PAPER – Construction Productivity – CONSTRUCTION PROJECT MANAGEMENT

9 Refining the citiBLOC index

Introduction

The previous chapter investigated the performance of selected high-rise construction projects completed between 2003 and 2012 in the five largest cities in Australia and the United States. Part of this research required the conversion of cost data into a comparable form. A standard basket of construction material, labour and plant, priced in each city, was used for this purpose. The value of the standard basket (defined as equal to one ‘citiBLOC’) became the unit of cost comparison. For example, if a building’s construction cost was AUD$10 million and the citiBLOC index for its location was AUD$10,000, then the ‘cost’ of the project would be computed as 1,000 citiBLOCs. Similarly, if a building was USD$16 million and the citiBLOC index was USD$8,000, then the ‘cost’ would be 2,000 citiBLOCs. The number of citiBLOCs can be used to translate local currency cost data into a comparable form. If the building floor areas were the same, then the US example would have a cost/m2 rate that was twice as expensive as the Australian example.

Cost conversion is critical to international performance comparisons. A reliable method for doing this is necessary. Purchasing power parity (PPP) is generally accepted as the appropriate philosophy, but there are a number of detailed approaches for determining indices. Which one should be used? How can we be sure that cost conversion using currency exchange rates is unreliable? How should construction-related PPPs be produced for different countries and/or cities? What other issues should be considered?

The aim in this chapter is to answer these questions. Five approaches are discussed and explored for a representative range of countries. Australia and the United States are used as examples of small and large developed countries, respectively, whilst Malaysia and India are used as examples of small and large developing countries, respectively. Base data is largely determined from public-domain cost information available as part of the Turner and Townsend International Construction Cost Survey 2012 (Turner and Townsend 2012) and citiBLOC calculations undertaken by the Centre for Comparative Construction Research (CCCR) at Bond University, Australia.

Background

Whenever the performance of the construction industry is called into question, the immediate reaction, and rightly so, is to attempt to benchmark performance against other countries. Performance is a complex issue, however, and includes multiple criteria such as cost, time and quality, to name but a few. While cost and time might appear reasonably straightforward, subjective issues such as quality are much more difficult to assess.

International cost comparison methodology is the focus of this chapter. Previous pricing studies have employed a range of methodologies, such as estimating the cost of identical standard projects (actual or hypothetical) or comparisons of functionally similar projects taking into account local practice or a combination of both. In any event, the question soon arises as to how to compare costs on an equal basis, since whenever different currencies are involved, the cost impacts cannot be immediately understood. Costs vary for a range of factors, not the least of which is time, but the issue of location is to be explored here and is central to the need to compare costs across national borders.

The exchange rate adopted to compare costs arising from projects in different locations is a critical factor for the usefulness of results that come from any international benchmarking study. Applying currency exchange rates is an obvious choice, but these change frequently and do not provide confidence that the relativity between construction industries in two different countries is actually being assessed. For example, the Asian economic crisis triggered in 1997 could be used to conclude that the dramatically lower cost of construction in some Asian countries, as calculated by falling exchange rates against their Western counterparts, was a result of increased competitiveness in-country. The reality was that the local industry had not changed, but the value assigned to projects that were under construction or previously completed had sharply declined.

The use of PPP as an alternative to traditional currency exchange rates is generally regarded as a superior approach (e.g. Rogoff 1996; Langston and Best 2005). PPP is an attempt to measure the economic well-being of people according to the country in which they reside. While not pretending to be an indicator of living standards, it does reflect the cost of living in-country and therefore forms a new baseline against which construction costs can be interpreted.

PPPs can be calculated at the value of a particular good or service or using a weighted basket of goods and services and can be expressed in relation to gross domestic product or income capacity. In fact, PPPs have been calculated using items that are available in most countries worldwide, such as via use of the Big Mac Index regularly compiled by The Economist magazine. There are grounds to suggest that an approach specific to construction goods and services would be preferable to one that is generic of entire economies (e.g. Walsh and Sawhney 2004).

