See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/222590683

Assessing Extension of Time delays on major

projects
Article in International Journal of Project Management January 2003
DOI: 10.1016/S0263-7863(01)00060-6

CITATIONS

READS

38

547

1 author:
Terry Williams
University of Hull
102 PUBLICATIONS 3,127 CITATIONS
SEE PROFILE

All content following this page was uploaded by Terry Williams on 22 August 2016.
The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.

WILLIAMS, T.M. (2003) Assessing Extension of Time delays on major projects.

International Journal of Project Management 21, 1, 19-26

Assessing Extension of Time delays on major projects

TERRY WILLIAMS
Strathclyde University

Department of Management Science, Strathclyde University, 40 George Street, Glasgow G1

1QE, UK

terry@mansci.strath.ac.uk
Phone: +44-141-548-3548
Fax: +44-141-552-6686

Assessing Extension of Time delays on major projects

This paper describes the standard methods currently available for assessing Extension of Time
delays on major projects, and issues around such assessment. Network-based techniques have
been much developed, and are powerful and credible tools for assessing the effect of a small
number of discrete impacts on a project, where major reactive management actions have not
been needed. The paper points out the problems inherent in such methods in other, more
complex situations, and describes the contribution that other methods using cause mapping
and System Dynamics can make. These however are also not universally useful, and the paper
describes how the two methodologies can be used together to produce useful analyses of the
impact of delays on a project.

Keywords: Network analysis, complex projects, CPM, system dynamics

EXTENSION OF TIME CLAIMS

Projects are classically defined by the need to complete a task on time, to budget, and with
appropriate technical performance / quality. In recent decades, projects have tended to
become more time-constrained (Clarke [1]), and the ability to deliver a project quickly is
becoming an increasingly important element in winning a bid. There is an increasing
emphasis on tight contracts, using prime contractorship to pass time-risk onto the
contractor, frequently with heavy liquidated damages (LDs) for lateness. Thus it is
becoming all the more important for a contractor, when faced with delays caused by the
client, to ensure (s)he claims for a suitable Extension of Time (EOT) to his contractual
finish-date, otherwise he will find himself subject to LDs for reasons within the clients
control, not within his own control. Cushman et al [2] says that (referring to US case law)
it is well established that in a construction contract, time is not generally of the essence
unless it is specifically unless it is expressly provided, and a contractors failure to complete
its work in accordance with the time requirements of the contract does not entitle the owner
to terminate the contract or excuse nonpayment, but it may expose the contractor to liability
for delay damages (our emphasis).

Now good contractors will try to avoid getting into claim situations where possible. Jergeas
and Hartman [3], for example, give a paper on how to avoid construction claims. They
suggest as general guide-lines: good record-keeping in case of a claim [although in the
complex situations of interacting effect discussed below, we dont always know in advance
what records to keep], knowledge of contract (often contractors dont know whats in it),
preservation of rights by filing notice of potential claims (for anything different from
anticipated, congestion, owner-supplied equipment late, requirement to stop or accelerate
etc); qualify change orders (i.e. dont waive rights by signing-off change orders, especially
when indirect/D&D costs might be significant); effective scheduling; proactive actions (e.g.
requesting EOT, stating up-front who is responsible for acceleration, recording
disagreements, etc etc). (Semple et al [4] gives the results of a pilot study of 24 construction
companies in Canada, describing the amounts of claims and delays, and the main causes
claimed).

