Niels Bohr once said: “It is difficult to make predictions, especially about the future.”
I say: Prediction is not about the future, it’s about the present.
Now that I’m fairly old, I can look back on a wide range of visions of the 21st century, many of which now seem obsolete. Here are just a few examples:
I recall the 1960s as being a period of techno-optimism. There were visions of a 21st century filled with space travel, supersonic airliners and flying cars.
The mood became more pessimistic during the 1970s. The famous Club of Rome report warned of overpopulation and resource depletion.
In the early 1990s, there was optimism that “The End of History” would usher in an age of peaceful democratic capitalism.
Late in the 1990s, worry increased that the 21st century would see dramatically rising global temperatures.
In the early 2000s, “The Clash of Civilizations” view became trendy, with a special focus on the clash between the West and the Muslim world.
Around 2010, there was an increasing view that this would be the Chinese Century.
A few years later, there were increasing fears of secular stagnation.
Today, people worry about the rise of nationalism, falling birthrates, and fear of unaligned AI.
What do all of these visions have in common? The only unifying thread that I can see is that they all reflect what was going on at the moment. That does not mean that they were all incorrect, rather that they correlate much more closely with conditions at the time then they do with conditions today.
Consider what was occurring when these predictions were made:
The 1960s saw transportation technology advancing dramatically over previous decades. It was easy (but wrong) to extrapolate those trends into the 21st century. Technology did continue to advance, but often in unexpected ways (biotech and computer chips.)
In the late 1960s, global population growth reached a peak of slightly over 2%/year, a rate that we may never see again. And the early 1970s saw disruptions in the supply of food and energy.
The early 1990s saw the collapse of communism in Eastern Europe and the Soviet Union.
By the late 1990s, there was increasing evidence of an upward trend in global temperatures. (Some cold winters in the 1970s had led to fears of a new Ice Age.)
The famous terrorist attacks of 9/11 led to a general view that we would see many more such attacks in the near future. The public felt quite vulnerable.
Expectations of a Chinese century peaked at the end of a multi-decade period of double-digit Chinese economic growth.
Fears of secular stagnation developed after a sluggish recovery from the Great Recession and a long period of near zero interest rates. Even I expected interest rates to stay fairly low.
So what’s in the news today? We see reports of a dramatic fall in fertility. We see news of a rise in nationalism. We see warnings that AI poses an existential threat. If there’s a unifying thread in all of these concerns, it might be described as “anxiety over the Great Replacement.” People worry that sharply declining birthrates in advanced countries will lead to our replacement with cheap labor from poor countries and increasingly sophisticated robots.
But how long before these current anxieties are replaced by an entirely new set of worries?
I won’t try to predict the future course of the 21st century, but I will try to predict the future course of 21st century predictions. Ten years from now, I expect the consensus forecast for remainder of the century will largely reflect the headline news stories of the 2030s. I predict that in the 2040s, forecasts for the remainder of the century will largely reflect the headline news stories of that decade. Ditto for the 2050s. What will those news stories be? I have no idea. But I strongly believe that predictions of the future will continue to reflect current events, and will mostly not reflect the actual future that plays out over time.
In other words, expect the unexpected. And then expect people to assume that the unexpected will become the new norm, until it is superseded by some other unexpected event.
READER COMMENTS
Richard W Fulmer
May 12 2024 at 6:37pm
In other words, people often try to predict the future by making straight line projections of current trends.
Early in the “electrical age,” there were concerns that the city of Chicago would soon need to be covered with power plants to meet projected demand. The city was “saved” by the introduction of turbines, which had much higher power-to-size ratios than did piston engines.
Similarly, someone back around 1900 calculated that everyone in America would need to become a telephone operator to meet rising demand for phone service. In effect, that’s exactly what happened. Dial telephones shifted the ability to connect phone calls from operators to callers.
New technology can redirect trends in unexpected ways.
Mark Z
May 12 2024 at 8:28pm
One could tell the exact opposite story with a few different anecdotes. Some scientists understood the greenhouse effect way back in the early 19th century. In the late 19th century Svante Arrhenius was already making reasonable predictions about how much the temperature would warm in response to increasing CO2. Fertility rates have been declining for decades – in some places maybe centuries – and public concern over declining fertility is quite old. It was one of the major concerns of the fascist movements in the 1920s. for example. It’s not some new discovery by demographers of the 2020s.
If you don’t have a good argument for why a troubling trend at the moment won’t abate reverse any time soon (I would include low fertility here), then cherry picking a few instances where people worried about some trend that later stopped or reversed just isn’t a substitute. There are also plenty of examples of initially worrying trends enduring for centuries.
