Book Review: Stubborn Attachments, by Tyler Cowen

A Vision for a Society of Free, Prosperous, and Responsible Individuals


I've had this book sitting in my Kindle list for a while. Something about the title made me stubborn too, and I went "I'll read it another time". But I'm glad I did, and annoyed I didn't do it as soon as I bought it, because for length:insights ratio it's one of the best books I've read.

Tyler lays his thesis out in 3 clear precepts.

  • Wealth is awesome. It's what enabled us to go from average age thirty to average age eighty, and air conditioning, and all the great things that makes us never want to take a time machine to go back in time.

  • Exponential growth is even more awesome. It's what allowed Sessa to request a humble doubling of of rice every chess square, and won more wheat than existed in the land. Einstein called compound interest magic, and we see this magic every day with compounding growth. We want it to be like the Knightian Crusonia plant that gives ever greater outputs each period.

  • Discount value for the distant future should be low - we should care about the future humans as much as we do today, or close to it. In Tyler's words:

The conclusion, if you believe all three precepts, is that you should try to increase economic growth as much as possible, since that's the best thing we can do as a civilisation. It's not just a good idea to make us wealthier, it'll make us healthier and wiser as well !

Ah but there's a caveat. It's not just growth at any cost that matters. We don't want to lose whole cities and countries to smokestacks and sweatshops.

So Tyler does say that the growth needs to be sustainable, by which he means both that it shouldn't hurt the environment and create horrible negative externalities, and also that human rights are important. These are relatively self evident normative ideas, albeit annoyingly flimsy ones in practice.

For instance, if China is building coal plants by the hundreds to grow faster, creating a clear negative externality for its current citizens, should that matter if the result is that a few decades hence we'll be able to see many more wealthier and happier Chinese? And consequently around the world?

Or, if we decide that human rights include the right to an education, housing and the internet. It's not that farfetched a notion in theory, just in execution. But the resultant redistribution of resources is under fierce contention precisely because we haven't been able to nail down a pure definition of what human rights we mean. If we could all congregate in a large enough agora maybe we can argue it out, but that seems difficult to put in practice.

Or, if a company discovers the secret behind Star Trek like replicators. It's going to revolutionise the world, and make us all live like Jetsons. The problem is that it comes from a horrendous mining process that'll kill Papua New Guinea as we know it. But it's a short term pain, as we only need to mine enough over a decade or two. The long term calculus, even if we discount the longer term benefits, is massive!

The necessary conditions of using undiscounted weights for the future is that the moral calculus shifts dramatically in favour of the future. And that's the right thing to do in most instances, just not in all. It's a form of arguing against infinity, which is a great way to go full Cantor and lose your marbles. Also, quoting:

Imagine that if in the days of antiquity Cleopatra had taken an extra heaping of dessert and said, well, this may lead to 5,000 people dying a horrible death in the year 2000, but that’s more important.

This is probably the most unconventional assumption in the whole book. For instance, assume that your actions today will cause your tenth generation down to have one untimely death. Is that a net negative we should worry about today? I don't know.

So the issue becomes that despite a clear direction, we have something that will undoubtedly get utilised by anyone in whatever direction they choose. Such is the burden of philosophy.

For instance, he proposes a sort of civilisational marshmallow test for us. Whenever there is a redistributive question, like how to split an apple amongst friends, if there is disagreement he suggests we maximise for more apples later on by planting the tree, rather than spend more and more energy thinking of this as a pure zero sum game.

The unspoken part of the book that excels is the one that looks at how to turn zero sum games into positive sum games. That also feels like the least analysed part of modern economics!

Also, for pure growth, should we be encouraging poor people to give their money to Bezos because he knows how to grow it faster? If not coerced, it's probably a great idea. If nothing else it'll change the fact that 90% of stocks are held by the top 10%. But it might also destabilise the growth apparatus in the longer run because it creates systemic risks. Considering the trade-off explicitly however is something we absolutely must do.


But now, let's take a closer look at how feasible this actually is. After all, we've gotten relatively good at identifying the most common 80% of human rights and sustainability as a planet. We might have disjoint arguments by nefarious actors, like the tobacco industry or the climate change denier industry, but over as period of a decade or two those also largely starts getting ironed out.

So we have a largely sensible world trying to implement a largely sensible policy. What can happen?

Let's start with the most visible problem - if we believe the whole philosophy as laid out in Part I, so what? We don't know what actions will actually bring us consistent economic growth. While we can make our best guess judgements on what's driven us thus far, the road ahead isn't nearly as clear.

One could argue that's also why Tyler is heavily interested in the world of Progress Studies, because you can't easily move forward without understanding the actual path dependence of progress.