PPPs are defined as exchange rates that replace traditional currency exchange rates by taking into account the differences in prices between countries (Pakko and Pollard 2003). They convert local costs into ‘international dollars’ compared to a nominated base country. The philosophy behind PPPs is the Law of One Price – namely that the cost of a good or service should be the same in different countries – else people would buy goods cheaper from one country and sell them at a profit in another.

Whether the Law of One Price holds for any particular item depends on the item meeting four basic criteria (UBS 2003:27). They are:

1. The item must be tradable.

2. There are no impediments to trade.

3. There are no transaction costs (such as transport) involved in trade of the item.

4. The item is perfectly homogeneous across all locations.

If all four criteria are met, then the price of the item should be the same in different places at the same time. In that case, the cost of an item in currency X should represent the same value as the cost of the same item in currency Y (Best 2008).

The United Nations–sponsored International Comparison Program (ICP) commenced in 1967 and now produces PPPs published by the World Bank Group for most countries on an approximate three-year cycle. These indices have been interpreted and extended to form the Penn World Table (PWT) produced by the University of Pennsylvania. The Eurostat-OECD joint program currently collects more detailed PPP data than the ICP, but for a much smaller set of countries. Indices for both ICP and Eurostat-OECD PPPs are expressed as a proportion of per-capita gross domestic product. The Union Bank of Switzerland (UBS) has also been producing PPP data since 1970, again approximately on a three-year cycle. They use a basket of goods and services and express their data in three forms (using a base for Switzerland, the United States or the Euro-zone, respectively). One criticism of these programs is the time delay between data collection and publication. Another criticism is the cost of the process.

Controversially, The Economist magazine has published an alternate PPP index based on the McDonald’s Big Mac hamburger price since 1986 for a number of countries. Known as ‘burgernomics’ (Lan 2003), this approach has moved from a lighthearted look at fast food metrics to a quite seriously debated topic (e.g. Pakko and Pollard 2003, who found a correlation of 0.73 between the PWT and the Big Mac Index using 2000 data). The Big Mac Index has the advantage that input data is relatively easy to collect and therefore enables it to be up to date and city specific. The Economist now publishes its index several times each year. Cumby (1996) found that when the US dollar price of a Big Mac is high in a country, the relative local currency price of a Big Mac in that country is likely to fall during the following year. The index has been employed to identify currency over and under valuations, although this is not a recommended use.

The reliability of various methods is unknown, as there is no correct value that each can be compared against other than monetary exchange rates, which are volatile and subject to influence from a number of external sources. Pakko and Pollard (2003:22) concluded that “it is interesting to find that the simple collection of items comprising the Big Mac sandwich does just as well (or just as poorly) at demonstrating the principles and pitfalls of PPP as do more sophisticated measures”. Ong (2003) concurred. But over the last decade in particular, attention has now turned to developing indices that are industry focused, such as comparing construction-related costs independent of general economy activity (Meikle 1990; Walsh and Sawhney 2004).

Approach

Five methods shall be compared as part of this study. Each will be applied to the four selected countries to highlight differences due to country size and affluence level. The methods are described briefly below:

1. Method 1: a mix of various base prices for the supply of construction labour and material, weighted according to their cost impact on building sites.

2. Method 2: an unweighted mix of composite prices for the supply and installation of various construction components commonly found in international building projects.

3. Method 3: a basket of location-specific common work items for material supply, labour charge-out rates and plant hire, averaged across the five largest cities by population.

4. Method 4: the current price of a standard specification McDonald’s Big Mac hamburger.

5. Method 5: currency exchange rates.

The methods are selected based on their ability to expose important issues related to international cost conversion. Other variations on these methods are also possible. A method based on pricing identical projects in different countries was not tested, as it was too time consuming and unlikely to lead to satisfactory outcomes given that individual items of work in some countries will not be representative of local practice.

Across all methods, one country is nominated as the base and the PPP indices for the other countries are computed as price relatives to this base. In this study, Australia is the selected base country on the basis that it has the lowest coefficient of variation (CoV) across its five largest cities, as determined by Method 3 (see  Table 9.9 ). The United States had the largest national variation.

