Engineering Practice

Updating the Plant

Cost Index
TABLE 1. STRUCTURE OF THE CECPI
A. Equipment Index,
includes the following sub-indexes:
• Heat exchangers and tanks
• Process machinery
• Pipe, valves & fittings
Changing ways of building plants • Process instruments
• Pumps & compressors
• Electrical equipment
are reflected as this widely used index • Structural supports & miscellaneous
B. Construction Labor Index
is brought into the 21st century C. Buildings Index
D. Engineering and Supervision Index
CE Plant Cost Index (composite)
William M. Vatavuk the data series and the relative ted in Figure 1. A cursory look at this
Vatavuk Engineering weights since 1982. That is a long figure tells us that none of the five in-
time to freeze an index, and does not dexes increased monotonically during
he Chemical Engineering Plant necessarily reflect recent changes. For these 37 years. They had their “ups

T Cost Index (CEPCI) is an estab-

lished institution. Since its intro-
duction in 1963 [1], it has been
published in each and every issue of
CE. For more than 37 years, chemical-
that reason, we have decided it was
time for the CEPCI to be revised
again, this time to bring it into the
twenty-first century. The structure is
not being changed, but many underly-
and downs,” but all of them were sig-
nificantly higher in 2000 than in 1963.
The Equipment sub-index exhibited
the most growth (approximately 340%)
over this period, followed by the com-
process-industry (CPI) professionals — ing details are being updated. posite CEPCI, and the Buildings,
engineers, managers, and technicians Engineering and Supervision, and
— have used the CEPCI to adjust The established structure Construction Labor sub-indexes, re-
process plant construction costs from Before describing the latest CEPCI re- spectively. The relatively steep rise in
one period to another. This index — vision, we need to review its structure the Equipment sub-index was partly
rather, indexes, as it consists of a com- and some history. Table 1 shows that due to sharp increases in fabricated
posite index and eleven sub-indexes — the composite index is built from seven equipment prices, which were driven,
has received such wide acceptance that component-indexes and four sub-in- in turn, by jumps in the costs of raw
it has even been written into construc- dexes, For all of these, values have materials (for instance, stainless steel).
tion-contract cost-escalation clauses. been reported running back to 1947. Because this sub-index contributed so
The CEPCI has a history of revi- Table 2 lists annual values of the four much (61%) to the composite CEPCI,
sions. Most of these have been cos- sub-indexes and the composite index the latter exhibited very similar behav-
metic, such as renaming the “Fabri- for the years 1963 to 2000. Compo- ior; the two curves are nearly parallel.
cated equipment” sub-index to the nent-index data and an extension back There is another reason for the rela-
more descriptive “Heat exchangers and to 1947 can be purchased from CE’s tively modest increases in the other
tanks” sub-index. The most recent, and editorial department. Each annual sub-indexes — the CEPCI productiv-
most significant, revision — more like index is the arithmetic mean of the ity factor. This factor, which discounts
an overhaul — occurred in 1982 [2]. monthly indexes. changes in the labor-cost components
The major changes included reducing Details of computing the values will in the sub-indexes, tends to dampen
the number of components from 110 to be discussed below. For now, we will increases in these sub-indexes. An
66, replacing many components with stick with an outline, as shown in index with large labor components,
more suitable ones, and revising the Table 1. The component-indexes are such as Construction Labor, is more
productivity factor downward from compiled, and with appropriate influenced by the productivity factor
2.50% to 1.75%. (See next page for an weighting factors, are added up to than those with less-significant labor
explanation of the productivity factor.) make up the Equipment Index. Inde- cost elements (such as Equipment).
The 1982 changes deliberately re- pendently, three other sub-indexes The five index curves change slopes
tained the structure of the CEPCI. are compiled, and with appropriate downward around the year 1982. This
The set of sub-indexes has remained weighting and normalizing, the four was the year of the last major CEPCI
unchanged for nearly four decades — sub-indexes add up to the CEPCI. revision. This revision involved a
decades that have seen major changes While Table 2 is informative, it is a change in the productivity factor, as
in the CPI, especially in plant design lot easier to see trends and relative well as major reductions in the num-
and operation. We have not changed changes among the indexes when plot- ber of index components. Taken to-
62 CHEMICAL ENGINEERING WWW.CHE.COM JANUARY 2002
 TABLE 2. ANNUAL PLANT COST INDEXES
Year Com- Equipment Construc- Buildings Engineering
posite tion and super-
CE Index labor vision
1963 102.4 100.5 107.2 102.1 103.4
1964 103.3 101.2 108.5 103.3 104.2
1965 104.2 102.1 109.7 104.5 104.8
1966 107.2 105.3 112.4 107.9 106.8
1967 109.7 107.7 115.8 110.3 108.0
1968 113.7 109.9 121.0 115.7 108.6
1969 119.0 116.6 128.3 122.5 109.9
1970 125.7 123.8 137.3 127.2 110.6
1971 132.3 130.4 146.2 135.5 111.4
1972 137.2 135.4 152.2 142.0 111.9
1973 144.1 141.8 157.9 150.9 122.8
1974 165.4 171.2 163.3 165.8 134.4
1975 182.4 194.7 168.6 177.0 141.8 FIGURE 1. The major indexes show an irregular rise over 37 years
1976 192.1 205.8 174.2 187.3 150.8
1977 204.1 220.9 178.2 199.1 162.1 of personal computers and ary 1947 and the index-update month
1978 218.8 240.3 185.9 213.7 161.9 other electronic tools, and The index is calculated monthly.
1979 238.7 264.7 194.9 228.4 185.9 less-evident (to non-con- That is why the average productivity
1980 261.2 292.6 204.3 238.3 214.0 struction professionals, at increase appears as p/12.
1981 297.0 323.9 242.4 274.9 268.5 least) innovations as the For instance, suppose that the raw
