved.
Another example is email and social media spammers, who impart a social cost and are hard for governments to contain due to the decentralized and global nature of such operations. The FTC can sue, but it can sometimes take years to bring a case, and recovery of ill-gotten gains is difficult, if not impossible especially if the defendant is not a US citizen. Criminal cases are more effective, but these can also take a long time and there is no guarantee the government will act. Awareness campaigns that try to warn the general public are expensive and offer no financial incentive. The potential pool of victims is huge, so creating a campaign to warn everyone would be prohibitively expensive and wasteful, especially for a private individual. Sure the government can spend millions on TV and radio ads, but an individual cannot.
A possible solution, similar to above, involves the dark web and some math. It’s counter-intuitive: make the problem worse in the short-run in order to fix it in the long-run. The returns to spam are diminishing, meaning that increasing the amount of spam on a social network or ISP by a factor of 100 does not mean 100x more sales, but maybe 2x more or even less, but it will lead to more complaints by users, bad press, and users defecting. Similar to selling an exploit on the dark net, one sells how-to guides on how to spam, with the intent in the long-run of reducing spam. Unlike spreading awareness, selling guides is profitable and more effective in terms of having a small captive audience. Having more competitors drives down profit margins for the spammers. Inundated with complaints, social networks and ISPs will respond, fixing the problem to a large degree, although containing it completely is often impossible. Email spam obviously still exists but it tends to be hidden in spam folders where it is seldom seen, or rather than indiscriminate is targeted based off of carefully curated and resold lists.
Email spam was so profitable early on, in the mid to late ’90s, that it lead to many syndicates copying it and considerably more spam for users, so by the mid 2000s ISPs out of necessity had developed algorithms to reduce it considerably and make it much harder for spammers to have their spam delivered and read (even if the spam does not bounce, it can be filtered, so the goal is to both have the email not be bounced and have it not filtered to the spam folder). Only later would the US government pass regulation to try to deter email spam, notably the CAN-SPAM Act of 2003, but this is ineffective against spammers not subject to US jurisdiction. Had it just been a handful of email spammers, people would have maybe gotten a single spam message a day instead of dozens, and ISPs would have not gotten as many complaints and bad press and would have been less compelled to act. This solution assumes that the corporations and governments are complacent and or cannot do anything.