Hide table of contents

Abstract

Here, I present GDP (per capita) forecasts of major economies until 2050. Since GDP per capita is the best generalized predictor of many important variables, such as welfare, GDP forecasts can give us a more concrete picture of what the world might look like in just 27 years. The key claim here is: even if AI does not cause transformative growth, our business-as-usual near-future is still surprisingly different from today.

Latest Draft as PDF

Results

In recent history, we've seen unprecedented economic growth and rises in living standards.

Consider this graph:[1]

 

How will living standards improve as GDP per capita (GDP/cap) rises? Here, I show data that projects GDP/cap until 2050. Forecasting GDP per capita is a crucial undertaking as it strongly correlates with welfare indicators like consumption, leisure, inequality, and mortality. These forecasts make the future more concrete and give us a better sense of what the world will look like soon. Abstract thoughts about utopia generate little emotional energy; I find these forecasts more plastic and informative, because GDP/cap is highly predictive of welfare.[2] GDP/cap's generalized predictive power helps us paint a more vivid picture of what the world will look like soon. The business-as-usual near future suggested by the data below could be seen as a soft lower bound on how much the world will change. And yet, this world still seems radically different from today.

Since the figures below are adjusted for purchasing power parity (PPP), you can compare the GDP/cap of a poorer country in 2050 with the GDP/cap of a richer country in 2020. For instance, between now and 2050, China's GDP/cap will go from $19k to $43k, which is similar to France's today. And so, by 2050, 1.3B Chinese people might enjoy a lifestyle not dissimilar to that of a typical French person today.

These GDP/cap are ~3x[3] higher than the median (i.e. typical) income, due to income inequality. Instead of downward adjusting them in our head we can also read them as an approximation of the typical combined consumption three person household.[4] And so, we can still look at these figures and imagine the median household income in Japan today and in China in 2050 to be ~$40k.

 

2020

  

2024

 

2030

  

2040

  

2050

 
 

Pop.

GDP

GDP/ cap

GDP/ cap

Pop.

GDP/ cap

GDP

Pop.

GDP/ cap

GDP

Pop.

