Here are ~32 ways in which I think US policy could be better.
Many of these are unrealistic changes. In some cases I think there is a corresponding marginal change that we should advocate for, but in some cases it’s just wishful thinking. You should read this post in the same spirit that you’d read a post about board games.
This list is written from a “dealist” perspective: I am trying to find policy changes that would leave almost everyone better off. If I was instead writing from a long-run utilitarian perspective, I would make much more dramatic recommendations.
I know only a little bit about each these areas, so many of my views are likely wrong and will change as I learn more. I would have put “I think” or “I suspect” in front of every sentence, but that got tedious. (I think that this list is mostly relatively orthodox amongst economists, though that doesn’t stop them from being wrong.)
We should mostly raise revenue via efficient taxes, particularly progress income taxes, consumption taxes (probably via VAT), and land value taxes. We should impose Pigovian taxes wherever appropriate and err on the side of overestimating externalities.
We should eliminate inefficient taxes (especially capital gains and corporate tax) even if they are more progressive than consumption taxes. We should usually achieve our desired level of redistribution by implementing efficient redistributive programs instead of inefficient progressive taxes. In some cases lower levels of inefficient but progressive taxes can outperform higher levels of regressive taxes, but for things like capital gains or corporate tax that seems quite unlikely.
Means of redistribution
Whatever level of redistribution we want, we should mostly reach it by handing people money; other policy decisions should be made primarily on the basis of (Kaldor-Hicks) efficiency, only modestly overweighting the welfare of less-well-off beneficiaries (with the maximum ratio corresponding to the social cost of raising $1 to fund more redistribution, which I expect is between $1 and $1.50). I expect that other spending will be redistributive on net, somewhat decreasing the need for direct transfers, but that direct transfers should still be a majority of the total redistribution.
Moving away from means-tested welfare would involve large tax increases. In practice we’d likely implement this as an income tax deduction (with many families now paying negative income tax), rather than having people separately giving and taking money from the government. The average tax burden wouldn’t change, but it would now be a more steeply increasing function of income (though there would be no implicit tax from benefits phasing out).
We shouldn’t pay people more if they live in areas with high cost of living. There is some luck-of-the-draw in where you are born that we’d like to insure against, but realistically that seems small compared to incentive effects, and it is much more efficient for welfare beneficiaries to move away from key economic centers.
Realistically if we implement redistribution via cash transfers then we probably need to ensure that bankruptcy is available for all debts, to limit the harms from borrowing against future transfer payments.
A common theme in this post will be: stop inefficient redistribution, instead raise taxes + increase transfer payments. I expect people on both the left and right will dislike this. (The other side will never give an inch on inefficient policies that benefit them, so we can’t give an inch on inefficient policies that benefit us…) As part of the ideal policy package, I’d want to lean more heavily on legislation that explicitly compromises. One kind of compromise is including some changes that increase and some that decrease overall redistribution, to increase efficiency while preserving overall redistribution, though I understand this is extremely difficult in practice. I also understand that large increases to taxes + transfers are not politically realistic even if the redistributive effects are offset elsewhere, but I do think that elites on both sides should push for marginal changes in that direction.
Interlude: levels of redistribution
I’d prefer the total amount of redistribution be in the ballpark of each person receiving ~$28,000/year (50% of per capita GDP) in total government benefits (including public services) with costs allocated based on ability to pay.
I don’t have a strong view about levels of redistribution between say 40% and 70% of GDP. Lowering redistribution far below 50% starts to have very large welfare costs with only modest gains for incentives/”just desserts,” while raising it too far above 50% starts to create large distortions without correspondingly larger welfare gains. I expect that reaching 50% redistribution would require government spending around 50-60% of GDP (currently it’s 40%), at least if we wanted to maintain current defense spending.
For many goods, the efficient price is too low to incentivize private production (private providers would charge a high price that resulted in underutilization): roads, transit, parks, infrastructure with economies of scale, and information goods like research, drugs, or media. In these cases there should be a mild presumption in favor of state provision or subsidies. There should never be bans on private competition—using a state-granted monopoly to subsidize an industry is much less efficient than subsidizing it directly.
Many goods are partially rivalrous—there are economies of scale and so the efficient price is too low to incentivize production, but also one person using the good hurts other users. We should typically charge a congestion price that actually reflects the externality. For example, the state should charge market-clearing tolls for roadways and subway systems (which will often be very low off-peak and relatively large on-peak). Charging a lower price and using queueing or lotteries may have some positive redistributive effect, but is also colossally inefficient.
