Saturday, November 21, 2009

Blame the Models? Take a Good Look in the Mirror

Is financial modeling to blame for the recent crisis? As always the fault does not lie with the tool but rather with the user. When was the last time you took a look at the validity of Net Present Value or Option Pricing?

The recent, or ongoing, financial crisis has been attributed, amongst other things, to over reliance on quantitative financial models which replaced good business judgment instead of supporting it.

The reason behind the failure of models wasn't simply poor modeling, for the most part. The main reason was poor business and risk management processes which placed blind faith in the models.

In this post I begin to explore the reasons why financial modeling increased the severity of the recent crisis and more importantly, what are the lessons we can implement to our personal finances?


The need for Financial Modeling

With the development of computing power advanced mathematical models enabled the creation of complex new financial instruments and professions such as financial engineering. Mathematicians and statisticians found their way to investment banks and hedge funds due to the increased demand and profitability of these endeavors.

These financial instruments included, for the most part, new underlying assets, that require complex financial models to model their pricing and behavior. Financial models which sadly broke during the recent crisis.

That is not to say that financial engineering and the use of models have been proven unsuitable for the financial markets. Much on the contrary, financial engineering and complex financial instruments help, in many cases, to increase the level of perfection in the market and ease the transfer of undesirable risks from one party to the other.

Models as a Representation of Reality

A Model, by its definition, is a representation of reality. Economics, for example, as any 1st year student knows is based on very simple models of production, supply, demand and price. The strength and importance of these models is their ability to explain economic behavior even under very simplistic assumptions.

Think of the prisoner's dilemma in game theory or Nash's equilibrium which are very simple models that won their authors Noble Prizes. The beauty of the model is its ability to represent reality with a very limited framework.

As modeling advances assumptions are slowly removed creating more and more complex models which require a better understanding of the mathematic complexity. The reason is simple, the more "free" parameters the model needs to explain the higher the complexity.

Naturally, when dealing with financial instruments and their fair value pricing, models need to explain as much of the price as possible to adequately represent the fair value of the instrument. As such, models have grown quite complex as room for assumptions is very small.


What went wrong?

Before we get too judgmental let's consider the simple, commonly used model of Net Present Value. Net present value is a model used to determine how much a certain stream of cash flows is worth today – or, the present value of the cash flow.

Net Present Value is commonly used to determine the economic sense behind undertaking certain projects and investments. All those in the finance profession, as well as personal finance enthusiasts have probably tried to determine the net present value of a certain undertaking. The problem starts when you dig deeper into the model.

Most people are not aware Net Present Value assumes the following due to the very central role of the cost of capital in the model (further reading on NPV is available here):

  • The existence of an efficient financial market – In order to price correctly an efficient financial market is required. NPV cannot be turned into value in the present if an IPO cannot be performed or valuated correctly.

  • Access to financial markets– Without access to an efficient financial market the entrepreneur, again, cannot transform NPV to value in the present.

  • The existence of diversified investors – Adequate pricing of the cost of capital can only be performed by ignoring the specific risk of the investment. A diversified investor, in finance, sees only the market risk when investing (the reason being a diversified portfolio will stay diversified even after investing in this particular project, for example).


When was the last time the Net Present Value calculation you had performed met these requirements? My guess is never.

The model is a helpful tool but the underlying assumptions and limitation cannot be ignored.

The reason behind why many of the financial models broke has to do with the assumption of liquidity and volume in the markets. Pricing models are built to price instruments under a normal market environment, as part of the normal course of business. When stressed scenarios occur the assumptions behind the models cease to exist and the model breaks.

  • Black and Scholes model very commonly used for option pricing assumes normal distribution of returns (Not true, obviously).

  • The CAPM Model (Capital asset pricing) which is a corner stone in asset allocation is based, again, on normal distribution of returns but also assumes, amongst others, the following highly debatable assumptions: Perfect availability of information, No taxes or transaction costs and a market portfolio which includes all types of assets (!).

With no liquidity in the markets strange phenomenon start to take place, phenomenon which cannot be analyzed through modeling but rather through good judgment and independent thinking.


The Importance of an Adequate Process

The role of an adequate risk management process cannot be overstated. Financial modeling is a tool, nothing more (and nothing less).

Like any tool the one who uses it needs to understand its capabilities and its limitations. Model limitations are inherent as they are, and only can be, a representation of reality.

Financial modeling will not go away. The need is too strong. What does need to happen is the implementation of robust risk management processes which will continue to remind what are the assumptions and limitation of the models and what the proposed mitigations against these limitations are.

The wisdom of the ancient Greek philosopher Socrates is a fitting quote: "I know that I know nothing" he said, not because he knew nothing because he didn't claim perfect knowledge. A humble state of mind is a very good start.

Related Posts:

No comments: