Stefan Wasilewski, ceo of Contingent Capital Corporation, discusses the need to model sustainability in order to convince shareholders of risk

Reprise
Back to Alan Greenspan (see SCI issue 118): in my opinion, he wasn't at fault, just deceived. Not by anyone in particular, but by his own paradigm of thinking. Though I believe his accumulated wisdom with regard to the global economy and the US banking system is great, his own paradigm of control relates to a generation ago.

It appears that we have a person who has led with honour, but the system either outgrew him or changed without anyone's approval. Greenspan's 'steel' assumptions (see his early career) were wrong because technology had replaced it without reporting correctly to him.

Control is all about setting, monitoring and responding to the correct parameters in order to effect a change in the system in view. The problem is, if you don't properly define the system and hence the parameters, how can you 'control' something? Also control means having the ability to affect the changes required.

The latter is very important because 'action' will be needed in a consistent and concerted manner if the effects are to be realised, thereby resetting the 'system' within the parameters you've set. In the end the economy is all about social networks and simple rules of a game, but not the network or game that has commonly been portrayed.

Any network based on the actions or results of 'rules' is based upon the 'GIGO principle - garbage in equal's garbage out'. The capital and credit markets essentially operate on a 'grund norm' of rules, commonly accepted processes and parameters that are iteratively built upon to create bespoke products. The problem is that if the basic rules rely upon interpretations from bad data, then all subsequent parameters are wrongly set and the processes output garbage.

For those of you too impatient to understand
What we're about to say is that, by changing the basis upon which we form enterprises and correctly mapping their functionality to a standard, we can bring more information into pricing credit by understanding a business' context and control methodology. It's not difficult. An initial paradigm shift analysis is easily scalable and explains why the misuse of data and analytical processes has resulted in repeated failures in the financial markets.

We will show that an 80-year-old science has:

• the mathematics and rigour to predict system failure, but finds a balance between autocracy and anarchy (democracy)
• the ability to control, but also to invest freely
• is a 'laissez faire' approach, but bounded by different control parameters at each 'system' level - i.e. each system is free to run itself, but clear parameters (boundaries) will be set before outside interference occurs to bring errors under control. It's like the UK's 'principal-based regulation' approach, but with continuous monitoring by regulators.
• can use all the same processes within credit analysis, but modify them to give a richer picture.

What about credit?
Let's consider a simple aspect of credit. Time: it's essential to the existence of the business, the context it's within and how management controls the enterprise. Wow, where do you get all that information?

Answer: with difficulty, because most businesses do not collect all of it. They are input, output, resources-based, with a bias to net profit and not how the company is operating.

So, how do you go about creating an easily understandable data set that fulfils these requirements? Miyamoto Musashi¹ had a good solution: if you're finding it difficult, change your spirit and the solution will become apparent. This means change your paradigm, stop looking at ancient reports and accounts, and start looking at what they're doing in the context of their business milieu.

By establishing a systematic approach to analysing a business and its context, data can be created that modifies existing credit indicators and therefore enhance the predictive ability of individual businesses and markets as a whole.

How does this apply to banking and capital markets?
Now let's move onto banks and the capital markets. If the credit markets are ill advised, think of what the banks' situation would be: fed incorrect data, investing other people's capital and ruled by emotion, the concept of 'a sustainable economy' is an oxymoron.

So, let's dissolve the problem, not solve it, by moving the paradigm of investment and control; instead of 'robbing Peter to gamble with Paul', maybe we should define what a system is and how each client fits within it. That's novel: it'll supply data for the business and its marketplace, along with the rate of change.

The market has already begun the paradigm change: instead of being 'earnings' based, it has started to look at what capital is required. In itself a good move, but without putting capital in its right place in a broad scheme of operations the same processes that brought the current problem will reassert themselves and hey presto, we're back to the same problem.

What we need is a proper definition of a business in functional terms, regardless of its internal processes - a kind of 'General Relativity Theory' for business. We need to be able to define a 'system' and how it interacts with its environment, so that we know how its management controls the business.

So, what are 'recursive boundaries'?
Is this a new thing, a difficult subject? No, the genesis of 'system theory' is over 80 years old, supplied by biologists, social anthropologists and the original computer nerds.

People like W Ross Ashby, W McCulloch, H von Foerster, Stafford Beer, G Pask and N Weiner started the ball rolling, but others like Frederick Vester, John Holland, Stuart Kauffmann and lately Duncan Watts have contributed complexity theory into the mix to bring the whole thing to a state where it moves from atom to universe in a series of nested organisations with clear boundaries and interactions. Add Eric D. Beinhocker and his 'The Origin of Wealth' and you have a direct application to economics.

Oh dear, a little complex? Not really; what they are saying is that 'nature' seems to operate by creating a set of nested organisms, each with defined boundaries that operate on simple rules. Each can combine with complementary others to create bigger organisms, like Russian Dolls; imagine looking at your body and then internally to your component parts, each an organism but together something greater.

So to the global economy: a set of inter-dependent organisms, with its own boundaries but dependent upon each other to survive. How well it does, depends upon the level of communication and internal control.

So, what are 'recursive boundaries'? The diagrams show what we mean in two ways, one visual (The 'Droste Effect': see Wikipedia) and the other economic (Exhibit 1).

Exhibit 1

Each enterprise becomes well formed when it sets boundaries to its internal processes; defines inputs, outputs and operational rules; and those parameters that confine it. When an enterprise is so defined, it is said to be a level of recursion from which you can say it operates within the next level up and has some that operate within it.