But EOT claims do happen, and they are often very difficult to prepare, both conceptually
and practically. Scott [5] describes a (UK) survey, and says that Claims for EOT appear on

the majority of major civil-engineering contracts, although acceleration claims occur much
less frequently. Arditi and Patel [6] say that any time-related claim situation needs to be
resolved with regard to three basic elements of time impact: causation, liability and
damages. Cushman et al [2], quoted above, go on to say (still referring to US case law) that
the proof of delay has evolved dramatically in the last decade as a results of sophisticated
owner requirements, evidential case law as described [earlier], and advances in computer
technologyThe contractor bears the burden of proving the extent of the delays for
which it seeks compensation and, in addition, the burden of proving the damages it incurred
as a result of such delays. They assume that there is a fixed unambiguous entity known as
Time Impact Analysis which can be used on each occasion an EOT claim arises but as
we shall in this paper, this is over-optimistic, and there are a number of issues in even quite
simply cases, and critical problems in complex projects. Alkass et al [7], who have a lot of
experience in such delay claims, talk about the effort involved: for example, they say that
70% of the effort in a claim is spent on searching and organizing information, and they
identify three problem issues: the proper classification of delay types (so that the right party
is credited), concurrent delays (which can have an effect on the overstatement of
compensation), and real-time CPM analysis (so you use the CPM in effect at the correct
time).

CPM, or the Critical Path Method, is overwhelmingly the standard approach for considering
the effects of delays on a project (Wickwire and Smith [8]), and forms the basis for
discussion of EOT claims (a standard text for example being Rubin [9]). There is a great
deal of work done in CPM analysis lets divide into the four different questions asked

What is the effect of client-induced delays on a project?

Which CPM should be used?

What is the effect of many non-concurrent delays excusable & non-excusable on a

project?

What is the effect of many delays excusable & non-excusable some of which are
concurrent on a project?

Then well see when and whether the methods work, and what additional methods are
needed.

Before we start that, we should mention the different categories of delay and entitlement.
Various references define very similar categories, again mostly referring to US cases.
Reams [10] defines

excusable / compensable

the clients fault, so the contractor gets extension of

time and delay damages

excusable / noncompensable

not the clients or the contractors fault, so the

contractor gets an extension of time but no delay
damages

non-excusable / noncompensable the contractors fault, so no extension of time, and

indeed the client can claim damages

King and Brooks [11] define (with US legal cases to illustrate each case):

delays that are not excusable, when the contractor may be liable in damages to the
owner, computed either based on actual damages suffered, or as contractually deliniated
liquidated damages.

excusable delays, not the fault of the contractor and may extend the time for contract
performance; these can be compenable or noncompensable (note that when each party
has contributed to a delay, each party then bears its own costs of the delay).

Note also that Cushman et al [2] in defining excusable delay, in J.D.Hedin Construction Co.
V. United States the Court of Claims said that the contractor was not required to have
prophetic insight and take extraordinary preventative action which is simply not reasonable
to ask of the normal contractor.

WHAT IS THE EFFECT OF CLIENT-INDUCED DELAYS ON A PROJECT?

If there is just one delaying event at issue, how is the liability calculated that is, how much
(if any) time-extension should be allowed?

McCullough [12] gives the obvious answer to this question but points out the difficulties
involved: The basis of most construction claims is a delay. CPM schedules are used to
evaluate the delay caused by specific impacts. This process is quite simple in theory but
extremely difficult in practice. In theory, a baseline schedule is developed based upon the
best estimates of achievable production and sequence of activities at the time. As job
conditions change, this baseline will be updated. To determine if a delay has occurred, the
baseline at the time of the impact is used, and the actual impact is inserted into the schedule
as a new activity or as a change to the existing one. The schedule is then recalculated, and
the change to the end date of the schedule is the delay attributable to the impact being
evaluated. He then goes on to identify some of the problems with this for example,

schedules are not in practice updated, so you have to recreate the as-built schedule at the
time of the impact from current diaries; quantifying the effect of the impact is not obvious,
and so on. (Reams [10] makes similar comments.). But these are practical problems the
essential technique of inserting an activity into the CPM seems fairly clear theoretically.

