Congestion pricing, algorithmic bias, price gouging—the world is full of false policy dilemmas. The only way out is up, using Thomas Schelling's 2d analysis.
The congestion charge, and then the Ultra Low Emission Zone, and then the extension of the ULEZ, are some of the best things ever done in London. I struggle with the pollution as it is; it would be unlivable for me here the way it used to be.
For cases where the “equity” concern relates to income, as in regressive taxation, I think this is a good way to think about these problems and something that’s frequently overlooked. For something harder to quantify directly, such as the equity issues related to COMPAS, what does a “two-dimensional” solution look like?
If we were to naively mimic the approach used when pairing regressive taxes with progressive distribution, I’d expect it to look something like, “we are going to implement COMPAS, because of how efficient it is, but we’ll also pair it with a flat six-month sentence reduction for all black defendants to account for its bias” - but this is obviously a terrible policy!
It feels like this just moves the efficiency-equity tradeoff into the higher dimension, no? It’s more equitable to implement the 2D policy, but it’s much harder to do, and in particular to coordinate the two policies. Take the congestion pricing case in 2D. Like how do you give vouchers to exactly those poor folks affected by the congestion pricing; that’s a massive info and coordination problem. Doing so would be more equitable but maybe the more efficient thing (in the sense that it’s more doable) is to just have an income based voucher.
This is great, but also reminds me of this Simpsons scene: "no that's the beautiful part - when wintertime rolls around, the gorillas simply freeze to death" https://www.youtube.com/watch?v=LuiK7jcC1fY
Yew-Kwang Ng (1984) makes a separate but related point: a less efficient but more equitable policy is always dominated by a more efficeint + tax & transfer scheme because the equitable policy also imposes the same disincentive effects of the tax & transfer scheme.
It seems to me, both when it comes to optics and computationally, that the problem is inventing more problems by attempting to cleave off the unfairness in the simple solution to the 1D problem.
Taking the congestion example, it would appear that morally, you just ("just") want to dissuade bad route choices by getting people to pay what hurts. And the problem with the flat fine is that it doesn't hurt everyone the same. And so, in a more utopian world, you could fine everyone "what hurts". Only you can't dint everyone's utility uniformly like that for reasons that should be obvious, and progressive taxation or whatever else in order to make up for that just feels like a different thing entirely even if it's meant to offset that.
Progressive taxation won’t equalize the hurt that people feel when they pay the fee, but it can help with the problem of overall hardship for the poor. Which is just to say: such hardship isn’t by itself a decisive objection to congestion pricing, whose distributive effects shouldn’t be evaluated only in isolation.
I don’t necessarily disagree, but I feel that it has much the same flavor of the senior going to the PCP, who prescribes them X, which has some side effect Y (high bp?), for which they then need (or it’s prudent to address using) Z, and so on. What is “decisive” vs “good because alleviates” often being a judgment call in practice across domains, ofc. And before long, the senior has 80 meds a day; cf a gov having a raft of additional or otherwise highly visible policies in place.
Great piece! Your series in general has prompted me to think about Abba Lerner’s underrated book, Everybody’s Business, which explains some of these points really well. It’s hard to find these days, and I thought it might be due a revival.
What’s valuable is not so much Lerner’s explanations as his fairly impeccable leftist credentials: he is, after all, the designer of a comprehensive market socialist system. And (relevant to the online left), he’s the inventor of functional finance—a precursor to MMT.
>But Yale’s Ian Shapiro reminds us that we have a second dimension to consider—the progressiveness of government spending.
I dont think this works. If you charge a less progressive tax but give it back in welfare spending to fix the equity, youre creating the same inefficiency-causing incentives that the progressive tax would. All that matters is the relationship of pre-tax income to actual total benefit. Like, if the progressive spending here is just transfers, thats obviously just accounting, but even if its a public good that just ends up being more important for low-income people, thats still the same marginal benefit of more work.
The idea is that progressive spending might be easy to do (free libraries and subsidies for rice and beans for everyone - but these benefits are worth more to the poor, for whom this is a large fraction of their consumption) while progressive taxation often requires huge amounts of paperwork from everyone to verify how much income and wealth they do or don't have.
"We could instead think in two dimensions. We adjust congestion pricing to whatever point best addresses pollution (etc.), and then we separately address the problem of poverty using a dedicated anti-poverty policy."
