Causes of delays in Saudi Arabian public sector construction projects

Abstract

Many public construction projects have been undertaken throughout the Kingdom of Saudi Arabia as part of the government's national development plans in the past three decades, with significant public expenditure involved. One of the critical problems concerning these projects is the frequent and lengthy delays that occur. In order to improve the situation, it is first necessary to identify the major causes involved. Several studies have already been reported which do this but all use different sets of variables. Also, none has attempted to identify the extent to which improvements are possible in practice. A new survey is reported that uses all the variables from the previous work and measures for both current degree of effect on delays and the extent to which each can be practically improved. These are contained in seven groupings: client, contractor, consultant, materials, labour, contract and relationship‐related causes. The survey covers a sample of 86 clients, contractors and consultants working in the Saudi construction industry. The analysis reveals some considerable heterogeneity between the cause groupings and respondent groupings in terms of means and correlations, apparently partly due to lack of knowledge of respondents and a tendency for the consultants to blame the contractors for the delays and vice versa. The main results, therefore, are disaggregated to reflect the views of each respondent group concerning each group of causes. In general however, it is found that the most influencing current cause of delay is the lack of qualified and experienced personnel—attributed to the considerable amount of large, innovative, construction projects and associated current undersupply of manpower in the industry.

Keywords:

• Delays
• public sector
• Saudi Arabia
• survey.

Notes

1. ANOVA is a parametric method of analysis for interval level data, while Likert scales are intrinsically ordinal. However, ANOVA is known to be robust in non‐parametric settings and is regularly used for Likert scales (e.g. Chan and Lee, 2008; Lam et al., 2006; Zou and Lee, 2008).