Knowledge Spillover: The Boon And The Bane Of Technological Progress

When I was a child, I was fascinated by motorsports. If you think about it, there is nothing unique about this situation; young boys are fascinated by race cars and race bikes more often than not. However, things started getting more interesting as I transitioned into a teenager.

I began understanding and appreciating the technical details behind the spectacle — the race to eke out every micro-second from the raging machinery, the electronics, the hydraulics, the aerodynamics, the strategy, etc.

As I grew into a young adult, it dawned on me that I could actually make more out of my passion for motorsports. I tried getting involved in the field as a professional and succeeded. Over the next few years, I was often involved in local racing events in one way or another.

Slowly but surely, it became clear that the local motorsport scene was nowhere close to the level of motorsport that I used to watch and admire on television as a child. In other words, I was far from state-of-the-art. I had to make a decision.

To pursue my career further in motorsports, I had to relocate to Europe, where all the tasty action was happening.

I will continue this story further in the essay. But for now, I wish to linger on a subtle point here.

Why is it that I had to consider relocation to pursue my passion?

Could the state-of-the-art not make its way to my local motorsport scene? Was that impossible?

Well, at least part of the challenge comes from a phenomenon known as knowledge spillover — the central topic of this essay. I will start by describing the origins of this concept, and then proceed to explain its evolution and implications on technological progress. Finally, I will touch upon how one could take advantage of this phenomenon and avoid its shortcomings. Let us begin.

The Origins of Knowledge Spillover

As far as I could track, the notion of knowledge spillover traces back to an English economist named Alfred Marshall, who theorised in the late 1800s how knowledge transferred between professionals.

Later on, economists Kenneth Arrow (whose work I am a fan of) and Paul Romer extended this notion and proposed that knowledge spillover happens between organisations within close physical proximity in a common industry.

Urbanist Jane Jacobs and Co. argued that knowledge spillover happens in between firms of different industries within close proximity as well.

While all of this might sound theoretical, these folks were trying to build a model to explain observations from the real world. So, what exactly did they observe? I should know from my practical experience, which ties back to the story that I paused earlier.

---

The Clustering of Knowledge

I mentioned that I had to consider moving to Europe if I wanted to pursue my career seriously in motorsport, as all the tasty action was happening there. What kind of action are we talking about here?

Well, for starters, if one wished to work in Formula 1, many Formula 1 teams back then were (and even now are) located in the United Kingdom within relatively close promixity to each other. This region is unofficially known as “Motorsport Valley” — the home of Red Bull, Mercedes, McLaren, Williams, Lotus, Force India, etc.

All of these teams conducted technical research in collaboration with local universities in the region and hired promising students via internships and work placements.

The effects of this reached beyond just motorsport. Famously, the Bloodhound SSC project that aimed to design and produce a land-based vehicle capable of travelling at 1000 Miles per hour happened at the Swansea University (among others; the project has had a troubled history of funding and marketing).

All of this knowledge could, in theory, diffuse away from this region. But it didn’t/doesn’t quite happen. Why is that? Well, one reason is how motorsport works — it is a sport, after all. This means any knowledge that could lead to a competitive advantage remains relatively secretive.

Even when employees leave a company for another, they get put on long “gardening leaves”, non-disclosure agreements, etc. And when they do switch, where do they move to? Their new employer is just across the turf within the valley.

I hope you can imagine how the term knowledge spillover comes into context here. But before we generalise this notion, the keen reader might note that Formula 1 is not the entirety of motorsport.

Apart from the United Kingdom, Germany has a strong motorsport/car culture going for it as well. Big brands like Porsche, Daimler/Mercedes, BMW, Audi, etc., are located in Germany and similar to the F1 teams, these companies do their cutting-edge technical research in collaboration with the local universities — Stuttgart, Esslingen, Aachen, Karlsruhe, etc. As you can imagine, these manufacturers have their own supply chain of interns and work placements from these universities as well. These feed into strong local German racing series like the VLN at the Nürburgring Nordschleife, DTM, etc.

We have so far covered the United Kingdom and Germany. What about their friends in Italy? Well, there exists a wealth of carbon fibre expertise in Italy — the likes of which service the motorsport needs of the United Kingdom and Germany. And of course, how can I forget Italy’s national pride? Ferrari.

And all of these clusters together power world series like the World Endurance Championship (Le Mans & co), World Rally Championship (even mighty Asian manufacturers like Hyundai and Toyota have their motorsport headquarters in Germany), etc.

I mentioned earlier that one of the reasons why knowledge doesn’t spread from these clusters is because of motorsport’s secretive nature. But what about other reasons? Let us look at another important reason next.

Network Effects

I just described how each motorsport cluster has its own supply chain of talent from local universities via internships and work placements. Does this mean that it is impossible to enter these clusters as an outsider?

Well, no. But it is significantly harder. And even when someone gains “insider” knowledge, it is hard for that knowledge to diffuse to other regions/clusters.