1982 314.0 336.2 263.9 290.1 304.9 implementation of modu- (unadjusted) increase in the Engi-
1983 317.0 336.0 267.6 295.6 323.3 lar construction tech- neering cost component from Febru-
1984 322.7 344.0 264.5 300.3 336.3 niques. However, Matley ary 1963 to August 2000 was 400% (or
1985 325.3 347.2 265.3 304.4 338.9 said that the factor, 4.00). Here is how to calculate the ad-
1986 318.4 336.3 263.0 303.9 341.2 “should not be considered justed change in this component: Sub-
1987 323.8 343.9 262.6 309.1 346.0 to account for changes in stitute an annual productivity growth
1988 342.5 372.7 265.6 319.2 343.3 productivity arising from rate of 2.2% or 0.022 (this value is ver-
1989 355.4 391.0 270.4 327.6 344.8 improvements in the qual- ified p.69) and the number of months
1990 357.6 392.2 271.4 329.5 355.9 ity of construction work- from January 1947 to August 2000 (53
1991 361.3 396.9 274.8 332.9 354.5 management… [nor 3 12 + 8 = 644). Equation (2) solves for
1992 358.2 392.2 273.0 334.6 354.1 should it] be regarded as a productivity factor of:
1993 359.2 391.3 270.9 341.6 352.3 reflecting productivity
1994 368.1 406.9 272.9 353.8 351.1 P.F. = 1/(1 + 0.022/12)644 =
changes due to advances
1995 381.1 427.3 274.3 362.4 347.6 0.3074 (2)
in the skill, experience or
1996 381.7 427.4 277.5 365.1 344.2 motivation of the work- The adjusted increase in the engi-
1997 386.5 433.2 281.9 371.4 342.5 force. Lastly, the factor neering cost component is 0.3074 3
1998 389.5 436.0 287.4 374.2 341.2 does not take into account 400% = 123%.
1999 390.6 435.5 292.5 380.2 339.9 regional variations in con- The exponent n in Equation (1) is
2000 394.1 438.0 299.2 385.6 340.6 struction wages” [3]. calculated from January 1947 instead
Indexes’ base: 1957-59 = 100.0 Let us do the math and of from February 1963, the month and
enter the productivity fac- year the CEPCI was introduced, be-
gether, these changes significantly af- tor into the calculation of the CEPCI cause 1947 is the year that the Mar-
fected the next 18 years of the CEPCI composite and its sub-indexes. It is shall and Swift (M&S; formerly Mar-
and its sub-indexes. There were other used on every labor-cost component shall and Stevens) Equipment Cost
macroeconomic trends that also influ- within a sub-index, component index, Index was started as a regular feature
ence the data; such as a decline in in- or the composite index. Take the raw of CE [4]. We suspect that Arnold and
terest rates over 20 years from 20 to change in a labor-cost component and Chilton derived the productivity fac-
2%. The inflationary 1970s have come multiply it by the productivity factor tor from this data, but they did not
and gone but left their mark. to obtain the adjusted change in that leave a paper trail.
component. This adjusted change is Despite the fact that the M&S index
Meet the productivity factor then an input to the calculation of the is based on used-equipment prices,
We have mentioned the CEPCI pro- appropriate CEPCI sub-index, and while the CEPCI is built around new
ductivity factor several times already the composite index. plant-construction costs, trends in the
without explaining what it is or how it The productivity factor is calculated M&S have been comparable to the
is calculated. In 1982 Matley [3] ex- via Equation (1): CEPCI. The quarterly M&S index is
plained that the productivity factor published on CE’s Economic Indica-
P.F. = 1/(1 + p/12)n (1)
“should be thought of as a technologi- tors page, along with the VAPCCI (see
cal productivity factor [that is] predi- where: P.F. = productivity factor (< 1) box, p. 69) and other indexes.
cated on advances in working tools and p = Annual growth in construction Two questions of engineering phi-
techniques.” These advances include labor productivity (fraction) losophy intrude here and need ad-
such obvious ones as the proliferation n = Number of months between Janu- dressing. Why do we use a productiv-
CHEMICAL ENGINEERING WWW.CHE.COM JANUARY2002 63
 TABLE 3. REVISED CEPCI COMPONENTS AND WEIGHT FACTORS
Component BLS index Component Component Component
Engineering Practice number weight group—Level I group—Level II
factor (weight factor) (weight factor)
Plates, carbon steel pcu3312#412 0.140
Plates, stainless steel pcu3312#45 0.110
ity factor at all? Why not just leave the
Pressure tubing, welded,
labor-cost components unadjusted?
carbon steel wpu10170628 0.043
According to Arnold and Chilton [5],
Pressure tubing, seamless,
“All cost indexes that have labor rates carbon steel wpu10170628 0.043
as components and that do not make Nonferrous mill shapes wpu1025 0.043
corrections for labor productivity have Plates, alloy pcu3312#431 0.110
built into them what index techni- Hot rolled bars, plates,
cians call an upward bias.” They felt and structural shapes pcu3312#4 0.110
that, without labor productivity cor- Storage tanks, 6,000 gal
rections, indexes will, over time, over- or less wpu10720104 0.027
state true (as opposed to apparent) Storage tanks, over
labor-cost changes. However, we 6,000 gal wpu10720104 0.025
should note that the productivity fac- Metal tanks wpu10720152 0.021
tor dampens labor-cost decreases, as Custom tanks, 3/4 in.
and less wpu10720136 0.082
well as increases. As mentioned ear-
Other custom tanks,
lier, Figure 1 displays the effect of field assembled wpu10720152 0.010
labor productivity on index trends. Petroleum storage tanks wpu10720104 0.010
Subtotal, heat exchangers
Assembling the CEPCI data and tanks – components —— 0.774
The productivity factor is one of 54 in- Subtotal, heat exchangers
puts that are used for calculating the and tanks – labor eeu31340006 0.226
updated CEPCI and sub-indexes. We Total for Heat
will discuss accessing the other key in- Exchangers and Tanks —— 1.000
puts used in their calculations and Plates, carbon steel pcu3312#412 0.105
how this calculation is performed. Sheets, hot-rolled,
carbon steel pcu3312#311 0.030
We said that each index and sub-
Cold-rolled sheets
index is the weighted sum of several
and strip pcu3312#7 0.010
components. Most of these components
Foundry and forge
correspond to Producer Price Indexes shop products wpu1015 0.060