GDP/ cap

GDP

US

334M

$20T

$60K

$68K

356M

$66K

$23T

374M

$76K

$28T

389M

$88K

$34T

Netherlands

17M

$1T

$54K

$59K

18M

$61K

$1T

18M

$71K

$1T

18M

$85K

$1T

S. Arabia

34M

$2T

$55K

$56K

39M

$70K

$3T

43M

$86K

$4T

46M

$102K

$5T

Germany

80M

$4T

$52K

$54K

79M

$59K

$5T

77M

$69K

$5T

75M

$82K

$6T

Australia

26M

$1T

$52K

$53K

28M

$58K

$2T

31M

$67K

$2T

33M

$77K

$3T

S. Korea

51M

$2T

$42K

$47K

53M

$50K

$3T

52M

$59K

$3T

51M

$70K

$4T

Canada

38M

$2T

$48K

$48K

40M

$53K

$2T

42M

$61K

$3T

44M

$70K

$3T

UK

67M

$3T

$45K

$47K

70M

$52K

$4T

73M

$61K

$4T

75M

$71K

$5T

France

66M

$3T

$44K

$48K

68M

$50K

$3T

70M

$56K

$4T

71M

$66K

$5T

Japan

125M

$5T

$40K

$43K

120M

$47K

$6T

114M

$54K

$6T

107M

$63K

$7T

Poland

38M

$1T

$31K

$39K

37M

$40K

$2T

35M

$53K

$2T

33M

$63K

$2T

Spain

46M

$2T

$39K

$42K

46M

$47K

$2T

46M

$53K

$2T

45M

$61K

$3T

Malaysia

32M

$1T

$32K

$31K

36M

$42K

$2T

39M

$54K

$2T

41M

$69K

$3T

Italy

60M

$2T

$39K

$45K

59M

$43K

$3T

58M

$48K

$3T

57M

$55K

$3T

Turkey

82M

$2T

$26K

$35K

88M

$34K

$3T

93M

$43K

$4T

96M

$54K

$5T

Russia

143M

$4T

$28K

$31K

139M

$34K

$5T

133M

$45K

$6T

129M

$55K

$7T

China

1403M

$27T

$19K

$20K

1416M

$27K

$38T

1395M

$34K

$47T

1348M

$43K

$58T

Thailand

69M

$1T

$19K

$19K

68M

$25K

$2T

66M

$34K

$2T

62M

$45K

$3T

Argentina

46M

$1T

$22K

$21K

49M

$27K

$1T

53M

$34K

$2T

55M

$43K

$2T

Mexico

135M

$3T

$19K

$21K

148M

$25K

$4T

158M

$32K

$5T

164M

$42K

$7T

Colombia

50M

$1T

$16K

$16K

53M

$21K

$1T

55M

$28K

$2T

55M

$38K

$2T

Iran

83M

$2T

$21K

$17K

89M

$27K

$2T

91M

$35K

$3T

92M

$42K

$4T

Indonesia

272M

$4T

$14K

$13K

295M

$18K

$5T

312M

$25K

$8T

322M

$33K

$11T

Brazil

216M

$3T

$15K

$17K

229M

$19K

$4T

236M

$25K

$6T

238M

$32K

$8T

S. Africa

57M

$1T

$14K

$13K

60M

$19K

$1T

63M

$28K

$2T

66M

$39K

$3T

Egypt

101M

$1T

$13K

$14K

117M

$17K

$2T

134M

$23K

$3T

151M

$29K

$4T

Vietnam

98M

$1T

$8K

$12K

105M

$12K

$1T

110M

$19K

$2T

113M

$28K

$3T

India

1389M

$12T

$8K

$8K

1528M

$13K

$20T

1634M

$18K

$30T

1705M

$26K

$44T

Philippines

108M

$1T

$10K

$10K

124M

$13K

$2T

137M

$17K

$2T

148M

$22K

$3T

Bangladesh

170M

$1T

$5K

$8K

186M

$7K

$1T

197M

$11K

$2T

202M

$15K

$3T

Pakistan

208M

$1T

$6K

$6K

245M

$8K

$2T

279M

$10K

$3T

310M

$14K

$4T

Nigeria

207M

$1T

$6K

$5K

263M

$7K

$2T

327M

$9K

$3T

399M

$11K

$4T

Mean/Sum

5851M

$116T

$28K

 

6251M

$34K

$155T

6545M

$42K

$202T

6740M

$51K

$259T

 

Discussion

In the GDP/cap forecasts above, one thing jumps out: in just 27 years, the world might be fundamentally changed. I find it illustrative to imagine yourself being 27 years older in a world where:

  • China's GDP is the largest in the world, and India's is as big as the US's. Indonesia's economy is the fourth largest.
  • Some large poorer countries like Nigeria still have relatively low incomes (~$10k/cap). India's GDP/cap will grow from $8k to $26k (like Mexico today) and China's from $19k to $43k (like Japan today).
  • A billion people in big rich countries (e.g. US, Western Europe) will be as rich as the very few millions of people living in small very rich countries today (~$80k/cap like in Switzerland, Norway).[5] Relative to population, these countries might have big influence on the world via science (where rich countries have more scientific papers per person[6]), or aid (e.g. the UK's aid budget is only 2x that of Norway's despite a 10x larger population[7]).
  • As world GDP doubles, the global average GDP/cap also doubles from ~$12k to ~$25k, which is still fairly poor, but getting towards an income that frugal people can live on in rich countries. Most humans will have relatively good healthcare, education, and travel, and high-quality food, work opportunities, and access to global knowledge. The recent book 'The World in 2050'[8] argues that in 2050, two-thirds of the world's population will be middle class or rich.

A richer world might also have:[9]

  • lower birth rates and fewer people
  • more democracy
  • more gender equality and other social and political rights
  • greater life satisfaction
  • enhanced social trust

Generally, in 2050 we should expect the wealth gap between richer and poorer countries to be smaller. Poorer countries have been catching up with richer ones since the '90s[10] (but see[11]). They grow faster than rich countries because of fast catch-up or 'copy-and-paste' growth, whereas richer countries at the frontier can't grow as fast anymore since ideas are getting harder to find. This leads to convergence in countries' wealth. Looking at this in action: just 27 years ago, Lithuania's and Estonia's GDP/cap (PPP-adjusted) were ~⅓ of Japan's, but, surprisingly, now they're the same.[12] We might be surprised again in 27 years. Crucially, the changes ahead seem radically more dramatic than the changes between 27 years ago and now.