I’d like to see states make estimates for the public value that these services are delivering, and closing them down if those estimates are lower than the cost (or auctioning off resources like roads and parks if the amount the private sector is willing to pay is more than those estimates). It’s OK if those estimates tend to be too optimistic, that would be much better than nothing.
Other public services
We should mostly redistribute via efficient cash transfers rather than in-kind transfers. If we are concerned that poor people make bad consumption decisions then it seems better to use nudges where possible; for example, people with low incomes might receive “default” health insurance, savings programs, or subsidized housing, but recipients should have the ability to opt out in favor of cash transfers which they can use as they see fit. As a general rule I don’t think we should try to override firmly expressed preferences of beneficiaries except when there is a very strong argument.
State services (including those aimed at poor recipients) should mostly compete on the market rather than receiving subsidies. If services can’t compete on the market, they should shut down.
Regulation and consumer protection
Regulators should ban products (and require vocational licensing) less often than they currently do. Marginal regulation today seems to be net harmful on average. (Often debate on the internet turns to whether regulation on net is good or bad, which is only really important if you are an ideological libertarian. Scaling down regulation on the margin is much more likely to be a good idea than dismantling entire regulatory agencies.)
In many cases regulators could replace bans with mandatory “would have banned” labeling laws, letting consumers know that regulators believe products are unsafe or ineffective. If we wanted to continue being relatively paternalistic rather than trusting consumers to make a decision about whether they think regulators are adding value, we could restrict would-have-banned products to “accredited consumers”—e.g. if you wanted to get your hair cut by an unlicensed barber, you could be forced to take a qualifying exam where you answer questions about the risks presented by unlicensed barbers and the nature of licensing requirements.
Where regulators do want to ban products and harmonization isn’t too important, we should prefer ban different things in different places so that we can continue to learn about products that we expect should be banned. Banning a product in 95% of places probably captures about 95% of the value of banning it outright; allowing it in 5% of places captures much more than 5% of the useful information from allowing it, especially if you randomize.
Regulation and externalities
Where possible regulators should impose taxes rather than banning behaviors. Taxes should be based on estimated externalities. We should err on the side of overestimating rather than underestimating externalities, since taxes also raise revenue and so allow us to lower other taxes.
We should separate agencies responsible for monitoring/enforcement and the agencies responsible for estimating externalities. E.g. we should separate the branch of the EPA that enforces environmental law and the branch that estimates social costs.
In general, we should be more willing to compensate the losers from regulatory changes, e.g. paying oil companies for the losses in value when carbon taxes are increased. This kind of compensation should be based on a conservative underestimate for the actual cost. If firms made investments after the social costs should have been clear, then there should be a weaker presumption in favor of compensation.
Regulation and mistakes
Many regulations that don’t address significant externalities or have a coherent paternalistic justification. Most price controls are a bad idea, price gouging should probably be legal, a lot of licensing has no plausible consumer protection justification, any time a regulator is setting a lower bound on prices it’s a really bad sign, etc. Forms of gambling that clearly don’t exploit problem gamblers (e.g. prediction markets involving accredited investors) should be legalized. Prostitution and organ markets should probably be legal.
In general proposed regulations should be supported by credible arguments that they would improve public welfare, and those arguments should be held to a higher standard than they are today. Some of the big failure modes here seem to be economic illiteracy, concentrated interests that push for self-serving regulation, and well-meaning people who are indignant but don’t fully understand the area; higher standards for arguments for regulations would help curb those problems, including protecting individuals from firms that invest in lobbying.
We should make a serious effort to let banks fail when they make bad or unlucky decisions. If we were able to make this change (which is harder than it sounds), then we could significantly weaken regulations on banks while probably having a slightly more stable financial system in the long run.
We should generally allow accredited investors to sign whatever crazy financial contracts they want; the SEC does not seem good at understanding which financial instruments could potentially create value, and they probably shouldn’t be in the business of making those decisions (and should instead focus more on preventing fraud, misrepresentation, and clear abuse). They can use “would have banned” labels too if they think that some contracts are a bad idea. It should be easier for people with moderate savings to become accredited investors.
