Cost overruns and delays are usually blamed on execution. The evidence says the fate of large projects is mostly sealed before ground is broken.
Every region building at scale carries its gallery of troubled giants. The refinery years late, the rail line double its budget, the stadium finished in a heroic sprint that quietly cost a fortune. Across the Gulf's development programmes, Africa's infrastructure push and Asia's urban expansion, capital projects of unprecedented ambition are underway, and the global base rates are sobering. Large projects overrun their budgets far more often than not, and the bigger they are, the worse the odds.
The instinctive explanation blames delivery. Contractors underperformed, conditions surprised, prices moved. Study the autopsies carefully, as researchers have across thousands of projects, and a different pattern dominates. Most overruns trace to decisions made before construction started. Optimistic estimates produced to win approval. Scope agreed before the ground was understood. Contracts that transferred risk to whichever party could least control it. Projects, in short, fail on day zero and spend the following years revealing it.
The mechanics deserve naming because they repeat everywhere. First comes strategic optimism, the systematic understatement of cost and time that occurs when the people estimating a project are the people who want it approved. It is rarely dishonesty. It is incentive structure, and it guarantees that budgets reflect hope rather than base rates. Second comes premature certainty, locking designs and announcing dates before geotechnical, regulatory and logistical realities are known, which converts every later discovery into a claim, a delay and a headline. Third comes risk misallocation, contracts drafted to push risk away from the owner and onto contractors who price it back with margin, dispute it when it lands, or fail under it entirely, at which point the risk returns to the owner enlarged.
The correctives are known and unglamorous, which is why they are unevenly applied. Reference class budgeting anchors estimates in the actual outcomes of similar projects rather than in this project's persuasive spreadsheet. Front end loading spends real money on studies, ground investigation and design maturity before commitment, the cheapest phase in which to discover problems. Contracting for alignment pays contractors to finish well rather than to claim well, through target cost mechanisms, shared pain and gain, and early contractor involvement in design. And modularity, breaking the giant into repeatable units delivered in sequence, converts a bet the company gamble into a learning curve, which is why programmes of repeated similar assets show dramatically better cost performance than singular monuments.
Owners hold one more underused lever, the strength of their own team. A capable owner organisation with continuity of leadership, authority to decide quickly and honest reporting lines catches drift early. A thin owner team supervising powerful contractors through layers of consultants discovers problems at the pace the contractors choose to reveal them.
For governments and boards approving the next wave of large capital commitments the implication is direct. The moment of maximum influence over a project's outcome is the moment of approval, when enthusiasm is highest and scrutiny feels like disloyalty. Institutionalise the scrutiny anyway. Demand base rate comparisons, independent estimate reviews and evidence of design maturity before dates are announced. A month of rigour at day zero routinely saves a year of recovery at year three, and unlike recovery, it never makes the news.
