Quantum of Solace

Finally, an Effective Means of Optimising Corporate Capital Structures

As global markets show few signs of calming down, treasurers struggling to optimise multiple financial instruments within their corporate capital structures have a new and cost-effective option: quantum computing.

If there is one treasury ability that will widen the eyes of corporate leaders, it’s being able to optimise the capital structure of the business. Find the right balance between debt and equity, and the company can enhance profitability and shareholder value while keeping its overall cost of capital down.

The issue with this is that the most advanced method of computing the right balance is based on heuristics. A heuristic is an assumption-based quantification that often uses rankings and weightings to estimate an objective solution. It calls upon past experience and reasoning to arrive at a ‘best-guess’ outcome but remains deeply flawed.

This is not the fault of the financial practitioner but of the systems they are given or have inherited. Nonetheless, when capital structure optimisation is reported to senior stakeholders in best-guess terms, its recipients, acutely aware of the need to get it right, may be entitled to feel somewhat on edge.

Heuristics may suit a small organisation that runs on a couple of debt and equity securities, where the treasurer is able to make perfectly satisfactory optimisation calls based on how those securities are impacted by market variables such as lending and interest rates, cash and debt position, contractual terms, and credit rating.

But when applied to a treasury that is overseeing 10 or more financial instruments as part of the organisation’s operations and growth strategy, these variables make optimisation exponentially more challenging, notes Ethan Krimins, CEO, Quantum Research Sciences (QRS), a Purdue University Quantum Institute affiliate.

As part of the rapidly evolving macro- and micro-economic conditions, Krimins understands that treasurers face a dynamic regulatory environment, evolving customer and supplier needs, and their organisations’ constantly shifting product and service lifecycles. To this can be added constant and intensifying commercial competition, often on a global scale, and the natural (and sometimes unnatural) fluctuations of physical and financial resources in the 24/7 global debt and equity markets.

Beyond faster horses

In light of this challenging set of uncertainties, the best-guess approach to managing corporate capital structures is surely an unacceptable risk profile of its own, suggests Krimins. But optimisation seems like an impossible goal. Indeed, adapting a typical large corporate’s complex capital structure in real-time response to a host of ever-shifting market variables is hardly, if at all, possible using classical computing techniques. A new way of thinking is needed.

Henry Ford, reflecting on his revolutionary Model T car, is often quoted as saying that if he’d asked the people what progress looked like, most would simply have said ‘faster horses’. While provenance of his quote is still subject to debate, the notion that pursuing existing thinking offers anything other than finite possibilities is not.

Classical computing, which encodes data in binary bits, has performance limitations that are now being reached. For most users, this may be sufficient, for now. But with the levels of complexity, and risk, being faced by many corporate treasuries today, the task of computing optimal capital structures is unachievable using this technology. Evolution, at least for complex tasks, has run out of road. It’s time for a revolution: quantum computing.

Hello quantum treasury

The name ‘quantum’ in this context is unfortunate, comments Krimins, because it suggests “futuristic and hyper-complex technologies”. But it is neither. He explains that while it has roots in quantum physics, the science of which is seriously challenging, quantum computing is both available and accessible. Indeed, QRS is already using live data in a quantum operational application for US Department of Defense supply chain optimisation.

Quantum computers use quantum bits (qubits), harnessing effects that exist at the level of atoms, electrons and photons to carry data. Various forms are possible, including neutral atom qubits that use lasers to charge neutral atoms with energy.

Unlike classical bits, whose state is either ‘on’ or ‘off’, qubits can exist in multiple states simultaneously. Their status is explained by the quantum mechanical phenomena of superposition: essentially it stores information as 0, 1 or a superposition of 0 and 1. It’s a construct that is notoriously difficult to grasp (especially as other states such as entanglement and parallelism muddy the waters). But just as few non-scientists can explain what electricity is, lack of knowledge does not preclude anyone from experiencing its benefits.

To dive a little deeper, in classical binary computing, tiny amounts of electricity are turned on or off, as all possible solutions for a given problem are explored. Known as ‘brute force’ computing, it relies on sheer processing volume to process every step of a calculation.

In stark contrast, quantum computing is naturally disposed to take the most efficient path. The hardware uses subatomic particles – the qubits referred to above – and groups of these (sometimes in excess of 1,000) in superposition can form multidimensional computational spaces. This ‘framework’ enables the computer to process multiple inputs at the same time. When manipulated by the hardware, the framework takes milliseconds to identify solutions that require the least amount of energy. In other words, these systems are naturally designed to be efficient in processing vast and complex datasets simultaneously.

