> The approach we have adopted in this study also has limitations. While MRIO analysis is a powerful method for estimating the paid labour and resource use associated with different consumption profiles, it functions at the sectoral level. The method is particularly well suited to spending categories that align directly with economic sectors, such as education and healthcare. However, it is less well suited to spending categories linked to specific products, such as appliances or mobility. Graeber(2018) argues that modern economies contain a substantial number of “bullshit jobs”. These are jobs that are not needed to meet people’s needs, but which serve the role of keeping people plugged into the economic system. Given that our analysis explores the impact of reducing consumption on a sector-by-sector basis, it is not able to capture this effect.
Not accounting for unnecessary labor in a calculation of required labor seems like a pretty extreme exclusion. Beyond the specifically mentioned bullshit jobs, where the job exists for its own sake; there is also considerable inefficiency simply because cheap labor is available.
For example there are various food items consumed in large quantities and available at low costs only because there are farmers willing to work for next to nothing to harvest and process them. But hand picked exotic fruits aren't really necessary for everyone to have enough to eat. If strawberry pickers got paid 20 times as much we would probably eat fewer strawberries, or become more tolerant of blemishes on mechanically picked strawberries. We would certainly not starve.
Likewise for clothing, manufactured goods, and many services you are probably spending a lot on labor intensive crap you don't need only because it is so cheap. Nearly every essential product has at least a substitute that is conducive to automation. Generally the demand for artisanal products is in part to demonstrate the disposable income that can be spent on such luxury; such demand need not be satisfied in a "minimum required" analysis. Besides automation there's also substitution for longer lasting, more durable goods - typically something that lasts 10 times as long does not take 10 times as much labor to produce.
The fact is that people got along just fine a few decades ago despite much less productivity per capita. While our standards of living have improved since then, those improvements don't fundamentally require much more per capita labor, and in many cases have actually decreased. It took 24 man-hours to produce a cheap car in 1975, in 2025 it's down to around 18 man-hours despite being safer, more fuel efficient, and lasting longer. An average US home in 1975 took 40000 man-hours to build, a modern average US home despite being substantially larger and with many comforts takes 5000 man-hours. In 1975 it took 2.5 man hours to produce a ton of wheat, now it takes 0.3 man hours.
A quick tally (admittedly based on google results) indicates that the major physical goods a pretty well off American consumes (900 kg food, 68 articles of clothing, new phone every year, new computer every 2, new car every 3, new 2500 sq ft house constructed every 30) only amounts to about 5 man-hours of labor per week whereas the paper is estimating it takes 13 man-hours per week to provide current typical levels of these goods for the UK. In a pretty damn good life scenario (UK levels of food consumption, 20 articles of clothing per year, electronics purchased half as frequently, new car every 7 years, living in apartments rebuilt every 50 years) this reduces to around 2 man-hours per week. I'd bet you could get under 1 man-hour per week with some basic optimization.
It's much harder to determine the labor required for services like education and healthcare. These are definitely bigger contributors, and likely not as over-inflated, but even if the service labor estimates are spot on, the elimination of so much labor estimated for physical goods alone would invert the conclusion of the paper. I'm not saying that my quicj analysis would stand up to rigorous scrutiny and the paper is wrong, just that its own analysis is insufficient to back up the claim it's making.
> The approach we have adopted in this study also has limitations. While MRIO analysis is a powerful method for estimating the paid labour and resource use associated with different consumption profiles, it functions at the sectoral level. The method is particularly well suited to spending categories that align directly with economic sectors, such as education and healthcare. However, it is less well suited to spending categories linked to specific products, such as appliances or mobility. Graeber(2018) argues that modern economies contain a substantial number of “bullshit jobs”. These are jobs that are not needed to meet people’s needs, but which serve the role of keeping people plugged into the economic system. Given that our analysis explores the impact of reducing consumption on a sector-by-sector basis, it is not able to capture this effect.
Not accounting for unnecessary labor in a calculation of required labor seems like a pretty extreme exclusion. Beyond the specifically mentioned bullshit jobs, where the job exists for its own sake; there is also considerable inefficiency simply because cheap labor is available.
For example there are various food items consumed in large quantities and available at low costs only because there are farmers willing to work for next to nothing to harvest and process them. But hand picked exotic fruits aren't really necessary for everyone to have enough to eat. If strawberry pickers got paid 20 times as much we would probably eat fewer strawberries, or become more tolerant of blemishes on mechanically picked strawberries. We would certainly not starve.
Likewise for clothing, manufactured goods, and many services you are probably spending a lot on labor intensive crap you don't need only because it is so cheap. Nearly every essential product has at least a substitute that is conducive to automation. Generally the demand for artisanal products is in part to demonstrate the disposable income that can be spent on such luxury; such demand need not be satisfied in a "minimum required" analysis. Besides automation there's also substitution for longer lasting, more durable goods - typically something that lasts 10 times as long does not take 10 times as much labor to produce.
The fact is that people got along just fine a few decades ago despite much less productivity per capita. While our standards of living have improved since then, those improvements don't fundamentally require much more per capita labor, and in many cases have actually decreased. It took 24 man-hours to produce a cheap car in 1975, in 2025 it's down to around 18 man-hours despite being safer, more fuel efficient, and lasting longer. An average US home in 1975 took 40000 man-hours to build, a modern average US home despite being substantially larger and with many comforts takes 5000 man-hours. In 1975 it took 2.5 man hours to produce a ton of wheat, now it takes 0.3 man hours.
A quick tally (admittedly based on google results) indicates that the major physical goods a pretty well off American consumes (900 kg food, 68 articles of clothing, new phone every year, new computer every 2, new car every 3, new 2500 sq ft house constructed every 30) only amounts to about 5 man-hours of labor per week whereas the paper is estimating it takes 13 man-hours per week to provide current typical levels of these goods for the UK. In a pretty damn good life scenario (UK levels of food consumption, 20 articles of clothing per year, electronics purchased half as frequently, new car every 7 years, living in apartments rebuilt every 50 years) this reduces to around 2 man-hours per week. I'd bet you could get under 1 man-hour per week with some basic optimization.
It's much harder to determine the labor required for services like education and healthcare. These are definitely bigger contributors, and likely not as over-inflated, but even if the service labor estimates are spot on, the elimination of so much labor estimated for physical goods alone would invert the conclusion of the paper. I'm not saying that my quicj analysis would stand up to rigorous scrutiny and the paper is wrong, just that its own analysis is insufficient to back up the claim it's making.