TITULO V BASE LIQUIDABLE
artículo 37.- Deducción por inversiones en activos no corrientes nuevos
The accuracy of the cost estimate was found to positively correlate with the stage of project planning, seemingly due to more accurate scope specification and more exten- sive data availability. Furthermore, the project organization should define the purpose of the estimate and the resources that can be allocated for the process of cost estimation. Considering that such a small share of “suspects” is converted to customers (Figure 20), the firms cannot afford to produce a detailed customer solution design and detailed esti- mate. AbouRizk et al. (2002) conducted a case study of the estimation accuracy bound- aries at different stages of project planning in construction projects. The result of their study is shown in Figure 26.
Figure 26. Accuracy of cost estimation at different stages of project planning.
The figure shows the observed accuracy boundaries at the stages of project planning. The strategic estimate is used in managerial purposes, for example in a milestone “go/no-go” decision (Doloi, 2013), and the observed accepted accuracy at that stage was ±50%. At the conceptual stage of the planning process, the accepted accuracy was about ±30%. At the planning, or so-called pre-design stage of the process, the accepted accuracy range was observed to be ±20%. Later, at the design and the implementation stages of the project, the accepted variance corridor was calculated to be ±10%. The implementation phase of project lifecycle does not belong to the preparation phase, but us often used in accuracy boundary setting, since the benchmark value for actual ex- penditure is obtained only after the project completion. The study was carried out based on the data from 213 construction projects, and the construction industry is notorious for cost overruns (Shehu et al., 2014; Kwon and Kang, 2018). The investigation on varying industrial standards is to be undertaken to gain an understanding of acceptable cost estimation accuracy boundaries with respect to the stage of the project lifecycle. Canadian Ministry of Transport and Infrastructure classifies their cost estimates based on the idea of accuracy boundaries into estimate levels. The estimate levels are concep- tual, planning, preliminary, design and pre-tender, which are associated with the level of completeness of project development (MOTI, 2013). The conceptual stage is associated with 0-2% of development completeness, planning – 1% to 15% completeness, prelimi- nary – 10% to 40%, design – 30-90% and pre-tender – 80% to 100% development com- pleteness. The accuracy ranges are defined with a 90% confidence level and are ±35% for conceptual and planning estimate level, ±20% for preliminary and design estimate levels and ±10% for pre-tender estimate level (MOTI, 2013). The accuracy boundaries for the earliest stages of project development are stricter than the ±50% observed by
AbouRizk et al. (2002). The two reasons for a narrower accuracy range may be the in- dustrial differences as well as the evolution and computerization of cost estimation prac- tices.
Taylor (2007, p.108) published the accuracy boundaries for the project data analyzed by the Project Management Institute. The three estimate categories that the author defined are a rough estimate, budgetary proposal, and definitive estimate. The accuracy ranges for the three classes respectively are -25% to +75%, -15% to +25% and -5% to +10%. The same ranges are collaborated by Kerzner and Kerzner (2013, p.681), whose guide- lines are also based on the PMI’s data. However, Kerzner and Kerzner (2013) addition- ally emphasize that the acceptable accuracy boundaries depend on the organization and the industry they operate in and are to be determined case-by-case. Humphreys (2005, p.59) carried out a review of standardized cost estimate classifications and accuracy boundaries. The author describes that the American National Standards Institute in 1991 defined three classes of estimates – order-of-magnitude, budget estimate, and definitive estimate. The respective accuracy boundaries of the classes are -30% to +50%, -15% to +30% and -5% to +15%.
Washington State Department of Transport employs a different set of guidelines for their estimation accuracy ranges. WSDOT (2015) defines five different estimate levels corre- sponding to the stage of the design completion. Planning includes stages of completion from 0-2% and 1-15%, scoping is a stage at 10-30% of development, design stage cor- responds to 30-100% design completion and PS&E (plans, specs, estimate) is the last stage of project development with 90%-100% degree of completion. The five accuracy ranges that correspond to each stage of the project design development are:
• Planning (0-2%): -50% to +200%, • Planning (1-15%): -40% to +100%, • Scoping (10-30%): -30% to +50%, • Design (30-90%): -10% to +25%,
• PS&E (90-100%): -5% to +10% (WSDOT, 2015, p.4-14).
