Print this article | Return to Article | Return to CFO.com
The financial crisis obliterated corporate forecasts. Now, CFOs struggle to assess what lies ahead.
Vincent Ryan, CFO Magazine
December 1, 2008
Former Federal Reserve chairman Alan Greenspan made plenty of news in October when he admitted before Congress that he had "found a flaw" in his model of financial reality, but another, less-publicized portion of his testimony should resonate with CFOs. The Fed's failure to anticipate the risks and consequences of securitizing and selling mortgages, Greenspan said, was due to poor forecasting.
"We're not smart enough as people," he said. "We just cannot see events that far in advance. There are always a lot of people raising issues, and half the time they're wrong."
Few if any CFOs have the luxury of throwing up their hands and citing human limitations as an excuse for their inability to foresee adversity. Indeed, uttering such words in public would immediately erode any confidence that a board and investors had in a finance chief. Suggest that you agree with Greenspan that forecasts often have no better than a 50-50 chance of being right and the impact on your career may be all too predictable.
And yet Greenspan may be right. For example, in a recent lawsuit filed by chemical maker Huntsman to force a merger with Hexion, Hexion's CEO testified in open court that forecasting raw-materials costs is almost impossible. According to documents supplied by Huntsman in a Securities and Exchange Commission filing, Hexion's method of forecasting raw-materials costs is to arbitrarily prognosticate that they will remain flat, pretty much ensuring the numbers will be wrong. (Hexion declined to comment, due to litigation.)
Meanwhile, forecasts of 2008 revenues and profits have missed the mark substantially at many organizations. That's not remarkable, given that the speed and severity of the third-quarter collapse caught nearly everyone off guard. What is surprising is the degree to which forecasting, a task that few companies have ever felt very confident about, is now so difficult that in a spot survey of 250 Website visitors (see "The Dark Side" at the end of this article), CFO found that 70 percent say they can't forecast more than one quarter out. (Of those, one-quarter said they can't forecast more than two weeks ahead and an equal number said "We're in the dark.")
But abandoning the prediction game is not an option. Companies are now redoubling their forecasting efforts, even though (and perhaps because) the horizon is brimming with unknowns. In a recent CFO Research survey, 41 percent of senior finance executives said they have strengthened scenario-planning procedures in light of the banking crisis. Retail organizations in particular are scrambling to manage inventory and adjust financial projections with the knowledge that holiday retail sales may decline for the first time in more than 10 years.
"Companies need to get the best perspective on what the future looks like so they can make actionable decisions and share information with investors and stakeholders," says Stephen Lis, partner in charge of business-performance services at KPMG.
The credit crisis and the downturn in world economies may offer a useful lesson: forecasting can't be just about number-crunching. Static spreadsheets filled with hundreds of line items that represent best guesses from operations just won't cut it if a company hopes to produce forecasts that aid quicker, fact-based decisions under stress.
Some finance departments are beginning to incorporate methods such as scenario modeling, sensitivity analysis, and contingency planning to help CFOs think through a wide-ranging set of potential situations, thus avoiding a monocular view of what's ahead. They are refreshing forecasts more frequently, homing in on a handful of measures that have a financial effect on the company (so-called driver-based forecasting), and doing more to provide synchronous information flow between finance and operations.
Taking the Long View
Most companies may feel that adapting any of these techniques, or attempting to refine their current approaches to forecasting, is not exactly top-of-mind at the moment. In an environment like this one, "the urgent drives out the important and they operate in panic mode, instead of developing a longer-term, more-rational understanding," says Nick Turner, co-president of Global Business Network, a scenario-planning consultancy and a member of the Monitor Group.
An exception is The Principal Financial Group, a life and health insurer as well as a large administrator of employer-sponsored retirement plans. The Principal has some exposure to mortgage insurers as well as a commercial real estate portfolio, but it has weathered the credit storm by cutting its dividend and suspending stock buybacks. It also employs a comprehensive forecasting process that includes short- and long-term components and incorporates what CFO Terry Lillis calls "stochastic" modeling — generating a host of scenarios that follow a random distribution.
The Principal develops five-quarter, 5-year, and 10-year forecasts that have a set of conservative baseline assumptions, such as the equity market growing 2 percent a quarter. Then it introduces variances to that baseline to understand the magnitude of the effect on operating earnings, sales, and assets under management — the key drivers. The forecasts are now revisited every two weeks. "You get into problems by trying to model too much minutiae," notes Lillis. "Your best models are those that identify four or five key drivers and focus on the interplay between them."
The stochastic element, which emphasizes randomness, enables The Principal to spot potential blowups. The modeling generates a wide distribution of scenarios, which Lillis uses to determine how much capital the company needs to hold in 99 percent of all possible outcomes, from the best to the worst cases of economic performance. Says Lillis, "It helps us deal with tail risk" — those potential futures that are several standard deviations from the mean. The model also projects items such as cash flow from investments versus the surrender of liabilities, critical for a company that uses asset-liability management.
While 10-year (and even 5-year) forecasts may appear anachronistic at a time when so many companies say they are "in the dark," some industries demand them. At Bonny Doon Vineyard near Santa Cruz, California, it can take 5 years to go from vine planting to bottling. CFO Lisa Kohrs employs a 10-year forecast of revenues and liabilities, factoring in items such as vineyard labor costs, cellar costs, and eventual selling prices in four different channels. To complicate matters, the business is managed as three different companies affected in different ways by environmental, regulatory, and economic forces. Current projects are as diverse as developing a new 300-acre estate vineyard and building a new tasting room.
Whether companies think long- or short-term, the ability to react quickly to events is really all that CFOs can ask of forecasting, say experts. An all-out drive for pinpoint accuracy, especially in light of current events, can be less helpful. The Principal's forecasters, for example, did not foresee the huge drop in equity indexes this past year. "The value of [forecasting] is directional," Lillis says. "If I say to our finance people that the best estimate of our earnings is not good enough, the question becomes, What can we do about it? What drivers do we have to change? Are they within our control? If so, do we pull the lever?"
Generating timely, reliable financial forecasts that help executive management implement decisions faster requires using true driver-based forecasting — tracking the operational measures (such as hours of temporary labor required and associated labor rates in a manufacturing plant) that have a decided financial effect, says Tony Levy, director of product marketing at software firm Cognos (now owned by IBM). Finance personnel have to think in terms of business factors instead of dry lines on a general ledger. For a telesales organization, for example, the drivers might be dollars per deal or the conversion rate of customers — measures that can be influenced by performance.
At W.W. Grainger, a $6 billion distributor of industrial supplies and equipment that averages 120,000 transactions per day, company operations include more than 600 branches, 18 distribution centers, and multiple Websites worldwide. Finance personnel are embedded in the company's business units, such as product management, business development, and marketing. The key to Grainger's business is high inventory availability and service levels at walk-in customer sites, as well as next-day delivery from distribution centers. By sitting alongside internal business partners, finance personnel get a much closer view of things like demand, product uptake, the success of new-product introductions, and supply-side trends. "We rely on sales-force input, marketing analytics, and supply-chain feedback that filters through to finance," says Ronald Jadin, Grainger's CFO.
Finance people can also draw a bead on crucial economic factors like inflation. They get the "micro" view on price increases that suppliers may be planning to pass along, and those increases are entered into Grainger's quarterly forecasts to supplement any macro analysis on inflation that Grainger gleans from economists, Jadin says. The quick relay of information also enables Grainger to better manage the price increases suppliers may try to pass on due to rising commodity costs. "We try to get them to hold off passing along inflationary pressure to us," Jadin says, "by limiting price increases to the annual publication of our catalog."
The company's forecasting process also focuses on contingency planning for a downturn. The company plans three to five major actions it might take if the economy were to soften. "The business units commit to it — that strips the emotion out," Jadin says. "If the problem arises, you just execute the plan."
For satellite services provider Hughes Communications, the most material determinant of the company's profitability is consumer uptake of services. Projections of profit and loss, cash flow, and the balance sheet depend heavily on a three-year model of new consumer subscriptions — the only forecast that Hughes updates every month. The forecast is so important that it can affect investment decisions such as whether or not the company should launch its own satellite (a $400 million prospect) or continue to lease transponders from others, says finance chief Grant Barber.
By staying close to call-center orders and new installations, Hughes took preventive action ahead of the economic downturn two months ago. The company gave Internet service consumers the option to pay equipment and installation fees over time instead of all upfront, which Barber says kept new installations from stalling. "We're constantly turning the knobs and making changes to the consumer models," he adds.
Good forecasts have long hinged on finance's ability to dig into operational metrics, but increasingly companies are realizing that communication, and data, need to flow both ways. "It doesn't make sense to sit in an ivory tower anymore and make projections," says Lillis. "You don't get buy-in from leadership."
Business-unit leaders, for example, can run "distress simulations" on their own, increasing awareness of how their performance alters corporate results. Park Nicollet Health Services, a Minnesota-based primary-care and hospital provider, uses a baseline run-rate model and a second version that executives of its 54 business units can tinker with, says CFO David Cooke. A primary-care-facility manager can change patient volumes and the mix of commercial versus Medicare payers, for example, to measure the effects on profit margin. "Last year we lost $84 million because [the Medicare and Medicaid mix] went up unexpectedly," Cooke says.
