4 ms·

This is absolutely incorrect, and you have no idea what you are talking about. There is no sensitive chaotic model behind this, it's completely straightforward

This is absolutely incorrect, and you have no idea what you are talking about. There is no sensitive chaotic model behind this, it's completely straightforward dynamics. The fertility rate is about 2 kids per woman now. But the generation of 20 year olds is much larger than the generation of 40 year olds. So even though the generation of 0 year olds is the same size as that of 20 year olds, in 20 years time we will have replaced a small 40 year old generation by a large 20 year old generation. There are only two factors here: Fertility rate per woman, and death rate. This prediction is not terribly sensitive to either, and both are easy to measure accurately and on quite stable long term trajectories: https://ourworldindata.org/fertility-rate https://ourworldindata.org/fertility-rate

jacquesm 7y ago
> There is no sensitive chaotic model behind this, it's completely straightforward dynamics. Sorry for the poor choice of words, no chaotic model was implied or intended. > Fertility rate per woman, and death rate. No, there is also technological development, wars and so on. So any prediction that looks 80 years out that only looks at fertility rate and death rate is going to be found wanting. It's a bit long for a long bet, but I'd be more than happy to take one for a 30% or more deviation from the prediction. The degree to which people are willing to believe simplistic models without further qualification is astounding to me, given the past 80 years in development of medical technology, electronics and our ability to wage war any predictions about the next 80 years should come with huge error bars. And that's before we get into the consequences of Climate change.
- Certhas 7y ago
  > No, there is also technological development, wars and so on. Which have an effect on fertility and death rate. There is literally nothing that matters to population dynamics other than fertility and death rate. We have extensive experience on that now: Technological and social development drives fertility to at or below replacement levels rapidly. This is the main factor. Importantly though, _this is not a prediction_. This reduction has already happened. So the bulk of the prediction is just playing out the consequences of developments we've already seen. Of course major events like climate change have the potential to invalidate this prediction. But that doesn't mean this prediction is useless. And this is would not be due to "sensitivity to initial inputs".
  - jacquesm 7y ago
    > Which have an effect on fertility and death rate. Yes, but given that they are unknowns you can't just extrapolate from the presently available data and call it a day. > But that doesn't mean this prediction is useless. It's an open door that the world population will stabilize at some point, it is also an open door that the speed with which the world population is increasing itself is reducing. It has to, simply because more affluent people will have fewer children because they are not looking at their children as their source of income when they are themselves unable to generate any. Children as a pension scheme is a very common thing still, but that is fortunately changing. > And this is would not be due to "sensitivity to initial inputs". To a degree it is: any error in the predictions for the next couple of years will result in a much larger error in the future, simply because of the fact that population increase is a non-linear affair to begin with, and that any med-tech development that causes people to live longer will also cause a large increase in the number of people alive at any one moment. Compounded those two could have an enormous effect. For reference, the world, 80 years ago, so just before WWII had only 2.2 billion people.