Within those constraints, what is clear however is that the effects of our actions have to follow at least a couple precepts:

  • There can't be prolonged periods of negative growth, or stasis, without getting back to a period with a higher growth rate that compensates for it. i.e., when in recessions or depressions it's not enough to get back to previously seen growth rates, we have to exceed it

  • The expansion in human rights, or our definition of sustainability, could become a net drag as it expands, circumscribing new areas of economic action

The first part of this is self evident. If the growth moves around stochastically but still centers on the same number, the end result is likely to be similar. But if it dips in a recession, and if we don't manage to get catch up growth at any epoch, then the long run growth is in trouble. That's just math.

The second one is a little bit more complicated. For one thing, whatever we define has to be deontologically expansive enough to help us contextualise human flourishing, but utilitarian enough that it doesn't stop us from doing what it takes to grow.

For instance, if there were the option to consign one country in Africa to hell for a generation to get us enough Vibranium and make the future worth it, that clearly seems deontologically problematic. But the utilitarian calculus is equally simple. This isn't a problem that can be easily squared.

And is that a fair characterisation? Let's have a look.

Human Rights

Human rights have had a chequered but overall positive trajectory in history. As time has passed, we've seen an expansion in what are considered human rights.

Documents asserting individual rights, such the Magna Carta (1215), the English Bill of Rights (1689), the French Declaration on the Rights of Man and Citizen (1789), and the US Constitution and Bill of Rights (1791) are the written precursors to many of today’s human rights documents. Yet many of these documents, when originally translated into policy, excluded women, people of color, and members of certain social, religious, economic, and political groups.

The UN charter of human rights states:

Recognition of the inherent dignity and of the equal and inalienable rights of all members of the human family is the foundation of freedom, justice, and peace in the world.

Is this an end point? I doubt it! Whenever the average condition of humanity improves, it's important to also push the base condition of humanity up. But human rights is not all that matters as guardrails for our growth.


When we talk about sustainable growth, that oftentimes includes a few newer considerations. For instance, from the world summit on Social Development:

The 2005 World Summit on Social Development identified sustainable development goals, such as economic development, social development, and environmental protection.

It's probably premature to think of just these as an endpoint. There are ongoing conversations that are specific to climate change effects, animal rights and capturing pollution externalities, all of which will meaningfully curtain the economic expansion possibilities.

This isn't just one sided though, these explorations might also find new areas of growth. Whether it's the Green New Deal or whether it's the endogenous growth driven by solar power, constraints can help us find new growth avenues.

The definition of sustainability though is itself a matter of contention. Which makes relying on "sustainable growth" a platform difficult, because we don't agree on the definitions. There's a medium term argument that suggests that over a period of a few decades we do end up agreeing on most of the precepts of sustainability, but it still leaves huge gaps in our efforts as it stands.

For instance, the fact that we can't come to a consensus on actions to be taken on climate change for decades means that extreme weather events rise and demolish parts of the economy that wouldn't have otherwise been destroyed. This means the task in front of us becomes greater, because we need to catch up and surpass that which has been destroyed to get back to the growth rates needed.

The thing is, these ideas about binding the growth with constraints around sustainability isn't new. John Stuart Mill had said the exact same thing during the Industrial Revolution. So had John Muir who emphasised the environmental aspects of growth. Friedrich Engels was appalled by the working conditions of the masses. And Thomas Malthus questioning our ability to solve the problems of exponential growth. And even the Luddites who questioned how the benefits should be distributed.


But that's not all. To grow as an economy is for everyone to benefit from said growth. In the long run, technological progress does trickle down, as can be seen by the ubiquity of cell phones as just one example. Here is a clear large-scale social justice question that we need to deal with. What's the optimum level of inequality we should accept as a society?

If the wealth is power-law distributed, through economic growth we pull up that graph (imagine someone attaching a string to that curve and gently pulling it straight up). You will have much richer rich people, a higher median income, a higher mean income, but still some ridiculously poor folks left behind at the back end of the curve.

Is that what we want? The right says it's a necessary precondition to determine individual perseverance. The left says it's what holds us all back, morally and economically. Either way, that question deserves some attention. And Thomas Piketty writes about how our worship of growth coming from capital (the infamous r > g) makes this inequality worse.

Tyler thinks that this argument doesn't hold water because of what I said above, that the poor of today seem vastly richer compared to the poor of 50 years ago. But even if that were true, we make decisions at a point in time, and telling people that their granddaughters will have a better life will not, by itself, help you pass any policy if you're simultaneously leaving them to their less-than-ideal circumstances.

The fact that our global inequality is declining rapidly at a time we're so worried about in-country inequality shows the nature of the problem.

This is going to be a fight that exists within the economic sphere as long as we have a power law wealth distribution. We could just help those at the far end of the distribution and forget about the rest, but that doesn't seem to be a politically feasible reality either. Which means we're stuck with another definitional argument for the long haul.

The concern that our overriding growth might have downsides is manifestly true, because nothing in the world has zero downsides. Understanding them and ameliorating their effects though, that's the whole ballgame.


