Does the theory of comparative advantage apply to horses?

by on December 5, 2012 at 12:13 pm in Uncategorized | Permalink

And if not, why not?  Where exactly does the model fail to apply to our equine friends?

Beginning in the late 19th century, and with increasing mechanization in the 20th century, especially following World War I in the USA and after World War II in Europe, the popularity of the internal combustion engine, and particularly the tractor, reduced the need for the draft horse. Many were sold to slaughter for horsemeat and a number of breeds went into significant decline.

The link for that is here, and for the pointer I thank Braden Anderson.

Brock December 5, 2012 at 12:19 pm

They could gain from trade, but not enough to pay their cost of upkeep. Like mine workers who end up deeper in debt to the company store each year.

Spencer December 5, 2012 at 12:21 pm

In the US before the horse was replaced by the tractor farmers had to allocate some 20% to 25% of their cropland to producing feed for their draft animals.

So a significant part of the sharp jump in farm productivity in the early 20th century may have only
reflected the shifting of this land from feeding horses and oxen to producing feed grains for people.

Cycles December 5, 2012 at 1:00 pm

And now farmers allocate cropland to produce ethanol fuel for their combines. The wheel turns.

tomhynes December 5, 2012 at 12:24 pm

I check MR to be intellectual, and but I usually have a time wasting reddit window open. After clicking on the link, I became very confused.

Peter Schaeffer December 6, 2012 at 1:30 pm


You are correct. The first link points to a reddit thread that starts with

“Trying to hire high-skilled workers at rock-bottom rates, is not a skills gap.”

Alex Godofsky December 5, 2012 at 12:24 pm

Where from the quoted section do you get the idea that comparative advantage doesn’t apply? Previously one of the applications of horses was for breeding more horses; when that became less valuable, horses were reassigned to other duties, like being food.

Alex Godofsky December 5, 2012 at 12:26 pm

Oh, I see the context now from the link. My response now looks very different.

gwern December 5, 2012 at 3:51 pm

Indeed! Of course, software doesn’t eat flesh, but that just raises the question: what are humans’ comparative advantage when we are out-thought and aren’t even worth eating…?

Fred Thompson December 5, 2012 at 12:32 pm

Tyler, This comment is a little cryptic. Are you referring to the decline in the number of horses or to the number of breeds? In either case, why would you expect comparative advantage to trump absolute advantage?

Floccina December 5, 2012 at 12:59 pm

One of their comparative advantages tasting good enough to eat.

j r December 5, 2012 at 1:10 pm

I admit that I read this post, along with the links, several times and I still don’t know what it means. I’m also not sure that is a bad thing.

Are you testing this out for a final exam question?

Seth C December 5, 2012 at 1:39 pm

Horses are not our friends.

dearieme December 5, 2012 at 1:53 pm

When the British Army left France and Belgium in 1918, did they sell their horses for the locals to eat?
They had already mechanised considerably so they presumably would have had lots of surplus horses.
Or perhaps the locals took advantage of lots of horses going cheap and carried on ploughing with them.

British schoolchildren who had never seen horsedrawn ploughing at home in the 50s used to be surprised to see it in action in Germany well into the 60s.

Rahul December 5, 2012 at 4:00 pm

Belgians love horsemeat don’t they?

dearieme December 5, 2012 at 5:25 pm

Rightly so, though I prefer kangaroo myself.

roystgnr December 5, 2012 at 3:14 pm

Every explanation of competitive advantage I’ve seen starts with workers’ production capabilities expressed as “X units of output A per hour or Y units of output B per hour”. In that situation any inequal ratio will mean that there’s a mutually beneficial trade of X for Y possible.

But what if your required input isn’t just “hours”, it’s “food” or “fuel” or some other limited fungible resource that has to be traded to the worker? I was born with a supply of hours I could work, but I wasn’t born with a supply of food to eat. If a robot can produce more of any X or Y per day than I could, using less energy than would have been necessary to produce my day’s meals, why would anyone with that energy trade it to me or to a farmer producing food for me, instead of to the robot?

