Model, Philosophy

Against Novelty Culture

So, there’s this thing that happens in certain intellectual communities, like (to give a totally random example) social psychology. This thing is that novel takes are rewarded. New insights are rewarded. Figuring out things that no one has before is rewarded. The high-status people in such a community are the ones who come up with and disseminate many new insights.

On the face of it, this is good! New insights are how we get penicillin and flight and Pad Thai burritos. But there’s one itty bitty little problem with building a culture around it.

Good (and correct!) new ideas are a finite resource.

This isn’t news. Back in 2005, John Ioannidis laid out the case for “most published research findings” being false. It turns out that when you have a small chance of coming up with a correct idea even using statistical tests for to find false positives can break down.

A quick example. There are approximately 25,000 genes in the human genome. Imagine you are searching for genes that increase the risk of schizophrenia (chosen for this example because it is a complex condition believed to be linked to many genes). If there are 100 genes involved in schizophrenia, the odds of any given gene chosen at random being involved are 1 in 250. You, the investigating scientist, decide that you want about an 80% chance of finding some genes that are linked (this is called study power and 80% is a common value) You run a bunch of tests, analyze a bunch of DNA, and think you have a candidate. This gene has been “proven” to be associated with schizophrenia at a p=0.05 confidence level.

(A p-value is the possibility of observing an event at least as extreme as the observed one, if the null hypothesis is true. This means that if the gene isn’t associated with schizophrenia, there is only a 1 in 20 chance – 5% – we’d see a result as extreme or more extreme than the one we observed.)

At the start, we had a 1 in 250 chance of finding a gene. Now that we have a gene, we think there’s a 19 in 20 chance that it’s actually partially responsible for schizophrenia (technically, if we looked at multiple candidates, we should do something slightly different here, but many scientists still don’t, making this still a valid example). Which probability to we trust?

There’s actually an equation to figure it out. It’s called Bayes Rule and statisticians and scientists use it to update probabilities in response to new information. It goes like this:

(You can sing this to the tune of Hallelujah; take P of A when given B / times P of A a priori / divide the whole thing by B’s expectation / new evidence you may soon find / but you will not be in a bind / for you can add it to your calculation.)

In plain language, it means that probability of something being true after an observation (P(A|B)) is equal to the probability of it being true absent any observations (P(A), 1 in 250 here), times the probability of the observation happening if it is true (P(B|A), 0.8 here), divided by the baseline probability of the observation (P(B), 1 in 20 here).

With these numbers from our example, we can see that the probability of a gene actually being associated with schizophrenia when it has a confidence level of 0.05 is… 6.4%.

I took this long detour to illustrate a very important point: one of the strongest determinants of how likely something is to actually be true is the base chance it has of being true. If we expected 1000 genes to be associated with schizophrenia, then the base chance would be 1 in 25, and the probability our gene actually plays a role would jump up to 64%.

To have ten times the chance of getting a study right, you can be 10 times more selective (which probably requires much more than ten times the effort)… or you can investigate something ten times as likely to actually occur. Base rates can be more powerful than statistics, more powerful than arguments, and more powerful than common sense.

This suggests that any community that bases status around producing novel insights will mostly become a community based around producing novel-seeming (but false!) insights once it exhausts all of the available true (and easily attainable) insights it could discover. There isn’t a harsh dividing line, just a gradual trend towards plausible nonsense as the underlying vein of truth is mined out, but the studies and blog posts continue.

Except the reality is probably even worse, because any competition for status in such a community (tenure, page views) will become an iterative process that rewards those best able to come up with plausible sounding wrappers on unfortunately false information.

When this happens, we have people publishing studies with terrible analyses but highly sharable titles (anyone remember the himmicanes paper?), with the people at the top calling anyone who questions their shoddy research “methodological terrorists“.

I know I have at least one friend who is rolling their eyes right now, because I always make fun of the reproducibility crisis in psychology.

But I’m just using that because it’s a convenient example. What I’m really worried about is the Effective Altruism community.

(Effective Altruism is a movement that attempts to maximize the good that charitable donations can do by encouraging donation to the charities that have the highest positive impact per dollar spent. One list of highly effective charities can be found on GiveWell; Givewell has demonstrated a noted trend away from novelty such that I believe this post does not apply to them.)

We are a group of people with countless forums and blogs, as well as several organizations devoted to analyzing the evidence around charity effectiveness. We have conventional organizations, like GiveWell, coexisting with less conventional alternatives, like Wild-Animal Suffering Research.

