Abstract
Construction cost and duration are two critical project indicators. It is acknowledged that these two indicators are closely dependent and highly uncertain due to various common factors and limited data for explanatory model calibration. However, the stochastic dependence underlying construction cost and duration is usually ignored and the subsequent probabilistic analysis can be misleading. In response, this study develops a Nataf distribution model of building cost and duration, in which the uncertainties of total cost, unit cost, and duration are respectively quantified by univariate distribution fitting, while their stochastic dependence is inferred by maximum likelihood estimation. This method is applied to the costs and durations of 77 China residential building projects completed between 2011 and 2016. The goodness of fit test illustrates that the data conform well to the developed model. The conditional distributions of cost and duration are then derived and the corresponding conditional expectations and variances are given. The results provide the distribution of building costs for a desired duration and the expected duration given a budget. This, together with the ability to update probabilities when new project information is available, confirms the potential of the proposed model to benefit precontract decision making from a risk perspective.
Introduction: The Independence Fallacy
In construction management, the adage "time is money" reflects a deep statistical reality. Project cost and duration are inextricably linked; shortening a schedule typically requires resource intensification, driving up costs, while budget constraints often lead to project delays. Despite this, traditional risk models frequently treat cost and duration as independent variables. This oversight can lead to significantly flawed risk assessments, as it ignores the non-linear stochastic dependence between these two critical indicators.
Furthermore, construction data rarely follows a standard normal distribution. Variables like total cost and duration are often skewed, requiring more flexible distributions like Weibull or Gamma for accurate modeling. This paper introduces a robust probabilistic framework using the Nataf distribution to capture these complex interdependencies.
Methodology: The Nataf Framework
The researchers employed the Nataf distribution method, chosen for its ability to handle non-normal marginal distributions while maintaining a clear structure for modeling dependence. The process follows a structured three-stage workflow:
- Univariate Fitting: Each variable (Total Cost, Unit Cost, Duration) is fitted to its best-performing distribution (e.g., Weibull, Gamma, or Log-Logistic) rather than forcing a one-size-fits-all model.
- Nataf Transformation: Diverse, non-normal variables are mapped into a joint standard normal space. This allows for the calculation of a correlation matrix that describes the underlying dependence.
- Maximum Likelihood Estimation (MLE): Dependency parameters are inferred from historical data to ensure the model reflects real-world behavior.
- Conditional Inference: The model enables "what-if" queries, allowing managers to derive the probability distribution of one variable given a specific value of another (e.g., cost distribution given a fixed deadline).
Case Study and Key Insights
The model was validated using a dataset of 77 residential building projects in China completed between 2011 and 2016. All costs were adjusted for inflation to ensure a consistent baseline.
The Cost of Speed
The study confirmed a significant negative correlation (-0.4949) between Total Cost (TC) and Duration (D) in the dataset. This highlights the financial burden of "crashing" a schedule. For example, shortening a project's duration from 837 days to 355 days caused the expected total cost to surge from ¥5.64 million to ¥19.58 million.
Sensitivity and Risk
A key finding was that Total Cost is more sensitive to changes in Duration than Unit Cost is. Moreover, the model quantifies the increased variance—and thus higher financial risk—associated with fast-tracked projects. Conversely, a restricted budget (at the 5th percentile) was shown to push the expected duration to 795 days, compared to just 392 days when the budget is ample (95th percentile).
Conclusion
The Nataf-based model provides a dynamic tool for pre-contract decision-making. By allowing for Bayesian-style updates as new information becomes available, it offers a live risk-management framework. While the current study focused on residential projects in China, the methodology is adaptable to other sectors. Future work could incorporate additional explanatory variables, such as project complexity or environmental factors, to further refine its predictive power.
Summary updated