Costs/m2 are collected for a range of building types and then adjusted by the derived PPPs. Rather than use one or more building types individually, a large range of building types is averaged to increase statistical reliability. The average cost/m2 for building in each country is then divided by the country PPP and multiplied by the base PPP, and the results are compared. A summary of each method per country is used to determine which method produces the lowest CoV and therefore which method best reflects the Law of One Price. The same approach was previously applied in Langston and Best (2005).

All data used in this study is from Turner and Townsend (2012) unless otherwise noted. Several errors found in this data are noted and amended. It is acknowledged that accuracy of this data is assumed to be acceptable for the purposes of testing the relative performance of the different methods. Nevertheless, accuracy of raw data is a limitation of international cost comparisons.

Results

Method 1 (L+M)

Turner and Townsend (2012) provide 5 items of construction labour and 11 items of construction material for Australia, the United States, Malaysia and India. These are national prices and do not reflect variations between different locations.  Table 9.1  shows the raw data obtained from this source.

Table 9.2  converts the raw data in  Table 9.1  to price relatives using Australia as the base. Labour and material means are then combined to form a single index. The ratio is derived from the data used in Method 3. The percentage of labour is computed as the sum of the four labour items in the citiBLOC basket divided by the sum of all 10 items plus 50% of the plant item. Similarly, the percentage of material is computed as the sum of the five material items in the citiBLOC basket divided by the sum of all 10 items plus the remaining 50% of the plant item. As the quantities used in Method 3 were set to ensure evenly weighted basket items, the ratios for Australia reflect the 50:40:10 design of the citiBLOC basket itself.

The definition of ‘mean’ used in this method and others in this study is the arithmetic mean. It would be inappropriate to use geometric mean, as the data is not a compounding time series. Arithmetic mean is also simpler to calculate and more understandable in practice.

Table 9.1    Method 1 raw data (Turner and Townsend 2012)

Table 9.1    Method 1 raw data (Turner and Townsend 2012)

	Australia (AUD)	United States (USD)	Malaysia (MYR)	India (INR)
LABOUR (/hour):
Group 1 tradesman	68	75	20	56
Group 2 tradesman	57	65	15	50
Group 3 tradesman	55	57	20	38
General labourer	38	53	10	20
Site foreman	72	77	55	80
MATERIAL:
Concrete 30 MPa (m3)	186	135	230	5,000
Reinforcement bar 16mm (tonne)	1,250	992	3,220	40,500
Concrete block 400 × 200 (per 1,000)	3,360	1,030	3,980	35,000
Standard brick (per 1,000)	541	350	400	6,250
Structural steel beams (tonne)	2,955	1,150	4,420	50,000
Glass pane 6mm (m2)	47	58	189	500
Softwood timber for framing 100 × 50mm (m)	29	7	34	275
Plasterboard 13mm (m2)	39	3	26	280
Emulsion paint (litre)	15	8	39	340
Copper pipe 15mm (m)	12	7	27	^ 435
Copper cable 3C+E 2.5mm PVC (m)	5	4	16	^^ 25
COMBINED MIX (%):
Labour	44.62	53.81	16.55	8.14
Material	55.38	46.19	83.45	91.86
sum	100.00	100.00	100.00	100.00

^ error fixed: /m not /kg ^^ error fixed: /m not /100m

Method 2 (composite)

Turner and Townsend (2012) further provide 19 composite items of construction for Australia, the United States, Malaysia and India. Again, these are national prices and do not reflect variations between different locations.  Table 9.3  shows the raw data obtained from this source.

Table 9.4  converts the raw data in  Table 9.3  to price relatives using Australia as the base.