All of these organizations need to justify their existence somehow. All of these blogs need to get shares and upvotes from someone.

If you believe (like I do) that the number of good charity recommendations might be quite small, then it follows that a large intellectual ecosystem will quickly exhaust these possibilities and begin finding plausible sounding alternatives.

I find it hard to believe that this isn’t already happening. We have people claiming that giving your friends cash or buying pizza for community events is the most effective charity. We have discussions of whether there is suffering in the fundamental particles of physics.

Effective Altruism is as much a philosophy movement as an empirical one. It isn’t always the case that we’ll be using P-values and statistics in our assessment. Sometimes, arguments are purely moral (like arguments about how much weight we should give to insect suffering). But both types of arguments can eventually drift into plausible sounding nonsense if we exhaust all of the real content.

There is no reason to expect that we should be able to tell when this happens. Certainly, experimental psychology wasn’t able to until several years after much-hyped studies more-or-less stopped replicating, despite a population that many people would have previously described as full of serious-minded empiricists. Many psychology researchers still won’t admit that much of the past work needs to be revisited and potentially binned.

This is a problem of incentives, but I don’t know how to make the incentives any better. As a blogger (albeit one who largely summarizes and connects ideas first broached by others), I can tell you that many of the people who blog do it because they can’t not write. There’s always going to be people competing to get their ideas heard and the people who most consistently provide satisfying insights will most often end up with more views.

Therefore, I suggest caution. We do not know how many true insights we should expect, so we cannot tell how likely to be true anything that feels insightful actually is. Against this, the best defense is highly developed scepticism. Always remember to ask for implications of new insights and to determine what information would falsify them. Always assume new insights have a low chance of being true. Notice when there seems to be a pressure to produce novel insights long after the low hanging fruit is gone and be wary of anyone in tat ecosystem.

We might not be able to change novelty culture, but we can do our best to guard against it.

[Special thanks to Cody Wild for coming up with most of the lyrics to Bayesian Hallelujah.]

Advice, Model

Context Windows

When you’re noticing that you’re talking past someone, what does it look like? Do you feel like they’re ignoring all the implications of the topic at hand (“yes, I know the invasion of Iraq is causing a lot of pain, but I think the important question is, ‘did they have WMDs?'”)? Or do you feel like they’re avoiding talking about the object-level point in favour of other considerations (“factory farmed animals might suffer, but before we can consider whether that’s justified or not, shouldn’t we decide whether we have any obligation to maximize the number of living creatures?”)?

I’m beginning to suspect that many tense disagreements and confused, fruitless conversations are caused by differences in how people conceive of and process the truth. More, I think I have a model that explains why some people can productively disagree with anyone and everyone, while others get frustrated very easily with even their closest friends.

The basics of this model come from a piece that Jacob Falkovich wrote for Quillette. He uses two categories, “contextualizers” and “decouplers”, to analyze an incredibly unproductive debate (about race and IQ) between Vox’s Ezra Klein and Dr. Sam Harris.

Klein is the contextualizer, a worldview that comes naturally to a political journalist. Contextualizers see ideas as embedded in a context. Questions of “who does this effect?”, “how is this rooted in society?”, and “what are the (group) identities of people pushing this idea?” are the bread and butter of contextualizers. One of the first things Klein says in his debate with Harris is:

Here is my view: I think you have a deep empathy for Charles Murray’s side of this conversation, because you see yourself in it [because you also feel attacked by “politically correct” criticism]. I don’t think you have as deep an empathy for the other side of this conversation. For the people being told once again that they are genetically and environmentally and at any rate immutably less intelligent and that our social policy should reflect that. I think part of the absence of that empathy is it doesn’t threaten you. I don’t think you see a threat to you in that, in the way you see a threat to you in what’s happened to Murray. In some cases, I’m not even quite sure you heard what Murray was saying on social policy either in The Bell Curve and a lot of his later work, or on the podcast. I think that led to a blind spot, and this is worth discussing.

Klein is highlighting what he thinks is the context that probably informs Harris’s views. He’s suggesting that Harris believes Charles Murray’s points about race and IQ because they have a common enemy. He’s aware of the human tendency to like ideas that come from people we feel close to (myside bias) – or that put a stick in the eye of people we don’t like.