(PPIs), updated and published monthly Industrial material-
by the U.S. Department of Labor’s Bu- handling equipment wpu1144 0.075
reau of Labor Statistics (BLS; Wash- Fans and blowers, except
ington, D.C.). According to the BLS de- portable wpu1147 0.025
finition, the PPIs, “track the Chem. ind. machinery wpu116604 0.250
average-change in net transaction Integral-hp motors and
prices that domestic producers in the generators pcu3621#2 0.035
mining, manufacturing, agriculture, Crushers, pulverizers and
screening machines wpu119202 0.150
and forestry sectors, as well as selected
Concrete ingredients wpu132 0.030
services industries, receive for the
Subtotal, Process
products that they make and sell.” The machinery – components —— 0.770
price quotations that the BLS uses to Subtotal, Process
build these indexes come from a statis- machinery – labor eeu31350006 0.230
tically chosen sample of representative Total for Process
transactions obtained from a statisti- Machinery —— 1.000
cally chosen sample of representative Plastic construction
producers in each 600 or so industries. products wpu0721 0.050
“In all, the PPI includes roughly Line and standard pipe,
carbon steel wpu10170618 0.400
100,000 price quotations from about
Mechanical tubing,
25,000 domestically producing estab-
carbon steel wpu10170629 0.100
lishments, resulting in the publication
Copper and brass mill
of approximately 13,000 indexes” [6]. shapes wpu102502 0.100
From these several thousand in- Concrete pipe wpu1332 0.050
dexes, 41 PPIs have been selected as Metal valves, except
inputs to the CEPCI and sub-indexes. fluid power wpu114902 0.200
These PPIs cover products as diverse Metal pipe fittings,
as carbon steel plates, fans and blow- flanges, and unions wpu114903 0.100
ers, concrete pipe, and lighting fix- Total for pipes, valves
tures. All of these items and many and fittings —— 1.000
Continues on adjacent page
more are key elements in a typical
64 CHEMICAL ENGINEERING WWW.CHE.COM JANUARY 2002
TABLE 3. REVISED CEPCI COMPONENTS AND WEIGHT FACTORS (Continued)
Component BLS index Component Component Component
number weight group—Level I group—Level II
factor (weight factor) (weight factor)
Sheets, hot-rolled,
carbon steel pcu 3312#311 0.057
CPI plant. The first columns in Table
Sheets, cold-rolled,
carbon steel wpu10170711 0.014
3 list these PPIs and their BLS code
numbers. The latter begin with the
Foundry and forge shop
products wpu1015 0.077 designation pcu or wpu.
Nonferrous mill shapes wpu11025 0.060 The CEPCI inputs also include 12
Copper and brass mill
labor-cost indexes (see Table 3).
shapes wpu1102502 0.053 These indexes are also compiled by
Integral-hp motors and BLS. They track changes in labor
generators pcu3621#2 0.036 rates, for such categories as Fabri-
Electronic components cated products and General building
and accessories wpu11178 0.400 contractors. In addition, special labor-
Metal valves, except cost indexes are included for technical
fluid power wpu114902 0.053 specialties, such as engineers and de-
Subtotal, Instrument and signer-drafters. Except for the last
controls – components —— 0.750 two, which are pcu8711#1 and
Relays & industrial pcu8712#4, all of the labor-cost index
controls – labor eeu31362506 0.063 numbers begin with the letters ecu or
Process control eeu. The large number of entries for
instruments – labor eeu31382306 0.187
products and labor results in a table
Subtotal, Instrument and that sprawls over three pages.
controls mfg. – labor —— 0.250
In Table 3, alongside each compo-
Total for process
instruments —— 1.000 nent and BLS index number is a com-
ponent-weight factor, which has been
Industrial pumps wpu114102 0.900
revised (or left unchanged) in this up-
Air compressors, stationary wpu114103 0.050
date. As the name implies, this factor
Gas compressors,
stationery wpu114104 0.025
weights the change in the component’s
price in proportion to its importance to
Other compressors and
vacuum pumps wpu114111 0.025 the sub-index or composite CEPCI.
Total for pumps Consider the first entry: Plates, carbon
and compressors —— 1.000 steel. This component is given a
Nonferrous wire and cable wpu1026 0.057 weight of 0.140. This means that
Lighting fixtures wpu1083 0.188 changes in the price of carbon-steel
Integral-hp motors and
plate accounts for 14% of the change in
generators pcu 3621#2 0.306 the “Heat exchangers and tanks” com-
Motors, generators, ponent-index. The subtotal of weight
generator sets wpu1173 0.043 factors equals 0.774. The balance
Transformers and power (0.226) is contributed by the labor for
regulators wpu1174 0.146 fabricating these components.
Switchgear, switchboard Moving one column to the right,
equipment wpu1175 0.260 “Heat exchangers and tanks” is listed
Total for electrical under the column heading Component
equipment —— 1.000 Group — Level I. The weight factor in
Prepared paint wpu0621 0.024 that column is 0.338. This weight factor
Hot-rolled bars, plates, denotes the contribution of “Heat Ex-
and structural shapes pcu 3312#4 0.406 changers and Tanks” to the Equipment
Concrete reinforcing bars, index, which is listed in the column
carbon steel pcu 3312#425 0.089 Component Group — Level II. Finally,
Concrete ingredients wpu132 0.129 notice that the weight-factor for Equip-
Insulation materials wpu1392 0.353 ment is 0.507, meaning that it con-
Total for structural tributes 50.7% to the composite CEPCI.
supports and The other three major sub-indexes do
miscellaneous —— 1.000
not have a Component Group — Level
TOTAL--EQUIPMENT —— —— 1.000
I. Buildings has just two components:
Construction materials “Construction materials” and “General
(special index) wpusi012011 0.530 ——
building contractors.” These contribute
General building
contractors eeu20150006 0.470 53% and 47% to the Buildings sub-
Total for buildings —— 1.000
index, which, in turn, contributes 4.6%
to the CEPCI. The four components of
Continues on next page
the Engineering and Supervision sub-
CHEMICAL ENGINEERING WWW.CHE.COM JANUARY2002 65
 OBSERVING THE FIVE-YEAR RULE
rule-of-thumb limit for cost escalation is five years. This ap- a lower dollar-per-ton production cost.