The people affected negatively don't believe a "dedicated anti-poverty policy" will ever happen. In practice, instead what will happen is that they'll be told that economic efficiency is good for everyone overall - if they oppose it, that's *Communism*, bad bad bad. That anti-poverty programs are nests of waste, fraud, and abuse, and really liberals are the ones to blame for not policing them better. That the correct strategy is "abundance" of getting government out of the way of businesses, etc.
Really, this is a very large portion of our politics, and arguably a big factor in why we have Trump.
How do governing bodies keep these policies bundled together over time? Five years later, if one policy is removed, does the other remain, either purposefully or by accident? Just wondering if there's precedent for managing this in a world where governments already seem to struggle maintaining the policy systems they create. Thanks for sharing!
Usually it’s debatable who benefits from a policy. In Seattle we just increased sales tax to pay for more midday bus routes. Who benefits? The poor? The wealthy retirees? No one? Even in congestion pricing, your strategy fails to account for non-drivers. Do they benefit from less noise pollution or are they harmed by more crowding on buses?
Also, there are other factors in policymaking than efficiency and equity. Suppose we raise a tax to fund space exploration. Your idea to remove inequity would be to lower other taxes on non-astronauts. Even if that were possible, it would simply amount to defunding other programs. That will just create additional inequities and inefficiencies to solve. We want space explorations to achieve civilizational goals, even if it is inequitable and presently inefficient.
Great essay. I've been arguing to progressive for a while, that rather than the democratic-socialist position of direct government intervention and ownership, that it's far more efficient to have a market economy with a tax-and-spend welfare state, and this will be much more likely to achieve their goals. But you made me realize that the taxing and the spending are actually two separate dimensions.
Politically though, a problem is that so often the tax code is manipulated in ways that hides spending and regulation and takes it "off budget" . You can pretend the market is free and then implement a tax penalty instead of a regulation, which is kind of what Trump is doing with import tariffs.
This is the sort of thing I wish was discussed in my public policy degree, instead of just endless kvetching about how terrible neoliberalism is.
The congestion charge, and then the Ultra Low Emission Zone, and then the extension of the ULEZ, are some of the best things ever done in London. I struggle with the pollution as it is; it would be unlivable for me here the way it used to be.
A great example of why "efficiency" really does matter. Pollution sucks! When we can get rid of it in cost-effective ways, we absolutely should!
For cases where the “equity” concern relates to income, as in regressive taxation, I think this is a good way to think about these problems and something that’s frequently overlooked. For something harder to quantify directly, such as the equity issues related to COMPAS, what does a “two-dimensional” solution look like?
If we were to naively mimic the approach used when pairing regressive taxes with progressive distribution, I’d expect it to look something like, “we are going to implement COMPAS, because of how efficient it is, but we’ll also pair it with a flat six-month sentence reduction for all black defendants to account for its bias” - but this is obviously a terrible policy!
It feels like this just moves the efficiency-equity tradeoff into the higher dimension, no? It’s more equitable to implement the 2D policy, but it’s much harder to do, and in particular to coordinate the two policies. Take the congestion pricing case in 2D. Like how do you give vouchers to exactly those poor folks affected by the congestion pricing; that’s a massive info and coordination problem. Doing so would be more equitable but maybe the more efficient thing (in the sense that it’s more doable) is to just have an income based voucher.
This is great, but also reminds me of this Simpsons scene: "no that's the beautiful part - when wintertime rolls around, the gorillas simply freeze to death" https://www.youtube.com/watch?v=LuiK7jcC1fY
Unrelated: but I had no idea Thomas Schelling looked like a goddamn boss. He looks like he’d play some badass in an action movie.
Historians estimate that at least three nuclear attacks in the 1950s were staved off by the sheer force of Thomas Schelling’s stare.
Yew-Kwang Ng (1984) makes a separate but related point: a less efficient but more equitable policy is always dominated by a more efficeint + tax & transfer scheme because the equitable policy also imposes the same disincentive effects of the tax & transfer scheme.
https://www.jstor.org/stable/560?seq=1
Love this. Thanks for cutting through the blah blah blah, Daniel.
It seems to me, both when it comes to optics and computationally, that the problem is inventing more problems by attempting to cleave off the unfairness in the simple solution to the 1D problem.