There are other characteristics of contextualizers. They often think thought experiments are pointless, given that they try and strip away all the complex ways that society affects our morality and our circumstances. When they make mistakes, it is often because they fall victim to the “ought-is” fallacy; they assume that truths with bad outcomes are not truths at all.

Harris, on the other hand, is a decoupler. Decoupling involves separating ideas from context, from personal experience, from consequences, from anything but questions of truth or falsehood and using this skill to consider them in the abstract. Decoupling is necessary for science because it’s impossible to accurately check a theory when you hope it to be true. Harris’s response to Klein’s opening salvo is:

I think your argument is, even where it pretends to be factual, or wherever you think it is factual, it is highly biased by political considerations. These are political considerations that I share. The fact that you think I don’t have empathy for people who suffer just the starkest inequalities of wealth and politics and luck is just, it’s telling and it’s untrue. I think it’s even untrue of Murray. The fact that you’re conflating the social policies he endorses — like the fact that he’s against affirmative action and he’s for universal basic income, I know you don’t happen agree with those policies, you think that would be disastrous — there’s a good-faith argument to be had on both sides of that conversation. That conversation is quite distinct from the science and even that conversation about social policy can be had without any allegation that a person is racist, or that a person lacks empathy for people who are at the bottom of society. That’s one distinction I want to make.

Harris is pointing out that questions of whether his beliefs will have good or bad consequences or who they’ll hurt have nothing to do with the question of if they are true. He might care deeply about the answers of those questions, but he believes that it’s a dangerous mistake to let that guide how you evaluate an idea. Scientists who fail to do that tend to get caught up in the replication crisis.

When decouplers err, it is often because of the is-ought fallacy. They fail to consider how empirical truths can have real world consequences and fail to consider how labels that might be true in the aggregate can hurt individuals.

When you’re arguing with someone who doesn’t contextualize as much as you do, it can feel like arguing about useless hypotheticals. I once had someone start a point about police shootings and gun violence with “well, ignoring all of society…”. This prompted immediate groans.

When arguing with someone who doesn’t decouple as much as you do, it can feel useless and mushy. A co-worker once said to me “we shouldn’t even try and know the truth there – because it might lead people to act badly”. I bit my tongue, but internally I wondered how, absent the truth, we can ground disagreements in anything other than naked power.

Throughout the debate between Harris and Klein, both of them get frustrated at the other for failing to think like they do – which is why it provided such a clear example for Falkovich. If you read the transcripts, you’ll see a clear pattern: Klein ignores questions of truth or falsehood and Harris ignores questions of right and wrong. Neither one is willing to give an inch here, so there’s no real engagement between them.

This doesn’t have to be the case whenever people who prefer context or prefer to deal with the direct substance of an issue interact.

My theory is that everyone has a window that stretches from the minimum amount of context they like in conversations to the minimum amount of substance. Theoretically, this window could stretch from 100% context and no substance to 100% substance and no context.

But practically no one has tastes that broad. Most people accept a narrower range of arguments. Here’s what three well compatible friends might look like:

We should expect to see some correlation between the minimum and maximum amount of context people want to get. Windows may vary in size, but in general, feeling put-off by lots of decoupling should correlate with enjoying context.


 Here we see people with varyingly sized strike zones, but with their dislike of context correlated with their appreciation for substance.

Klein and Harris disagreed so unproductively not just because they give first billing to different things, but because their world views are different enough that there is absolutely no overlap between how they think and talk about things.

One plausible graph of how Klein and Harris like to think about problems (quotes come from the transcript of their podcast). From this, it makes sense that they couldn’t have a productive conversation. There’s no overlap in how they model the world.

I’ve found thinking about windows of context and substance, rather than just the dichotomous categories, very useful for analyzing how me and my friends tend to agree and disagree.

Some people I know can hold very controversial views without ever being disagreeable. They are good at picking up on which sorts of arguments will work with their interlocutors and sticking to those. These people are no doubt aided by rather wide context windows. They can productively think and argue with varying amounts of context and substance.

Other people feel incredibly difficult to argue with. These are the people who are very picky about what arguments they’ll entertain. If I sort someone into this internal category, it’s because I’ve found that one day they’ll dismiss what I say as too nitty-gritty, while the next day they criticize me for not being focused enough on the issue at hand.

What I’ve started to realize is that people I find particularly finicky to argue with may just have a fairly narrow strike zone. For them, it’s simultaneously easy for arguments to feel devoid of substance or devoid of context.

