INDICES: their use and abuse

The term Cost Index is defined by the American Association of Cost Engineers as: `a number which relates the cost of an item at a specific time, to the corresponding cost at some arbitrarily specific time in the past. A series of numbers recording changes in values since some predetermined date when the base value was equal to 100.'

A cost index thus provides a comparison of cost changes year to year for a fixed quantity of goods or services. It allows the cost engineer a means of forecasting the cost of a similar design from the past to the present or future period without going through a detailed estimating exercise. Providing the engineer uses his or her discretion in choosing the proper index, a reasonable approximation of cost would result. Extrapolation though a time-series analysis of cost indices is possible for future periods.

One of the first traceable references to the use of indices was by an Italian, G R Carli, who in 1750 used index numbers to investigate the effects of the discovery of America on the purchasing power of money in Europe. Other pioneers include Sir George Shackbough Evelyn (1751-1801), whose base date was 1550 = 100. His first record was the year 1050 = 26, concluding at 1800 = 562!

In more recent times, Enady and Boyd published an article in a Cost Engineer of 1964 `Indices and erection cost of UK chemical plants.' This is known in the trade as the ACE Cost Indices Plants A, B, C & D.

FROM THE PAST TO THE FUTURE

So, as we can see, indices are not new. But how do we use them? As explained on this page last month, if a previous design cost (Cr) is known, then present cost (Ca) can be determined by multiplying the original cost by the ratio of the present index (Ic) to the index value applicable when the original cost was obtained (Ir). Or:

Ca = Cr(Ic/Ir)

The selection of the appropriate index to upgrade an estimate is clearly of extreme importance. The cost engineer should study the format of the index: for example, its elements of design, location, and weighting of the elements in the current index and the individual elements indicated in the original estimate.

If major items are ignored in the estimate, the composition of the index may have to be adjusted accordingly. Beware of an index that is only published infrequently, as this seldom reflects all the factors to be considered, such as technological progress or local and special conditions.

Typical questions that can be asked of cost engineers by `management' and `client' are:

{{What is a cost index?What is its origin?How is it used?How reliable is it?Why not use our own?}}

We have already answered the first question, but the point always to bear in mind is that the cost index is the choice of the individual or as laid down in company policy.

As far as the UK is concerned, two sources of reliable data for process plant are the CEI of the Association of Cost Engineers, published bi-monthly in the Association's house journal - and, of course, our own Predict Indices published each month on the facing page. In both these cases, their weighting for engineering disciplines are published as separate indices as well as an overall index.

As to reliability, most of the raw data is obtained from the Central Statistical Office, either as published or privately purchased. The exception is CEI, where the labour element is obtained from BEMA with adjustments to productivity being made by the ACE from experience in the industry.

International indices can be obtained for most countries and we publish a set of indices from some 14 countries with raw data being sourced from the OECD in Paris. The USA is well served by Chemical Engineering's `Economic Indicators'.

Ultimately, however, reliability is in the hands of the user. Indices are published with the best of intentions to aid the user, and ignore factors such as RPI (retail price index) which has its own weighting and basket of contents adjusted politically.

In the panel we present an example of a calculation for the future escalation of a known project costing £100million.

CALCULATION OF FUTURE PROJECT COST

Consider a project with a known current cost of £100million. How do we calculate the cost of a similar project at some time in the future? Start by analysing the component elements of the project, their costs as a percentage of overall project cost, their dwell times (i.e. how long into the project before they become active), and the actual working or production times for each:

{{ Cost Dwell Working (%) (months) (months)

Design & engineering 15 36 12Materials 50 39 12Labour 25 47 18Sub-contracts 10 47 18}}

The period of time starts for each component at a common point, referred to as the estimate base date. The actual working or production time for each component commences at the end of its respective dwell time. All times for escalation are calculated to the `cost centroid' of the component, that is the time when half of the cost will have been spent. These cost centroids relative to the base date are calculated as follows:

{{Design: 36 months + 0.5(12months) = 3.5 yearsMaterials: 39 + 0.5(12 = 3.75 yearsLabour: 47 + 0.5(18) = 4.75 yearsSub-contracts: 47+ 0.5(18) = 4.75 years}}

If we know that the current value of the project is £100million, and the rates of escalation are: Design 15%pa; Materials 6%pa; Labour 10%pa; and Sub-contracts 8%pa; then the total project cost, including escalations, will be:

{{Design: 100x0.15x(1.15)3.5 = £17.79mMaterials: 100x0.50x(1.06)3.75 = £62.21mLabour: 100x0.25x(1.10)4.75 = £39.31mSub-contracts: 100x0.10x(1.08)4.75 = £14.41mTotal estimated cost = £133.72m}}

The calculated escalation to add to the project is, therefore, £33.72million.