VAR modelling for non-financials

An economic consulting firm owned by Marsh & McLennan Companies, the insurance, asset management and consulting giant, has developed a model to measure the risk of corporate cashflow shocks. The firm, White Plains, New York-based National Economic Research Associates (Nera), was prompted to develop its model, called Cash Flow at Risk (C-FaR), by inquiries from Marsh’s consulting clients, who wanted to know if there was a parallel to value-at-risk for non-financial corporations.

“These people want to know the probability of their running out of cash and being unable to fund investment expenditures,” says Jeremy Stein, a professor of economics and corporate finance at Harvard University and a consultant to Nera. The firm plans to market the model primarily to Marsh’s existing client base.

C-FaR, launched in January, provides a probability distribution of company cashflows one quarter or one year in the future. Stephen Usher, a consulting economist who helped devise the product, says the model will be valuable for companies seeking to stress-test their capital structures. “They can ask whether, in a one-in-20 worst-case scenario, they can continue to fund their value-creating investments,” he says. It can also be used to enhance disclosure and as a tool for doing cost-benefit analyses of risk management policies.

Rather than analysing and aggregating individual risks upward to come to an estimate of a company’s total cashflow risk, Nera has built its model using a top-down approach based on the behaviour of comparable companies. “Operational risk and strategic errors tend to dominate the problems for companies’ cashflows.

Can you build up from the bottom a list of all the risks facing a company and then model them? We decided that this was an intractable problem, and you could miss an important risk,” Usher says. Stein adds: “Some of a company’s exposures are to measurable things like exchange rates or currencies, but how do you capture the risk for Dell that demand for computers will go down?”

Nera has built the C-FaR model’s distribution using 11 years of data on corporate cashflow behaviour, obtained from Compustat, Standard & Poor’s database of public company financials. Nera uses four criteria to find a proper set of comparables for the company being analysed: market capitalisation, profitability, industry riskiness and stock price volatility. Stein says the result provides a level of predictive information similar to an insurance company’s actuarial table.

Currently, the main product in the market for measuring risk to corporate earnings and cashflow is the RiskMetrics Group’s CorporateMetrics system.

CorporateMetrics uses the bottom-up approach that Nera is eschewing. It uses a VAR-style methodology to analyse the effects of changes in market variables such as foreign exchange rates, interest rates, commodity prices and equity prices on a company’s earnings and cashflow.

Stein notes that there is a trade-off between the two approaches. Nera’s top-down approach may incorporate more potential risks at the expense of predictive accuracy for a specific company. Meanwhile, the bottom-up approach used by CorporateMetrics has a high degree of predictive accuracy for specific risks as they relate to a specific company, but it may miss non-market risks like declining demand for a company’s products. “Perhaps there is no one right way to do it,” he says, adding that the two approaches may be useful reality checks on one another.