I think one way that you can make arguments with friends more productive is explicitly lay out the window in which you like to be convinced. Sentences like: “I understand what you just said might convince many people, but I find arguments about the effects of beliefs intensely unsatisfying” or “I understand that you’re focused on what studies say, but I think it’s important to talk about the process of knowledge creation and I’m very unlikely to believe something without first analyzing what power hierarchies created it” are the guideposts by which you can show people your context window.

Economics, Falsifiable

You Might Want To Blame Central Banks For Poor Wage Growth

The Economist wonders why wage growth isn’t increasing, even as unemployment falls. A naïve reading of supply and demand suggests that it should, so this has become a relatively common talking point in the news, with people of all persuasions scratching their heads. The Economist does it better than most. They at least talk about slowing productivity growth and rising oil prices, instead of blaming everything on workers (for failing to negotiate) or employers (for not suddenly raising wages).

But after reading monetary policy blogs, the current lack of wage growth feels much less confusing to me. Based on this, I’d like to offer one explanation for why wages haven’t been growing. While I may not be an economist, I’ll be doing my best to pass along verbatim the views of serious economic thinkers.

Image courtesy of the St. Louis Federal Reserve Bank. Units are 1982-1984 CPI-adjusted dollars. Isn’t it rad how the US government doesn’t copyright anything it produces?

 

 

When people talk about stagnant wage growth, this is what they mean. Average weekly wages have increased from $335 a week in 1979 to $350/week in 2018 (all values are 1982 CPI-adjusted US dollars). This is a 4.5% increase, representing $780/year more (1982 dollars) in wages over the whole period. This is not a big change.

More recent wage growth also isn’t impressive. At the depth of the recession, weekly wages were $331 [1]. Since then, they’ve increased by $19/week, or 5.7%. However, wages have only increased by $5/week (1.4%) since the previous high in 2009.

This doesn’t really match people’s long run expectations. Between 1948 and 1973, hourly compensation increased by 91.3%.

I don’t have an explanation for what happened to once-high wage growth between 1980 and 2008 (see The Captured Economy for what some economists think might explain it). But when it comes to the current stagnation, one factor I don’t hear enough people talking about is bad policy moves by central bankers.

To understand why the central bank affects wage growth, you have to understand something called “sticky wages“.

Wages are considered “sticky” because it is basically impossible to cut them. If companies face a choice between firing people and cutting wages, they’ll almost always choose to fire people. This is because long practice has taught them that the opposite is untenable.

If you cut everyone’s wages, you’ll face an office full of much less motivated people. Those whose skills are still in demand will quickly jump ship to companies that compensate them more in line with market rates. If you just cut the wages of some of your employees (to protect your best performers), you’ll quickly find an environment of toxic resentment sets in.

This is not even to mention that minimum wage laws make it illegal to cut the wages of many workers.

Normally the economy gets around sticky wages with inflation. This steadily erodes wages (including the minimum wage). During boom times, businesses increase wages above inflation to keep their employees happy (or lose them to other businesses that can pay more and need the labour). During busts, inflation can obviate the need to fire people by decreasing the cost of payroll relative to other inputs.

But what we saw during the last recession was persistently low inflation rates. Throughout the whole the thing, the Federal Reserve Bank kept saying, in effect, “wow, really hard to up inflation; we just can’t manage to do it”.

Look at how inflation hovers just above zero for the whole great recession and associated recovery. It would have been better had it been hovering around 2%.

It’s obviously false that the Fed couldn’t trigger inflation if it wanted to. As a thought experiment, imagine that they had printed enough money to give everyone in the country $1,000,000 and then mailed it out. That would obviously cause inflation. So it is (theoretically) just a manner of scaling that back to the point where we’d only see inflation, not hyper-inflation. Why then did the Fed fail to do something that should be so easy?

According to Scott Sumner, you can’t just look at the traditional instrument the central bank has for managing inflation (the interest rate) to determine if its policies are inflationary or not. If something happens to the monetary supply (e.g. say all banks get spooked and up their reserves dramatically [2]), this changes how effective those tools will be.

After the recession, the Fed held the interest rates low and printed money. But it actually didn’t print enough money given the tightened bank reserves to spur inflation. What looked like easy money (inflationary behaviour) was actually tight money (deflationary behaviour), because there was another event constricting the money supply. If the Fed wanted inflation, it would have had to do much more than is required in normal times. The Federal Reserve never realized this, so it was always confused by why inflation failed to materialize.

