Attractive profits

“Attractive profits disappear at one stage in the value chain because a product becomes modular and commoditized, opportunities arise at the other end for proprietary products” [1]. Clayton Christensen became famous for his theories explaining disruptive technologies. In this article we will discuss how the market for railway control systems is a failed market, which cannot adopt disruptive technologies.

Christensen’s theory is that as a new product appears and functionality is still poor, the best companies rely on proprietary architectures. Its greater design freedom helps to improve the product’s performance. The upstream or downstream parts in the value chain are commoditised: cheap and adapted around the proprietary product.

Over time, this split of what is commodity and proprietary shifts along the value chain. Once product performance satisfies the customers, a modular architecture becomes the best solution. The lower prices and faster times to market compensate for the reduced design freedom. Then the nice profits flow to a critical component in the modular architecture, away from the integrators. A good example is the computer industry where the pendulum has kept swaying back and forth:

• mainframes and first micro-computers – proprietary
• personal computer – modular computer around a proprietary CPU (Intel)
• smartphone – proprietary around a modular CPU (Qualcomm or Apple)

In short: the market rewards the companies solving difficult problems. No company can sit on past merits, they must tirelessly look for what clients will need next, or else end up as an undifferentiated commoditised player.

Unfortunately, in the railway sector, business models have not changed for over 30 years. This endangers the technological evolution in the sector.

Systems architecture & software eating the world

“Software is eating the world” is a famous sentence in the Silicon Valley, pointing to how software is gradually becoming crucial to all economic activities. The beauty of software is that it crystallises the knowledge of an organisation or person, into something that executes a job. Software has large upfront costs for development but then costs nothing to replicate, thus the more an industry can reuse and share code, and so build upon the existing knowledge blocks, the faster it will progress. Unfortunately, in the rail sector, the adoption of open source or shared development of software is incipient.

Why does this happen? One reason is historic. With relay technology and older electronics, the functionality was hard-coded in the hardware, so hard and expensive to change. Monolithic systems as these lend itself to being built by just one company. And indeed, the main industrial companies were set up as integrated companies to produce an integrated product. The form fit the function. Modularity and upgradeability were not in focus then. Modern IT is different, behaviour is coded in software, and software is much more flexible, with architectures emphasising modularity and loose coupling, which enables reusability and independence. Nevertheless, integrated companies are optimal for integrated products, so that is what they build. In a new paradigm of modular architecture, integrated companies would have to change to remain competitive. So why should they help this change come about?

Indeed, after using computer-based systems for 30-40 years, monolithic systems still dominate in the railway. As railways are critical for society, governments pay what is needed to keep the core network going. State subsidies cushion market pressures. The problem with monoliths is that reuse of software and independent evolution is near impossible. Added to other disadvantages, such as early obsolescence of products [3]. Although there are railway initiatives to modularise the core railway control systems, they operate on a shoestring and with aggressive opposition from the industrial companies. Moreover, the railways depend on them for their technical knowledge.

Market failure in railway control systems

In a normal market, this backwardness would be an opportunity for new companies. But a Mckinsey article [2] gives some hints at why this will not happen in the railway: “Taken together, high asset specificity, intensity, and durability often cause high switching costs for both suppliers and customers.” In the railway sector:

• assets are specific to sector and often country, or region;
• durability of assets reaches multiple decades;
• the development process is capital-intensive, due to tedious regulatory approval and the prevalent monolithic architectures;
• In addition, the market is small, with significant state presence and thus rife with protectionism.

So railways and industrials are in a tight mutual dependency, where there are high investment risks for industry, who are weary to invest without guarantees and railways have limited offer for critical products.

Railway control systems market: before vs now

It was not always like this. Until the 90’s, the balance of power in the sector was very different. Railway companies were designing and engineering trains and signalling systems in close collaboration with industrial companies [4]. The railway reforms of the 90’s which have forced the separation of operation and infrastructure, have also led to a dramatic culling of the in-house expertise of railway companies. Ironically, the reform meant to increase competition in the railways, destroyed a public monopoly to create a private oligopoly.

Railway value chain before

In the 30 years since, industrial companies bought and merged with competitors, increasing their market share (horizontal expansion). They also expanded vertically, from their manufacturing know-how into complete design, engineering, system integration and maintenance (not depicted here for simplicity reasons). They have also expanded from supplying components for one major subsystem, to providing both trains and complete trackside control systems.

Railway value chain today

Vertically integrating companies (to have a company owning several parts of the supply chain), has higher costs and risks, but amongst other advantages, it increases their market power by raising barriers to entry [2]. In Europe, we are moving to have three large pan-European groups designing and supplying trains plus signalling equipment. The market power of these companies is significantly larger than that of the largest suppliers 30 years ago.

Competition authorities in the EU seem to be only concerned about the horizontal reach, e.g. % of a product market held in one country, but watched silently as companies became integrated across the entire railway stack.

Conclusion

Summing up: for complex products or supply chains, architectures swing between proprietary and modular. The continued use of monolithic products in the railways prevents it from entering into the age of software. The railway reforms of the 90’s created a market failure larger than the one it was supposed to fix. The passivity of competition authorities only made the problem worse.

Given that the railway sector is tightly regulated and with significant state intervention, politics needs to admit that the railway reforms did not make the railway system more competitive. They rather introduced a lot of entropy, which after 30 years has not yet subsided and stunned the technological progress in the sector.

How the sector could change is a topic for further discussion. A few questions need to be asked: where in the value chain are the greatest opportunities, and do competitors have the freedom and the means to improve their products? Train operation does not score very well in any of these issues.

Some railway infrastructure managers are investing in reacquiring the lost technological know-how to become smarter buyers. It is the less disruptive path, but it is a long one, where a few companies invest and all are set to gain. It demands decade-long investment and stability which is at odds with political cycles.

References

[1] Christensen, Clayton, Raynor, Michael E.: The Innovator’s Solution: Creating and Sustaining Successful Growth, Harvard Business Review Press, 2013

[2] Stuckey, John; White, David: When and when not to vertically integrate, Mckinsey quarterly 8/1993 – https://www.mckinsey.com/business-functions/strategy-and-corporate-finance/our-insights/when-and-when-not-to-vertically-integrate

[3] Muttram, Rod: The many aspects of architecture and their impact on system performance: Part 2, IRSE News, 6/2021, p. 20-25

[4] Bantleon, Dieter, et al.: Integrated Railway System – Target for railway industry and railways: wake-up call for more cooperation. Railway update 1/2022, p. 25-31