Technological progress often boils down to getting even the minute details right. And with the specialised nature of most jobs and roles these days, without the strong network integration offered by an established knowledge cluster, inefficiencies start seeping in and leading progress astray.

This essay is supported by Generatebg

Again, this is not impossible; it is just significantly harder. On the contrary, the network effects ensure that it is significantly easier to progress faster by relocating to or near an already established knowledge cluster; the knowledge spillover is more efficient and happens faster. This is the reason why even strong automotive manufacturers like Hyundai and Toyota have their rally divisions based in Europe.

Now, to drive this point home, let us venture outside of motorsports and generalise this phenomenon. The world-famous Silicon Valley was a deliberate attempt from silicon-based chip manufacturers back in the day to establish a kowledge cluster for efficient knowledge spillover. Although this worked brilliantly, it is ironic that the present-day silicon-manufacturing knowledge cluster is largely concentrated in Taiwan (with TSMC).

However, the old Silicon Valley remains famous for something else — technological progress via tech startups. Because of the current knowledge spillover and clustering in this geographical region, startups in this region tend to have an edge. This is why many venture-backed startups go to great lengths to relocate to this region.

Moving on to another example, back in the day, if you wanted to “make it” in the field of world english cinema, Hollywood was the place to go to. As a result, a strong network of artists and professionals existed in this region, which in turn attracted more talent to the region like dominoes. This made the knowledge spillover and access to talent and resources very efficient and concentrated. I go could go on and on, but you get the picture.

So far, what I have established is that knowledge spillover is governed strongly by network effects. While this helps faster technological progress, it also locks/restricts this progress to local clusters.

What can one do if one is not a part of this cluster? Well, it is time to turn once again to my personal experience.

---

Swim Against the Current — Yes or No?

I worked in motorsport for a few years and decided to move on. And when I did, I moved into software development. I wanted to found a tech startup and realised how much harder it is to swim against the current.

The lack of a knowledge cluster meant that I had to make mistakes that were probably made by millions of human beings before me, yet not communicated to me. In other words, valuable time was lost. In startup-land, this meant slower progress.

Relocating to silion valley would have helped, but that would have also come at a huge cost. So, what can one do in such a situation? Well, my answer was to not swim against the current, but improvise.

Luckily, we live in an era where digital networks reign supreme — cue in forums, social media, and whatnot! All I had to do was shift my mindset from playing catchup to actively being resourceful. It took a while, but I eventually found myself in the midst of all of the progress. A few years back the state of the art moved via scientific papers/research submitted to journals.

Had someone told me that scientific research in the future would happen via discord channels or twitch streams, I would not have believed it. Yet, here we are.

Now, don’t get me wrong. A tech startup in Silicon Valley happens to have access to all these opportunities just as a tech startup elsewhere. So, there is an inherent loss in edge by not being in a physical cluster where knowledge spillover is definitely faster and more efficient. But the situation is not as dire as before if one is willing to be more resourceful.

Having said that, knowledge spillover and clustering does have some downsides. This is when it might be well worth swimming against the current.

For instance, I firmly believe that a lot of the present-day research, funding, and technical resources are overinvested at the time of writing this essay in the field of Artificial Intelligence (if you are reading this essay at a later time, I am pretty sure that there is a global trend hogging all the glory in your timeline).

While I believe that Artificial Intelligence is genuinely a good thing in the long run, if you or I have deep knowledge of a tech branch outside the scope of Artificial Intelligence, I would wager by staying far away from the aforementioned tech clusters. If you or I were part of the cluster and were highly optimistic about what we’re working on, we are likely to hear something along the following lines from our peers:

“Really? Don’t tell me you haven’t thought about integrating AI into your tech yet. Surely, you are missing out!”

This in turn might seed doubts in our minds. This is how networks work; they accelerate knowledge spillover and progress, but they also tend to increase and accelerate bias towards trends.

If what you are working on has little to do with the hottest “trend” in your core field or industry and you are confident about your expertise, I believe it is in your best interest to swim against the current.

To be fair, it is difficult to know beforehand and easier to know in hindsight. Either pathway could be a mistake; we simply cannot be certain. Again, this is where I feel today’s digital clusters/networks could be an advantage. If tuned properly, they give us the option to moderate the bias towards trends from knowledge spillover.

There is a large caveat here, though. The “if tuned properly” part is no walk in the park! Today’s social media and network-based resources are designed to be predatory. So, one needs to be careful. I could go deeper into this, but I feel that it is a topic for another day.

For now, I hope you found this essay informational or entertaining — or perhaps both!

---

If you’d like to get notified when interesting content gets published here, consider subscribing.

Further reading that might interest you: 

• How To Understand Computational Irreducibility?
• Polls Mislead Most Of The Time. Here Is Why?
• Fermi Problems: How To Deal With Huge Numbers?