There is a generally accepted rule that delays to non-critical activities dont matter (unless
the delay makes them critical). For example, Kraiem and Diekmann [13] state that
concurrent delays on non-critical paths are not considered, because a delay that occurred
on a non-critical activity does not participate in delaying the completion of the project.
Yogeswaran et al [14] looks at claims for extension of time under excusable delays,
analysing 67 civil engineering projects in Hong Kong; looking at non-critical activities, they
say that an excusable delay to a non-critical activity does not give rise to an extension of
time to the date for completion unless the delayed period exceeds the float available to the
non-critical activity. Cushman et al [2] similarly say that For purposes of determining
whether the project has been delayed and for purposes of apportioning delays, only delays
on the critical path of the project figure in the analysis because, by definition, delays not on
the critical path will not delay the completion of the project [reference to G.M.Shupe Inc. v.
United States]. However, we will go on to discuss below that this is a simplistic and overdiscrete understanding of CPM. Indeed, it is worth noting here that the latter authors
(Cushman et al) go on to say there are many possible critical path schedules for any
particular project, depending on variables such as projected or actual material, labor and
equipment resources, and depending on the sequence of construction preferred by the
superintendent in charge. In addition, the as-planned schedule almost always differs from
the as-built schedule, because as the project moves forward contingencies arise that delay
some activities and accelerate others so that the critical path changes.

WHICH CPM SHOULD BE USED?

Which version of the CPM network should be used? Reams [10] says that classical methods
(e.g. Rubin [9] ) use the as-planned schedule. He discusses all the work involved in
preparing an as-planned schedule for court, either when an as-planned schedule already
exists (drawing on the then standard book, Hohns [15] who says that the schedule must be
complete, and also error-free, and also points to cases where the use of the schedule for a
claim was rejected because it did not reflect the sequence of work as actually intended and

performed) or where it must be developed as a good guess of what should have happened.
But then he points out that the schedule will have changed before the delay (due to errors,
contractors changes, clients scope-changes, etc etc), so it isnt correct as it stands but must
be modified a contractor may be held to have not acted prudently if it fails to reflect this
new knowledge in subsequent project plans.

Scott [5] says that in the US, the general approach is from Wickwire and Smith [8], using an

as-built schedule and then removing all of the delays due to the owner, whereas under the
UK Institution of Civil Engineers Conditions of Contract (Institution of Civil Engineers
[16]), the supervising engineer uses his best judgement to decide EOT claims (the
established procedure in the USA [of using as-built CPM schedules for claims] is almost
unheard of in the UK). He uses some extremely simplistic examples (based on two-activity
networks), but even for these there wasnt agreement between practitioners where liability
lay or what should be claimed so clearly the issues are non-trivial. (There also wasnt
agreement about critical-path issues, so clearly the difficulties here are recognised.)

In fact, to explain the complete course of events in a project, a number of CPMs are
required. Arditi and Patel [6] use five:

The as-planned schedule

The as-built schedule

The Owner-accountable schedule, which is essentially the as-planned schedule plus

delays that can be attributed to the client

Adjusted schedules a series of schedules to explain how the as-planned turned into
the as-built so, first as-planned, then as-planned plus the first delay, used to quantify
the effect of the second delay, then as-planned plus the first two delays and so on

As-projected schedule used mid-project, which is as-built up to now and as-projected

for the remainder of the project.

WHAT IS THE EFFECT OF MANY NON-CONCURRENT DELAYS EXCUSABLE

& NON-EXCUSABLE ON A PROJECT?

Where there are many delays to a project, some of which can be blamed on the client and
some the fault of the contractor, it is necessary to identify the effect of what can be blamed

on the client. To simplify things initially, let us assume that these delays are nonconcurrent.

Bordoli and Baldwin [17] summarise methods to look into this situation, starting at the
level of bar-charts and scatter-diagrams, which although simple can still have a high visual
impact. They look at:

the as built network, showing the story of what happened

then the as built subtracting impacts network subtracts the impacts to give a nondisrupted programme

then the baseline adding impacts adds the impacts to a baseline.

and finally a window analysis, referring to Galloway and Nielsen [18] since the
as-planned network is updated throughout the project (and thus the critical path), the
delay-impact assessment is only carried out within each window between major
updates.