A plies to all escalations of capital or operating costs. There are

sound reasons for this particular time limit. Over periods of up
to five years, the differences between actual prices of equipment
Other changes were due to outside factors. Among the most in-
fluential of these were the many health, safety, and environmental
regulations that were imposed on the CPI. These included rules af-
and labor and those predicted by a cost index have been found to fecting such diverse matters as the height of catwalk railings, the
be small relative to the inherent error in most budget-level esti- monitoring of wastewater pH, and the control of fugitive VOC
mates (±20 to ±30%)[10]. Over longer periods, these deviations (volatile organic compound) emissions from relief valves. To com-
can become much larger. That implies that the CEPCI can (and ply with these regulations, CPI plants had to modify their
should) be used to adjust costs — but only if the adjustment period processes, install monitoring and control equipment, and hire ad-
is limited to five years. ditional professional and support personnel. For this reason alone,
An example will be used to as background for discussion. Let us a comparison of the 1970 and 2000 LOx plants would fall in the
say that a 700-t/d liquid oxygen (LOx) plant was built in 1970 for “apples-and-oranges” category.
$10 million. How much would it cost in 2000 to construct a 700- But let’s get hypothetical for a moment; suppose that the 1970
t/d LOx plant? First, read the annual average composite CEPCI’s and 2000 LOx plants were absolutely identical in all respects.
for 1970 and 2000 from Table 1. These values are 125.7 (1970) Even then, the CEPCI-escalated cost would not be accurate.
and 394.1 (2000). Then multiply the 1970 cost by the ratio of Even though these two plants would contain exactly the same
these two CEPCI’s to obtain the plant cost in 2000. Inflationary number and types of all parts, the prices of this equipment and,
changes in the buying power of the dollar are built into the index, more importantly, the labor required to install it, have escalated
so don’t make unnecessary currency corrections. at different rates. The major change to the composition of the
Cost = $10,000,000 3 (394.1/125.7) = $31,300,000 (3) CEPCI is to increase the weight of the costs of labor and cut the
This cost is not likely to be accurate. To begin with, during the in- weights of the equipment costs. Specifically, in the old CEPCI,
tervening thirty years a host of changes occurred. Some of these Equipment contributed 61% to the composite index, while in the
changes involved the liquid-oxygen production process. There are new CEPCI it contributes just 51%. At the same time, labor’s
steps made more efficient, steps omitted and steps added. The net weight factor increased from 32% in the old index to 45% in the
result was an improvement in the production efficiency, resulting in new CEPCI. ❒