Sovereign wealth fund
When governments can borrow cheaply, they should probably take the opportunity to borrow and invest. For example, the government currently borrows over 30 years at real interest rate of <1% / year, which seems well below investment returns over the next 30 years (current stock yields are >3%, high-quality bonds are around 2.5%). A $10T domestic investment would make the US government own about 10% of the stock market and 10% of the bond market, and would net something like 2%*$10T= $500B/year of revenue, around $1,500 per year per citizen.
As long as a sovereign wealth fund is a modest fraction of the total market, it could be managed passively without much efficiency loss. Realistically this is probably the only way to avoid a huge mess. A large passively managed wealth fund would definitely drive up returns in finance by effectively outsourcing management to the financial sector. Probably a larger effect would be depressing returns to capital and increasing returns to labor.
Owning something like 10-15% of the stock market would have dramatic effects, so you’d want to introduce such a policy gradually. In practice the total upside is probably not large enough to outweigh the considerable risks of mismangement or something unanticipated going terribly wrong, so you’d want to stop before getting up to levels this large even if everything seemed to be going well.
Minimum wage, labor laws, unions
A minimum wage transfers some wealth from capitalists to workers. But it appears to be very inefficient compared to taxes+redistribution. There is a lot of debate about whether the minimum wage costs enough jobs that it is an obviously terrible policy (it probably doesn’t), but there is little debate about whether it costs enough jobs that it is worse than taxes+transfers (it probably does). The cost of the minimum wage is quadratic in the distance from the market-clearing wage, so we should be especially unhappy with large minimum wages (I consider the enthusiasm for a $15 minimum wage one way in which the left in the US has gone off the rails). Minimum wages slightly above market-clearing wages can be efficient for the same reason, but in light of labor market frictions (that “barely employed” people are not in fact indifferent about losing their jobs) I think raising existing minimum wages is usually a bad idea. (Though Sanders’ “Stop BEZOS” proposal is way worse.)
In the absence of a minimum wage, most labor laws are probably inefficient. In a reasonably functional market the employer has a choice between offering lower benefits or lower salaries, and the employer’s choice here is basically aligned with social interests—they should offer lower benefits if and only if employees prefer the money to the benefits. If you find it distasteful that employees would prefer the money, then redistribute more, most labor law isn’t an efficient form of redistribution though. Similarly for job stability and unemployment insurance. (Of course if you have a minimum wage you are effectively forced to regulate every aspect of minimum wage jobs. That’s another reason to prefer transfers to a minimum wage.)
Unions probably shouldn’t receive special legal protections; improving the bargaining power of workers has good redistributive effects, but probably leads to less efficient employment contracts. I don’t have a strong view on right-to-work laws, or on the more general question of whether unions should be regulated as cartels; I think this would probably improve efficiency, but the gains might be small and there would be other costs.
We’d like to redistribute money from healthy people to sick people (for similar reasons that we’d like to redistribute from wealthy to poor), at least in cases where there is no moral hazard. Private health insurance can’t really do this for pre-existing conditions. It’s hard to directly transfer money from healthy to sick people, since it will introduce incentives to look sick. Transferring health care to sick people is more attractive because then if you fake being sick all you get is some health care that you didn’t actually need.
I think the total welfare gain achievable by this kind of redistribution is not huge; in particular, I think it’s small relative to the size of health care spending in developed countries, suggesting that we should mostly optimize health care for efficiency rather than prioritizing this kind of redistribution. Even as such, it probably makes sense for the state to give sicker people money to cover healthcare expenses.
In general the costs of subsidized healthcare grow a lot as you approach 100% of costs (since consumption choices become increasingly inefficient), and the benefits don’t grow correspondingly, so it’s probably not a good idea to try to get to 100% coverage rather than to just view the policy as softening the blow of pre-existing conditions. It doesn’t seem efficient for the state to provide health care itself. I’d prefer structure transfers by choosing a list of conditions to reimburse and a total reimbursement for each, and then reimbursing a fraction of medical costs up to that limit.
We should be significantly more liberal with bail and parole. We should probably have more lenient sentencing across the board. We should use criminal fines more extensively, since they are a form of punishment with very low deadweight loss, though we should adjust fines to make them more appropriate as a punitive measure—they should scale with the income or wealth of the accused, and revenue should not go to the jurisdiction that collects the fine.
I think that the central bank should print enough money that (nominal) gross national income follows a steady and predicable trajectory, say growing by 5% / year + population growth. We should subsidize forecasts of gross national income to help the fed hit these targets, e.g. by issuing bonds linked to GNI or subsidizing prediction markets.