Predicting cash flow is an important element in Gwinnett County's $1.5 Billion Water and Wastewater Capital
Improvements Program (CIP). Being able to accurately forecast cash flow reduces numerous problems in
prioritizing, scheduling, funding, and completing essential projects. Gwinnett's Department of Water Resources
(DWR), in order to meet the infrastructure demands of a rapidly growing Metro-Atlanta community, has many
hundreds of projects within its CIP. With such an enormous workload and strain on available capital, DWR sought
to achieve better predictability of project cash flow. Because of the quantity and complexity of projects included in
the CIP, a Primavera P3 master scheduling program was used and cash flow projections were developed on a
project-by-project basis. In each project, a cash flow distribution similar to a Gauss-Laplace normal probability
distribution was initially used. Through trial and error, the probability distribution was often modified to
accommodate front-loading or back-loading depending on project type. The resulting information was then used to
consider various financial options to make funds available to complete the projects. Although these cash flow
forecasts seemed to at least get us into the ballpark, these predictions were often overly optimistic, fraught with
technical concerns such as a "bow-wave" effect, and relied heavily on individuals' reasoning and experience. The
results were often inaccurate forecasts which led to under-expenditure of bond funds and a growing desire for a
more accurate model. By analyzing historical cash flow patterns and broadly accommodating the variables, DWR
developed normalized cash flow curves for various project types. These curves yield better predictive models,
thereby improving the forecasting capability of DWR. This paper discusses the analyses and presents the
predictive models used. Includes figures.
| Edition : | Vol. - No. |
| File Size : | 1
file
, 1.1 MB |
| Note : | This product is unavailable in Ukraine, Russia, Belarus |
| Number of Pages : | 8 |
| Published : | 11/01/2008 |
