Fund Reserves - Laserfiche WebLink

Home Browse Search

the"Triple-A"approach to uncertainty: GFOA has developed a quantitative accept,assess, and augment3 The way to decide the model of the likelihood and mag- Accept. Uncertainty is inevitable, nitude of potential risks, covering a appropriate amount for your ten-year period.A relatively long time- and people are generally overcon- frame is necessary because man risks fident in their ability to predict the local government is to better rY Y governments are concerned about future while underestimating uncer- understand your risks. occur infrequently, so a ten-year time- Triple-A method is to accept that tainty. Therefore, the first part of the frame provides a better opportuni- ty to see the potential effects. The we are subject to uncertainty and to model also assesses the combined effect of all the risks the broaden our expectations for what could occur.For example, another recession will eventually occur.The Great Recession government faces. was of historic magnitude, but we should not assume that The combined effect of the risks a government is subject to is the worst possible recession we could experience. Many has to be considered because the combined effect can look economists think that the downturn could have been far very different from each risk in isolation. worse if the federal government hadn't intervened as it did. Augment. The third part of Triple-A is about raising our Few people would argue that political parties have become expectations of risk and making sure we don't underestimate more effective since 2007,so it is not difficult to imagine them risk by getting distracted by "average" numbers. The key is failing to work together to successfully to respond to a future thinking in probabilities. recession. GFOA uses a technique called "Monte Carlo analysis" Assess.We need to first assess which risks we are subject to build probabilistic models. This is the same technique to. For most local governments,the most salient risks will be used by insurance companies, aerospace manufacturers, recessions and natural disasters like earthquakes,hurricanes, and other industries that require a detailed understand- tornados,wildfires,floods,extreme snowfall,etc.There is also ing of risk. This method has been around since the 1950s, the specter of man-made disasters such as hazardous material but up until recently it required powerful computers and spills or the closure of a major employer/tax producer. specialized software. Now it now can be used by anyone Next,we need to assess the likelihood and potential magni- who has access to Microsoft Excel, thanks to the work of tude of these risks.There are many potential sources of data ProbabilityManagement.org, which has developed open- to examine, including: source technology that allows Monte Carlo to run on Excel with no macros and no add-ins. Monte Carlo simulates thou- ■The local government's own experience. sands of possible versions of the future.We can then examine ■The experience of other local governments where those scenarios to see how often different levels of damage extreme events have occurred. occur. For example, if we run 1,000 scenarios, and in 100 ■National institutions that track data about extreme events. of these scenarios a local government incurs $20 million or The United States Geological Survey has data about earth- more in damages within five years,then there is a 10 percent quake likelihood, for instance. chance(100 in 1,000)of incurring at least$20 million in dam- ■Staff judgment. Of course,we must take steps to counter- ages within five years.Monte Carlo analysis allows us to easily act the shortcomings of human judgment,as we described see and analyze the more extreme potential outcomes. earlier. For instance, knowing that human judgments are Knowing the chances of incurring different amounts of typically 50 percent too narrow, as noted in the example damages allows a local government to judge its appetite for described earlier, GFOA doubled the range of the fire risk and size its reserve accordingly. For example, if there department's estimates to complement the data we had. were only a 1 percent chance of incurring risks totaling $50 You can also break estimation problems down into very million or more in ten years, most local officials would prob- small parts that are easier for staff to estimate and then ably be comfortable with keeping reserves at less than $50 add these small estimates together. million.As we get into less extreme scenarios,public officials 12 Government Finance Review I June 2019