Look up 'recursion' in Wikipedia and you'll find plenty of references: it's the mathematical approach we want, along with a discussion on reverberating networks. The formal mathematics is also there and a clue to how 'set theory' plays a part in analysing networks. It really only works when you've identified self-similar functional wholes because - as the system closes and becomes internally self-regulating - you are left with inputs and outputs; the rest is a black box for which its function and parameters become discernable from the analysis of past activity.

Well-formed boundaries are good
By simplifying the information and determining the links within the system and its environment, a 'Sensitivity Model' [The Art of Connected Thinking: Frederic Vester] can be created for the processes and the functional parts compared to a general map [The Viable System Model: Stafford Beer].

If the boundaries are not formally set, then we have open networks that can range from chaotic to complex. Neither is bad; it just depends upon yours and the market's level of risk taking capacity because it will certainly be volatile. But that's a market for you - one person's meat etc.

If the market is not deep enough or confidence is removed in the fundamentals, then a catastrophic drop in trading occurs as value becomes impossible to set. It doesn't mean that the economy's bad, just that the reason why it over-heated must be removed or dampened, so that the underlying cashflows of good business restore confidence.

This will happen in a 'complex' market because they have inherent rules that trigger a stabilisation process around a particular point or 'attractor'. You may not like it because another shock needs to occur before normal business is resumed and the attractor removed. A stochastic process is one whose behaviour is non-deterministic in that a system's subsequent state is determined both by the process's predictable actions and by a random element.

'Chaotic' systems in economies are generally bad for you, so you don't want that unless it is localised. They are deterministic systems highly sensitive to initial conditions.

So, well-formed boundaries are good for you? Yes, because, while nothing is a given (remember: correlations are dynamic, conditional and unknowable - see SCI issue 118), the inherent stabilisation processes within a well-managed enterprises anticipates oscillations and compensates by smoothing results over time.

Rating agencies, system theory and 'The Crunch'
You cannot productise the economy and that's what I believe the process of rating a business does. Notwithstanding the fact that the assumptions are opaque, the focus inappropriate and the data invariably wrong, the whole concept of an unregulated, for-profit entity managing a crucial part of the global economy is wrong.

Having dealt with most of them, I think individuals tried to do a good job but the approach was wrong. How can you assess the viability and sustainability of a business without a formal structure of both the business and its market? To add to this, they only use out-of-date data in accounting formats that are themselves regulatory nightmares.

In the early 1990s I ceased creating structured finance products because the reliance on accounting or regulatory arbitrage was leading to the 'robbing Peter to gamble on Paul' routine. Insurance companies and banks were leveraging client's money to gamble on a system they did not understand.

Anyone going to a casino knows the house always wins; they just kid themselves that they're better, when they're only lucky [Fooled by Randomness: Nicholas Taleb]. Another reason was that these products have little or no embedded liquidity; rather, relying on derivative trades with complex, opaque rules: a lawyer's heaven.

The rating agencies are a service, not a business, and should therefore be regulated by government. However, first they should restructure how they form opinions of a business and its viability, for a triple-A rating is a measure of longevity, as well as risk price. But, as we've seen, governments can fail and we have a global economy, so individual enterprises must be measured in time, space and connectivity for risk.

Where's the driver of the car?
A system is like a car; its functional parts may be well honed and coordinated, but what use is it without the driver that must gauge the road, and what use is the road without the reason to travel along it?

System theory is not just theory; it's been applied to businesses, banks and indeed countries. Frederic Vester's book gives some good examples and Stafford Beer's 'Diagnosing the System' can show you how to compare your business to a general map.

Both are good maps to guide you, but you are the driver! This is what you, the manager, will be measured on - not whether you can confound investors by P&L massage: cashflow is still king.

Systems models and Gödel's incompleteness theorem
I have a paper to write that compares the application of processes and models in the insurance and capital markets. It will fundamentally say that, without a formal understanding of what a system is, how can processes within it be properly regulated?

All processes are rules and Gödel basically said any such system you set up can be circumnavigated: so don't set up complex regulations, but monitor formal functional systems for flows and events that can lead to 'tipping points'.

For example, the financial industry effectively became the gate-keeper to capital for business, but - without a formal understanding of viability - the natural process of birth, develop, die for businesses was interrupted because the 'birth' bit was still-born and now 'venture capital' is an oxymoron.

Human nature took over and VC funds became structured financiers and then hedge funds, leveraging equity risk against rated debt with return expectations so high that they left the real economy behind. Everyone wants returns at 40% ROE, but if the economy has a real return below 10%, where's the other return coming from? Equity risks were being taken at debt prices.

Regulators cannot regulate if they're not in the market. The data flow and discrete decisions that control a market network are 'complex', need a formal structure to model and are partly path-dependent. I hope some sense will come to the UK and the Bank of England is given control over some of the essential capital markets, while the FSA regulates the rest. As I mentioned before, a regulator must be part of the system in order to control it.

Business managers are at the front line, but not for themselves, for all stakeholders. It's easy to convince oneself as a manager that 'shareholders want returns', but a director also has an obligation to disclose the risk levels at which they are allocating their capital. Again, if you cannot model sustainability, how can you properly convince shareholders of the risk/return?

Hope
It is not without hope that we face the current market forces of 'the crunch', but if we are to avoid the continual repetition of boom-bust, we need better models to measure capital risk, and the tools are there. Am I going to tell you how to do it? No, because there's a lot of money to be made in bringing the first really Operational Risk Metric to market. But if you want help, please call.

Footnotes

¹A Book of Five Rings (Go Ring No Sho). A book on strategy written by the samurai warrior Miyamoto Musashi circa 1645.