This set off the perfect storm that led to the long recovery after the recession. Inflation didn’t drive down wages, so it didn’t make economic sense to hire people (or even keep as many people on staff), so aggregate demand was low, so business was bad, so it didn’t make sense to hire people (or keep them on staff)…

If real wages had properly fallen, then fewer people would have been laid off, business wouldn’t have gotten as bad, and the economy could have started to recover much more quickly (with inflation then cooling down and wage growth occurring). Scott Sumner goes so far to say that the money shock caused by increased cash reserves may have been the cause of the great recession, not the banks failing or the housing bubble.

What does this history have to do with poor wage growth?

Well it turns out that companies have responded to the tight labour market with something other than higher wages: bonuses.

Bonuses are one-time payments that people only expect when times are good. There’s no problem cutting them in recessions.

Switching to bonuses was a calculated move for businesses, because they have lost all faith that the Federal Reserve will do what is necessary (or will know how to do what is necessary) to create the inflation needed to prevent deep recessions. When you know that wages are sticky and you know that inflation won’t save you from them, you have no choice but to pre-emptively limit wages, even when there isn’t a recession. Even when a recession feels fairly far away.

More inflation may feel like the exact opposite of what’s needed to increase wages. But we’re talking about targeted inflation here. If we could trust humans to do the rational thing and bargain for less pay now in exchange for more pay in the future whenever times are tight, then we wouldn’t have this problem and wages probably would have recovered better. But humans are humans, not automatons, so we need to make the best with what we have.

One of the purposes of institutions is to build a framework within which we can make good decisions. From this point of view, the Federal Reserve (and other central banks; the Bank of Japan is arguably far worse) have failed. Institutions failing when confronted with new circumstances isn’t as pithy as “it’s all the fault of those greedy capitalists” or “people need to grow backbones and negotiate for higher wages”, but I think it’s ultimately a more correct explanation for our current period of slow wage growth. This suggests that we’ll only see wage growth recover when the Fed commits to better monetary policy [3], or enough time passes that everyone forgets the great recession.

In either case, I’m not holding my breath.

Footnotes

[1] I’m ignoring the drop in Q2 2014, where wages fell to $330/week, because this was caused by the end of extended unemployment insurance in America. The end of that program made finding work somewhat more important for a variety of people, which led to an uptick in the supply of labour and a corresponding decrease in the market clearing wage. ^

[2] Under a fractional reserve banking system, banks can lend out most of their deposits, with only a fraction kept in reserve to cover any withdrawals customers may want to make. This effectively increases the money supply, because you can have dollars (or yen, or pesos) that are both left in a bank account and invested in the economy. When banks hold onto more of their reserves because of uncertainty, they are essentially shrinking the total money supply. ^

[3] Scott Sumner suggests that we should target nominal GDP instead of inflation. When economic growth slows, we’d automatically get higher inflation, as the central bank pumps out money to meet the growth target. When the market begins to give way to roaring growth and speculative bubbles, the high rate of real growth would cause the central bank to step back, tapping the brakes before the economy overheats. I wonder if limiting inflation on the upswing would also have the advantage of increasing real wages as the economy booms? ^

Economics, Politics

You’re Doing Taxes Wrong: Consumptive vs. Wealth Inequality

When you worry about rising inequality, what are you thinking about?

I now know of two competing models for inequality, each of which has vastly different implications for political economy.

In the first, called consumptive inequality, inequality is embodied in differential consumption. Under this model, there is a huge gap between Oracle CEO Larry Ellison (net worth: $60 billion), with his private islands, his yacht, etc. and myself, with my cheap rented apartment, ten-year-old bike, and modest savings. In fact, under this model, there’s even a huge gap between Larry Ellison with all of his luxury goods and Berkshire Hathaway CEO Warren Buffett (net worth: $90.6 billion), with his relatively cheap house and restrained tastes.

Pictured: Warren Buffett’s house vs. Larry Ellison’s yacht. The yacht is many, many times larger than the house. Image credits: TEDizen and reivax.

Under the second model, inequality in new worth or salary is all that matters. This is the classic model that gives us the GINI coefficient and “the 1%”. Under this model, Warren Buffett is the very best off, with Larry Ellison close behind. I’m not even in contention.

I’ve been thinking a lot about inequality because of the recent increase in the minimum wage in Ontario. The reasons behind the wage hike – and similar economic justice proposals (like capping CEO pay at some double-digit multiple of worker pay) – seem to show a concern for consumptive inequality.