Table 9.2    Method 1 price relatives (Australia = base)

	Australia (AUD)	United States (USD)	Malaysia (MYR)	India (INR)
LABOUR (/hour):
Group 1 tradesman	100.00	110.29	29.41	82.35
Group 2 tradesman	100.00	114.04	26.32	87.72
Group 3 tradesman	100.00	103.64	36.36	69.09
General labourer	100.00	139.47	26.32	52.63
Site foreman	100.00	106.94	76.39	111.11
mean	100.00	114.88	38.96	80.58
MATERIAL:
Concrete 30 MPa (m3)	100.00	72.58	123.66	2,688.17
Reinforcement bar 16mm (tonne)	100.00	79.36	257.60	3,240.00
Concrete block 400 × 200 (per 1,000)	100.00	30.65	118.45	1,041.67
Standard brick (per 1,000)	100.00	64.70	73.94	1,155.27
Structural steel beams (tonne)	100.00	38.92	149.58	1,692.05
Glass pane 6mm (m2)	100.00	123.40	402.13	1,063.83
Softwood timber for framing 100 × 50mm (m)	100.00	24.14	117.24	948.28
Plasterboard 13mm (m2)	100.00	7.69	66.67	717.95
Emulsion paint (litre)	100.00	53.33	260.00	2,266.67
Copper pipe 15mm (m)	100.00	58.33	225.00	3,625.00
Copper cable 3C+E 2.5mm PVC (m)	100.00	80.00	320.00	490.00
mean	100.00	57.76	192.21	1,720.81
COMBINED MIX (%):
Labour	44.62	61.82	6.45	6.56
Material	55.38	26.58	160.39	1,580.69
sum	100.00	88.40	166.84	1,587.25

Method 3 (citiBLOC)

Method 3 is the citiBLOC index applied in the previous chapter. The five cities forming the Australian average are Sydney, Melbourne, Brisbane (including Gold Coast), Perth and Adelaide (listed in decreasing order of population size). In the United States, the cities are New York, Los Angeles, Chicago, Houston and Philadelphia. In Malaysia, the cities comprise Kuala Lumpur, Johor Bahru, Ipoh, Kuching and Georgetown. Finally, the cities forming the India average are Mumbai, New Delhi, Bangalore, Kolkata and Chennai. The total value of the citiBLOC basket in each of these cities is shown in  Tables 9.5  to  9.8 . The proportional mix of material, labour and plant by city is also provided based on the quantity of each work item. For the Australian base, the total value of each item is approximately equal value. The national figure is assumed to be the average of prices from the five largest cities.

Table 9.3    Method 2 raw data (Turner and Townsend 2012)

	Australia (AUD)	United States (USD)	Malaysia (MYR)	India (INR)
SUPPLY AND INSTALL:
Excavate basement (m3)	29	10	28	750
Excavate footings (m)	71	8	33	475
Concrete in slab (m3)	256	171	339	6,500
Reinforcement in beams (tonne)	2,442	2,100	3,578	56,000
Formwork to soffit of slab (m2)	120	71	51	525
Blockwork in wall (m2)	131	75	68	1,075
Structural steel beams (tonne)	6,298	2,855	6,632	82,500
Precast concrete wall (m2)	299	120	n/a	8,250
Curtain wall glazing incl. support system (m2)	1,051	775	408	9,250
Plasterboard 13mm thick to stud walls (m2)	30	34	92	2,250
Single solid core door incl. frame/hardware (no)	765	2,100	816	27,500
Painting to walls primer + 2 coats (m2)	15	10	8	200
Ceramic tiling (m2)	109	90	112	2,800
Vinyl flooring to wet areas (m2)	71	64	100	2,100
Carpet medium tufted (m2)	80	44	92	2,350
Lighting installation (m2)	81	83	153	2,900
Copper pipe 15mm to wall (m)	60	32	80	^ 600
Fire sprinklers (m2)	46	30	102	825
Air conditioning including main plant (m2)	300	172	306	3,625

^ error fixed: /m not /kg

Australian data was extracted from Cordells online cost database. American data was extracted from RS Means online cost database. Malaysian and Indian data was provided by personal contacts in-country.  Table 9.9  converts the raw data in  Tables 9.5  to  9.8  to price relatives, again using Australia as the base.

