Market risk is the risk of loss related to changes in the market value of a portfolio of financial instruments.
Risk, whether it be market risk or otherwise, has two components that must both exist for there to be a risk (according to Holton):
- an exposure: in this case the holding of a portfolio of financial instruments
- an uncertainty: in the case of market risk, the uncertainty lies in the future evolution of the market value of the instruments held. In the event of adverse market developments, the investor is exposed to a loss instead of the expected profit.
Market risk and economic risk should not be confused. Economic risk management is based on long-term investment planning and uses ROI (return on investment) as a fundamental indicator. Market risk management is short-term management (which does not necessarily mean short-termism), in the sense that it aims to avoid substantial losses in the long term by limiting them over a succession of short-term periods.
We immediately notice that risk and result are closely linked. Zero risk does not exist when there is an investment, and as a general rule, the potential gain increases with risk taking. However, there are more or less efficient investment choices in terms of the benefit/risk ratio. The set of portfolios that maximize the potential outcome for a given level of risk (or minimize the level of risk for a given potential outcome) is called the "efficient frontier".
As for the result, it is closely linked to the valuation of the financial instruments. From this valuation will ultimately result the financial flows generated by the position held, financial flows that directly impact the income statement. This is why risk management is ultimately about asking questions such as:
- What is my position worth today, compared to what it was worth at the time of the acquisition? (unrealised gain or loss)
- What will my position be worth tomorrow, in a week, in a month, all other things being equal?
- What would my position be worth if this or that market indicator changes?
Valuation of financial instruments
Market risk measurement methods therefore require the ability to measure a fair market value (mark-to-market) of the assets held. The calculation of a market value is possible for assets for which there is an active and liquid market. In the absence of such as market, mathematical models (mark-to-model) are used and the question of liquidity risk adds to the management of the market risk itself. Indeed, the market value of the instrument only makes sense if it is actually possible to liquidate the instrument at that price, or if it is certain that the instrument will be held until maturity.
For derivatives, one of the main distinctions with regards to valuation is whether or not the financial instrument is linear. An instrument is said to be linear if its payoff function (result at maturity based on the market value of the underlying) is linear or quasi-linear. Such instruments include spot, forward and futures positions. Non-linear instruments include options, exotic derivatives and bond options (hybrid products).
Market risk factors
The market data, which generate market risk for the trading portfolio and on which its valuation depends, are usually classified according to asset classes:
- Foreign exchange
- Raw materials
In general, instruments can be categorised by asset class according to the type of market risk to which they are exposed. However, a device is rarely "chemically pure" in terms of the risk factor to which it is exposed. For example, a bond is classically classified as an "interest rate instrument", but a bond denominated in a currency other than the portfolio's reference currency also contributes to currency risk. In addition, a corporate bond is exposed to a downgrade risk of the rating of its issuer, which adds a credit component to the risk...
Credit risk is traditionally considered separately. However, in view of the "marketisation" of credit risk, and the existence of instruments specific to this new market such as credit default swaps, the risk of loss related to changes in credit spreads can be considered both a credit risk and a market risk.
Market indicators can be imported directly from various sources provided by financial information broadcasters such as Bloomberg or Reuters. The problem with these data is that, however recent they may be, they only provide a picture of a past situation. However, to measure the risk, it is necessary to anticipate the behaviour of these indicators in the coming days.
For this, we will have to build a model of the behaviour of market indicators, either on the basis of historical data, or on parametric functions, or more often on a combination of the two. For example, the future behaviour of a market variable may be based on its past behaviour, but with a higher weighting for recent fluctuations.
Measuring market risk
To sum up what has just been said, measuring market risk involves market indicator models as well as financial instrument valuation functions.
It is then necessary to define a risk metric and calculate this indicator for the portfolio in question. A measure of risk is obtained by the result of this calculation. It is immediately clear that the methodological choices, both in terms of the definition of the indicator chosen and the way it is calculated, have a decisive influence on the measurement of risk.
A good indicator is one that gives coherent measures of risks (according to Hull), i.e.:
- If portfolio A consistently produces a lower result than portfolio B, then its risk measure must be higher than that of B
- If an amount of cash K (risk-free asset) is added to the portfolio, then its risk measure must decrease by the same amount (K)
- If the value of a portfolio is multiplied by an amount M, then its risk measure must be multiplied by the same amount (M)
- If two portfolios A and B are merged, then the risk measure of the resulting portfolio must not be greater than the sum of the initial risk measures of A and B.
Some risk indicators can be more or less well suited and consistent, depending on the nature of the financial instruments held.
As mentioned above, risk has two dimensions, a quantitative dimension (how much I risk losing) and a probabilistic dimension (with what probability). Risk indicators can quantify one or the other aspect, or a combination of both.
Moreover, a portfolio rarely consists of single instrument, but as we have just seen, the risk indicator is more or less well suited to the asset class (interest rate, equity, foreign exchange, commodities, credit) and the linearity (or not) of the instrument. Hence the search for synthetic indicators, which led to the concept of Value at Risk, or V@R. These synthetic indicators also need to take into account the correlations between the different market variables involved in the calculation. The correlation between two variables indicates their tendency to move simultaneously, in the same direction or in opposite directions (correlation does not necessarily mean causality).
Finally, it is important to take into account the use that will be made of the risk measure thus produced. The front office is looking for precise, single-instrument, reactive models. The risk department will adopt a more conservative and comprehensive view of the level of risk and is looking for synthetic indicators, which also have the advantage of facilitating dialogue with the bank's management.
On the Web
- The Risk Management and Financial Institutions book, author John C. Hull. If you lack time or money, you can find slides here.
- Value at Risk: Theory and practice - Glyn Holton