Values and Culture

What will people's values be in 2050? Since 1980, the West has moved away from 'obedient values' towards 'emancipative values' and became more internally homogenous.[13] Overall, the world is moving towards Western values, but non-Western nations are adopting different values at different speeds. Concretely, nations are de-emphasizing religion and moving from valuing obedience to valuing independence relatively quickly, but they are slower to adopt Western secular-rational and self-expression values, such as beliefs about homosexuality, abortion, divorce, prostitution, euthanasia, childhood obedience, and suicide. This leads to convergence of values in some areas and divergence in others.[14] 

Much of the variation in values between societies can be explained by GDP/cap differences and boils down to two broad dimensions:[15]

  1. Traditionalist vs secular-rational: Traditionalists value their country, god, authority, obedience, and marriage. As countries' wealth increases, citizens' sense of existential security increases, moving them towards secular-rational values.
  2. Survival vs self-expression: Survival values prioritize security over liberty. Those with survival values tend to be more homophobic, uninterested in political action, distrustful of outsiders, and less happy. As people transition from industrial to knowledge societies, their sense of agency increases and they move towards self-expression values. 

Since richer people tend to favor secular-rational and self-expression values over traditionalist and survivalist values, we should expect people in 2050 to be more secular and more interested in self-expression than people today.

GDP growth has been argued to lead to 'Greater opportunity, tolerance of diversity, social mobility, commitment to fairness, and dedication to democracy.'[16] (but also more meat consumption[17]).

In 2050, people will be older, on average. From 1970 to 2022, the global median age has already increased from 20 to 30 years; it's projected to further increase to 36 by 2050. Unexpectedly, Chinese people will be older than people in the US and the EU.[18],[19] People become slightly more conservative as they age.[20]

In 'The World in 2050', McRae argues that ideas will flow more from east to west. China and other Asian nations' technological advancements will contribute to broader 'Easternisation', signifying a global power shift toward the East, just as the world was Europeanized in the 19th century and Americanized in the 20th century. Now, in the 21st century, due to fast catch-up and population growth, the world is being 'asianized'[21]).

McRae also claims that the anglosphere will become even more important: English-speaking countries will make up 40% of the world's GDP, and populous countries on many continents will be English-speaking (Americas: US; Africa: Nigeria; Asia: India; Oceania: Australia; Europe: British Isles).

Growth could be much faster

As mentioned above, this data offers a soft lower bound on how much the world could change over the next 26 years. In contrast, however, some have speculated that future growth could be extremely fast. For example, MacAskill writes:

'Given that growth rates have increased 30x since the agricultural era, it's not crazy to think that they might increase 10x again; but if they did, the world economy would double every 2.5 years.'[22]

Or relatedly, Karnovsky's 'Most Important Century' hypothesis[23] claims:

  1. The far future is very unfamiliar and could come surprisingly fast.
  2. AI might drive a productivity boom this century that could advance AI even further and lead to a long-lasting, galaxy-wide civilization that could be anything from very utopian to very dystopian.
  3. We might shape this transition and massively impact many future people, if we can make sense of the situation enough to find helpful actions. But we aren't ready for this.
  4. Despite these seemingly 'too wild to take seriously' claims, there are many reasons we could be living in wild times, and we should be ready for anything.

But here, in the business-as-usual near-future scenario, nothing has to go 'zoom'; this 'can go on', perhaps with only modest innovation, even without a transformative AI productivity boom and even if ideas are getting harder to find[24] or Chinese growth slows down.[25],[26] Such a future might even be recognizable to us today because we know what rich countries look like.

I am not arguing that the business-as-usual scenario is more likely than a transformative AI scenario that makes the world economy double every 2.5 years—I am merely making a weaker claim by suggesting that this 'lower bound' world is already surprisingly different from today (we could call it wild). Indeed, it might be this intermediate step of continued 'business as usual' growth that will lead to transformative AI (see below). This means that GDP/cap projections can also help us get a better sense of what the world could look like in the lead-up to transformative AI. And we have a better sense of how far the world will change, since we know how countries changed when they went from $10k to $40k per capita GDP. Experts suggest that there's more than a 10% chance of transformative AI by 2036 and a ~50% chance by 2060.[27] In the absence of catastrophes, it seems like that the inverse probability - 50% chance that we will be in live in this non-TAI world until 2060.

Figure from[28]. Two trajectories: increase followed by plateau; or stasis at close to the current level

 

Implications for AI

GDP forecasts inform forecasts of global challenges (e.g. climate models depend on growth).[29] For instance, both OpenPhil and Epoch estimate that the most expensive AI training run will reach some limit based on economic constraints (i.e. 1% of US GDP).[30] I'll now discuss how AI might influence future GDP/cap figures, and vice versa.