Table 9.4    Method 2 price relatives (Australia = base)

	Australia (AUD)	United States (USD)	Malaysia (MYR)	India (INR)
SUPPLY AND INSTALL:
Excavate basement (m3)	100.00	34.48	96.55	2,586.21
Excavate footings (m)	100.00	11.27	46.48	669.01
Concrete in slab (m3)	100.00	66.80	132.42	2,539.06
Reinforcement in beams (tonne)	100.00	86.00	146.52	2,293.20
Formwork to soffit of slab (m2)	100.00	59.17	42.50	437.50
Blockwork in wall (m2)	100.00	57.25	51.91	820.61
Structural steel beams (tonne)	100.00	45.33	105.30	1,309.94
Precast concrete wall (m2)	100.00	40.13	n/a	2,759.20
Curtain wall glazing incl. support system (m2)	100.00	73.74	38.82	880.11
Plasterboard 13mm thick to stud walls (m2)	100.00	113.33	306.67	7,500.00
Single solid core door incl. frame/hardware (no)	100.00	274.51	106.67	3,594.77
Painting to walls primer + 2 coats (m2)	100.00	66.67	53.33	1,333.33
Ceramic tiling (m2)	100.00	82.57	102.75	2,568.81
Vinyl flooring to wet areas (m2)	100.00	90.14	140.85	2,957.75
Carpet medium tufted (m2)	100.00	55.00	115.00	2,937.50
Lighting installation (m2)	100.00	102.47	188.89	3,580.25
Copper pipe 15mm to wall (m)	100.00	53.33	133.33	1,000.00
Fire sprinklers (m2)	100.00	65.22	221.74	1,793.48
Air conditioning including main plant (m2)	100.00	57.33	102.00	1,208.33
mean	100.00	75.51	118.43	2,251.00

Method 4 (Big Mac Index)

The price of an international standard commodity can be used as a handy PPP since, theoretically, it is of equal value in all countries. Yet this is unlikely for a number of reasons, including distortions to free trade, different market contexts and demand, and local availability of key resources. A McDonald’s Big Mac is an example of a standard commodity. This product is used here to compare against the industry-specific PPPs. The price of a Big Mac in each country for 2012 is provided in  Table 9.10 . Data was sourced from The Economist magazine and reflects the mean national price for the year.

Table 9.5    Method 3 raw data (Australia)

Table 9.6  Method 3 raw data (United States)

Table 9.7  Method 3 raw data (Malaysia)

Method 5 (Currency exchange rates)

Currency exchange rates are not PPPs. However, this method is included to benchmark results against a popular, albeit inappropriate, strategy for international cost comparisons. The exchange rates at 30 June 2012, sourced online ( www.oanda.com/currency/classic-converter ) are shown in  Table 9.11 .

Table 9.8    Method 3 raw data (India)

Comparison of methods

Turner and Townsend (2012) provide construction costs/m2 for nine different building types in each of the four selected countries. These are listed in  Table 9.12 . Rather than compute each building type separately, the mean of all building types was adopted to provide a more stable cost benchmark.

Table 9.9    Method 3 price relatives (Australia = base)

	Australia (AUD)	United States (USD)	Malaysia (MYR)	India (INR)
LOCATION:
City 1 (largest)	103.26	106.93	124.55	1,320.16
City 2	97.54	85.59	109.21	1,204.83
City 3	102.94	90.26	126.41	1,336.12
City 4	98.37	70.05	139.87	1,396.47
City 5	97.88	96.38	120.63	1,376.47
mean	100.00	89.84	124.13	1,326.81
CoV	2.85%	15.19%	8.90%	5.63%

Table 9.10    Method 4 raw data

	Australia (AUD)	United States (USD)	Malaysia (MYR)	India (INR)
INTERNATIONAL STANDARD PRODUCT:
Big Mac hamburger	4.80	4.20	7.35	84.00

Table 9.11    Method 5 raw data

	Australia (AUD)	United States (USD)	Malaysia (MYR)	India (INR)
30 JUNE 2012:
Currency exchange rate (compared to USD)	0.9844	1.0000	3.1989	56.225

Equivalent costs/m2 are shown in  Table 9.13 . The CoV can be used to rank the reliability of each method. The lower the CoV, the more the method reflects the Law of One Price philosophy.