That is to say, the prevailing narrative around inequality is that it is bad because:

  1. Rich people are able to consume in a way that is frankly bananas and often destructive either to the environment or norms of good governance
  2. Workers cannot afford all basic necessities, or must choose between basic necessities and thinking long term (e.g. by saving for their children’s education or their own retirement)

Despite this focus on consumptive inequality in public rhetoric, our tax system seems to be focused primarily on wealth inequality.

Now, it is true that wealth inequality can often lead to consumptive inequality. Larry Ellison is able to consume to such an obscene degree only because he is so obscenely wealthy. But it is also true that wealth inequality doesn’t necessarily lead to consumptive inequality (there are upper middle-class people who have larger houses than Warren Buffett) and that it might be useful to structure our tax policy and other instruments of political economy such that there was a serious incentive for wealth inequality not to lead to consumptive inequality.

What I mean is: it’s unlikely that we’re going to reach a widely held consensus that wealth is immoral (or at what level it becomes immoral). But I think we already have a widely held consensus that given the existence of wealth, it is better to wield it like Mr. Buffett than like Mr. Ellison.

To a certain extent, we already acknowledge this. In Canada, there are substantial tax advantages to investing up 18% of your yearly earnings (below a certain point) and giving up to 75% of your income to charity. That said, we continue to bafflingly tax many productive uses of wealth (like investing), while refusing to adequately tax many frivolous or actively destructive uses of wealth (large cars, private jets, private yachts, influencing the political process, etc.).

Many people, myself included, find the idea of large amounts of wealth fundamentally immoral. Still, I’d rather tax the conspicuous and pointless use of wealth than wealth itself, because there are many people motivated to do great things (like curate all of the world’s information and put it at our fingertips) because of desire for wealth.

I’m enough of a post-modernist to worry that any attempt to create a metric of “social value” will further disenfranchise people who have already been subject to systemic discrimination and fail to reflect the tastes of anyone younger than 35 (I just can’t believe that a bunch of politicians would get together and agree that anyone creates social value or deserves compensation for e.g. cosplay, even though I know many people who find it immensely valuable and empowering).

That’s the motivation. Now for the practice. What would a tax plan optimized to punish spurious consumption while maintaining economic growth even look like? Luckily Scott Sumner has provided an outline, the cleverness of which I’d like to explain.

No income tax

When you take money from people as taxes, then give it back to them regardless of how hard they work, you discourage work. It turns out that this effect is rather large, such that the higher income taxes are, the more you discourage people from working. People working is a necessary prerequisite for economic growth and I view economic growth as largely positive (in that it is very good at engendering happiness and stability, as well as guaranteeing those of us currently working the possibility of retiring one day and generating revenues for a social safety net) and therefore think we should try and tax in a way that doesn’t discourage this.

No corporate tax

Another important component of economic growth is investment. We can imagine a hypothetical economy where absolutely everything that is produced is consumed, such that much is made, but nothing ever really changes. The products available this year will be the products available next year, at the same price and made in the same factory, with any worn-down equipment replaced, but no additional equipment purchased.

Obviously, this is a toy example. But if you’ve bought a product this year that didn’t exist last year, or noticed the cost of something you regularly buy fall, you’ve reaped the rewards of investment. We need people to deliberately set aside some of the production they’re entitled too via possession of money so that it can instead be used to improve the process of production.

Corporate taxes discourage this by making investment less attractive. In fact, they actively encourage consumptive inequality, by making consumption artificially cheaper than investment. This is the exact opposite of what we should be aiming for!

Interestingly, there have been a variety of report positive results of the recent cut in corporate tax rates in the US, from repatriation of money for US investment to bonuses for workers.

Now, I know that corporate taxes feel very satisfying. Corporations make a lot of money (although probably less than you think!) and it feels right and proper to divert some of that for public usage. But there are better ways of diverting that money (some of which I’ll talk about below) that manage to fill the public coffers without incentivizing behaviour even worse than profit seeking (like bloated executive pay; taxing corporate income makes paying the CEO a lot artificially cheap). Corporate taxes also hurt normal people in a variety of ways – like making saving for retirement harder.