AI causes growth: Most top economists think that AIs such as ChatGPT have a measurable impact on national innovation,[31] and many agree that AI will substantially boost EU/US per capita income over the next 30 years—perhaps more than the internet.[32],[33] Forecasts suggest that AI may increase the US growth rate by 0.6-3.6% until 2040[34] and annual global GDP by 7%.[35] Other economists argue against this and estimate an upper bound on GDP gains might be more modest at ~1.7% over the next 10 years.[36]

Growth might speed up AI development: If India's GDP/cap really grows from $8k to $26k (equivalent to Russia's today) and China's from $19k to $43k (equivalent to Japan's today), and if India and China have, relative to their population, as many researchers as Russia and Japan do today, then that would double today's global research force of 10M.[37] Demographics suggests that China will have ~11M college graduates of which ~half were STEM graduates per year until 2040, resulting in ~80M new STEM workers, and then decline due to falling fertility (cf. the US has a total of 16M graduates).[38]

Global R&D spending is currently ~$1T[39], and this might double if GDP more than doubles. Until 2015, the percentage of GDP spent on R&D was roughly constant at ~2%, but it's now 2.5%.[40] Even if returns to research diminish and new ideas are getting harder to find,[41] more funding might still speed up research effectively. Doubling the number of workers also increases the rate of idea production, and thus innovation. The non-rivalry of ideas means that someone can use an idea without impinging on someone else's use of that idea (e.g. your use of the chain rule doesn't prevent other people from using the chain rule). And so for R&D, doubling the inputs to production more than doubles output, as we adopt innovations created by others. Surprisingly, this effect is very large even if new ideas are getting much harder to find.[42] 

A recent Epoch report[43] argues that the mid-20th century economic slowdown follows from the standard economic growth model and the demographic transition, which decoupled population growth from economic growth, leading to declining fertility and growth rates. In contrast, AI workforces can expand rapidly, as hardware manufacturing is not constrained like human population growth, and software can be copied cheaply, which allows the 'population' of AI workers to grow and innovate at the same time. Unless AI-driven growth is slowed, AI labor might increase the speed of growth a lot until physical bottlenecks stop this at very high levels of growth by current standards.

The three main first-order inputs for AI progress are:[44]

  1. training data
  2. compute
  3. algorithmic efficiency

Doubling the global research force will likely mean more scientists and research engineers working on improving algorithmic efficiency and driving down prices for compute. Globally, ~2% of World 'GDP' or ~$2.4T/year will be spent on all R&D (~20% is spent on basic R&D, which has increased over time) [45],[46],[47] As a rule, the richer the country, the higher the share of GDP it spends on R&D.[48] For instance, US R&D is 3% of GDP.[49] The EU aims to increase combined public and private investment in R&D to 3.5% of GDP by 2025.[50] The UK wants to increase it to 2.4% of GDP by 2027.[51] (see roadmap[52]). NATO spends more than half of the ~$2T spent on the military globally.[53] NATO countries have agreed to spend 2% of GDP on defense by 2024; at least 20% of the 2% target should be spent on major equipment, including related military R&D. These international agreements are pegged to GDP. This means that their budgets (e.g. military, research) are coupled so that they increase in lockstep with GDP. If we grow GDP, which economies optimize for,[54] we automatically increase these budgets, which creates risks from emerging technologies like synthetic biology and AI. These dynamics are relevant to the principle of differential technological development.[55]

Will growth slow?

'Before the Industrial Revolution, it took many centuries for the world economy to double in size; now it doubles every twenty-five years [...] [Over the last fifty years, global GDP / person grew by 2% / year, and all major geographic areas are experiencing significant growth. Growth economists surveyed thought that this trend would stay broadly the same, at 2.1% / year]'[56]

World GDP growth seems to be slowing down from ~5% per year in the 1960's, to ~4% in the 70s, to ~3% in '80s, '90s, '00s, to ~2.5% in the 2010s. I naively extrapolated this trend[57] and found ~2% in 2030 and 1% in 2050:

This trend might continue, albeit more gradually:[58]

By 2027 GDP/cap growth in the median rich country will be <1.5% a year. In some countries, such as Canada and Switzerland, it might be ~0%.[59] Forecasts predict that growth will continue to slow for 'structural reasons such as aging populations, shifts from goods to services, slowing innovation, and debt.'[60] It might also be because scientific progress is slowing.[61] IT–related productivity increases explain 75% of all recent US productivity growth,[62] with similar results globally.[63] 