Its natural energy efficiency means a quantum computer with 250 qubits consumes around 10 kilowatts of electricity. This is less than 0.05% of the 21,100 kilowatts required to operate Frontier, the world’s most powerful supercomputer. And although quantum computers generate some heat, their total system power demand is significantly less than equivalent classical computing.

In use, the aim is processing extremely large datasets. Here, the key differentiator between quantum and classical computing is speed. “When it comes to optimisation, a quantum computer is faster than a classical computer,” states Krimins. “It’s not that a classical computer can’t do it, but a complex corporate capital structure equation would take weeks, months or even years to solve classically. The same calculation run on a quantum system would take milliseconds.”

Fastest wins

Under rapidly evolving macro- and micro-economic conditions – multiple variables – quantum computing speed permits optimisations that are more precise, more current, and more frequent. It enables treasury, for example, to reassign securities while markets, economies and industries are changing, rather than after-the-fact.

In other commercial settings, where the number of variables may extend to hundreds or thousands, even the current iteration of quantum computing hardware has not yet evolved to the point where it could successfully optimise. However, with most large corporates juggling around 15 to 20 assets within these conditions, quantum is perfectly capable.

“The number of corporate assets typically held fits well with quantum: it’s not so large that it’s overrun, but it’s large enough that it’s too much for a classical computer,” explains Krimins. “What’s more, because the global stock and debt markets are effectively open 24/7, treasury has an opportunity to run an asset optimisation, acquire an almost instant answer, and go to market either in their own time zone or somewhere else in the world, and quickly change their mix of assets to optimum effect.”

Making that dramatic leap

While it’s important to state that a quantum solution cannot achieve perfection – nothing can – it will deliver, from real-time information, an optimised answer that is way beyond the best-guess compromise of the heuristic.

The quantum computer system itself is a highly specialised piece of technology. It is usually the preserve of dedicated and scaled-up organisations such as IBM and Google. However, accessing quantum computing for treasury is no different than accessing classical servers, in that cloud-based connectivity with the service provider is the norm.

The QRS setup process for customers has two steps: custom coding and computer selection. Because every company’s needs and expectations around capital structure optimisation are different, a tailored equation and corresponding quantum computer programme will be developed, explains Krimins. Once coded, the most appropriate quantum computer is selected to run that programme, choosing one from around two dozen options located around the world. “We find the best match for each company, based on a combination of their chosen criteria such as accessibility, speed, price, and reliability.”

How the investment pays for itself

A cost-comparison may be front of mind for many prospective quantum users: it sounds expensive, and compared with building a heuristic within today’s classical approach indeed there is a premium.

“Yes, it is going to initially cost more to optimise on a quantum computer,” admits Krimins. “But asset allocations can be updated daily, hourly, or more frequently if needed. The savings treasury will achieve from not being locked into long-term securities, and being able to trade these, and get out of positions based on changes in national government fiscal policies, or events that are taking place in the world, will easily make up for the extra cost differential.”

The QRS assumption is that today’s heuristics offer an accuracy rate of between 70% to 75%: in other words, they have an error rate of between 25% to 30%. “That error rate, and corresponding room for improvement, suggests the magnitude of the opportunity for QRS customers,” says Krimins. Indeed, a medium-size corporate may have a USD-equivalent of debt and equity assets in the tens of millions. A large corporate will hold assets in the hundreds of millions, and a global giant may be in the billions. “A 25% or greater saving on the debt burden in the capital structure of any of these companies will more than pay for their investment in quantum computing,” he notes.

History in the making… and more

If the initial costs can be recouped, then the only real downside of quantum computing in treasury is that its implementation takes time. “This is not off-the-shelf software. It has to be custom-developed for each company,” states Krimins.

“It requires collaboration between the company, its treasury team, and us, to develop the software. It might take three to six months to code, and have the processes and user interface built, so the treasury team can use it on a daily basis.” But once that work is done, it will rarely need to be revisited. And even then, he says these are “easy tweaks”.

In the field of optimisation, quantum computing has been proven. “It’s an exciting time for this industry,” declares Krimins. “If we can find more treasurers to do this, they’re going to be among the first people in history to leverage this technology.” But more than making history, if it means delivering commercial advantage in challenging markets, it’s a true demonstration of treasury’s value.