TABLE 3. REVISED CEPCI COMPONENTS AND WEIGHT FACTORS (Continued) TABLE 5. NORMALIZATION
Component BLS index Component Component Component FACTORS
number weight group—Level I group—Level II Subindex or component Normalization
factor (weight factor) (weight factor) index Factor
Administrative support, Heat exchangers and tanks 4.550
including clerical ecu11142I 0.060 ———— Process machinery 5.264
Engineering pcu8711#1 0.325 Pipes, valves, and fittings 5.484
Designer/drafter pcu8712#4 0.395 Process instruments 4.499
Executive, administration Pumps and compressors 6.665
and managerial ecu11112I 0.220 Electrical equipment 3.393
Total for engineering Structural supports
and supervision ———— 1.000 and miscellaneous 4.244
General building Equipment 4.911
contractors eeu20150006 0.467 ————
Buildings 5.764
Heavy construction
Engineering and supervision 11.190
contractors eeu20160006 0.317
Construction labor 9.779
Special trade contractors eeu20170006 0.217
Composite CEPCI 6.363
Total for
construction labor ———— 1.000
GRAND TOTAL—CE PLANT COST INDEX: 1.000 is the base value of the Producer Price
Index. This base value is the arith-
TABLE 4. DATA FOR MARCH 2001 - BUILDINGS metic mean of the twelve monthly val-
Component Weight Base Current PPI Ratios x 100 ues of this PPI for 2000. Thus, we can
Factor PPI PPI Unadj. Adj. Weighted say that the base date of the revised
Product CEPCI is Year-2000. On Table 2 this
Construction material 0.530 144.1 142.3 98.8 98.8 52.34 is the last line augmented with the
General. bldg. values for the component-indexes. We
contractors 0.470 17.22 17.50 101.6 30.9 14.52
believe that selecting a year’s worth of
Total — before
normalization: 66.86 data allows for a broader base than a
Total — after single month.
normalization (normalization factor = 5.764): 385.4 At this point, the data entry is cor-
rect, but the results make no sense,
index contribute to that category, while tered into an Excel spreadsheet that is unless the data are normalized. The
the Construction Labor sub-index has used to calculate the composite CEPCI PPIs do not all start at 100 in
three components. The “Engineering and sub-indexes. 1959–1961, which is the historical
and supervision” and “Construction base of the index. (PPIs are given a
labor” sub-indexes account for 15.8% Normalizing the index value of 100 on the month that the se-
and 29.0% of the composite CEPCI, re- Also entered into this spreadsheet, in ries begins.) Also, labor costs are
spectively. These weight factors are en- protected squares, for each component given in units of dollars per hour.
66 CHEMICAL ENGINEERING WWW.CHE.COM JANUARY 2002
 TABLE 6. DISCONTINUED BLS PRODUCER PRICE INDEXES
AND THEIR REPLACEMENTS
Discontinued Producer Price Index Replacement Producer Price Index
Number Name Number Name
wpu10170611 Standard pipe, carbon wpu10170618 Line and std. pipe and oil
country tubular goods, etc
2001 is exactly equal to the value of
wpu10170622 Pressure tubing, welded, wpu10170628 Pressure tubing, carbon
carbon
the corresponding old index for Janu-
wpu10170624 Mechanical tubing, wpu10170629 Mechanical tubing,
ary 2001. This normalization process
welded, carbon carbon ensures that the old and revised in-
wpu114105 Other compressors wpu114111 Other compressors and dexes will have a seamless transition.
vacuum pumps If that was a little hard to under-
wpu10720112 Bulk storage tank, 6,000 wpu10720104 Storage and other stand, we present a simple example to
gallons or less non-pressure tanks illustrate the CEPCI calculation
wpu10720113 Bulk storage tank, over wpu10720104 Storage and other process. Consider the Buildings sub-
6,000 gallons non-pressure tanks index for the March 2001. The compo-
wpu10720138 Custom tanks, 3/4 in. wpu10720136 All other tanks and nents, their weight factors, and their
and less vessels,custom-fabricated
at factory PPIs are shown in Table 4. The weight
wpu10720147 Petroleum storage tanks wpu10720104 Storage and other factors are in Table 3 and the base PPI
non-pressure tanks values for 2000 have been stored. The
wpu10720148 All other customized wpu10720152 Metal tanks and vessels, current PPI data were read from the
tanks, field assembled custom fabricated and BLS website: www.bls.gov/data/home
field erected .htm. The unadjusted ratio is (current
wpu11660412 Chemical mixers wpu116604 Chemical industry PPI)/(base PPI). General Building Con-
machinery
tractors is a labor cost, so it is multi-
pcu3312#41511 Structural steel shapes pcu3312#4 Hot rolled bars, plates,
and structural shapes
plied by a productivity factor (0.3040 in
wpu114904 Parts & attachments for wpu114903 Metal pipe fittings, flanges,
March 2001).
valves & fittings and unions The Weighted Product is the ad-
pcu3312#453 Stainless steel plates pcu3312#45 Plates and structural justed PPI ratio multiplied by the
shapes, stainless weight factor. The sum of these
weighted products is shown both be-
TABLE 7. WEIGHT FACTORS AS REVISED fore and after normalization. Again,
Component or component group PPI No. Weight factor the normalization (splicing) factor is
original Revised the ratio of the Buildings sub-index
Prepared paint wpu0621 0.028 0.024 calculated via the old CEPCI to the re-
Hot-rolled bars, plates & structural shapes pcu3312#4 0.382 0.406 vised sub-index. Finally, the post-nor-
Concrete reinforcing bars, carbon steel pcu3312#425 0.077 0.089 malization value, 385.4, is the Build-
Concrete ingredients wpu132 0.117 0.129 ings sub-index for March 2001. The
Insulation materials wpu1392 0.396 0.353 splicing factors are stored in the
Engineering pcu8711#1 0.330 0.325 spreadsheet and they were worked
Designer-drafter pcu8712#4 0.470 0.395 out to make sure that the index con-
Executive, administrative, and management ecu11112I 0.140 0.220 tinues without a discontinuity. Table
General building contractors ecu20150006 0.334 0.467 5 is a list of normalization factors.
Heavy construction contractors ecu20160006 0.333 0.317
Special trade contractors ecu20170006 0.333 0.217 What was revised?
Heat exchangers and tanks grouping 0.370 0.338 There are more compelling reasons to
Process machinery grouping 0.140 0.128 revise the CEPCI than the centennial
Pipes, valves and fittings grouping 0.200 0.190 of this magazine in 2002 or the start of
Process instruments grouping 0.070 0.105 a new century. Progress over the last
Pumps and compressors grouping 0.070 0.064 18 years is the main impetus for a
Electrical equipment grouping 0.050 0.070 fresh look. Here are some of the rea-
Structural supports and miscellaneous grouping 0.100 0.105 sons for an update, and some of the ac-
Equipment grouping 0.610 0.507 tions put into the revised index.
Buildings grouping 0.070 0.046 1) The BLS no longer reports 14 of
Engineering and supervision grouping 0.100 0.158 the PPI inputs to the old CEPCI. Con-
Construction labor grouping 0.220 0.290 sequently, each of these inputs had to
be frozen at the value last reported by
There has to be some computational made for all the components of a sub- BLS. Needless to say, a frozen PPI
adjusting to make the series continue. index, the results are added. This sum contributes nothing to the index-up-
Every month, the latest values of becomes the current value of the sub- dating process. We were able to find
the component PPIs are entered into index. Lastly, each of the sub-indexes suitable replacements for these dis-
the spreadsheet. For each component, is multiplied in the spreadsheet by a continued PPIs. These replacements
the ratio of the current PPI and the “normalization factor” that has been are listed in Table 6.
base PPI is multiplied by the weight selected such that the resulting value 2) Two of the labor categories in the
factor. When this calculation has been of each revised sub-index for January old CEPCI (Draftsman and Typist) are
CHEMICAL ENGINEERING WWW.CHE.COM JANUARY2002 67
 Engineering Practice

