Payment delay as the dominant predictor of schedule overrun in construction Projects
This study employs a novel SHAP-attributed Explainable-AI framework across three major Iraqi construction projects to empirically demonstrate that payment delays are the dominant predictor of schedule overruns, thereby providing the first evidence-based justification for prioritizing procurement reforms in Iraq's public construction sector.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine trying to bake a massive cake for a thousand people, but every time you need to buy flour, sugar, or eggs, you have to wait for the head chef to sign a piece of paper. If the chef is slow, the ingredients never arrive, the bakers get bored and leave, and the cake ends up being late and half-eaten. This is essentially what happens in the world of construction, but instead of cakes, they are building skyscrapers, housing complexes, and bridges. For a long time, people in Iraq have suspected that the government is the "slow chef," causing projects to drag on forever because payments for work done are delayed. However, in the world of science, "suspecting" isn't enough; you need proof. This is where a branch of computer science called Artificial Intelligence (AI) steps in. Think of AI as a super-smart detective that can look at thousands of clues at once to figure out exactly what caused a problem. But there's a catch: sometimes these AI detectives are "black boxes," meaning they give an answer but won't explain why. To fix this, scientists use a special tool called SHAP (which sounds like a friendly name but stands for a complex math method). SHAP acts like a magnifying glass, breaking down the AI's decision to show exactly how much each clue—like bad weather, late materials, or slow payments—contributed to the final mess. The big question is: Is the slow payment the main villain, or is it just one of many equally bad guys?
In this study, a researcher named Layth Shakir from the University of Baghdad decided to put this detective work to the test on three massive construction projects in Iraq. These weren't just small houses; we are talking about a $772 million commercial tower, a $10.1 billion residential development with over 100,000 homes, and a $4.7 billion marine infrastructure project. Together, these projects were worth over $15.57 billion. The researcher built a two-layered AI system to act as a digital watchdog. The first layer, a "CNN-LSTM," was like a security camera that watched the construction sites in real-time, looking for anything weird happening, like a sudden stop in work. The second layer, a "Transformer" model, was the crystal ball that tried to predict when the project would be finished and if it would run out of money or time.
Once the AI had made its predictions, the researcher used the SHAP magnifying glass to ask: "What was the single biggest reason these projects were late?" The answer was surprisingly clear and consistent. Across all three very different projects, the "Payment Delay Index" (PDI)—a score measuring how late the government was paying the builders—was the number one predictor of schedule overruns. It wasn't just slightly important; it was the heavyweight champion. The AI gave the payment delay a "importance score" (called a phi value) of 0.313 on average. The next closest suspect, the "Critical Path Delay Index" (which tracks the most important steps in the building process), only scored 0.214. This means the payment delay was about 46% more influential in causing delays than the next biggest factor.
The study suggests that this isn't just a problem with one specific building or one unlucky contractor. Because the result was the same for a commercial tower in Baghdad, a massive housing project, and a marine project in Basra, it points to a deeper issue: the way the Iraqi government handles payments is a structural part of the system, like a slow engine in a car that makes the whole vehicle run poorly, no matter who is driving. The researchers also found that when they used this AI system to help manage the projects, the overall "health" of the construction projects improved by an average of 14.7 points on a new scale they created called the Iraq Construction Health Index (ICHI).
The paper explicitly rules out the idea that payment delays are just a minor issue or that they are equally important as other risks like material shortages or workforce problems. While those things do matter, the data shows they are often just side effects of the main problem: not getting paid on time. When the money is late, builders can't buy materials, and subcontractors can't pay their workers, causing a chain reaction that stops the work. The researchers are quite sure about this ranking because they used a statistical method called "bootstrapping" to double-check their work, and the results held up every time. However, they are careful to note that this is a "proof-of-concept," meaning it's a very strong test run that proves the idea works, but it's not yet a guarantee that every future project will be fixed just by using this tool. They also admit that because they only looked at three projects, they can't say for sure that this applies to every single construction project in Iraq, but the evidence is strong enough to suggest it's a major systemic issue.
In the end, this research provides the first hard, data-driven evidence that if Iraq wants to stop its construction projects from being late, the most effective place to start is by speeding up the payment process. It's like realizing that the cake is late not because the oven is broken or the bakers are unavailable, but because the person holding the key to the pantry is taking too long to open the door. By using advanced AI to shine a light on the problem, the study offers a clear path forward for fixing the system.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.