Serving a higher percentage of government-insured patients (which is a current trend) means taking steps to boost productivity — asking doctors to see two more patients a day, for example, Cooke says. New software (from Cognos) may help, because it will give business leaders the ability to input data directly rather than route it through finance. That may allow the company to move more quickly as it tackles important challenges such as staffing (it can take a year to recruit a surgeon) and determining whether it should consolidate facilities.
While companies are developing a new appreciation for greater hands-on involvement from business managers, they do see limits, particularly when it comes to tying forecasts to compensation. Mueller Inc., a supplier of metal roofing and steel buildings, recently took such responsibility away from individual managers. "We were very poor at forecasting," says Phillip Arp, CFO of the Ballinger, Texas, company. "Managers couldn't back up forecasting with anything concrete and they felt pressure to give a number higher than the prior year, so we removed the performance contract piece."
Now, individual managers are responsible for executing against a matrix of recommendations based on branch sales levels that Mueller's finance unit provides. "They receive a set of tools that gives them guidance on their potential profitability, which includes margin, compensation, head count, asset levels, and mainline expenses," Arp says.
Age of Certainty
If the downturn in the economy does anything for corporate forecasting, it might be to sway organizations to become habitually conservative in their financial planning.
Apache Corp., a $10 billion energy exploration and production outfit, "constantly reassesses what the future looks like" based on the latest oil and natural-gas pricing and other macro inputs, says Roger Plank, the company's finance chief. But the modus operandi is to "plan for the worst and hope for the better." In its three-year projections, Apache uses a lower-price forecast for oil and gas than market pricing indicates, to prevent having capital spending outrun cash flow if oil prices underperform the futures market. "When I started with Apache, every year there was an escalating assumption, but prices didn't cooperate," Plank says. "The rate of return didn't turn out to be real." Forecasting more conservatively has allowed Apache to build up a $1 billion cash position.
The downside risk is slow growth, but Plank thinks it's a risk worth taking. "Anyone can assume higher prices and borrow against that, but with today's prices that doesn't make sense," he says. "It's worth missing out on a little of the upside in order to be flexible and have money when others don't."
Vincent Ryan is a senior editor at CFO.
Mapping the Future
Can intuition succeed where facts fail?
Multiple standard deviations from the mean are by definition rare, so it's not surprising that a world banking crisis caught many companies unawares, even if it seems wholly predictable in retrospect. Nor is it unexpected that CFOs are now scouring the horizon for more such "black swans." Supplementing straight-line extrapolations with scenario planning — the kind of forecasting that emerged from Royal Dutch Shell's planning department in the 1970s — can help executive teams challenge deeply held beliefs and assumptions, and cut down on blind spots about the future.
Scenario planning creates a set of stories, or narratives, that aid executives in thinking about "multiple, plausible, relevant, challenging futures and how a company might respond if they unfold," says Nick Turner of Global Business Network, a consultancy. "It's thinking about the horizon beyond the horizon." Scenario planning is not an attempt to predict the future or write contingency plans for every possible outlier; it's more about highlighting the limits of executives' perceptions around a core business question — like whether a firm should expand into China.
"Scenarios are really about helping to inform intuition," Turner says, instead of "just looking at the data and the facts." They help companies plan strategic options in light of what future events might imply for their industry, clients, and products. "Scenario planning is like fast-forwarding five years and writing the history of the period — how events unfolded and what the actors were doing," says Turner.
Recently, Global Business Network produced a set of scenarios for the political and economic fallout of the global financial crisis. The consultancy created a matrix defined by two key uncertainties: the depth and length of a global recession and the nature of political leadership during the crisis. The four resulting possibilities range from a scenario in which house prices recover and a mild recession results in global financial markets remaining decentralized, lightly regulated, and highly liquid (a scenario now increasingly less plausible, unfortunately) to a story line called "the global unraveling," in which a long, deep recession occurs and "outrage about the socialization of economic losses provokes populist, protectionist, and anti-immigration sentiments" among nations.
If scenario planning sounds resource-intensive, it can be. A company may have 20 to 30 people constructing scenarios simultaneously. And a broad, diverse group of corporate leaders must be engaged: these include decision makers, knowledge-holders, and "the creative and the curious," Turner says. "If you just listen to yourself and the same people you always talk to, you'll miss the big idea — often people won't want to consider certain possibilities, because they don't fit their 'mental maps.'"
Senior sponsorship is essential, and even the consultants caution against leaning too heavily on consultants. Much of the value in scenario planning is the "journey, the conversations with colleagues about what the future may hold," Turner stresses. As one futurist has said, the forecaster's task is to map uncertainty. In this business climate, scenario planning may be the closest thing to a GPS you're likely to find. — V.R.