As investors, key questions we are always thinking about are “What is the competitive moat for a project?” and “How sustainable are the cashflows?”. In infrastructure, the key investment characteristics are high upfront capital costs to build physical assets, low incremental marginal costs per user, high operating margins, being a price setter rather than a price taker and long term cashflows (often inflation linked). These characteristics are usually due to one of the following,
· Geographic monopolies. Due to an asset’s location or physical network, it is able to prevent competition. For example, if you want to fly to Adelaide on a scheduled passenger flight – you have to fly to Adelaide Airport.
· Regulatory monopolies. These infrastructure assets are regulated by Government with the allowable prices effectively set by Government. Examples include electricity and gas transmission and distribution networks.
· Government or corporate offtake. These infrastructure assets have some level of availability linked cashflows that are well defined. Project owners are effectively only taking execution risk versus the price they have bid. Examples are PPPs and highly contracted power generation assets. Assets with no offtake, for example merchant electricity generation, don’t really fall within the definition of core infrastructure.
A fascinating contrast is to compare the investment moats of physical infrastructure assets against today’s hot investment sector, technology (or software). Are the tangible infrastructure moats that different to the intangible moats in the technology space? In our view, the investment moats are more similar than you would otherwise think. Both have relatively high initial starting costs and low incremental marginal costs.
The main investment moat in technology is the phenomenon commonly known as a network effect. That is the utility a user derives from a good or service is dependent on the numbers of other users on the network. The more users there are on the platform, the greater the benefit to the users, the greater the competitive position of the platform and the higher the operating margins. This in effect creates a flywheel, which is considered the wholly grail of technology investing. So what are some of these network effects?
· Platform network effects. These are platforms (or intermediaries) that provide significant additional features and benefit to users that only increase as the number of users in the ecosystem increase. Common examples are Google, Microsoft, Shopify, Mastercard/Visa or any SaaS business.
· Marketplace networks effects. These are traditional two-sided marketplaces with supply and demand. These are difficult to start but once created have one of the strongest flywheels. Suppliers want a marketplace with significant numbers of buyers, and buyers want a marketplace with a large selection of quality suppliers. Examples are Amazon, Ebay, Alibaba, Shopee and even Tinder.
· Social network effects. The ability to communicate, interact and connect with others. These network effects are built of the identity of the user to perform or communicate some function with other users. As the network/metaverse builds it can become difficult to leave due to the number of social connections in the network. Examples are Facebook, LinkedIn, Twitter and Pinterest.
There are probably many other types of network effects we have missed. The key dynamic with all of them is that there is a significant competitive moat that can potentially be achieved through scale. The only issue is getting there. Here in lies the distinction between the two types of investing. In infrastructure large capital costs are spent with the future competitive position known with a reasonably high degree of certainty. Whereas in technology there is no economic moat to begin with and the future competitive position is path dependent on growing into network effects either by delivering great product or efficient marketing spend. Those technology businesses that make it to the endgame demonstrate significant monopolistic power as they do in core infrastructure.
The durability of established infrastructure moats is observable given the characteristics mentioned previously. However, it would be naïve to assume these moats can last forever. Technological change can render a monopoly irrelevant over time. For example, the canal barons of the 18th and 19th century did not last as the steam train was invented. A more contemporary example in Australia is the monopoly pipeline position of APA which is starting to look weaker today with the decline of fossil fuels (and, hence, their desire to diversify into electricity).
The durability of a network effect in technology investing is complex as it is mostly unobservable until achieved. To achieve a durable network effect usually requires growing as fast as possible at the expense of short-term profit. This can be very polarising amongst investors particularly those tied to traditional valuation metrics. Not all high growth technology moats are equal, and some will prove to be illusory. Some issues to think about,
· How reoccurring is the revenue? How integrated into processes is the platform? What is the switching cost?
· What is the value proposition or unit economics to users on each side of the platform/marketplace?
· How easy is it to multi-tenant across platforms? For example, how easy is it to list a product or service on an alternative platform?
Even when market dominance is achieved there can be pressure from Governments to regulate anti-competitive behaviour. Given the strong similarities between infrastructure and mature technology businesses we make the following curious observations:
· What does this imply about earnings multiples? Is infrastructure overvalued versus big tech? Should core infrastructure assets trade at the rich EBITDA multiples of circa 20x today? Some forward EV/EBITDA multiples in tech: Microsoft 22x, Amazon 19x, Google 16x, Facebook 13x, Mastercard 26x, Visa 23x, Apple 20x.
· Listed big tech has the potential to be as interest rate sensitive as is infrastructure. Both sectors are long duration and interest rate exposed. This is potentially a new correlation dynamic that should be considered at a whole of portfolio level for funds with high allocations to both.
· Why are the capital structures so different? Technologies companies have traditionally had minimal debt on their balance sheets (and any they do have is mainly used to buy back stock). Infrastructure on the other hand is usually highly levered with physical asset offered as security.
Our key point is that these sectors may be more similar than you would otherwise think.
Comments