But there are clear flaws with these so that they developed a multi-step methodology for
assessing the events in all their implications. Based on the as-planned network, for each
event the methodology:

identifies progress at event date and updates the network using the progress data

simulates the event in the updated network

considers mitigating actions (which we will discuss further below)

then the analysis in the end shows the extension of time increasing event-by-event, and
these extensions can be summed for excusable and non-excusable events.

This is clearly a good method which seems to get around a lot of problems. However, it
does not cope with concurrent delays, and also (and crucially, as we shall see later),
mitigation actions are not as simple as those modelled in the Bordoli and Baldwin paper,
and indeed can often have quite non-intuitive effects. Let us look at the first problem, then
move onto the latter.

WHAT IS THE EFFECT OF MANY DELAYS EXCUSABLE & NON-EXCUSABLE

SOME OF WHICH ARE CONCURRENT ON A PROJECT?

When both client and contractor delay the project concurrently, the legal position, at least in
the US, is essentially that the client is not eligible for extension of time or damages.

Cushman et al [2], for example, say that If the contractor would have been delayed in any
event by causes within its control, that is, if there was a concurrent nonexcusable delay, the
general rule is that it would be inequitable to grant the contractor either an extension of time
or additional compensation (with rule stated in Klingensmith Inc. v. United States). . On
the other hand, when the owner and contractor concurrently delay the work, and
responsibility for the delay cannot be apportioned, the contractor is generally not liable for
liquidated damages.

Curiously, Bartholemew [19], discussing Kraiem and Diekmann [13] says that The writer
totally disagrees with the authors premise that concurrency is relevant to the contractual
significance of delays. Contractual significance can only be judged by the effect delays
have on: (1) Extending overall project completion; (2) extending interim milestone dates;
or (3) by rendering certain following activities more costly (in the case of certain
compensable delays). CPM network analysis permits this judgement to be made. However,
he does go on to say When a number of such delays are involved the cumulative effect is
usually the important issue, then goes on to discuss his usual way of working: An asplanned schedule seldom constitutes the criterion for measuring actual fulfillment of the
work as the authors claim.the usual situation is the retrospective analysis case depicted by
the authors; so his method for analysing the effect of delays is to draw the network as
actually happened, then remove the delays to see the effect of all of them (or look at
removing them one by one in some order). Logcher [20] points to a court case doing exactly
that. Kraim and Diekmann [21] end up the discussion with the equation:

New contract duration = as-planned duration + excusable delays+ compensable

delays + concurrent delays

So how are the delays combined? A well-known paper here is by Alkass et al [22], who
define the combination of delays (quoting Rubin [9]) thus:

if excusable and nonexcusable delays occur concurrently only an EOT is granted

if excusable compensable and excusable noncompensable delays occur concurrently

only an EOT is granted, no damages

if two excusable compensable delays occur concurrently an EOT and damages is

granted

They then describe the use of four types of schedule:

as-planned

adjusted (as-planned adjusted for change orders, construction changes, delays,

acceleration)

as-built

entitlement schedules show the original construction completion dates, how these dates
have been impacted by excusable delays, and the projected completion dates given the
remaining work. Final entitlement schedules reflect the original, adjusted and actual
completion dates used to establish the total time that the contractor or the owner is
entitled to for compensation.