Scott Sumner
May 12 2024 at 10:58pm
The last few years have seen a dramatic fall in fertility in many countries, and a dramatic increase in concern about this issue. Between 1980 and 2008, US fertility was trending slightly higher, and there was relatively little concern expressed about the issue.
https://www.macrotrends.net/global-metrics/countries/USA/united-states/fertility-rate#:~:text=The%20fertility%20rate%20for%20U.S.,a%200.11%25%20increase%20from%202020.
“If you don’t have a good argument for why a troubling trend at the moment won’t abate reverse any time soon ”
I think you misread the post—I never made any such claim.
Ahmed Fares
May 12 2024 at 9:01pm
France’s baby bust
Another good article below by David P. Goldman that makes the same case, i.e., fertility correlates with religiosity.
The Fecundity of Faith
Rajat
May 13 2024 at 8:19am
I don’t deny that the US is a leader in global thought-trends and so may be ahead on a few things – and that AI in particular could upset many forecasts – but when I think about how I would predict the remainder of this century, I think I would largely draw on all the fears you identified as emerging over the last 25 years or so. That is, I expect: (1) Rising global temperatures, and probably to higher levels than are being aspired to (although I accept that there’s a chance that geoengineering will save the day); (2) An ongoing ‘clash of civilisations’ between the non-Muslim and Muslim worlds (as has been the case for centuries); (3) China dominating the world economy in aggregate size if not in GDP per capita; (4) Ongoing secular stagnation, if measured by rates of real economic growth rather than by real interest rates. In other words, I expect governments to continue to reallocate societal resources away from productive activities and towards unproductive activities; (5) Falling fertility and populations. There’s also a bit of the Lucas Critique or NGDP futures-targeting critique one could make about your claim in this post, in that both policymakers and markets react to the concerns of the day to help ameliorate and profit from them (via votes or money, respectively). Therefore, it tends to happen that the most salient concerns of a particular time (like predictions of a major depression or hyperinflation) don’t come to pass because they contain the very seeds of their own diminution.
Scott Sumner
May 13 2024 at 12:58pm
I don’t see it as just a matter of will or will not certain trends continue, but rather will these be the dominant trends of the 21st century. I forgot to mention that in the 1980s there were predictions that Japan would overtake the US, at least in per capita terms.
James H
May 13 2024 at 11:21am
Just want to be pedantic and note that the quote is misattributed to Niels Bohr. It is from Danish humorist Storm P.
It’s Difficult to Make Predictions, Especially About the Future – Quote Investigator®
Trina Halppe
May 13 2024 at 11:22pm
What do you think of as a low interest rate?
Rates are not as low as the near zero rates of 2015 but still kinda low. The Bernanke talk about the global saving glut was in March 2005. So March 2005 marks a somewhat arbitrary benchmark for a low interest rate. The 10 year TIPS rate was 1.8% that month and the 20 year TIPS rate was 1.9%. This month the 10 year TIPS rate is 2.2% and the 20 year TIPS rate is 1.4%. So one of them increased 40 bp and the other decreased 50 bp. On those numbers, the global saving glut still exists and the past 19 years didn’t prove him wrong.
World population is still growing even with the worry about declining fertility. Even if the world population starts to decrease I think most people would still expect global savings to grow for several years because of the increase in the average income and average age and lifespan.
As for AI, I think hourly productivity will increase but per worker productivity will increase much less than than that because we will want to take some of the benefit in leisure. Some people will start working at an older age and stop working at a younger age so per person productivity will be less than per worker productivity. So ironically, the more effective the AI, the bigger the gap between how amazed we are and how much per capita GDP grows.
Scott Sumner
May 14 2024 at 8:19pm
I agree that rates are comparable to 2005, but they are higher than during the 2010s.
David S
May 14 2024 at 3:37am
One area of prediction that has fallen off the radar is risk of nuclear conflict at any scale. All the crowing about AI and climate change distracts from the simple fact that if Putin decides to make just one phone call then we’re all in a terrible place. The brutal logic of MAD still gives us some protection from a total exchange, but it’s pretty easy to imagine scenarios of more limited use by Putin, or Xi, or Iran, or the Orange Messiah—although in the case of Trump I like to think that there’s a better chance that some military office will have the guts to stop him.
Thomas L Hutcheson
May 14 2024 at 12:13pm
No predictions are easy but at least one can discount non-conditional predictions, those that do not specify what policies are expected to be in place that are consistent with those productions.