TABLE 8. NEW CE PLANT COST INDEXES: JANUARY – SEPT. 2001

Index/ Month Jan. Feb. Mar. Apr. May Jun. Jul.* Aug.* Sept.*
CE Plant Cost 395.4 393.6 394.4 393.2 393.8 393.9 393.8 394.2 393.7
Equipment 439.6 437.0 437.7 437.7 436.5 436.5 436.8 436.8 436.1
• Heat exchangers 367.7 365.6 365.7 360.3 361.3 361.3 363.6 363.0 364.6
and tanks
• Process machinery 441.6 441.1 442.1 441.5 442.1 443.3 442.6 440.9 431.1
• Pipes, valves 551.2 544.1 545.8 551.6 549.0 547.8 547.0 540.4 542.4
and fittings
• Process instruments 368.1 365.8 365.8 364.5 363.8 362.9 361.0 360.2 357.1
• Pumps and 672.6 670.3 673.0 673.3 675.8 674.0 673.3 693.1 695.3
compressors
• Electrical equipment 340.2 340.8 341.7 342.1 342.6 342.0 341.8 341.1 341.1
• Structural supports FIGURE 2. The revised CEPCI shows a
416.3 414.2 414.3 413.4 413.6 415.9 415.7 415.8 416.6
and misc. slight tendency to decline while the old
Construction labor 300.6 300.0 301.1 299.1 300.8 301.5 301.2 305.3 304.7 index is rising during 2001
Buildings 385.4 385.6 385.6 385.2 389.1 389.1 386.6 387.8 387.4
Engineering 341.3 341.1 342.5 343.7 343.4 342.8 342.2 342.2 341.5
and supervision
* The index values for this month are preliminary

TABLE 9. COMPARISON OF CE PLANT COST INDEXES: JAN.–SEPT. 2001

Month CE Index Equipment Const. Labor Buildings Engr.and Supv.
revised older revised older revised older revised older revised older
Jan. 395.4 395.4 439.6 439.6 300.6 300.6 385.4 385.4 341.3 341.3
Feb. 393.6 395.1 437.0 438.8 300.0 301.4 385.6 385.7 341.1 341.0
Mar. 394.4 394.3 437.7 438.1 301.1 300.0 385.6 383.9 342.5 341.7
Apr. 393.2 394.5 436.5 439.6 299.1 297.6 385.2 382.7 343.7 341.4
May 393.8 395.4 436.5 439.7 300.8 299.8 389.1 386.6 343.4 342.0
June 393.9 395.1 436.5 439.0 301.5 300.2 389.1 386.2 342.8 341.7
July(P) 393.8 395.3 436.8 439.1 301.2 301.6 386.6 385.2 342.2 341.5 FIGURE 3. The revised Equipment
Aug.(P) 394.2 396.3 436.1 439.7 305.3 303.9 387.8 386.7 342.2 342.1 subindex shows a steady decline while
the old Equipment subindex is nearly
Sept.(P) 393.7 396.8 435.7 439.9 304.7 305.4 387.4 387.4 341.5 341.8 constant during 2001

no longer relevant. These days, draft- Table 7 displays the weight factors creases from 0.22 to 0.29. The rela-
ing is typically done via CAD (com- that were changed, showing both the tively slow growth in equipment
puter-assisted drafting) programs, not old and revised values. We had to ad- prices (as tracked by the PPIs) over
on blueprints, and today’s engineering- just components (such as Prepared the past few decades account for much
design firms employ hardly any typ- Paint), Level I component groups of the drop in the equipment compo-
ists. PCs on everyone’s desk have re- (Heat Exchangers and Tanks), and nent weight factors. This slow price
placed most typists. Hence, we Level II component groups (Equip- growth is due to improvements in do-
replaced these labor categories with ment). The new weight factors for mestic fabrication processes, the im-
those that more accurately reflect the each entry (whether revised or left un- porting of foreign-made equipment,
21st-century labor mix. These new job changed) are listed with their position and other factors. On the other hand,
categories are in Table 3 under the En- on Table 3. For instance, the revised the relatively large increases in labor
gineering and Supervision sub-index. weight factor for Heat Exchangers costs (both technical and non-techni-
3) Another area needing moderniza- and Tanks (0.338) tells us that this cal) are attributable to the usual influ-
tion was the set of component-weight component group accounts for 33.8% ences — inflation, skilled personnel
factors. For the most part, these re- of the Equipment sub-index, not shortages, and labor-management
flected the composition of the typical 33.8% of the composite CEPCI. bargaining agreements.
CPI plant circa 1960 or 1970. We sur- The differences between the origi- 4) Lastly, the annual productivity
veyed roughly twenty CPI companies, nal (1982) and revised (2001) weight growth rate was revised to better re-
engineering firms, index publishers factors are not large. Still, there is a flect both short- and long term changes
and technical organizations to obtain definite trend away from equipment- in construction-labor productivity. As
the information necessary to update oriented components and toward this growth rate is the key input to the
these weight factors. As a result of their labor-cost-oriented components. The productivity factor, it had to be selected
responses, we modified all of the Level I Equipment sub-index weight factor with care. Unfortunately, while the
and Level II weight factors, and several decreases from 0.61 to 0.507, while BLS compiles labor-productivity data
of the component weight factors [7]. the factor for Construction Labor in- for 100% of the manufacturing sector,
68 CHEMICAL ENGINEERING WWW.CHE.COM JANUARY 2002
 THE VAPCCI INDEXES
TABLE 10. VATAVUK AIR POLLUTION CONTROL COST INDEXES
Several cost indexes appear on the Economic Indica-
tors page of this magazine. One set of that relies Year
heavily on BLS inputs consists of the Vatavuk Air Pol- Control device
1994 1995 1996 1997 1998 1999 2000
lution Control Cost Indexes (VAPCCIs). These are cus- Carbon adsorbers 101.2 110.7 106.4 104.7 103.6 100.8 108.0
tom-designed for adjusting the equipment costs of
Catalytic incinerators 102.0 107.1 107.0 107.7 106.5 102.9 114.3
air-pollution control systems. They were created in
Electrostatic precipitators 102.8 108.2 108.0 108.8 109.2 101.2 101.1
1994 and first published in CE in late 1995 [11].
These quarterly indexes presently cover 11 control- Fabric filters 100.5 102.7 104.5 106.2 109.5 111.7 113.0
equipment categories. The first-quarter of 1994 is the Flares 100.5 107.5 104.9 105.8 103.6 99.4 104.3
base date for the VAPCCIs. All indexes have been ar- Gas absorbers 100.8 105.6 107.8 107.6 109.7 110.9 112.9
bitrarily assigned a value of 100.0 for that base date. Mechanical collectors 100.3 103.0 103.3 103.9 111.0 119.6 121.8
Table 10 lists the annual VAPPCIs for 1994 through Refrigeration systems 100.5 103.0 104.4 106.1 107.6 105.7 106.1
2000. Each annual index is the average of the quar- Regenerative thermal
terly indexes. Year 2001 can be found on the back oxidizers 101.4 104.4 106.3 107.9 108.9 108.1 109.0
page of this issue. Reference [11] is posted at Thermal incinerators 101.3 105.9 108.2 109.4 110.5 108.1 107.9
www.che.com/CEEXTRA. ❒ Wet scrubbers 101.3 112.5 109.8 109.0 109.7 108.8 113.8

FIGURE 4. There is no discernable dif- FIGURE 5. There is no significant dif- FIGURE 6. There is a no systematic dif-
ference in Construction Labor subindex ference in Buildings Cost subindex in ference in Engineering and Supervision
in 2001 between the revised and the 2001 between the revised and the older Cost subindex in 2001 between the re-
older version version vised and the older version

it does not compile construction-labor places the 1.75% rate used in the old While it’s a bit risky to draw infer-
productivity data. Our efforts to obtain CEPCI computation. (Coincidentally, ences from preliminary results, we can
these data from other sources, govern- the 2.2% rate is close to the arithmetic indicate some trends, or lack thereof.
mental and private, failed. mean of 1982’s 1.75% and the original The CEPCI composite decreases from
However, based on BLS’s recom- 2.5% rate used [1963 to 1981] in the a January high of 395.4 to 393.7 in
mendation, we have decided to use the CEPCI calculation.) September. But in the meantime, it os-
labor productivity for the Total Non- cillates within that range. The Equip-
farm Business sector. According to Something old, something new ment sub-index shows a modest de-
BLS, this sector, “represents changes In this ending section, we present the cline while Construction Labor
in the productive efficiency of a sector new indexes and display them side- increases. Meanwhile, the Buildings,
that includes service-producing indus- by-side with the indexes they’ve re- and Engineering and Supervision sub-
tries and the construction and mining placed. The achieved objective has indexes stay almost constant during
industries, as well as manufacturing” been to improve and update the these nine months.
[8]. The annual productivity growth CEPCI but not change its basis. This is interesting, especially when
(measured in output per hour) for the For this comparison we prepared we remember that these last three sub-
Private Nonfarm Business sector was Table 8. This table lists the CEPCI indexes are heavily weighted with
2.3% for the 1995–98 period and 2.2% composite, along with its sub-indexes labor-cost components. That is, any
for the 1947–98 period [9]. (The most and component-indexes for the changes — increases or decreases — in
recent year for which these productiv- months of January through Septem- these labor components have been dis-
ity data are available is 1998.) Thus, ber 2001. Note that the indexes for counted by the productivity factor. De-
the short and long term compound July and August and September are spite that dampening effect, the sub-in-
growth rates for this sector are virtu- still preliminary. That is, at the time dexes have grown, while the composite
ally equal. We have selected the 2.2% of the writing of this article (mid-De- CEPCI and Equipment sub-index have
annual growth rate to use in the pro- cember) the BLS inputs for the decreased. A good part of that decrease
ductivity factor calculation. This re- months following June were not final. is due to the steep drop during this pe-
CHEMICAL ENGINEERING WWW.CHE.COM JANUARY2002 69
 Engineering Practice