And they describe six methods all tested on a common simple test case (consisting of 10
activities, adapted from Kraiem and Diekmann [13]):
i) the global impact technique adds together all delays; this clearly over-estimates the
total delay
ii) the net impact technique just puts all the delays onto a bar-chart, then calculates the
net delay (i.e. taking into account concurrency); this doesnt take into account delaytypes.
iii) the adjusted as-built CPM technique this takes the as-planned and puts into it the
delays as activities to end up with the as-built. Alkass et al discuss problems again by
not taking into account delay-types (and a discussion of the problem of claimants
tying their delays to non-critical parts of the network).
iv) the but for or collapsing technique, which takes the as-planned CPM and adds the
delays that one party is willing to accept, to give an acceptable total delay. However,
this doesnt take into account changes in CPM during the course of the project so a
delay might be on the as-planned critical path, but not on the as-built critical path.
v) the snapshot technique looks at the period between two snapshots from a snapshot
time, it takes actuality (durations & relationships) up to a second snapshot-times and
applies that to the as-planned, then compares the two end-dates; but this doesnt
consider delays prior to the first snapshot.
vi) the time impact technique, similar to the snapshot technique but looks at just one
delay, or delaying event.
The authors concludes that (i)-(iii) are very simplistic; (iv) is better but is only done once;
(v)-(vi) are good but dont classify delays by fault implying that more analysis is needed.
They then go on to describe a new, seventh, method

the isolated delay type the authors say that this tries to combine the approach of (v)(vi) with the scrutinizing approach of (iv). The method looks at time periods, then
applies only relevant portions of delays so looking at project completion dates before

10

or after shows change, and the discrepancy is attributed to the delays that were
incorporated.

PROBLEMS WITH CURRENT CPM METHODS

However, work in which the author has been involved, in particular his work with Eden,
Ackermann and Howick on post-project litigation claims (summarised in Eden et al [23])
has shown a number of problems with current CPM methods when dealing with certain
projects. These seem to be particularly relevant for complex projects (Williams [24] defined
a complex project in terms of structural complexity, the number and interdependence of
elements (following Baccarini [25]) and uncertainty in goals and means (following Turner
and Cochrane [26]).). They are perhaps also more relevant to more volatile phases of
projects (e.g. more so in an engineering phase than a construction phase).

There are three critical issues, with two additional problems also to note

(i)

The first, and perhaps most obvious, issue is that many delays or disruptive effects
impact many activities simultaneously a problem both in modelling projects posthoc and in pre-project risk analysis (Williams [27]). For example, a Change Order
might affect many design activities simultaneously and might bring additional interrelationships between the activities. Or to take Bowers [28] example, when
developing an aircraft, the loss of a test aircraft would cause disruption to each of the
airframe development, avionics development and engine development simultaneously
(and different effects to each one). To evaluate the impact of one such effect can be
carried out in a CPM by considering all of the impacts on each activity in the network;
however, when a number of effects occur and they overlap in time, then more
sophisticated analyses or simulations are required.

(ii)

Where there is a significant number of delays rather than simply assessing the effect
of a single event methods like Bordoli and Baldwin [17] become impractical. But
more than this, the delays will usually lead to effects which would not be obvious
from simply looking at the network. In particular, they lead to the softer (humanoriginated) type of effects. For example, (Eden et al [23]) disruptions might cause demotivation or loss of morale, which will affect productivity and hence cause further

11

delay. Or, there might be schedule pressure (what happens to productivity when there
is particular pressure to achieve a very tight deadline?), or stress, or effects on teamwork etc. These are clearly very important effects to recognise when managing
projects (the importance of recognising - and then modelling - these effects was
discussed in a recent NATO Workshop (Williams [29]).

