Stefan Wasilewski, ceo Contingent Capital Corporation, discusses a bridge to sustainable economic metrics

Reprise
In two previous articles (see SCI issues 118 and 125) I suggested that Alan Greenspan was not at fault for making a mistake, but that the world had moved on without informing him of fundamental changes in the manner and content of modern reporting processes - especially with regard to their use in risk assessment. I think we should now move on to someone else whose world view has changed: Jack Welch, former chairman and ceo of General Electric (USA).

The articles discussed the fact that correlations in business are dynamic, conditional and unknowable; but not unmanageable. The latter was justified by the belief that one could establish a functional control methodology (a map of sorts) to correctly identify the current state of the businesses market context and how we should respond.

It was also mooted that establishing these protocols across a business would facilitate better capital risk management and in future stabilise earnings. It did not guarantee a riskless state but shifted the focus away from short-term earnings towards business sustainability.

The paradigm shift required was explained in general terms and didn't go into how existing risk tools need to be augmented to this new approach. In this article I'd like to suggest how these new and developing concepts can be used alongside existing risk metrics.

For those with little time: don't trash the 'quants' or calculators yet
It is important and you should read further, but essentially the argument is: commonly-used statistical processes assume 'homogeneous' data and conditions to achieve meaningful results, but we know that businesses and the economy don't perform in a regular fashion. Only by creating the conditions upon which these assumptions are based can they be used reliably so that data can be provided that is timely, contextual and therefore useful.

How do we achieve it? By ensuring each enterprise can cope with market volatility, has real-time data and is functionally well structured. You do this by identifying the internal processes and mapping them to a consistent functional model.

Once in situ all the existing financial/risk tools can be used with one modification: a variable created to describe the state of a business with respect to its environment. This is because the functional approach creates the mathematical conditions needed for traditional pricing assumptions to work.

Therefore, by taking our 'map' of a well-behaved organisation, we can compare how local businesses should operate, then how they fit within the economic environment and so on until we encompass the globe. Although we set different parameters at each level at which control is exercised, we allow freedom of investment choice but guide the economy to meet our expectations. All this can be done using simple tools and benefits from modern internet communications and open-resource applications because 'the map' is the same, even though the processes may be different.

Jack Welch, 6-Sigma and current thinking
I choose Jack Welch because of an article in the Financial Times¹ ruing the focus on short-term profit and share price gains. To quote, he said: "On the face of it, shareholder value is the dumbest idea in the world. Shareholder value is a result, not a strategy . . . Your main constituencies are your employees, your customers and your products."

We agree on one thing: 'shareholder value is a result not a strategy'. But using inappropriate methodologies to achieve your strategy will largely repeat the current problems and its beauty is in the 'eye of the beholder'. The irony for me is that I believe GE US created a lot of the impetus to innovate financial products² using many of the instruments at fault today for that (shareholder value) same reason and that his nickname, 'neutron Jack', is at odds with how he treated a 'main constituent': employees.

Another issue I have is his pioneering use of 6-Sigma as a financial markets risk-tool when its real application is in manufacturing, where it should have stayed - but even then seriously modified. In particular, I refer to the application of the Normal Distribution within 6-Sigma, and most risk metrics, on the assumption that the data is homogeneous and used without reference to the contextual economic dynamics of the business.

Some would say you can apply modern copula techniques to achieve a confident result, but their own assumptions regarding correlations are suspect. What is at issue here is not the tools but the data and how it is derived. Homogeneous data comes from consistent production processes within a stable system, which imputes the concept of time and therefore expected results.

For those that aren't familiar with it, 6-Sigma is a methodology created to track the defects in linear production processes. As Wikipedia quotes it: "6-Sigma seeks to identify and remove the causes of defects and errors in manufacturing and business processes. It uses a set of quality management methods, including statistical methods, and creates a special infrastructure of people within the organisation ('Black Belts') who are experts in these methods. Each 6-Sigma project carried out within an organisation follows a defined sequence of steps and has quantified financial targets (cost reduction or profit increase)."

The italics are mine and designed to highlight the focus on 'processes' and their financial result. My issue lies in the word 'organisation', which is never properly defined. What's more, throughout 6-sigma literature the focus on 'quality, process and management' never relates to the contextual environment of the business sell or the necessary functions required within all 'organisations'.

It uses an aspect of linear programming within a bounded space, which is great if you're living in one, but we don't. Look up; the illusion of a sky is just that. In reality our environment goes out to the stars, the most important of which is 93 million miles away and whose operation is still mainly unknown to us.

