Can zero-sum games negate productivity gains?

[Estimated time: 10 mins]

The unusually good pictures in this post were all drawn by my very talented friend James. You can find more of his drawings at Non-Zero-Sum Games, where he writes about game theory, economics, philosophy, AI and more!

We usually assume that increasing productivity will make us better off. Broadly, I think that’s true. Countries that are more productive are richer, and richer countries tend to do better on many other indicators of well-being, such as health, education, and food security. 

However, our happiness depends on multiple things, and some of these may be subject to zero-sum competition. In this case, generalised productivity improvements may not make us better off. Productivity gains will save us resources, but if we just spend those resources competing harder in zero-sum competitions, we won’t be any happier. 

This post looks at whether zero-sum competitions might be preventing us from unlocking the full benefits of productivity gains. It has three parts:

• The first part defines “productivity” and explains why I think increasing productivity is generally good. 
• The second part looks at what we need to be happy. 
• Lastly, I explain how zero-sum competitions over things like status and land might negate some of the benefits of productivity gains.

In my second post over at Non-Zero-Sum Games, I look at how we might be able to extract ourselves from zero-sum competitions. Doing this would allow us to more fully enjoy our productivity gains. 

Productivity gains let us produce more with less

The term “productivity” is often misunderstood. As I’ve said before, increasing productivity does not mean “making more stuff”. When economists talk about increasing productivity, they mean making the same amount of stuff with fewer resources. Those resources might include time, energy, capital or natural resources. In that case, provided you’re producing something that is net good (so not, like, superviruses or animal suffering), increasing productivity will almost always be good. 

Of course, since you’re using fewer resources to make stuff you might then choose to scale up production—that is, make more stuff with the same amount of resources. But you don’t have to. You could just keep making and consuming the same amount, and just enjoy the resources you’ve saved from increasing productivity. So productivity improvements could, in theory, allow us to conserve natural resources without sacrificing consumption and give us more time for leisure and sleep. 

In an essay called “Economic Possibilities for Our Grandchildren”, John Maynard Keynes surmised that if productivity kept rising, most people could be working just 15 hours per week by 2030, while enjoying a standard of life 4 to 8 times higher than when Keynes was writing in 1930. Our problem would then be how to spend all our leisure time. 

We’re now in 2025. Keynes’ prediction that the standard of life would increase by 4 to 8 times has proven remarkably accurate.¹“Standard of life” as measured by real GDP per person. See Beretta et al (2024) and Crafts (2022). Yet it’s safe to say his prediction about our working hours will not come to pass. So why not?

To answer why we’re not yet working 15-hour weeks, we must look at what all our work is for. What do we need to be happy? This will then help us understand how increasing productivity can make us better off—and when it will fail to do so. 

What makes us happy?

Happiness depends on multiple things. In 1943, Abraham Maslow tried to capture what humans needed to thrive and came up with the following hierarchy of needs:

• physiological needs — food, water, shelter, rest
• safety — security of body and resources 
• love and belonging — intimate relationships, friends
• esteem — prestige, status, confidence
• cognitive — books, learning 
• aesthetic — beauty, art
• self-actualisation — meaning and purpose
• transcendence — spiritual needs 

Some of these needs are met by standard, rivalrous goods. Food is rivalrous because if I eat a burger, you can’t eat that same burger, so people end up competing over food. Luckily, the food supply is not fixed. Thanks to increases in agricultural productivity, we can produce food today with far fewer resources than we did a century ago. Productivity has therefore significantly relaxed the competition for food, particularly within developed countries. 

Some of the other needs in Maslow’s hierarchy are non-rivalrous, meaning they aren’t subject to the ordinary laws of scarcity. When I get more rest, it doesn’t take away from your ability to rest. You becoming more self-actualised in no way takes away from my ability to self-actualise. The primary way in which productivity can help us meet these needs is by giving us more time. 

This post focuses on our needs for status and land. Status is one of our “esteem” needs, but can also impact our need for “love and belonging”. Land affects our physiological need for “shelter” as well as “safety”. Both status and land are rivalrous, meaning we compete over them. However, unlike food, the supply for both status and land is relatively fixed.²My second post takes a closer look at how fixed the supply actually is, especially for land. And when supply is fixed, competition becomes zero-sum.