No inheritance tax

This is another example of artificially making consumption more attractive. Look at it this way: you (a hypothetical you who is very wealthy) can buy a yacht now, use it for a while, loan it to your kids, them have them inherit it when it’s depreciated significantly, reducing the tax they have to pay on it. Or you can invest so that you can give your children a lot of money. Most rich people aren’t going to want to leave nothing behind for their children. Therefore, we shouldn’t penalize people who are going to use the money for non-frivolous things in the interim.

A VAT (with rebates or exemptions)

A VAT, or value added tax, is a tax on consumption; you pay it whenever you buy something from a store or online. A “value-added” tax differs from a simple sales tax in that it allows for tax paid to suppliers to be deducted from taxes owed. This is necessary so that complex, multi-step products (like computers) don’t artificially cost more than more simple products (like wood).

Scott Sumner suggests that a VAT can be easily made free for low-income folks by automatically refunding the VAT rate times the national poverty income to everyone each year. This is nice and simple and has low administrative overhead (another key concern for a taxation system; every dollar spent paying people to oversee the process of collecting taxes is a dollar that can’t be spent on social programs).

An alternative, currently favoured in Canada, is to avoid taxing essentials (like unprepared food). This means that people who spend a large portion of their money on food are taxed at a lower overall rate than people who spend more money on non-essential products.

A steeply progressive payroll tax

If income inequality is something you want to avoid, I’d argue that a progressive payroll tax is more effective than almost any other measure. This makes companies directly pay the government if they wish to have high wage workers and makes it more politically palatable to raise taxes on upper brackets, even to the point of multiples of the paid salary.

While this may seem identical to taxing income, the psychological effect is rather different, which is important when dealing with real people, not perfectly rational economics automata. Payroll taxes also make tax avoidance via incorporating impossible (as all corporate income, including dividends after subtracting investment would be subject to the payroll tax) and makes it easy to really punish companies for out of control executive compensation. Under a payroll tax system, you can quite easily impose a 1000% tax on executive compensation over $1,000,000. It’s pretty hard to justify a CEO salary of $10,000,000 when it’s costing investors more than a hundred million dollars!

Scott Sumner also suggests wage subsidies as an option to avoid the distortionary effect of a minimum wage [1], a concept I’ve previously explored in depth and found to be probably workable.

A progressive property tax

Property taxes tend to be flat, which makes them less effective at discouraging conspicuous consumption (e.g. 4,500 square foot suburban McMansions). If property taxes sharply ramped up with house value or size, families that chose more appropriately sized homes (or could only afford appropriately sized home) would be taxed at lower rates than their profligate neighbours. Given that developments with smaller houses are either higher density (which makes urban services cheaper and cars less necessary) or have more greenspace (which is good from an environmental perspective, especially in flood prone areas), it’s especially useful to convince people to live in smaller houses.

This would be best combined with laxer zoning. For example, minimum house sizes have long been a tool used in “nice” suburbs, to deliberately price out anyone who doesn’t have a high income. Zoning houses for single family use was also seized upon as a way to keep Asian immigrants out of white neighbourhoods (as a combination of culture and finances made them more likely to have more than just a single nuclear family in a dwelling). Lax zoning would allow for flexibility in housing size and punitive taxes on large houses would drive demand for more environmentally sustainable houses and higher density living.

A carbon tax

Carbon is what economists call a negative externality. It’s a thing we produce that negatively affects other people without a mechanism for us to naturally pay the cost of this inflicted disutility. When we tax a negative externality, we stop over-consumption [2] of things that produce that externality. In the specific case of taxing carbon, we can use this tax to very quickly bring emissions in line with the emissions necessary to avoid catastrophic warming.

I’d like to generalize this to Pigovian taxes beyond carbon. Alcohol (and other intoxicants), sugary drinks, and possibly tobacco should be taxed in line with their tendency to produce costs that (in countries with public risk pooling of health costs) are not borne by the individual over-consuming. I do think it’s important to avoid taking this too far – it’s reasonable to expect people to cover their negative externality, but not reasonable to punitively tax things just because a negative externality might exist or because we think it is wrong or “unhealthy” to do it. Not everything that is considered unhealthy leads to actual diseases, let alone increased healthcare costs.

A luxury goods tax

This comes from a separate post by Scott Sumner, but I think it’s a good enough idea to mention here. It should be possible to come up with a relatively small list of items that are mostly positional – that is to say that the vast majority of their cost is for the sake of being expensive (and therefore showing how wealthy and important the possessor is), not for providing increasing quality. To illustrate: there is a significant gap in functionality between a $3,000 beater car and a $30,000 new car, less of a gap between a $30,000 car and a $300,000 car and even less of a gap between the $300,000 car and a $3,000,000 car; the $300,000 car is largely positional, the $3,000,000 car almost wholly so. To these we could add items that are almost purely for luxury, like 100+ foot yachts.