The most interesting part of the book, to me, remains the assumption that putting one's blinkers on and focusing solely on longer-term growth, and therefore finding new technology that can help humanity 'level up', will help us avoid several of the more mundane conflicts that plague us.

This is where the book also takes religious overtones. The point is that if you believe in this one thing, it will set you free from immediate, rather mortal considerations. You might not get into heaven, or your children, or your children's children. But eventually, everyone will have a seat at the golden table, and isn't that worth fighting for?

Considering we're the species that fought long and bloody wars for eventual reward in heaven, the answer seems to be yes. But considering we're also the species that created monuments to our vanities, the answer seems to be no. Like the best religious ideologies, it defies easy categorisation.

(In a funny way, this has rather amusing parallels to Marx, who also held the belief that growth will set us free, albeit in a more roundabout and vague fashion. But I'll freely admit this is a parallel that amused me in reading it, but you shouldn't read too much into it. No Straussian takes for you.)

But if we buy the fact that we do need to boost economic growth and make us all happy in the long run, and accept the epistemological limitations that constrain our actions, what exactly can we do? Are we just doomed to spend our days wondering if we made the right choice and furiously arguing with everyone about what "sustainable farming" means?

My reading is that if we want to maximise economic growth, we have to remove the obstacles to that growth. Some thoughts from a Straussian reading of the text.

  1. We have an immediate need to make our institutions less ossified and sclerotic. That stands for the regulatory bodies who are behind the industries they regulate, the consumer service bodies who take pleasure in denying service, the financial bodies riddled with perverse incentives, and the political bodies where zero-sum gamesmanship is the rule!

  2. We have an obligation to massively increase our expenditure in finding the next S curve. While we do have plenty of runway within our current paradigm, especially to increase efficiency by using technology, to find the next big innovative leap would require simultaneous shifts in multiple industries - materials, health, energy and more.

  3. Those were goals. In terms of process, it behoves us to work much much harder to experiment. Since there are no clear pathways to follow, we have to experiment much more with all parts of our system, the funding, the execution, the organisational setups, the information flows, the talent identification, styles of management and incentive structures, to try new ways of unlocking existing potential.

  4. "Step up" technologies, which is any technology that fundamentally increases our capabilities as a species, should get even more support. Healthcare, AI and robotics all fall into this cohort, because all of them essentially make us all supermen and superwomen.

If economic growth is good, then the good news is that there is a north star we should aim for. That might be the biggest benefit. What this also means is that efforts to "bump up" our growth by things like tax policies are unlikely to succeed, unless they create a Romer-style growth that creates more ideas which increases the steady state growth itself.

While the literature says we have been seduced by the easy measures of GDP, turns out there is an endogenous reason why that's so seductive. Because it matters! Whether we want to believe a particular (rather simple) model or not, the argument that we should only do actions that will create a higher steady-state level of growth ends up meaning that we should prioritise technology (in the economics sense) at almost every point. Ideas are the only currency worth a damn.

And if in the long run the growth is what really matters, then at any given point in time, our question has to be "how do we grow faster". It sounds obvious, but I'm not sure it's even close to what we do today.

For instance, almost every governmental department is sclerotic and under all kinds of political power struggles. Even the Pentagon, which has endless money, has the problem of too much, with strict strictures on where they can spend their multiple billions. We still argue about where we should have spent the money in the 1960s. We argue if Bell Labs was actually helpful or not, and if we can replicate it. It's unclear that our capital allocation for research efficiency has been well researched, or if it itself is particularly efficient.

Suppose someone made you Progress Czar. Can you credibly argue you know what to do with it? After all, public research funding is around $130Bn dollars in the US (roughly the same as the total Venture Capital funding), and private research funding adds another astronomical sum to it. Since picking projects to allocate capital to would be wasteful, you'd either end up recreating a similar type of bureaucracy as exists today (give funding to those that give funding to those that give funding to scientists), or you try and create as many experiments as possible to fund crazy projects. Should you focus on nuclear science? Nanomaterials? Healthcare? AI? Robotics? Space? Renewable Energy? The answer to all the above is yes, we should probably fund all of them, and try to minimise the ossification that results in overburdening the system with administrative and clerical oversight. Better said than done!

Like most technocratic philosophies, if there are questions around the actions or policies, the answer is to try and get some data to analyse them. For externalities or redistribution, assess whether they are large enough to negate the benefits of growth. If they aren't, well, we have to live with them. If they are, stop them. While we figure that out, we experiment. And push funding into the sciences and humanities so that the arc of our growth bends ever higher!

Ultimately, the book serves as either a straightforward philosophical text that speaks to the necessity and benefits of economic growth, and as a starting point to ask some of the more poignant questions of our age. At just slightly north of a 100 pages it's something you can polish off in an afternoon. And you absolutely should!