    - xorfish 7y ago
      > Children as a pension scheme is a very common thing still, but that is fortunately changing. It is still the case, just on a larger scale.
      - jacquesm 7y ago
        Good point.
    - beat 7y ago
      It's more likely the rate of birth rate decrease will increase, rather than decrease. The countries hitting birth rate decreases over the last 50 years are seeing birth rate drops in a generation that took a century in Europe - and remember, not that long ago, Europe once had birth rates, illiteracy rates, and poverty as high as the poorest countries in the world today. Moreover, the major causes of birth rate decrease, like vaccination and refrigeration (which strongly impacts food supply), aren't going away. Look at a country with a high birth rate today, and you'll still find people walking for miles with buckets full of the day's water for their household. And anywhere that indoor plumbing and clean tap water is universal, the birth rate is down to replacement level. There are fundamental causes for birth rate decrease, and they're not going away. Frankly, I think you're clinging emotionally to the Malthusian model, not rationally.
Retric 7y ago
Look at the error bars for the U.N. estimates. 95% chance between 9 billion and 13 billion 80 years from now is not exactly a narrow band. https://population.un.org/wpp/Graphs/Probabilistic/POP/TOT/900 https://population.un.org/wpp/Graphs/Probabilistic/POP/TOT/9... One extreme unknown is global life expectancy. New diseases, or new cures for existing diseases can make a huge difference. Add in wars, economic growth, and a host of other factors making long term estimates are hard. Just consider how accurate past population estimates have been.
- Certhas 7y ago
  Actually my reading is that the uncertainty here is driven entirely mostly the uncertainty in the fertility rate. People living longer has a relatively linear effect on population, fertility is much stronger. By eyeball this uncertainty projection for fertility seems extremely high, given the relatively smooth and predictable decline we've seen in the past [1]: https://population.un.org/wpp/Graphs/Probabilistic/FERT/TOT/900 https://population.un.org/wpp/Graphs/Probabilistic/FERT/TOT/... I spend a bit of time looking for the publications underlying this model right now, but couldn't find any. But accepting the error bars doesn't change the story. Which is that this isn't that much of a prediction. The main driver behind the population dynamics of the next century is the decline in fertility that has already happened. It's not a highly uncertain prediction that can turn out to be vastly different based on minor changes in the initial conditions. But if we underestimate the decline of fertility in Africa (where it is changing the most at the moment [2]) over the next decade or two it could easily end up being off by 10% or more. It's all fairly predictable. [1] For example the 95th percentile prediction sees a sudden bump in Asia, taking it up to fertility levels not seen since the early 2000s by the mid 2020s, despite a perfectly smooth decline since: https://population.un.org/wpp/Graphs/Probabilistic/FERT/TOT/935 https://population.un.org/wpp/Graphs/Probabilistic/FERT/TOT/... [2] https://population.un.org/wpp/Graphs/Probabilistic/FERT/TOT/903 https://population.un.org/wpp/Graphs/Probabilistic/FERT/TOT/...
  - Retric 7y ago
    On it’s own WWII killed off around 3% of the world population. The plague may have reduced the world population from an estimated 450 million to 350–375 million in the 14th century. Long term estimates need to consider such black swans as a real possibility. Relative amounts of war and disease in Africa over the next 20 years will have a real impact.
mensetmanusman 7y ago
Less than 2 in most developed countries if you don’t count the newly minted citizens.