riod in some of the steel-related PPIs pricing schemes. Over one extended pe-
that are key inputs to the Equipment riod, the CEPCI might track these
components. These include such com- prices quite closely. But over the very
ponents as stainless steel plates next period, it might deviate from them
(dipped 7.4%) and carbon steel sheet significantly. Needless to say, do not es-
(declined by 6%). Because the Build- calate plant costs blindly via the CEPCI
ings, Engineering and Supervision, (or any other index) if you can access
and Construction Labor sub-indexes current costs. Still, when there is not
are influenced much more by changes enough time or resources to obtain the
in labor costs than in basic steel prices, latest costs, the CEPCI and its compo-
they usually increase over time. nents — easy to use, easy to under-
Finally, the revised and old (previous) stand, and custom-designed for the CPI
CEPCIs for January through Septem- — provide an excellent substitute. ■
ber 2001 are compared in Table 9. This Edited by Peter M. Silverberg
table, along with Figures 2 through 6,
displays the respective CEPCI compos- References
ites and the four major sub-indexes. No- 1. Arnold, T.H. and Chilton, C.H., New Index
Shows Plant Cost Trends, Chem. Eng., pp.
tice that the January entries in Table 8 143–148, February 18, 1963.
are identical. This is not mere coinci- 2. Matley, J., CE Plant Cost Index — Revised,
Chem. Eng., pp. 153–156, April 19, 1982.
dence. As explained earlier, the revised
3. Ibid.
CEPCI composite and sub-indexes for
4. Stevens, R.W., Equipment Cost Indexes for
January have been purposely equated the Process Industries, Chem. Eng. , 54, 11,
with the their counterparts in the old pp. 124–126, November 1947.
5. Arnold and Chilton, Op Cit.
CEPCI. This normalization has been
6. Kowal, J., U. S. Bureau of Labor Statistics,
done to effect a seamless transition be- Producer Price Index Analysis and Public In-
tween the two CEPCIs. formation, E-mail sent August 27, 2001.
While there isn’t space in this article 7. The following individuals kindly provided
(via private communications) helpful sugges-
to compare all five indexes, we can tions for improving the weight factors:
focus on one, the CEPCI composite. • William Haselbauer (Lyondell, Houston, Tex.)
From equality in January-March, the • John Hollmann (American Association of Cost
Engineers, Morgantown, W. Va,)
respective composites begin to diverge. • Bernard A. Pietlock (DuPont Automotive,
By September, the composites have Wilmington, Del.)
drifted away from each other by 2.9 • Richard Vishanoff (Marshall & Swift, Los An-
geles, Calif.)
points. The revised index shows a little
8. Usher, L., U.S. Bureau of Labor Statistics,
deflation and the old index shows a lit- Division of Industry Productivity Studies, E-
tle inflation. We can allow one obser- mail sent February 28, 2001.
9. U. S. Bureau of Labor Statistics, “Multifac-
vation: The old and new CEPCIs are tor Productivity Trends, 1998,”
responding to the same economic in- (http://stats.bls.gov/news.release/
prod3.nr0.htm), September 21, 2000.
puts — for instance, the PPIs and pro-
10. Merrow, E.W., and Hollmann J.K,. Control-
ductivity factor. But, because of the re- ling Project Costs, Chem. Eng., 108, 12,
visions we’ve made, the indexes are pp.76–80, November 2001.
11. Vatavuk, W., Air Pollution Control — Esca-
responding to them differently. late Equipment Costs, Chem. Eng., 102, 12,
pp. 88–95, December 1995. A copy of this ar-
ticle has been posted at www.che.com/CE-
Some final thoughts EXTRA
We hope that the revised CEPCI is a
much better fit than the previous ver- Author
sion, with respect to how accurately it William M. Vatavuk is pres-
ident of Vatavuk Engineering
tracks changes in CPI plant-construc- (3512 Angus Road, Durham,
tion costs. Nevertheless, the CEPCI N.C. 27705-5404; Phone: 919-
489-8810; Fax: 413-638-1336;
cannot be applied unthinkingly; there Email: williamvatavuk@msn
are limitations to its use. .com), a consulting firm spe-
cializing in air-pollution-con-
In general, the CEPCI can be used trol technology and cost
analysis. He retired from the
confidently, to escalate plant costs, but U.S. EPA’s Air Programs Of-
only for periods no greater than five fice in 1999 after a thirty-year
career. Vatavuk is the author of two books and
years (See box p. 66). Such limitations dozens of technical articles on cost analysis for
merely remind us of the true nature of pollution control. He created and regularly up-
dates the “Vatavuk Air Pollution Control Cost
indexes. That is, like all other indexes, Indexes” which appears monthly in CE. Vatavuk
the CEPCI is merely a model — a rep- has a B.E. in pollution engineering from
Youngstown State University and is a registered
resentation of equipment and labor professional engineer in North Carolina.

70 CHEMICAL ENGINEERING WWW.CHE.COM JANUARY 2002