Possibilities and Probabilities: Forecasting Your Future Financial Needs
What does it cost to operate and maintain an industrial facility over its lifetime? If you guessed 6 to 10 times what you paid for it, you would be correct. If you are surprised, you are typical of most business owners and management teams. The next question that most decision makers ask is related to the magnitude and timing of the expenditures.
The best forecasting approach in situations such as the post-COVID19 recession is to use spreadsheet models with Monte Carlo simulations. According to the international risk standard, ISO-31000, an approach using Monte Carlo simulations is the one of the most strongly applicable approaches for quantitatively accounting for uncertainty and risk. A prediction on the back of a napkin or a forecast using a spreadsheet with best-guess point estimates will not get the job done – at least in situations with complexity (multiple parts) and high amounts of future uncertainty.
What is this Monte Carlo stuff?
This is a common question that is asked by a wide range of business managers and with wide ranges of educational backgrounds. First, it originated in the 1940s by a nuclear physicist working on the atomic bombs that ending World War II. That makes it old. And probably smart (technical) too.
Second, it has nothing to do with gaming and a lot to do with statistics. The name comes from one of the creators and named for the habit of gambling through the ages at Monte Carlo. In essence, if we play enough hands, shoot enough dice, and spin the roulette wheel many times, then our results (possibilities) can be associated with frequencies (probabilities) that are driven by statistics and physics.
Third, it is not that hard to do. There are a number of affordable spreadsheet add-ins or simulations can be organically developed in Excel. The computer has taken the manual simulation aspects out of it. Monte Carlo methodologies went “quite” until computing became cheap in the latter 1990s.
What is wrong with traditional point estimates?
The short answer is nothing when situations are simple and there is limited uncertainty. This is not the case with most industrial facilities where there are thousands of parts and equipment and where products, operating environments, and maintenance practices are regularly changing.
The long answer to what is wrong with traditional point estimates is multi-fold and has several important ramifications. First, it is left up to the forecaster to select the value of each input variable. Most technical professionals tend to be conservative in order to avoid shortfalls or simply being wrong. This conservatism in the input parameters leads to overstated financial needs and/or early spending needs.
Second, the need to have really good estimates of input parameters leads to collecting vast amounts of data. Collecting data takes time and money – usually lots of both. It can be assumed that developing probability distributions takes time and money too; however, the difference is that existing history can be used or approximated so that a ‘spin of the wheel’ can be generated. Decision makers, not analysts, can then decide on the importance if, and what types, of data is needed to improve the decision.
Third, forecasts based on point estimate inputs generate a single output. The result is a ‘beat the budget’ which is the product of the analyst rather than the probabilities and probabilities. With a Monte Carlo approach, we understand the full range of probabilities and, more importantly, inform us of the chance of success if we only have a fixed amount of money for our facility.
Summing It Up
The best forecasting approach in situations such as the post-COVID19 recession is to use spreadsheet models with Monte Carlo simulations. No one can predict when a piece of equipment will exactly fail, the degree to which operating conditions will change based on market conditions, or the amount of organizational capacity that will be available to maintain our systems. The possibilities and probabilities are marked with high degrees of complexity and uncertainty. Our methods for forecasting our financial needs must be aligned to meet the challenge.