We did not turn horses into glue
Addressing a recent, popular misconception
There is a common misconception promoted by certain experts that due to the adoption of the automobile, horses became obsolete and so we sent them to be rendered into glue. This did not happen.
For an example, see Noah Opinion.
Specifically Noah’s claim is referring to horse pulled transport for cities. The type of which produces mounds of poop.
The problem with his claim is that it’s wrong. It did not happen.
Horse population decline happened because we stopped breeding so many horses!
What actually happened to horses:
- Horses got different jobs.
- A lot of them died in the world wars.
- A lot were worked until the end. Then they were sent for rendering.
- A lot of them were slaughtered for meat by extremely hungry people.
- The horse population is now growing rapidly.
In France, even after widespread adoption of automobile transport in cities, the horse population grew!
Where does this misconception come from?
This misconception comes from the phrase or a phrase like it about old horses being turned into glue.
This is a figurative expression, meaning something that departs from the literal (what actually happened) to create vivid imagery to emphasize a deeper point. In reality, horses unfit for work were mostly slaughtered.
Using Google Search with time filters, at some point this expression mutated in the common consciousness to the current usage of:
“businesses didn’t keep using horses for transporting goods once trucks were invented, even though horses had a comparative (not absolute) advantage in pulling things. Horses were sent to the glue factory. Not a pleasant thought for human workers.” https://www.nytimes.com/2024/03/22/opinion/ai-jobs-comparative-advantage.html
Which is a literal expression. The invention of trucks meant we sent horses (healthy, prime age by implication) to the glue factory. Followed up with: “not a pleasant thought for human workers”.
Which is wrong. This did not happen.
Who is responsible?
Tracing it back, the switch from the figurative to literal (incorrect) expression is probably due to Tyler Cowen’s 2013 blogpost: Why the theory of comparative advantage is overrated.
Note that Tyler is using the figurative here! ‘so to speak’. So, he is correct. It’s people quoting him, like Noah and the nytimes, who are incorrect.
Why the horse population declined
I’ll expand on this later but the three word summary is: War, and Stalin.
Horse population peak by country
I read a bunch of house commons reports, newspaper clippings, etc. to get basic stats for when horse population peaked by country. The data is not great, for which I blame communists and fascists.
- US (25mn) peaks in 1915.
- UK (3mn) peaks in 1914.
- France (3mn) peaks ~1930.
- Germany (4mn) peaks either ~1930 or 1939.
- USSR (30mn) peaks ~1920s.
- Argentina (4.5mn) peaks ~1960s.
- China (11mn) peaks ~1985.
Note: I use ‘horse’ loosely here. The stats are terrible. Usually it means horses, mules, ponies, etc. I apologize. Limitation of the data.
Why the horse population declined by country
- US exported a lot of prime aged horses to Europe in WW1. The rest seems to be aging with little breeding due to fast mechanization.
- UK exported a lot of prime aged horses off the Home Island in WW1. About 500k died in the war, the remainder were left where they were. So many horses were exported, some breeds became endangered. I believe this sped up mechanization due to absence of horse labor.
- France lost a lot of horses in WW1. Rebuilt their numbers to a peak ~1930 as shown by the 1927 horse census. Maintained numbers through breeding until WW2. Decline then comes from war, and hungry people wanting meat. Rapid later decline due to little breeding, during and post WW2.
Note that this graph, while interesting, is misleading due to missing data points. The horse population of France declined dramatically from 1914 to 1918, before rising again.
- Germany lost a lot of horses in WW1. They also stole a lot of horses, and disrupted breeding programs such as the Belgians. Germany has historically subsidized horse production, and it seems as though the Nazis expanded horse breeding after coming to power, although I can’t find a clear source for the latter. Postwar decline due to low breeding, and slaughter for meat for hungry people.
- USSR decline is Stalin. The numbers are awful, both in the sense that I don’t trust them and that the horse population went from ~30mn to ~13mn due to mismanagement of the economy. Looking at horse population trajectory in other countries, both developed and less developed, had it not been for Stalin’s bungled collectivization, the horse population of the Soviet Union would likely have risen substantially. 50mn+ horses by 1950 seems plausible to me.
- Argentina. I have no idea what happened here. They seem to have peaked very late, primarily raising horses for meat. The decline (I think) is due to horse meat falling out of fashion, and less breeding.
- China. Peak in the 1980s is due to getting richer, which trends with more horses, best viewed as useful inputs for agriculture, decline after 1980s is due to getting much richer, which trends with mechanization. I’m not sure what happened to the horses here, whether aging or slaughter. I tend towards slaughter based on my personal bias.
Misc
Horse population an all time peak
There are about 10 million horses in the US today. We are already back to 1870 numbers, and more than twice as high as 1840 numbers.
Horse R&D Interrupted by War
The productivity of horses is not static. Equipment to use horses has improved over time, and horses themselves have been breed to be much more useful.
Get John Friedman’s stories in your inbox
Join Medium for free to get updates from this writer.
This is apparently the type of horse the Mongols (1300s) rode.
This is the Cleveland Bay, almost wiped out by World War One.
World War one interrupted horse breeding programs.
Horse breeding farms are important, because they create specialized horses for specific tasks.
Motor was invented far before the automobile, and horse population rose
Horses were already being replaced in 18th century UK by mechanization. Horse population dramatically expanded.
Slow obsoletion?
Horses are useful! They are very good for agriculture and as China’s mid 1980’s peak shows, they remained useful agricultural equipment a century after the invention of the automobile. Even rich countries like France relied on them for agriculture well into the 1950s.
Recovering horse populations
Graphs are my own, as are mistakes.
Horses better off now than before
Horses are doing great. Their US population has 5xed in the past 60 years. Now, they do recreation instead of heavy labor as they did in 1920.
Retirement of horses from agriculture during the decline
From reddit:
Relevance to artificial intelligence discussion
Important figures keep claiming that horses were sent to the glue factory after the invention of the automobile, and comparing humans to horses in the age of artificial intelligence. Which people take to mean terminator style human genocide.
“if it all ends up where humans like horses end up in the glue factory, man…” (Thiel, 2024)
But this did not happen to horses. Their population is rising, and is higher then it was in the mid 1850s. They are treated better than ever.
Why is this misconception thriving?
It’s a good talking point. No one arguing about AI cares what actually happened to horses with regard to motorization. They just want to talk about the possibility of AI replacing humans.
Did I make mistakes here?
Yes. The data is awful. Also I got lazy and mixed certain definitions. I am hoping that some kind soul on the internet will complain about my data because I got something wrong. Then we can become friends, because I like having correct data.
For example, the title. YES WE TURNED HORSES INTO GLUE. But we did this to old horses after extracting many years of labor. I am hoping the reader can make the distinction between the AI pundits point and mine.
This seems a little pedantic?
Yes. I care very much about what actually happened, in great detail. Consider this my interview for any future data position.