Method 1 has an issue in determining the proportional mix of labour and material, which varies between low-labour and high-labour countries and is difficult to quantify. Method 2 overcomes that problem but also embeds issues of productivity rather than being merely a cost conversion tool (i.e. higher comparative costs may indicate poor productivity, complex design or expensive resource inputs). Method 3 solves both of those shortcomings and has the added advantage of being an effective locality index. Method 4 delivers a quick and up-to-date general PPP index, but it may not always be applicable or defendable for construction work (particularly for developing countries). Method 5 is problematic, as currency conversion can change due to global macroeconomic factors irrespective of local purchasing power.

Table 9.12    Cost/m2 for various building types (Turner and Townsend 2012)

	Australia (AUD)	United States (USD)	Malaysia (MYR)	India (INR)
BUILDING TYPE:
Residential	1,990	1,649	1,700	29,250
Commercial	2,660	2,182	3,167	34,333
Warehouses	963	1,262	2,067	33,333
Retail	2,727	1,283	3,333	27,167
Hotels	3,702	2,158	5,767	53,667
Hospitals	3,806	2,898	2,300	29,000
Schools	2,500	1,885	1,650	28,000
Car parks	997	1,008	1,450	31,000
Airports	6,565	3,550	5,000	65,000
mean	2,879	1,986	2,937	36,750

Table 9.13    Cost/m2 price relatives (Australia = base)

Method 3 has the lowest coefficient of variation (12.48%) and is recommended as more likely to reflect construction-related PPP. Method 5 has the highest coefficient of variation (64.35%) and should not be used. Method 4 appears to produce results slightly superior to Methods 1 and 2 in the countries studied. Figure 9.1 shows the profile of the various methods.

Figure 9.1    Comparison of construction price relatives

The profile of methods is the same regardless of the accuracy of the average costs/m2 in each country. Further testing is nevertheless warranted, including a broader range of countries studied.

Discussion

In this chapter, citiBLOC indices for cities were calculated as an evenly weighted basket of construction material, labour and plant. The quantities for each item were chosen to enable the items to have equal influence. This is certainly the case for Australia, but other countries show that the balance is lost. This appears to be a disadvantage of Method 3 and may introduce bias in the results. However, the quantities chosen are irrelevant to the outcomes.

Table 9.14    Unit rates for citiBLOC for Australia

The adoption of price relatives means that the price of a particular item is compared to the base, and the quantity multiplier cancels out. Table 9.14 shows the calculation for Australia and the base for each of the 10 items in the citiBLOC basket against which unit rates in other countries are compared.

The unit rates for each country are expressed as a ratio of the Australian average and converted to a base of 100. The procedure is illustrated in Table 9.15 using India as an example. The difference between 1,324.63 (Table 9.15) and the previous figure for India of 1,326.81 (see Table 9.9) is merely rounding error. Therefore the use of quantities as part of Method 3 does not introduce bias and is really only done to ensure that prices are reflective of a given scope of work. Note that the proportional mix of material, labour and plant (see Table 9.8) cannot be calculated from unit rate data.

Table 9.15  Example price relatives for India (based on unit rates)

Conclusion

The introduction to this chapter raised a number of questions. Which PPP method should be used? How can we be sure that cost conversion using currency exchange rates is unreliable? How should construction-related PPPs be produced for different countries and/or cities? What other issues should be considered?

The answers to these questions are now clear. The recommended construction-related PPP method is identified as Method 3. Known as citiBLOC, it uses a basket of 10 common construction items (5 material, 4 labour and 1 plant) priced per city, with the country average being computed from the five largest cities (where applicable). This method has the lowest CoV of the five methods tested and therefore is more likely to reflect the Law of One Price underpinning the PPP philosophy. It is also shown that current exchange rates have the highest CoV, which demonstrates why they should not be used for international cost comparisons. The procedure for calculating citiBLOCs for different locations is outlined, including proof that the quantities assigned to the basket make no difference to the outcome and are used purely to signify pricing context.

The citiBLOC index is refined in this chapter. One citiBLOC is defined as the equivalent of AUD$10,000 in 2012 measured across the five largest cities. It needs to be computed each year. This value typically rises each year due to the effects of inflation. It can be used as a benchmark of construction cost for any location, whether within a particular country (i.e. regional comparisons) or across different countries (i.e. international comparisons).