So why do we place our faith in formulas that produce erratic results at the core of our modern financial risk pricing? It's because if we didn't make the assumptions that derive them, the mathematics would be intractable and we wouldn't have the culture we see today.

Whether you talk about 'ergodic systems', the 'second law of thermodynamics' or simply 'homogeneous data', you are addressing the same concept - 'a closed space can be mathematically well-behaved' or 'short hand formulae's get good results most of the time'. The core of most economic risk pricing relates to the same assumption.

Also, as we noted in the last article (SCI issue 125), 'Time' is an essential element in finance because we assume things are going to happen in 'such and such a well-behaved way' such that our expectations are met. Credit assessment/pricing implicitly embodies this in its definition of a sustainable business when it 'rates' an entity.

Well, you can see where I'm going. We don't live in economically well-behaved boundaries and have not taken into account rapid changes in environment, with the result that the data is not 'homogeneous'. And, as our metrics are therefore wrong, there's no data to support us in finding a solution to our current problems.

Building a bridge to another paradigm
The articles' new paradigm had us building sets of 'self-similar' organisations, composed of common functions but different product processes, which took resources as input and products as output for onward use. In between we created a 'black-box' that was consistent and well formed.

These could then combine with others using the same functional map such that they created another stable entity but a different level of activity. The focus was on sustainability; a time-frame to match our expectations and/or performance; and deriving information about the state of the businesses, market or economic context, as well as the internal performance of management.

Strange; in defining this approach we create the environment that traditional metrics need in order to make their 'assumptions' of bounded and/or well-behaved spaces hold true. At the same time, we establish a better understanding of our environment as it relates to our own performance.

I used the italics again to highlight the key issues. 'Functional map' refers to a consistent set of 'functions' needed to manage the 'processes' of the organisation. The 'map' also includes the required connections to maintain internal and external communications.

In order to keep this consistency, the individual and group structures must be similar - which has ramifications for conglomerates as it demands different parameters and hence capital to be set against the risks. It also means that regulatory authorities need to be 'in the market' in order to control the market. Hence the Bank of England needs an operative arm again to monitor the flows of national debt and product creation.

Recursion revisited
By using the 'functional map' we build a consistent operational ability: a generally bounded organisation or 'system'. When several of these 'systems' find a common thread, another 'functional map' can be created to make sure that they again provide consistency.

It is this 'nested' or Russian doll-like approach we call 'recursion'. By setting different performance parameters at each level or boundary at which corrective management is needed, we can control how the whole operates.

We can therefore describe the practical aspects of recursion as follows:

Level 1: Assume we start with a bus manufacturer. Producing the buses has certain processes and we organise it properly as above. The output, buses, need local operators whose business model is different but organisationally can be the same; they buy the buses and operate them.

Level 2: The local council needs to make sure that the fares of the operators are 'fair' to the locals, which is intertwined with the economy of the area and the manufacturer.

Level 3: A bank invests in the manufacturer, operator and other businesses within the local area.

Level 4: The government needs to make sure that all local councils are regularly managed along with the capital needed to support the banks.

So how do we implement this approach? Well, we have a good functional map in a product called the Viable System Model³ and we can more than ably derive the necessary processes and communications network using two investigative processes - 'the sensitivity model' by Frederic Vester and 'syntegration' by Stafford Beer. Each compliments the other, but the latter is the forum in which the former derives the complete set of processes and business model.

It is making sure that the boundaries are clearly defined and the control parameters set appropriately that is the skill of the new organisational paradigm. The output is better: more consistent data with a clearer understanding of the real dynamics because most people are astounded at the outcomes as they fly in the face of their common assumptions of how their organisation should perform.

Next article: practical examples?
In the next article we will look at real world examples: pricing the creditworthiness of a business and the rating of a complex bank. However, we have described a framework within which we can build organisations that can adapt to market volatility and show consistency such that the 'systems' create reliable data that is transparent.

We don't have to throw out the calculators or 'quants' because, as long as we operate within the functional framework, the mathematics will become stable enough to use similar metrics. What's more, we will have another metric - one that reflects the activity of a business within its environment.

But, you say, we can't all follow the same route. That's fine; businesses emerge all the time and risk appetite varies. As long as the overall is managed in the context of the three generally accepted paradigms - ordered, complex and chaotic - we can build a model that caters for all but the truly catastrophic. But at least these will become more rare.

Footnotes

¹ 'Welch condemns share price focus', by Francesco Guerrera in New York. Published: 12 March 2009 18:13
² Some would say to mask lacklustre performance elsewhere and keep the earnings high
³ Viable System Model: Stafford Beer - John Wiley & Sons