It is hard to predict the construction cost of a building or an infrastructure, due to the differences in design, scale, site conditions, choice of materials, inflation, etc. Traditionally, it is the professional knowledge of quantity surveyors to keep the records of the bills of quantities (BQ) of each available projects for the estimation of construction costs of future project, subject to subjective adjustments such as the time factor.
For preliminary estimation, QS would normally rely on some rule-of-thumb metrics for benchmarking, such as unit cost per CFA (construction floor area). Thus, some QS firms would report the average unit cost per CFA at different cities quarterly (Figure 1).
However, it is just a broad brush figure for benchmarking, without much information on the design, scale and site conditions, etc. So far, there have been very few empirical studies on the accuracy of construction cost (tendering price) prediction, probably due to the lack of market data. In general, tendering results would not be publicly available, and would only be accessible by the responsible QS firms and their clients. Thus, the data are fragmented and cannot be consolidated.
Fortunately, there is another set of data publicly available on all government construction projects, at the Legislative Council homepage, due to the process of budget clearance required to be passed at the Finance Committee of the Legislative Council. The following is a recent example 130KA — Immigration Headquarters in Tseung Kwan O (Legislative Council, 2019a). The standard format of the paper submitted would include the Project Scope and Nature (showing the site area and the net operational floor area), the Financial Implications (construction costs breakdown by trade) (Figure 2), it calculates the unit construction cost per CFA at $37,697 per sq. m. in MOD prices. The Design Drawing is shown in Figure 3.
However, it is hard to tell whether the estimation is accurate or not, as there is little feedback to the Legislature on the actual awarded tender price of each project. (My recent study found a systematic overestimation of the government request on construction cost in comparison with the actual awarded tender price, see Yiu, 2019)
Council members sometimes would also ask for further information on the benchmarking the construction costs of the government project. This case is an example. The government has tabulated 5 government building projects to provide a comparison of construction costs. The table (Figure 4) is extracted as follows:
Certainly, the differences in scale, design, materials, and site-specific conditions are not controlled. Yet, it provides a good attempt for benchmarking of unit construction cost per CFA, by controlling (1) similar type of construction — high rise government office constructed by Design and Build contracts, (2) control time factor by discounting to 2018 prices, (3) control size factor by calculating unit cost per CFA, and (4) similar facilities of each project. The results show a very consistent unit cost ranges from $26,128 to $33,740 per sq. m. CFA.
All these data are based on the government’s estimates rather than actual awarded tender prices. It has to be cross-checked with the actual to see how accurate the estimates are.
Admittedly, unit construction cost for government projects may not be relevant for private projects, due to various reasons. It may require another study to identify the systematic discrepancy in cost estimates between public and private projects, if any.
Worse still, the government does not take initiative to provide similar benchmarking information for other projects, because no Council Members asked for. There is also no database consolidating all the cost information submitted to the Legislative Council, though I am building a small-scale one. This database if available would be of utmost importance for monitoring government expenditures as well as doing research on the estimation of construction costs. I am also looking for an AI system to predict automatically the construction cost of government projects, it would save a lot of manpower, especially QS efforts, in preparing the estimations. If you are interested in this proposal, welcome to lead.
Arcadis (2019) Quarterly Construction Cost Review Q1 2019 — Hong Kong and China, https://www.arcadis.com/en/asia/our-perspectives/research-and-publications/quarterly-construction-cost-review/
Legislative Council (2019a) 130KA — Immigration Headquarters in Tseung Kwan O, Public Works Subcommittee Paper, PWSC(2018–19)39, Jan 30. https://www.legco.gov.hk/yr18-19/english/fc/pwsc/papers/p18-39e.pdf
Legislative Council (2019b) Immigration Headquarters in Tseung Kwan O, Public Works Subcommittee Paper, PWSC160/18–19(01), Mar 28. https://www.legco.gov.hk/yr18-19/english/fc/pwsc/papers/pwsc20190220pwsc-160-1-e.pdf
Yiu, C.Y. 姚松炎 (2019) 系統性高估工程價格, vocus June 20. https://vocus.cc/infrastructure/5d0b4b24fd89780001231ada