Special Evaluation Study on Project Cost Estimates

Evaluation Document | 31 July 2004
This special evaluation study aimed to determine the major causes of cost underruns and overruns of ADB projects, assess the contribution of each major factor to cost variations and recommend ways to improve the reliability of cost estimates.

The large number of Asian Development Bank (ADB) projects that have had either significant cost underruns or overruns has been an issue in ADB operations for years. Since 1995, 35% of projects supported by ADB loans had major total cost underruns and 9% had overruns. Cost underruns usually translate into loan cancellations. Over the past 5 years (1998-2002), an average of $1.22 billion in loan proceeds was cancelled annually. Amounts canceled have almost doubled in the past 5 years, from $0.62 billion for 62 loans in 1998 to $1.16 billion for 69 loans in 2002.

Loan savings and cancellations due to cost underruns have an impact on both developing member country (DMC) development and ADB operations. For DMCs, loan savings could have financed additional projects, and for ordinary capital resources (OCR), loans may incur unnecessary commitment fees. For ADB, underruns and overruns may indicate an inefficient allocation of scarce resources.

OED conducted a study to

  • determine the major causes of cost underruns and overruns of ADB projects
  • assess the contribution of each major factor to cost variations, particularly non-robust project cost estimations at loan appraisal
  • recommend ways to improve the reliability of cost estimates, and to streamline guidelines to address and share the risk of unpredictable changes in project costs.

Summary of findings

  • In the initial years, cost overruns predominated. But since the 1980s, cost underruns have become predominant. Important to the shift seems to have been (i) the smaller proportion of civil works components in later projects, (ii) a gradual shift in sector composition of ADB's portfolio, and, possibly, (iii) ADB staff's experience with the preparation of supplementary loans.

  • Over the years, capital investment components in projects have become more standardized and better estimated.

  • Overall, cost estimation is better than in ADB's first 10-15 years, but a tendency remains to overdesign certain smaller or softer components, and to overestimate the costs of equipment and materials. This more than counterbalances the general underestimation of the cost of civil works.

  • Project designs and project cost estimates can be improved, but that partly depends on resources. Nevertheless, unforeseeable factors are important in cost variations, and some degree of cost variation is likely to remain.


  • To improve estimation of processes for project costs, it is essential that adequate resources, in both time and money, be allocated to conduct the required analyses. Better project design and cost estimates can be expected to lead to smoother implementation, fewer loan cancellations, and more positive results from loan portfolios.

  • Improved estimates of project costs is possible with (i) more rigorous consideration of inflation and depreciation factors in the estimates; (ii) more thorough consideration of sectoral and country experiences with cost variations; (iii) estimating and verifying the real commitment to training and institutional development components; and (iv) more realistic estimation (i.e., less underestimation) of civil works components of projects, as well as operation and maintenance components. Better estimates may, paradoxically, lead to more projects with cost overruns. This may require a simpler and quicker process for preparation of supplementary loans, and a more sympathetic ADB attitude toward their approval.

  • Paying more attention to project administration will reduce cost underruns and overruns. Projects mired in implementation problems should be reviewed early, with a view to possible cancellation of part of the loan. The interests of executing agencies to avoid such cancellations should be weighed against the borrowers' interests in reducing the debt burden. Savings can then be used for purposes that yield a return, rather than being tied up nonproductively.

  • Savings in contingencies could be reduced by providing an option to the government/executing agency not to borrow for contingencies (in some projects, the government has exercised such an option). This option should be given only when the borrower has demonstrated its capacity and commitment to finance contingencies when they arise.

  • Periodic updating of cost estimates will help in timely assessment of loan savings. Consistent with ADB's new thrust toward results-based management, standard operating loan administration procedures should require the executing agencies to annually update project cost estimates, which ADB would review thoroughly. This would probably require dedicated resources of qualified ADB staff and consultants.


  • Executive Summary
  • I. Introduction
  • II Objectives and Scope
  • III. Literature Review
  • IV. Major Findings
  • V. Conclusions and Recommendations
  • Appendixes