Energy is fairly straightforward; why are infrastructure assets different?
Infrastructure deals simply have more moving parts than small energy projects, usually due to various loans and reserve accounts.
Toll roads are built to last longer than energy projects. So infrastructure debt structures can be longer (30+ years). The longer the timeline, the more opportunities to amortize debt or refinance.
Infrastructure deals also can assume more debt simply because the nature of the projects involves more stable cash flows and less risk. Energy projects usually involve some element of market risk whereas toll roads are seen as essential infrastructure. Stable cash flow is key to opening the bond market for infrastructure deals from which energy projects are generally excluded.
For example, lenders clamored to help New York’s JFK Terminal 6 ($4.2 billion) achieve a capital structure of roughly 75% debt with bond tenors extending beyond 35 years. Could you imagine that much debt spread over a 35-year bond solution for a partially contracted wind deal?? Probably not…
The cash waterfall
Understanding how to model both energy and infrastructure projects involves answering a deceptively simple question: “what is the rate required to generate revenue that a.) pays off debt; and b.) provides equity with a healthy return?”
Unpacking that question involves a walk down the cash waterfall: how much cash flow available for equity (CFAE) do I need to generate an IRR sufficient to justify the investment?
Because equity has no claim to the asset should it go bankrupt, equity’s cost of capital is higher than that of a lender, who can claim and then resell the asset under duress.
As a result, equity will want to minimize its risk adjusted investment total. And we can see this relationship visually when we notice that CFAE sits at the very bottom of the waterfall.
And, if you keep unpacking the waterfall by moving backwards up from the bottom of the waterfall (CFAE -> CFADS -> net cash flow…), you’re left with the original question: how do we size the start of the waterfall, namely revenue?
Energy vs. infrastructure – revenue
For energy deals, contracted revenue comes from a power purchase agreement (PPA). On the infrastructure side, this is usually called an availability payment.
Both operate in very similar ways. For energy companies, the projects are selling electrons to an off taker at a volumetric rate. Easy enough.
On the infrastructure side, the projects are being paid to be available. What does that mean?
Think: what does a bridge sell? Crossings. What does an airport sell? Enplanements. What does a toll road sell? The ability to drive. The utility of these outputs boils down to availability.
Specifically, users are paying to ensure the thing they’re using (bridge, airport, toll road) is available to be used.
Energy vs. infrastructure – debt
Two factors drive the differences in availability of debt for infrastructure vs. energy: 1.) moving parts; 2.) project length.
Let’s look at how a solar plant operates.
Once the concrete has been poured to hold the racks and the inverters have been affixed to the photovoltaic (PV) panels, in most cases, not much else happens. Maybe there’s a snowstorm and our O&M contractor needs to come and brush that off. Or maybe a cow bumps a panel out of alignment or a goat nibbles on an inverter wire. Otherwise, the sun shines and electrons flow.
Solar panels usually have a rated life span of between 25 and 30 years. However, PPA durations have been shortening lately, as big off takers diversify their power sources. And because banks will want to see long-term contracted cash flows before committing large sums in bank debt to solar deals, generally this means debt tenors have shortened commensurately.
Contrast that solar project’s relative operational stability with the movement on a toll road. Once the concrete is poured, there’s near constant abuse applied by cars and trucks of varying sizes and weights, some of which will have accidents, create skids, hit embankments, and maybe even explode and disable entire lanes.
As a result, the need for O&M is nearly constant. However, because they’re just concrete slabs and asphalt, toll roads can usually last for between 30-50 years.
While drivers can make day-to-day decisions on whether to pay tolls for faster access to their destinations, well-placed and well-constructed toll roads become essential in a way that affords them flexibility and generous terms from debt providers.
Essential service, combined with high capex and opex along with relatively stable cash flows starts to sound a lot like bond financing.
So what?
Now that we’ve established all the ways that a solar project is NOT a toll road, what’s the point?
Infrastructure projects get longer tenored debt facilities than solar projects and require more reserve accounts to handle all the vagaries of providing their services.
How we link together the logic of sizing multiple facilities and linking their amortization and refinancing requires a careful walk backward through the cash waterfall.
But, we hope we’ve demonstrated that, at its core, the modeling for infrastructure and solar projects is pretty similar.
Our template
The Orbis infrastructure template can be used to evaluate up to seven different tranches of debt, including a TIFIA loan.
We’ve included a fun macro that looks at how an equity owner might pay themselves a dividend by issuing more debt, all while not blowing through their debt service covenants (sometimes called a “dividend recapitalization” or “div recap” for short).
We hope you find this useful and look forward to hearing how you’re applying some of these lessons in your own modeling.