How zero-sum games can negate productivity gains

While productivity has increased a lot over the last hundred years, those increases have occurred unevenly across the economy. Manufacturing has seen substantial productivity gains, so standard goods like food, clothing and home appliances have gotten much cheaper. This is great!

However, there’s a limit to how happy these goods can make us because of diminishing returns. Going from “no food” to “some food” will make you much happier than going from “some food” to “so much food I really need to stop overeating”. 

Okay, so we can’t become happier by consuming ever-larger quantities of standard goods. Still, increasing productivity means we won’t have to use as much time and energy to produce those goods. So can’t we enjoy all that extra time for rest and self-actualisation? 

Not quite. The problem is that resources we save from productivity improvements in one area may just get diverted to another. And if that “other area” is a zero-sum game, we’ll be no better off. This may be clearer with an example:

Example: Productivity gains with zero-sum competition

Let’s assume that a hundred years ago, working a 40-hour week got you exactly:

• 20 units of standard goods like food, clothing, and home appliances. 
• 20 units of housing. 

Each labour hour in 1925 therefore gave you 1 unit of either standard goods or housing. 

Now, assume that productivity in standard goods quintuples by 2025, so that each labour hour generates 5 standard goods units. For 20 labour hours, you now get 100 units of standard goods. 

However, we shouldn’t expect the same productivity gains for housing, as the supply of land is relatively fixed. Let’s assume for now that the productivity of housing stays the same.

You might think that, since we’ve gotten so much more efficient at making standard goods, we can work less. If we want to consume just 60 units of standard goods (three times as much as in 1925), that should cost only 12 hours of work (as 60 / 5 = 12) instead of 20 hours.

Does that mean we can now enjoy more consumption plus an extra 8 hours of leisure each week?

Not so fast. At least, not if you want a house to live in. Because the price of land has now gone up! Since everyone else has figured out they don’t need to work 20 hours per week to buy all the standard goods they could want, they’re now channelling all those extra hours towards bidding up land prices. 

End result? You’re back to working more or less 40 hours per week. You get more standard goods for your efforts (60 units instead of just 20), which is nice, but no more leisure. 

Admittedly, this is a hypothetical and highly simplified example. But there is some data to suggest the underlying dynamic exists. Between 1900 and 2003, the average American household’s share of spending on food and clothing declined considerably, from over half their budget to less than 20%.³BLS 100 Years of U.S. Consumer Spending: Data for the Nation, New York City, and Boston. At the same time, the share of spending going to housing grew from less than a quarter of their budget to around a third.⁴The shares have not changed significantly since 2003. See BLS Consumer Expenditures – 2023. 

Eagle-eyed readers may notice that the increase in spending on housing for American households has not been nearly as dramatic as in my hypothetical example above. That’s because my hypothetical example illustrates what could happen if we threw nearly all our savings from productivity gains into a single zero-sum game (land). It’s like a “worse case scenario” where productivity gains are almost fully negated (“almost” because we still increase our consumption of standard goods).

Thankfully, things in real life aren’t quite as bleak. As I explain in my second post, there are ways to get out of this zero-sum competition and we have used some of them. 

Can we do anything about this?

Returning to the title of this post, “Can zero-sum games negate productivity gains?”, I would say the answer is “Yes”. Zero-sum games can negate productivity gains. But they don’t have to.  

To be clear, I am not saying that the past was better, or that productivity gains over the last 100 years have all been for naught. There are still many benefits to being able to manufacture goods using fewer resources. Increasing productivity is a great first step. I just don’t think we can stop there. Productivity gains may simply allow us to throw more resources into competing over land and status. We won’t realise the full benefits of productivity gains if we’re still stuck in some zero-sum games. 

We therefore need to find ways to extract ourselves from zero-sum games. That is the focus of my second post, which you can find over at Non-Zero-Sum Games.

What do you think of my arguments? Share your thoughts in the comments below!

• 1
  “Standard of life” as measured by real GDP per person. See Beretta et al (2024) and Crafts (2022).
• 2
  My second post takes a closer look at how fixed the supply actually is, especially for land.
• 3
  BLS 100 Years of U.S. Consumer Spending: Data for the Nation, New York City, and Boston.
• 4
  The shares have not changed significantly since 2003. See BLS Consumer Expenditures – 2023.
• 5
  This is a strong argument for a land value tax such as that advocated under Georgism. It won’t change the nature of the zero-sum competition, but it can at least redistribute those windfall gains.