While IT, new tech and an interconnected world may accelerate growth further, the effect might not be so huge: true, while big ideas advance the knowledge frontier, which transformed the West from 1870-1970, now growth has slowed (cf The Great Stagnation), and progress is harder with fewer low-hanging fruits, more complex knowledge and societal stagnation.[64] 

But even if GDP percentage growth slows, wellbeing growth can still speed up, for two reasons:

  1. The dollar GDP/capita growth rate goes up. Even if we adjust for inflation, Americans had ~$600 more every year in the 60s, but now have ~$900 more per year in the 2010s.[65] 
  2. The next $1k/cap increase in a country at $10k/cap is worth 10x as much as in a country with $100k/cap, because, the utility gained from increasing consumption from $10k to $11k is much greater than the utility gained from increasing consumption from $100k to $101k, even though the absolute dollar amounts are the same.[66] As more countries experience catch-up growth and lift large populations out of poverty, the total utility experienced by the world's population could increase a lot, even if the overall rate of global growth slows, since utility gains from improving living standards at lower levels of income are much larger than the utility gains from equivalent absolute increases at higher levels of income.

Methods

Spreadsheet here.

I used three datasets for this analysis: PwC's 2017 simplified Solow growth model, Gapminder's 2021 model, and Goldman Sachs' 2022 model.

  1. The main source is PwC's 2017[67] simplified Solow growth model, which takes into account demographics, human capital like education, growth in the physical capital stock, and technological progress, which drives improvements in total factor productivity (TFP) adjusted (TFP roughly gauges if we get more efficient and not just use more labor).
  2. Gapminder's 2021 model uses IMF forecasts until 2026, and then assumes that all countries converge to a common global growth rate of 2% per year. Correlations between their forecasts and PwC's are high (>.9), but with some differences; for example, GapMinder estimates that in 2050 India's GDP will be $28T (with a GDP/cap of $17K), while PwC's model estimates $44T ($26K).[68] In the last column in the spreadsheet, I averaged the two models for 2050.
  3. Goldman Sachs' 2022 model uses a simple Cobb-Douglas model to estimate real GDP (per capita). Correlations between this model and the previous two are high (>.9). But what jumps out is that real GDP growth lags quite far behind PPP-adjusted growth. For instance, China's real GDP/cap in 2050 is only $25K (compared to $43k PPP-adjusted, similar to Japan's today), which isn't even as high as Korea's today ($29K). Put differently, the difference between real and PPP-adjusted GDP/cap means that (for instance) for China, its population in 2050 might have a similar quality of life to Japan's today, but even by 2050, they might not be able to afford as many imports as Koreans today.

'Persistence' of growth

As always, note that these models are very crude and should not be taken literally, but they're preferable to leaving our assumptions about the future unarticulated, fuzzy, and vague.It is better to be wrong than vague, and to state explicit assumptions that can be questioned and falsified (as the common aphorisms in statistics go: 'Truth will sooner come out of error than from confusion', 'All models are wrong, but some are useful', 'You're not even wrong', 'It is better to be roughly right than precisely wrong', 'Better clear than clever', etc.). So, our models might be off, but they help us think through relevant considerations and formalize our intuitions. This very much applies here, since these models will very likely be off.

Some argue that it's hard to forecast GDP[69] and population growth[70], despite knowing all essential variables. But others show that, strikingly, some growth might be surprisingly persistent: in many rich countries, GDP/cap from 100 years ago predicts GDP/cap today well within a few percentage points, as average growth is usually ~2%.[71]

Relatedly, even though GDP doesn't measure welfare perfectly, it is one of the best predictors of welfare.[72] See also further links on GDP as a proxy for welfare below.

Future Research

Here are some research questions in this area that might be useful:

 

Appendix: Further reading

The World in 2050

In 1996, McRae published 'The World in 2020'. I only skimmed it but it seems like it makes some good predictions (e.g. Brexit, working from home, 'video phones', online retail), though crucially it seems to underestimate how central the internet was going to be. Now, 25 years later, he just published 'The World in 2050'[74]. Tl;dr:

  1. In 2050, Two-thirds of the world's population will be middle class or rich
  2. Ideas will flow more from east to west.
  3. The anglosphere will become even more important: In 2050, the most populous countries on every continent will be English-speaking: Americas; US. Africa; Nigeria, Asia; India. Australia; Oceania. Europe; British isles. English-speaking countries will make up 40% of world GDP.
  4. China — the world's largest economy — moves from aggression to cooperation because its population gets older.
  5. The EU diverges into a core and a periphery.
  6. India will become a great power, but also face challenges, and its future is uncertain.
  7. Africa will become more important but finding jobs for its many young people will be its greatest challenge.
  8. Globalization changes its direction from moving goods to moving ideas and money.
  9.  New technologies will change our lives again.
  10. A more harmonious relationship between humankind and our planet will arise –- we'll grow sustainably.