Now, if everybody was sensible and if it was possible for everyone to diversify their total investments (*including* diversification out of their own “human capital”), then I’d say “bring on the robots!” – we’d all be richer in both tangible wealth and leisure time afterwards. But even if everybody was enough of a futurist to want to hedge against the endgame of the robotics revolution, it’s not possible to diversify out of your own human capital, for both legal and market reasons. And so every unexpected new development of work-saving automation is a loss of wealth to those people whose human capital value previously depended on their ability to fill that newly mooted need for work.

What’s been different in the past is that the phrase “depended on” has always been flexible – farmers could become factory workers, factory workers could become office workers, etc. Machines have become superior to humans at some tasks and humans have moved to other tasks instead, and it’s been a big win for everybody. But what happens when machines become superior to humans (or even just 20% of humans) at every task that other humans consider worth trading for food?

somaguy December 5, 2012 at 5:59 pm

“But what happens when machines become superior to humans (or even just 20% of humans) at every task that other humans consider worth trading for food?”

Maybe communism was just a century or so too early?

roystgnr December 6, 2012 at 10:18 am

The trouble is that, even if capitalism *stops* working (for a definition of “working” which includes “people without diversified capital or charity don’t have to find a peaceful place to lie down and die”), that doesn’t suddenly make communism *start* working. A century or so should have been enough to fix the system; instead it was just enough for a few dozen megadeaths, and the only “fixes” seem to have been of the form “become more capitalistic”.

Clark M December 5, 2012 at 4:40 pm

In a similar vein, FDR created the federal Drought Relief Service during the Dust Bowl, which aimed to relieve farmers of their starving herds and help stabilize market prices. The agency’s most common solution was to buy the cattle from the farmers, slaughter them, and bury them in a ditch.

Anonyxmousse December 5, 2012 at 4:47 pm

People have much farther to fall than horses before they run out of marketable skills, especially in serving other humans. There are a lot of services that most humans prefer be done by other humans vs. “capital” ie robots. E.g. childcare, cooking, haristyling, etc. Imagine a world where the smart people make lots of money and everyone else is their servant, like back in the days of the feudal lords – this is already happening to some extent – cf Manhattan banker’s families with nannies, cooks, house managers. Also, a lot of people will always be good at selling things & services the rich people like to buy – this already happening now with the increase of the tourism & artisan economies all around the world. A lot of locally produced art, food, clothes, and crafts are only mediocre but people like having them be locally produced in great varieties, wherever they happen to be. The rich will continue to pay for more videos, music, writing, sports, and games, created by other humans, as more people will have jobs in some form of entertainment, also already happening. We should be OK unless robots can start coming with new ideas, funny jokes, art, and/or business plans all on their own. Also, the welfare state will just grow, expanding along with our ideas about who is unable to work because they are ‘disabled’.

Slocum December 5, 2012 at 5:22 pm

“Also, a lot of people will always be good at selling things & services the rich people like to buy – this already happening now with the increase of the tourism & artisan economies all around the world.”

Yes — Neal Stephenson covered that nicely in ‘The Diamond Age’:

“Well, each phyle has a different way, and some ways are
better suited to making money than others, so some have a lot of
territory and others don’t.”
“What do you mean, a different way?”
“To make money you have to work hard—to live your life in a
certain way. The Atlantans all live that way, it’s part of their culture.
The Nipponese too. So the Nipponese and the Atlantans have as
much money as all the other phyles put together.”
“Why aren’t you an Atlantan?”
“Because I don’t want to live that way. All the people in
Dovetail like to make beautiful things. To us, the things that the
Atlantans do— dressing up in these kinds of clothes, spending years
and years in school—are irrelevant. Those pursuits wouldn’t help us
make beautiful things, you see. I’d rather just wear my blue jeans
and make paper.”
“But the M.C. can make paper,” Nell said.
“Not the kind that the Atlantans like.”
“But you make money from your paper only because the
Atlantans make money from working hard,” Nell said.
Rita’s face turned red and she said nothing for a little while.
Then, in a tight voice, she said, “Nell, you should ask your book the
meaning of the word discretion.”