Taking the congestion example, it would appear that morally, you just ("just") want to dissuade bad route choices by getting people to pay what hurts. And the problem with the flat fine is that it doesn't hurt everyone the same. And so, in a more utopian world, you could fine everyone "what hurts". Only you can't dint everyone's utility uniformly like that for reasons that should be obvious, and progressive taxation or whatever else in order to make up for that just feels like a different thing entirely even if it's meant to offset that.
Progressive taxation won’t equalize the hurt that people feel when they pay the fee, but it can help with the problem of overall hardship for the poor. Which is just to say: such hardship isn’t by itself a decisive objection to congestion pricing, whose distributive effects shouldn’t be evaluated only in isolation.
I don’t necessarily disagree, but I feel that it has much the same flavor of the senior going to the PCP, who prescribes them X, which has some side effect Y (high bp?), for which they then need (or it’s prudent to address using) Z, and so on. What is “decisive” vs “good because alleviates” often being a judgment call in practice across domains, ofc. And before long, the senior has 80 meds a day; cf a gov having a raft of additional or otherwise highly visible policies in place.
Great piece! Your series in general has prompted me to think about Abba Lerner’s underrated book, Everybody’s Business, which explains some of these points really well. It’s hard to find these days, and I thought it might be due a revival.
What’s valuable is not so much Lerner’s explanations as his fairly impeccable leftist credentials: he is, after all, the designer of a comprehensive market socialist system. And (relevant to the online left), he’s the inventor of functional finance—a precursor to MMT.
>But Yale’s Ian Shapiro reminds us that we have a second dimension to consider—the progressiveness of government spending.
I dont think this works. If you charge a less progressive tax but give it back in welfare spending to fix the equity, youre creating the same inefficiency-causing incentives that the progressive tax would. All that matters is the relationship of pre-tax income to actual total benefit. Like, if the progressive spending here is just transfers, thats obviously just accounting, but even if its a public good that just ends up being more important for low-income people, thats still the same marginal benefit of more work.
The idea is that progressive spending might be easy to do (free libraries and subsidies for rice and beans for everyone - but these benefits are worth more to the poor, for whom this is a large fraction of their consumption) while progressive taxation often requires huge amounts of paperwork from everyone to verify how much income and wealth they do or don't have.
There's a well-known problem right here:
"We could instead think in two dimensions. We adjust congestion pricing to whatever point best addresses pollution (etc.), and then we separately address the problem of poverty using a dedicated anti-poverty policy."
The people affected negatively don't believe a "dedicated anti-poverty policy" will ever happen. In practice, instead what will happen is that they'll be told that economic efficiency is good for everyone overall - if they oppose it, that's *Communism*, bad bad bad. That anti-poverty programs are nests of waste, fraud, and abuse, and really liberals are the ones to blame for not policing them better. That the correct strategy is "abundance" of getting government out of the way of businesses, etc.
Really, this is a very large portion of our politics, and arguably a big factor in why we have Trump.
How do governing bodies keep these policies bundled together over time? Five years later, if one policy is removed, does the other remain, either purposefully or by accident? Just wondering if there's precedent for managing this in a world where governments already seem to struggle maintaining the policy systems they create. Thanks for sharing!
Very good. Very well written to boot
Usually it’s debatable who benefits from a policy. In Seattle we just increased sales tax to pay for more midday bus routes. Who benefits? The poor? The wealthy retirees? No one? Even in congestion pricing, your strategy fails to account for non-drivers. Do they benefit from less noise pollution or are they harmed by more crowding on buses?
Also, there are other factors in policymaking than efficiency and equity. Suppose we raise a tax to fund space exploration. Your idea to remove inequity would be to lower other taxes on non-astronauts. Even if that were possible, it would simply amount to defunding other programs. That will just create additional inequities and inefficiencies to solve. We want space explorations to achieve civilizational goals, even if it is inequitable and presently inefficient.
Great essay. I've been arguing to progressive for a while, that rather than the democratic-socialist position of direct government intervention and ownership, that it's far more efficient to have a market economy with a tax-and-spend welfare state, and this will be much more likely to achieve their goals. But you made me realize that the taxing and the spending are actually two separate dimensions.
Politically though, a problem is that so often the tax code is manipulated in ways that hides spending and regulation and takes it "off budget" . You can pretend the market is free and then implement a tax penalty instead of a regulation, which is kind of what Trump is doing with import tariffs.