As can be noticed, the ranges defined by WSDOT are not symmetrical, indicating that the projects are more prone to cost overruns. Additionally, the magnitude of the range of accuracy is substantial, compared to that defined by MOTI (2013) and AbouRizk et al. (2002). The reason behind the larger spread is the length of forecasting. Compared to the investigated construction projects that lasted around 3 years (AbouRizk et al., 2002), the forecast at the planning level of estimate carried out by WSDOT is 20- and 10-year plans respectively. It follows, that the accuracy of the estimate is dependant on the length of the project.
Queensland Government (2017) sets yet again a different standard for the accuracy of their cost estimation for transport and main roads related projects. The four levels of cost estimate accuracy boundaries are defined at the four stages of project development: concept, early design, design, and implementation. The concept phase has an estima- tion accuracy boundary of -15 to +20%; the early design stage is expected to range in accuracy from -10% to +15%. The design stage has boundaries from -5% to +10%, and at the implementation phase, the range of estimates is set at -2.5% to +5% (Queensland Government, 2017, p.5). The intervals set by Queensland Government are also asym- metric, expecting more cost overruns rather than cost savings.
Department of Energy (DOE, 2011) presents another classification employed in the cost estimate classification and accuracy range boundary setting. DOE defines five classes of estimates corresponding to the degree of the project definition. Each of the classes is assigned with a low and high range of expected accuracy, which is defined at a 50% confidence level. The accuracy levels are set as follows (DOE, 2011, p.15):
• Class 5 (0-2%): Low range -20 to -50%; High range +30 to +100%, • Class 4 (1-15%): Low range -15 to -30%; High range +20 to +50%, • Class 3 (10-40%): Low range -10 to -20%; High range +10 to +30%, • Class 2 (30-70%): Low range -5 to -15%; High range +5 to +20%, • Class 1 (70-100%): Low range -3 to -10%; High range +3 to +15%.
This performance standard is generalized, thus is includes the high and low ranges of estimation accuracy, instead of a single accuracy range. The guidelines are provided for energy and construction projects, although, it is mentioned that the guidelines are of a recommendatory character rather than mandatory. The ranges of estimation accuracy were developed in 1997 by AACE International as an elaboration of the previous stand- ard proposed by ANSI in 1991. AACE International is the Association for the Advance- ment of Cost Engineering and has issued a series of recommended practices for cost estimation (AACE International, 2016). The summary of all the reviewed accuracy boundaries is presented in Appendix B.
Overall, the asymmetric range of estimation accuracy boundaries suggests that in gen- eral, the projects tend to face uncertainties leading to cost overruns rather than cost savings. Love et al. (2013) state that normal distribution is not applicable to model the cost overruns since normal distribution assumes that the values are symmetrically situ- ated around the mean. Cost overruns are somewhat random and in order to model cost overruns, probability density functions are often used (Love et al., 2013). The probability density function is a basis of the Monte Carlo method, which is used in cost estimation. The deduced tendency of cost overruns being more likely than cost underruns, GAO
(2009, p.38) collaborates that the cost estimate tends to increase throughout the project’s lifecycle. It is noteworthy that Taylor (2007) defines an acceptable variance corridor of ±10% deviation from the baseline. Taylor (2007) claims that if the cost variances through- out the project exceed 10%, it typically cannot be recovered to attain the variance at completion of less than 10%.
The accuracy boundaries are required for setting contingencies. Contingency and man- agement reserve are used to alleviate the negative effect of anticipated project variance. Venkataraman and Pinto (2008, p.97) express contingency as “a cushion against time and money variances”, although, the researchers also mention that the customers often see contingency as additional expenditure rather and may get an impression that the seller organization is incapable of carrying out a high-quality estimation nor producing a detailed design on time. Hence, it becomes of utmost importance for the seller organi- zation to make sure that the contingency percentage is established correctly and corre- spondingly to the degree of uncertainty associated with a particular stage in a project’s lifecycle.