Dismalist December 5, 2012 at 4:50 pm

Too much of comparative advantage lately: Comparative advantage exists given technology. When technology changes, comparative advantage changes. I guess the terms of trade moved against the horse. So what?

Slocum December 5, 2012 at 5:16 pm

Well, obviously it’s always been the case that some people are unable to produce enough value to pay for their keep (small children, the elderly, the disabled) for example. But humans are much more adaptable and much less costly to feed than draft horses. And it’s just not true that ‘A computer can do every job a human can do except make decisions.’ It’s not even close to true.

So Much for Subtlety December 5, 2012 at 5:43 pm

I would like to find a Western worker who was as cheap to feed as a horse.

I also agree it is not true that computers can do everything humans can. But they are working on it. We will have working sexbots any day now.

Xiaoding December 5, 2012 at 8:06 pm

“And it’s just not true that ‘A computer can do every job a human can do except make decisions.’ It’s not even close to true.”

What do you do?

I work with tech, all day, and I can tell you, it’s very, very close.

Surgeons, truck and taxi drivers, road workers, roof workers, crop pickers, factory workers…all out of work within ten years. It’s happening right NOW.

Martin Ford December 5, 2012 at 9:59 pm

I actually wrote a blog post on comparative advantage vs. machines a while back and used the example of oxen v. tractors.

The post is here:

Not sure if I got it right, but my main points were these:

1. Comparative advantage works in a practical sense only if the absolute advantage is not too extreme. In other words, you have to “get in the ballpark” in what ever area you are “least bad” at. Otherwise, comparative advantage may be correct in theory but mean little in practice.

2. It matters that machines/technology can be easily replicated. Comparative advantage is tied closely to opportunity cost. If a person or nation chooses to do one thing, then they must necessarily forgoe doing something else. Hence the incentive to specialize. This is not so true for machines/technology. Tractors, for example, can be replicated. If you have a large farm, you do not have to choose where to use your tractor; you can obtain as many as you want. If there were some limit on the number of tractors that could be built/sold then we might expect to see horses/oxen still used in developed countries (perhaps on smaller farms, where they might enjoy a comparative advantage.)

This is complicated by the fact that replicating tractors is not free and requires resources. Not sure exactly how to factor that in. On the other hand, many technologies (software, for example) can be replicated at very low marginal cost.

If, someday, artifical intelligence matches or exceeds the capability of the average worker, and if it can be cheaply replicated and scaled out (almost certainly that will be the case), I do not think comparative advantage will guarantee most people a job or a meaningful income. That will be a problem.

The Economist (Free Exchange) disagrees, however:

I think this will be hugely important question at some point….

chuck martel December 6, 2012 at 12:27 am

Draft horses, and horses in general, have significant health issues, some that have had major economic and political impacts historically. Marlborough’s victory over the French at Blenheim was due in part to the rampant disease glanders among the French horses and the Panic of 1873 was precipitated by the Great Epozootic of 1872, when almost every horse in the US and Canada became infected with equine influenza, bring much of land transport to a halt. While tractors might bread down, horses keel over and die. At this very moment, race horses that have been stabled at Hawthorne race course in Chicago are being denied admittance to the Fairgrounds in New Orleans and other tracks because of an outbreak of equine herpes.

Curly December 6, 2012 at 2:08 am

If they had a comparative advantage, they were insufficiently smart, knowledgeable, or motivated to discover what it was. That information isn’t free, and they and we were unable or unwilling to pay the price of discovering it.

Chris Prottas December 6, 2012 at 8:58 am

The great horse recalculation with horse breeds understood as skilled labor with very sticky skills. Sure, ominous in its implications for short-term fate of existing mature labor force (but did we need horsemeat to tell us that?)

Saturos December 6, 2012 at 9:43 am

“Where exactly does the model fail to apply to our equine friends?”

Oh, just little things, like consent…

Comments on this entry are closed.

Previous post:

Next post: