Monday, September 17, 2012

Can catch-up growth take us to the stars?

Will our civilization ever be able to colonize the stars and avert astronomical waste? Will we create computer programs more intelligent and energy-efficient than ourselves, enabling much larger and smarter sapient populations? We don't know exactly how hard it will be to engineer interstellar probes, or build AI, and we probably won't be sure until we actually do so.

However, we can shed some light on the question of whether humanity will ever be able to colonize the stars by asking how existing methods and technologies could increase our capacities, if they were deployed widely and to their limits. Here's a thought experiment: if we imagine that we were magically frozen in roughly our current technological regime for a time, long enough for Malthusian population growth and competition, how much would our economic and scientific production grow? By Malthusian, I mean that population would keep increasing until food costs started to price people out of reproduction, with higher-income folk reproducing more, and institutions that lead to high incomes spreading through migration, imitation or conquest.

Below the fold, I consider several dimensions where existing systems could simply be scaled up to increase global output and R&D: bringing poor countries up to the standards of rich countries, increasing population, and increasing average human capital within countries to near the level of the best-endowed households. Collectively, I estimate they could increase global R&D efforts by more than one hundred fold.

Swedish Earth
Most obviously, the CIA world factbook reports that 2011 gross world product (GWP) per capita was around $12,000 (all PPP). In contrast, Sweden had a per capita GDP of $40,900, and is neither an oil-state nor an unrepresentative city-state. Bringing up the world average to Sweden's level would increase total GWP from $80.33 trillion to 273.91 trillion. Sweden spent 3.62% of its GDP on R&D in 2009, and if this was scaled up to the world, global R&D spending would rise to $9.9 trillion, from $1.276 trillion in 2009, an increase of almost eight times.

56 billion vegetarians
If we allow for increase in total population even greater gains are accessible, as the human population catches up with the maximum sustainable at our level of technology. In a Malthusian world where the economically productive population keeps increasing its demands will drive up the costs of food (until people can no longer afford to have more children than are needed to replace themselves). Higher costs would incentivize vegetarianism (animal feed consumes a vast portion of world crop production), the abandonment of wasteful biofuels, expansion of croplands, reduction in food wastage, and the use of expensive methods to increase yields. Existing selective breeding techniques continue to deliver substantial gains in yield, absent any further advances. Combined with bringing worldwide agricultural productivity up to frontier (e.g. U.S., Australian) levels, an eightfold population increase is plausible (see here for some more back-of-the-envelopes).

Household upbringing and genetics
Even within advanced countries, there are large variations in income, education, and intelligence that run in families, due to upbringing, social capital, genetics, and other factors. Even in Northern European countries, which have higher intergenerational income mobility than, for example, the United States, twin studies find a heritability of income close to 0.5. If embryo selection or other reproductive biotechnologies were used to enhance productivity (or even just donor sperm and eggs), making the typical genome as prone to productivity as the right tail of today's distribution, average income could easily be doubled. Malthusian competition would then strongly select for use of such enhancement, since food prices would in this model rise until only the enhanced could earn enough to afford them, and enhancement would become near-universal. Superior household upbringing practices would spread through differential reproduction and imitation.

This analysis is complicated by the fact that measurements of individual income and education neglect certain externalities. On the one hand, to some extent intelligence and education are used in zero-sum games to allocate resources or jobs to oneself from others. On the other hand, in international regressions variables like education and intelligence are associated with much larger variations in income than they are within countries. Part of this is doubtless reverse causation (increased wealth boosts education and intelligence), but part of it plausibly reflects positive externalities.

Moreover, to the extent that scientific progress depends disproportionately on the very able, e.g. those with 1 in 10,000 cognitive test scores, who are much more likely to achieve tenure at top universities or acquire patents, effects on scientific output would be larger than for typical income.

I may say more about these complications in future posts, but regardless, in combination with the above factors, this world would have more than a hundredfold the R&D spending of our own, with a superabundance of hardworking, well-educated geniuses to employ with those budgets.

The Industrial and Scientific Revolutions are a few hundred years old. The Earth has a billion years before changes in the Sun lay it to waste. While fossil fuels are being depleted, solar energy exists today at only a few times the price, and can last throughout that time. Increasing timescales by orders of magnitude, in combination with greatly increased R&D per unit time, makes it still more likely that if humanity survives it would be able to reach a 'posthuman' condition as Nick Bostrom defines it in his paper on the future of humanity.

Some caveats
This world would need to make greatly increased use of renewable energy, desalination, recycling, and the like. Catch-up growth has been erratic and uneven, although massive, over the last century of economic history. Populations are currently shrinking in rich countries while they increase in poor ones, and Malthusian pressures are affected by welfare states, migration policy, warfare, and other factors. Catastrophes or some processes of social decay could derail progress, and the artificial stagnation discussed above is unrealistic.

Nonetheless, this sort of analysis helps clarify the burden for claims along the lines of "sure, colonizing the galaxy is physically possible, but it almost certainly won't be achieved, not in a million years."

No comments: