Recently I re-discovered an interesting concept called path dependence. It came to me after reading the following Tweet, by journalist Jon Stone:
Can you imagine the hysterical reaction if someone had suggested the creation of public libraries today. “For free? how are you going to pay for that, STALIN?”
Considering our current governments which are quite conservative, the idea of public libraries can be preposterous. Most of the parties in power (e.g. in the US or the UK) would not consider providing such a non-vital good (books in this case) to people for free. So how is it possible that public libraries do in fact exist, even though the reigning parties would not consider it good policy? One way of explaining this is with the concept of path dependence. There are a number of different definitions of path dependence, but the one I use here is the idea that a specific product, process or even specific beliefs are in place because of historic use.
This is an incredibly interesting concept, because it can illustrate why something is in place today (such as public libraries) even though it goes against what we currently think is a better alternative or even the best option. In this article, I’ll illustrate the concept of path dependence with three examples.
Path Dependence in Hardware: The QWERTY Keyboard
One of the most famous examples of path dependence is the QWERTY keyboard. Chances are, you are reading this article on a device that uses QWERTY; simply look at the first letters on your keyboard.
This keyboard was created specifically to fix certain errors present in earlier keyboards. These earlier typewriters could jam when specific keys were hit quickly after another — the QWERTY design, created at the end of the 19th century, positioned these keys away from one another, so that the jams didn’t occur.
The story goes that the Dvorak layout*, patented in 1936, was much more efficient than QWERTY. Mr. Dvorak specifically designed this layout for efficiency, and typists trained with Dvorak are said to be up to 74% faster than those using QWERTY.
If this is the case, why then do we currently use the QWERTY keyboard? The reason is path dependence. Once a few companies started manufacturing QWERTY keyboards, and a few people started using them, the standard became an actual standard. If you see people in your vicinity use a QWERTY keyboard, and it’s cheaper than a Dvorak one, the choice is clear.
Even if you know that a Dvorak keyboard is more efficient, it’s not just making the switch itself that is difficult. Certainly, learning to type on a new keyboard layout will be a challenge, but it is doable. However, the path dependence here comes more from network effects, i.e. that everyone uses a QWERTY keyboard. This means for instance that anytime you use a device of someone else, you’ll have to go back to your old QWERTY ways, which makes it even more difficult to make a permanent switch to Dvorak.
Interestingly, even though everyone uses QWERTY vs Dvorak as the main example of path dependence, I do want to let you know that it has received a lot of criticism. Already back in the 1990s, two academics published an article named “The Fable of the Keys“, which demonstrated that there were actually very few efficiency differences between the two keyboard lay-outs. But whether or not it is fully accurate, the story is interesting and does a good job at illustrating path dependence.
* Note that Dvorak is not fully capitalized here, because unlike QWERTY, it was named after its inventor August Dvorak.
Path Dependence in Urban Geography: Cotton in Lancashire
Another interesting example of path dependence I found relates to urban geography, and the UK cotton industry. Blooming in the 19th century, the cotton industry was mostly focused around Lancashire, a county near Liverpool and Manchester. There were two reasons for the cotton industry’s presence here: skilled workers and natural resources (water, coal) that were heavily used in the production process.
Interestingly though, according to economists Crafts and Wolf, skilled workers and natural resources could also be found elsewhere. In other words, there was no specific reason for the cotton industry to be in Lancashire. And perhaps other locations would be even better suited for cotton manufacturers and traders, such as those with better access to London and other capitals.
So the reason that cotton was situated in Lancashire, according to these two economists, was again this network effect. Once people started manufacturing and processing cotton in Lancashire, they attracted others. Suppose for instance that you were interested in taking part in the cotton industry, as a worker, a manufacturer or trader. Whatever your job was, it made sense to go to Lancashire and set up shop there — after all, that’s where you could find work.
And so again, just like with the QWERTY keyboard, this example illustrates that even though better potential options exist, a ‘lesser’ option is taken, and it becomes increasingly difficult to from this option to a better one.
Path Dependence in Business: Fossil Fuels
A last example I want to mention is that of fossil fuels. By now you should have a fair idea of what path dependence is, so you’ll probably see how fossil fuels are relevant here.
Over the years, it has become clear that renewable energy is the better choice. There is no end to these energy sources, and pollution is kept to a minimum. Nonetheless, fossil fuels currently still make up between 75% and 80% of our total global fuel consumption.
One reason for this is that fossil fuels were and still are a cheap source of energy. But considering the climate crisis and all the natural and societal effects fossil fuels (and CO2-emissions) have, there are obviously better outcomes. So the reason for the prevailing power of fossil fuels can be found in path dependence; if you’re someone looking for a job, or a company looking for a market to sell to, you’ll simply have better chances in the oil industry than in the wind farm industry. Or in other words, because initial investments in (extracting and using) fossil fuels were made, it has become easier for others to follow the same path.
Breaking Away from a Dependent Path
So what can we do ‘against’ path dependence? As you’ve seen, path dependence often means that there are better potential options, possibilities or outcomes, that we do not or cannot take. And while this may not such be a big issue for our keyboard habits or the location of the cotton industry, in the case of fossil fuel use, we need to find a way to deal with this.
In my view, there are a couple of ways to do so. First, just like when making unrestricted choices, become aware that you’re on a ‘dependent path’. This awareness should come with the realization that change can be incredibly hard. The unfortunate thing of path dependence is that the problem at hand is now ‘nobody’s fault’. After all, it often relates to in-action (e.g. keyboard users didn’t want to start using the DVORAK keyboard, and keyboard manufacturers thus didn’t want to sell them), and the challenges are enormous. So even for those who do want to make a difference, it’s incredibly difficult.
Nonetheless, change is possible. So second, in some cases there is a first-mover advantage. This means that who ever moves first has an advantage over those they are in competition with. Take the fossil fuels example; those who are the first to move to renewables will have an edge over the competition. This applies to business (e.g. Tesla), but also to governments; countries with the right infrastructure for renewable energy will be a step ahead and attract more investments and jobs.
And third, even when there isn’t a first-mover advantage, there may be first-movers. Consumers, governments and companies may take a specific option that is not the cheapest. You may start to drink oat milk, even though it tastes slightly worse and is more expensive. By making this specific choice, you show others what’s possible. Hopefully, this may lead others to follow, and help us break away from the path we’re currently on.
This reminds me of the exploration – exploitation balance one has to find in AI algorithms (a popular example being a Mars rover that needs to find food).
Once a food source has been found, it will be exploited and exploration might not happen anymore, even though there might be better options around.
Also reminds me of the small book Who Moved My Cheese? (https://www.amazon.com/Moved-Cheese-Spencer-Johnson-M-D/dp/0743582853), worth a read!