It’s necessary to keep this list small and focus on truly grotesque expenditures, lest we turn into a society of petty moralizers. There’s certainly a perspective (normally held by people rather older than the participants) in which spending money on cosplay or anime merchandise is frivolous, but if it is, it’s the sort of harmless frivolity equivalent to spending an extra dollar on coffee. I am in general in favour of letting people spend money on things I consider frivolous, because I know many of the things I spend money on (and enjoy) are in turn viewed as frivolous by others [3]. However, I think there comes a point when it’s hard to accuse anyone of petty moralizing and I think that point is probably around enough money to prevent dozens of deaths from malaria (i.e. $100,000+) [4].

Besides, there’s the fact that making positional goods more expensive via taxation just makes them more exclusive. If anything, a strong levy on luxury goods may make them more desirable to some.


As I’ve read more economics, my positions on many economics issues have shifted in a way that many people parse as “more conservative”. I reject this. There are a great many “liberal” positions that sound good on paper, but when you actually do the math, hurt the poor and benefit the rich. Free trade makes things cheaper for all of us and has created new jobs and industries. A lot of regulation allows monopolies and large companies to crush any upstart rivals, or shifts jobs from blue collar workers making things to white collar workers ensuring compliance.

It is true that I care about the economy in a way that I never cared about it before. I care that we have sustainable growth that enriches us all. I care about the stock market making gains, because I’ve realized just how much of the stock market is people’s pensions. I care about start-ups forming to meet brand new needs, even when the previous generation views them as frivolous. I care about human flourishing and I now believe that requires us to have a functioning economic system.

A lot of how we do tax policy is bad. It’s based on making us feel good, not on encouraging good behaviour and avoiding weird economic distortions. It encourages the worst excesses of wealth and it’s too easy to avoid.

What I’ve outlined here is a series of small taxes, small enough to make each not worth the effort to avoid, that together can easily collect enough revenue to ensure a redistributive state. They have the advantage of cutting particularly hard against conspicuous consumption and protecting the planet from unchecked global warming. I sincerely believe that if more people gave them honest consideration, they would advocate for them too and together we could build a fairer, more effective taxation system.

Footnotes:

[1] A minimum wage can make it impossible to have Pareto optimal distributions – distributions where you cannot make anyone better off without making someone else worse off. Here’s a trivial example: imagine a company with two overworked employees, each of whom make $15/hour. The employees are working more than they particularly want to, because there’s too much work for the two of them to complete. Unfortunately, the company can only afford to pay an additional $7/hour and the minimum wage is $14/hour. If the company could hire someone without much work experience for $7/hour everyone would be better off.

The existing employees would be less overworked and happier. The new employee would be making money. The company could probably do slightly more business.

Wage subsidies would allow for the Pareto optimal distribution to exist while also paying the third worker a living wage. ^

[2] Over-consumption here means: “using more of it than you would if you have to properly compensate people for their disutility”, not the more commonly used definition that merely means “consuming more than is sustainable”.

An illustration of the difference: In a world with very expensive carbon capture systems that mitigate global warming and are paid for via flat taxes, it would be possible to be over-consuming gasoline in the economics sense, in that if you were paying a share of the carbon capture costs commensurate with your use, you’d use less carbon, while not consuming an amount of gasoline liable to lead to environmental catastrophe, even if everyone consumed a similar amount. ^

[3] For example, I spent six times as much as the median Canadian on books last year, despite the fact that there’s a perfectly good library less than five minutes from my house. I’m not particularly proud of this, but it made me happy. ^

[4] I am aware of the common rejoinder to this sort of thinking, which is basically summed up as “sure, a sports car doesn’t directly feed anyone, but it does feed the workers who made it”. It is certainly true that heavily taxing luxury items will probably put some people out of work in the industries that make them. But as Scott Sumner points out, it is impossible to meaningfully fix consumptive inequality without hurting jobs that produce things for rich people. If you aren’t hurting these industries, you have not meaningfully changed consumptive inequality!

Note also that if we’re properly redistributing money from taxes that affect rich people, we’re not going to destroy jobs, just shift them to sectors that don’t primarily serve rich people. ^