Economics

GDP as a proxy for welfare

AI

Forecasting

 

Fiction

 

 

Appendix: Causal Model Between Growth, Liberal Democracy, Human Capital, Peace, and X-Risk

Below I have a rough and unfinished conceptual dynamic causal model between growth, liberal democracy, human capital, peace, and x-risk, that summarizes the evidence for their relationships.

Economic Growth causes…

  1. Growth causes peace
  2. Growth causes stability: growth shocks predict violent conflict in a country [80]
  3. Does growth cause democracy?
    1. Growth causes dedication to democracy, greater opportunity, tolerance of diversity, social mobility, commitment to fairness
    2. Historically the effect was large, but today the richest autocracies are 2x the chance to democratize as the poorest countries (2% vs. 4% per year)
    3. Another paper says the evidence is mixed
    4. Wikipedia on democracy and economic growth
  4. Growth causes existential security

Democracy causes...

  1. Democracy causes peace & stability:
    1. Democratic peace—the idea that democracies rarely fight one another—has been called 'the closest thing we have to an empirical law in the study of international relations' and the relationship is ~5x 'as robust as that between smoking and lung cancer'.
    2. Does Democracy Bring International and Domestic Peace and Security?
  2. Democracy causes growth (⅓ the effect of human capital)
    1. Democracy and Economic Growth: a Literature Review: the paper examines channels through which democracy could affect growth.
    2. Does Democracy Cause Economic Growth, Stability, and Work for the Poor?
  3. Human Capital:
    1. Does Democracy Increase Global Health? 
    2. Does Democracy Cause Work for the Poor? 
  4. Existential Security:
    1. Does Democracy Cause Stability?
    2. Do Democracies Perform Better Combating Climate Change?
  5. Democracy:
    1. (Successful) Democracies Breed Their Own Support 
    2. Are Democracies Better for Social Protection of the Poor, Gender Equality, and Social Cohesion?
    3. 'Chance of conflict declines by ~1% points with 1 SD decline in high fertility (~2 births / woman). 1 SD increase in low secondary education decreases the chance of conflict by about ~1 percentage points, while 1 SD increase in high secondary education decreases conflict by ~1% points. These represent relatively sizable marginal effects considering that the average chance of dyadic conflict is ~3%. 1 SD change in the human capital composite variable, low or high, is about 2x as large as the effect of 1 SD change in low democracy.'
  6. Peace causes Growth: 'civil war reduces annual growth by 0.01–0.13 percentage points, and high-intensity inter-state conflict reduces annual growth by 0.18–2.77 percentage points'
  7. Peace does not cause democracy [citation trail]

Human capital causes…

  1. Human capital causes growth (3x effect size of democracy)
    1. Increasing the tax rate on human capital income from 50% to 55%, reduces the total stock of working human capital by 5% and the growth rate by 6%. After 30 years, this results in 3.8% lower GDP. 
    2. 'Strong evidence of a significant positive long-run effect of both public and private R&D on TFP and of a greater effect of public R&D than private R&D' 
  2. Human capital causes peace

Peace & stability causes...

Note that everything is correlated and an in-depth evidence review of 'Does X cause Y?' argues that 'We ultimately need to choose between (a) believing some overly complicated theory of the relationship between X and Y, which reconciles all of the wildly conflicting and often implausible things we're seeing in the studies; (b) more-or-less reverting to what we would've guessed about the relationship between X and Y in the absence of any research.'

 


  1. ^
  2. ^
  3. ^
  4. ^
  5. ^
  6. ^
  7. ^
  8. ^
  9. ^
  10. ^
  11. ^
  12. ^
  13. ^
  14. ^
  15. ^
  16. ^
  17. ^
  18. ^
  19. ^
  20. ^
  21. ^
  22. ^
  23. ^
  24. ^
  25. ^
  26. ^
  27. ^
  28. ^
  29. ^
  30. ^
  31. ^
  32. ^
  33. ^
  34. ^
  35. ^
  36. ^
  37. ^
  38. ^
  39. ^
  40. ^
  41. ^
  42. ^
  43. ^
  44. ^
  45. ^
  46. ^
  47. ^
  48. ^
  49. ^
  50. ^
  51. ^
  52. ^
  53. ^
  54. ^
  55. ^
  56. ^
  57. ^
  58. ^
  59. ^
  60. ^
  61. ^
  62. ^
  63. ^
  64. ^
  65. ^
  66. ^
  67. ^
  68. ^
  69. ^
  70. ^
  71. ^
  72. ^
  73. ^
  74. ^
  75. ^
  76. ^
  77. ^
  78. ^
  79. ^
  80. ^
  81. ^

129

1
0

Reactions

1
0

More posts like this

Comments11
Sorted by Click to highlight new comments since:

Great analysis, Hauke! Strongly upvoted.

One note. Higher real GDP per capita is associated with greater human welfare as you point out, but I think the impact on animal welfare is unclear due to the meat eater problem. I believe boosting economic growth leads to more animal suffering nearterm, as there is a correlation between meat supply per capita and real GDP per capita:

I suspect boosting economic growth decreases animal suffering longterm via making it peak earlier, and therefore decrease faster. Relatedly, many countries have decoupled economic growth from CO2 emissions.

However, I think greater economic growth would lead to a greater relative decrease in the cost-effectiveness of human welfare interventions than in that of animal welfare interventions. Humans currently prioritise human welfare over animal welfare, so increases in purchasing power are mostly spent on human welfare, which makes them become relatively less neglected. In any case, this does not affect my prioritisation much, as I already consider the best animal welfare interventions to be 1.44 k times as cost-effective as GiveWell's top charities.

Excellent point- I do cite the article from Our World in Data "Meat consumption tends to rise as we get richer", that includes the figure you pasted. 

I agree that we should try to decouple this trend - I think the most promising approach is increasing alternative protein R&D (GFI.org is working on this).

Great post, thank you for writing it. I particularly appreciate the sketch of your causal model summarizing the evidence for the relationships between growth, liberal democracy, human capital, peace, and x-risk; I think you deconfused me quite a bit w.r.t "the world getting better" with that.

But even if GDP percentage growth slows, wellbeing growth can still speed up, for two reasons:

  1. The dollar GDP/capita growth rate goes up. Even if we adjust for inflation, Americans had ~$600 more every year in the 60s, but now have ~$900 more per year in the 2010s.[65]

At first blush this seemed wrong, since I tend to assume log utility to simplify modeling. Your footnote linking to Basil Halperin's essay does note that this assumption is in fact an assumption, and that if utility is in fact "less curved" than log we can still have wellbeing growth speedup with GDP % growth slowdown (this was probably obvious to you but isn't clarified in the text):

But Halperin's essay doesn't actually explore whether this "less curved than log" assumption is true. Do you happen to know of empirical evidence for it?

The only reference I've come across provides evidence against it – it's this section in Sam Nolan's Quantifying Uncertainty in GiveWell's GiveDirectly Cost-Effectiveness Analysis. Quote: 

Isoelastic vs Logarithmic increases in consumption

GiveWell measures GiveDirectly's cost-effectiveness by doublings of consumption per year. For instance, Increasing an individual's consumption from $285.92 to $571.84 for a year would be considered 1 unit of utility. 

However, there is a bit of ambiguity here. What happens when you double someone's consumption twice? For instance, $285.92 to $1,143.68 a year? So, if increasing consumption by 100% has a utility of 1, how much utility does this have? A natural answer might be that doubling twice creates two units of utility. This answer is what the GiveWell CEA currently assumes. This assumption is implicit by representing utility as logarithmic increases in consumption.

However, how much people prefer the first double of consumption over the second is an empirical question. And empirically, people get more utility from the first doubling of consumption than the second. Doubling twice creates 1.66 units rather than 2. An isoelastic utility function can represent this difference in preferences.

Isoelastic utility functions allow you to specify how much one would prefer the first double to the second with the  parameter. When , the recipient values the first double the same as the second and is what the current CEA assumes. When , recipients prefer the first doubling in consumption as worth more than the second. Empirically. GiveWell recognises this and uses  in their calculation of the discount rate:

Increases in consumption over time meaning marginal increases in consumption in the future are less valuable. We chose a rate of 1.7% based on an expectation that economic consumption would grow at 3% each year, and the function through which consumption translates to welfare is isoelastic with eta=1.59. (Note that this discount rate should be applied to increases in ln(consumption), rather than increases in absolute consumption; see calculations here)

I've created a desmos calculator to explore this concept. From the calculator, you can change the value of  and see how isoelastic utility compares to logarithmic utility.

GiveDirectly transfers don't usually double someone's consumption for a year but increase it by a lower factor. So an isoelastic utility would find that recipients would gain more utility than logarithmic utility implies.

Changing  from 1 to 1.59 increases GiveDirectly's cost-effectiveness by 11%. Reducing the cost of doubling someone's consumption for a year from $466.34 to $415.87.

To be clear, I don't know anything else about this topic; I was just surprised/skeptical to see the wellbeing growth speedup point.

Edited to add: I was trivially wrong, sorry, I misread Sam's post. Quoting the relevant section from the passage above: 

When , recipients prefer the first doubling in consumption as worth more than the second. Empirically. GiveWell recognises this and uses  in their calculation of the discount rate

So this is evidence for your point, not against.

Thanks! Excellent comment.

My ambition here was perhaps simpler than you might assumed: my point here was just to highlight an even weaker version of Basil's finding that I thought was worth highlighting: even if GDP percentage growth slows down a smaller growth rate can still mean more $ every year in absolute terms.

Sorry I also don't know much more about this and don't have the cognitive capacity right now to think this through for utility increases and maybe this breaks down at certain ηs. 

Maybe it doesn't make sense to think of just 'one true average η', like 1.5 for OECD countries, but rather specific ηs for different comparisons and doublings. 

There was a related post on this recently - would love for someone to get to the bottom of it.

Appreciate the pointer to that post! That's basically the sort of thing I'm looking for (more reading material for 'gears'). And thanks again for writing the main post.

This is great!


Just wanted to flag that the image of your results doesn't include the last few spreadsheet rows, so it misses out on GDP results for India, which then gets mentioned in the discussion

This all seems extremely optimistic. I don't see the words 'environment', 'externalities', 'biodiversity', or 'pollution' mentioned at all, let alone 'geopolitics', 'fragmentation', 'onshoring', 'deglobalisation' or 'supply chains'. And nothing about energy demands and cost of extraction. Based on upbeat consultancies biased models that always conclude things are good for business? I'll be extremely surprised if this 'lower bound' scenario is even the upper. 

Fair point, but as I wrote, this is just the optimistic the 'business as usual' boring scenario in the absence of catastrophes (e.g. a new cold war). I still think it's a somewhat likely outcome.

On environment / energy: perhaps we'll decouple growth and environmental externalities.

The models from consultancies are based on standard growth models and correlate strongly with IMF projections.

Executive summary: Even without transformative AI, GDP per capita forecasts suggest the world in 2050 will be radically different from today, with major implications for global welfare, values, culture, and AI development.

Key points:

  1. By 2050, GDP per capita in China, India, Indonesia and other developing countries will grow dramatically, lifting billions out of poverty and shifting global economic power to Asia.
  2. A richer world in 2050 will likely have lower birth rates, more democracy, greater gender equality and life satisfaction, and shifting values from traditionalist to secular-rational and self-expression.
  3. Faster economic growth could substantially accelerate AI development by increasing the global research workforce and R&D spending. However, growth rates may also slow due to factors like aging populations.
  4. While GDP is an imperfect welfare measure, it remains one of the best predictors. A richer world, even with slower GDP growth, could see large increases in total global welfare.
  5. Key uncertainties include how growth will impact values and culture, whether growth will speed up or slow down, and how much it will accelerate AI development. More research is needed on these questions.

 

 

This comment was auto-generated by the EA Forum Team. Feel free to point out issues with this summary by replying to the comment, and contact us if you have feedback.

Hey Hauke, really interesting and important - thanks!

 

A quick note:

The next $1k/cap increase in a country at $10k/cap is worth only 1/10th as much as in a country with $100k/cap, because, the utility gained from increasing consumption from $1k to $10k is much greater than the utility gained from increasing consumption from $11k to $20k, even though the absolute dollar amounts are the same.

The claim in the part I bolded should be reversed, no? E.g. Increasing $1K where GDP/cap is $100K is worth 1/10th as much...etc.

good catch! fixed this it should be:

"The next $1k/cap increase in a country at $10k/cap is worth 10x as much as in a country with $100k/cap, because, the utility gained from increasing consumption from $10k to $11k is much greater than the utility gained from increasing consumption from $100k to $101k, even though the absolute dollar amounts are the same.

Curated and popular this week
Relevant opportunities