jacquesm 7y ago

> There is no sensitive chaotic model behind this, it's completely straightforward dynamics. Sorry for the poor choice of words, no chaotic model was implied or intended. > Fertility rate per woman, and death rate. No, there is also technological development, wars and so on. So any prediction that looks 80 years out that only looks at fertility rate and death rate is going to be found wanting. It's a bit long for a long bet, but I'd be more than happy to take one for a 30% or more deviation from the prediction. The degree to which people are willing to believe simplistic models without further qualification is astounding to me, given the past 80 years in development of medical technology, electronics and our ability to wage war any predictions about the next 80 years should come with huge error bars. And that's before we get into the consequences of Climate change.

Certhas 7y ago
> No, there is also technological development, wars and so on. Which have an effect on fertility and death rate. There is literally nothing that matters to population dynamics other than fertility and death rate. We have extensive experience on that now: Technological and social development drives fertility to at or below replacement levels rapidly. This is the main factor. Importantly though, _this is not a prediction_. This reduction has already happened. So the bulk of the prediction is just playing out the consequences of developments we've already seen. Of course major events like climate change have the potential to invalidate this prediction. But that doesn't mean this prediction is useless. And this is would not be due to "sensitivity to initial inputs".
- jacquesm 7y ago
  > Which have an effect on fertility and death rate. Yes, but given that they are unknowns you can't just extrapolate from the presently available data and call it a day. > But that doesn't mean this prediction is useless. It's an open door that the world population will stabilize at some point, it is also an open door that the speed with which the world population is increasing itself is reducing. It has to, simply because more affluent people will have fewer children because they are not looking at their children as their source of income when they are themselves unable to generate any. Children as a pension scheme is a very common thing still, but that is fortunately changing. > And this is would not be due to "sensitivity to initial inputs". To a degree it is: any error in the predictions for the next couple of years will result in a much larger error in the future, simply because of the fact that population increase is a non-linear affair to begin with, and that any med-tech development that causes people to live longer will also cause a large increase in the number of people alive at any one moment. Compounded those two could have an enormous effect. For reference, the world, 80 years ago, so just before WWII had only 2.2 billion people.
  - xorfish 7y ago
    > Children as a pension scheme is a very common thing still, but that is fortunately changing. It is still the case, just on a larger scale.
    - jacquesm 7y ago
      Good point.
  - beat 7y ago
    It's more likely the rate of birth rate decrease will increase, rather than decrease. The countries hitting birth rate decreases over the last 50 years are seeing birth rate drops in a generation that took a century in Europe - and remember, not that long ago, Europe once had birth rates, illiteracy rates, and poverty as high as the poorest countries in the world today. Moreover, the major causes of birth rate decrease, like vaccination and refrigeration (which strongly impacts food supply), aren't going away. Look at a country with a high birth rate today, and you'll still find people walking for miles with buckets full of the day's water for their household. And anywhere that indoor plumbing and clean tap water is universal, the birth rate is down to replacement level. There are fundamental causes for birth rate decrease, and they're not going away. Frankly, I think you're clinging emotionally to the Malthusian model, not rationally.

Retric 7y ago

Look at the error bars for the U.N. estimates. 95% chance between 9 billion and 13 billion 80 years from now is not exactly a narrow band. https://population.un.org/wpp/Graphs/Probabilistic/POP/TOT/900 https://population.un.org/wpp/Graphs/Probabilistic/POP/TOT/9... One extreme unknown is global life expectancy. New diseases, or new cures for existing diseases can make a huge difference. Add in wars, economic growth, and a host of other factors making long term estimates are hard. Just consider how accurate past population estimates have been.

Certhas 7y ago
Actually my reading is that the uncertainty here is driven entirely mostly the uncertainty in the fertility rate. People living longer has a relatively linear effect on population, fertility is much stronger. By eyeball this uncertainty projection for fertility seems extremely high, given the relatively smooth and predictable decline we've seen in the past [1]: https://population.un.org/wpp/Graphs/Probabilistic/FERT/TOT/900 https://population.un.org/wpp/Graphs/Probabilistic/FERT/TOT/... I spend a bit of time looking for the publications underlying this model right now, but couldn't find any. But accepting the error bars doesn't change the story. Which is that this isn't that much of a prediction. The main driver behind the population dynamics of the next century is the decline in fertility that has already happened. It's not a highly uncertain prediction that can turn out to be vastly different based on minor changes in the initial conditions. But if we underestimate the decline of fertility in Africa (where it is changing the most at the moment [2]) over the next decade or two it could easily end up being off by 10% or more. It's all fairly predictable. [1] For example the 95th percentile prediction sees a sudden bump in Asia, taking it up to fertility levels not seen since the early 2000s by the mid 2020s, despite a perfectly smooth decline since: https://population.un.org/wpp/Graphs/Probabilistic/FERT/TOT/935 https://population.un.org/wpp/Graphs/Probabilistic/FERT/TOT/... [2] https://population.un.org/wpp/Graphs/Probabilistic/FERT/TOT/903 https://population.un.org/wpp/Graphs/Probabilistic/FERT/TOT/...
- Retric 7y ago
  On it’s own WWII killed off around 3% of the world population. The plague may have reduced the world population from an estimated 450 million to 350–375 million in the 14th century. Long term estimates need to consider such black swans as a real possibility. Relative amounts of war and disease in Africa over the next 20 years will have a real impact.

mensetmanusman 7y ago

Less than 2 in most developed countries if you don’t count the newly minted citizens.