Andrei Kozlov repeatedly said that he considered risk managment to be one of the most promising professions in the financial industry. As on many other issues, Andrey Andreevich was a thinker well ahead of his time.

The global crisis was caused by poor risk management and by the very agencies that are supposed to measure risk. Indeed the very companies that were supposed to have the most sophisticated risk management systems in the world and which invented risk management techniques failed. Yet the crisis ironically demonstrated how much we rely in the modern financial word on risk management and on rating agencies.
The most successful risk managers and investors use the techniques of risk managers, but do not rely on them. However, far too many risk managers rely on the techniques without thinking. They are like Churchill's political opponents who used statistics in the way a drunk uses a lamp post, for support rather than illumination. It is not sufficient to gather data, manipulate the data using sophisticated algorithms to get results, and then believe the results without thinking. All mathematical formulae depend on assumptions and simplifications. The more useful formulae work because in normal circumstances the assumptions are met. Yet it is vital for a risk manager that uses techniques, such as Value At Risk, or option pricing formula, to demonstrate without question that for the data he is investigating, Guassian distributions are appropriate. This is not always the case!

The logic underlying structured products, such as sub-prime mortgage securitisations, is economically sound. But there must be independence between the borrowers and the behaviour of the borrower should not be affected by the conditions of the mortgages. In fact, it seems that a very high percentage of borrowers came from two US states, and mortgage issuers were inventing terms that would make rational people act in strange ways after two or three years. Those risk managers who investigated the assumptions did not buy sub-prime paper once they discovered the problems. And now we know there was at least one hedge fund that recognised the flaws and cooperated with Goldman Sachs to profit from the weaknesses in the system.

Even so, weakness in a single market should not have created a global crisis of such colossal dimensions. The entire market collapsed because the volumes of new exotic financial instruments were so large that as soon as flaws appeared in the structure of the markets, key institutions crumbled under the weight of the products they had sold. The exotic financial instruments are new, the financial markets are new, but the phenomenon of human arrogance and pride is as old as mankind.

How many times has humanity found a way to build and then created constructions so big they failed? We have discovered technologies that enable us to create skyscrapers and bridges. They are incredibly useful under normal circumstances. But engineers cannot always know all the factors that lead to failure. There are examples of buildings collapsing because of movements in the ground, or ships that broke apart because of stresses due to holes drilled for bolts that were supposed to strengthen the structure. And what happens when there are earthquakes or terrorist attacks? Whenever there is a failure, there will always be someone who says “don't build skyscrapers” or “cities are bad”. Indeed we could all live in the country in small houses, and we would not have objects for terrorist attacks or earthquake disasters. A logical solution, but hardly realistic for it ignores all the reasons why we benefit from living in cities and working in large buildings.

Given all the evidence that has appeared in the congressional record about the rating agencies, it seems clear: the big rating agencies were at fault for putting revenue before analytical quality, and they were at fault for claiming that the reputation they had earned from their bond ratings could be transferred to structured products and other types of rating.

There are those who say “eliminate the credit rating agencies”, or equivalently “brokers, hedge funds and banks should do all their own credit analysis”, which is really the same assertion in a different form. This is – to my mind – the same as saying “don't build skyscrapers”. Credit rating agencies have done a very good job of creating the idea that what they do is mystical or divine, something no one else can do. But they are not! Assessing credit risk is what every banker does before giving a loan, or the treasurer of a major company before putting his company's funds through a bank.

Rating agencies are the product of the economic law of specialisation – it is cheaper and more efficient to give a specific task to a dedicated professional or company, who supplies many customers, than to hire the professional for your own firm.

The problem in the rating business that caused the crisis is that a monopoly was created for two firms and the firms exploited the monopoly. Nor is it the first time the beneficiaries of a monopoly have sought to retain the monopoly or tried to prove it is “natural” or beneficial. And the solution to monopolistic aggression has been demonstrated historically: foster competition. More credit rating agencies, not fewer. This sometimes requires government action to create the conditions in which a free market can develop. In this respect, Russia seems to be learning the lessons of the crisis.

Do we need more regulation of the financial markets? Yes and NO. The financial markets, like the economy, is a complex dynamic system. All such systems, without exception, must have some form of control mechanism in order to stay within a set of behaviours that is optimal. All complex dynamic systems can exhibit behaviours which are suboptimal and can even lead to the destruction of the system.

The need to find optimal control systems occurs in engineering, and has been studied in much greater detail by engineers than by economists. In general, it has been found that a system needs a control mechanism that reacts gently to small movements from optimality and reacts with increasingly stronger actions the further the system moves from optimal behaviour. We know this from our own experience. If we are driving in a car straight along a smooth road, we need almost no effort to keep the car going straight. As soon as the car hits a hole, we have to react quickly to control the car. If the road is bumpy, even more effort is needed. But once we are back on a level road, we can relax. By contrast, we know that when a child gets in a car and moves the wheel strongly from side to side on a smooth road, instead of travelling smoothly, the car can actually go out of control.

The same applies for an economy. Strong firm regulation is needed when there is a shock – such as a crisis. But almost no regulation is needed when the economy is growing smoothly. If there is over-regulation during a time of smooth growth, then the economy can be pushed into suboptimal behaviours, and even forced into decline. So the answer is yes, there must be the potential of strong regulation. And NO, strong regulation is not needed once the crisis has receeded.

The world will need credit rating agencies, and risk managers, and financial regulations, and exotic financial instrument more and more. They will become as ordinary as auditors, and consultants, and police officers, and health regulations are now. And never forget, there was a time when auditors and police officers and health regulations were innovations and there were people who believed strongly our society did not need them.