(iii) Thirdly, most of the methods outlined above essentially assume an inactive
management. That is, the effect of events are modelled but it is assumed that the
network stays the same and management take no actions in response, which is clearly
wrong (Williams [30], Jorgensen and Wallace [31]). This alone can bring the results
from such methods into question. (The same criticism can be levelled against standard
time-risk-analysis models, which again essentially model a helpless management
watching to see the out-turn of their project without intervening to bring late-running
projects under control (Williams [32])).
Some authors in these areas do allow some management reaction. For example, King
and Brooks [11] defines excusable delays, as above, and say that they can lead to
constructive acceleration, when a contractor is forced to increase the pace of work to
meet a project schedule that has not been extended because of excusable delays; they
also list typical cost categories (e.g. jobsite support, escalation labour etc, unabsorbed
home office delay). Arditi and Patel [6] assume simple acceleration rules, such as
Unit Acceleration Cost and so on. And these ideas can go straightforwardly into
Bordoli types of simulation
But even individual actions are not this simple and combinations of actions have
consequences that these methods cannot model. If acceleration is attempted by taking
on more personnel, then often the effect is less than expected. Or it may have
ramifications that are costly for example, having hired extra labour, it is rarely easy
to hire and fire at will; so management might need to keep workers idle, or work on
tasks in a non-optimal sequence to keep the workforce occupied, or even work on
tasks before being given client approval. Indeed, such an action might even be
counter-productive (the well-known $2,000 hour effect Cooper [33]).
This last point leads on to a difficult issue when claiming. If management have taken
some action in response to the claim, was the action optimum - or even reasonable?
Indeed if management action in accelerating projects is often counter-productive
(Cooper [33], Howick and Eden [34]), is it reasonable for management to take it at
all!?

12

There are two additional problems.

Most of the authors above distinguish between delays to activities on the critical path
(which are deemed to be important) and delays to other activities (which are deemed to
be unimportant) (Kraiem and Diekmann [13], Cushman et al [2], Yogeswaran et al
[14]). However, with resource-constrained networks, even theoretically the idea of
criticality is undefined (Bowers [35]). And practically, even where specific resourcetypes are not defined, often there will be implicit resource constraints (e.g. a senior
engineering can only look after so many tasks at one time). This means that the standard
analyses can become meaningless as soon as there are significant resource-constraints
which is the case for many (if not most) large projects.

Finally, the effect of delays is not a simple function of their duration. We should
remember that sometimes an apparently small and unimportant delay can cause
significant costs. For example, a 3-month delay in a North Sea project which causes a
weather-window to be missed, could lead to a one-year delay to the overall project as
the project has to wait, idle, for the next summer. In the case of Murphey versus US
Fidelity, reported by Arditi and Patel [6], a contractor sued a supplier for losses due a
delay in delivery; here, despite efforts to expedite, the work ran into the winter and
hence costs overran; the court found in favour of the contractor.

The problems and issues have given rise to the use of causal-mapping (to answer the
question why?) combined with System Dynamics (SD) (to answer the question how
much?) to answer questions such as the effect of a delay, or a set of delays, upon a project,
as described in papers such as Ackermann et al [36] and Williams [37]. Such a process starts
by using interviews and workshops to develop cause maps of the perceived causes of effects in
the project and their causal links (in these papers, specialist software (Decision Explorer) is
used to help develop, structure, share and analyse the extensive maps). Following the
development of the cause map model, analysis shows the nature of the feedback dynamics and
other endogenous effects, enabling an Influence Diagram and thus the skeleton of a System
Dynamics (SD) model to be developed. SD is a quantitative analysis technique that could be
argued to naturally follow on from the use of cause mapping and feedback; it has a track record
since 1964 for explaining and modelling the systemic effects in projects (see a discussion in
Rodrigues and Bowers [39]). Because of its explanatory power (from the endogenous view it
takes of a system), it has had particular application in the post-mortem analysis of projects; an

13

early success was the Ingalls Shipbuilding case [40], and the method has been used on a number
of such litigations [37]. The cause-map and SD model are developed in parallel during the
process of analysis [36]. Such a process can (as described in [38])

show causality: the systemic inter-relationships which caused the various ramifications
of the delays to build up

show responsibility: show the party causing the initial causes of the delays (as opposed
to the immediate causes, which might simply be reactions to previous effects)

calculate the quantum of the effects in other words, show the size of cost- and timeoverrun resulting from a particular delay or set of delays.

One of the key aims of developments in this area is to provide transparency in the claims
process. Zack [41] says that while clients and contractors used to do things in partnership,
they now play out claimsmanship. He lists 11 claims games commonly played by
contractors on public projects (note that some of these are not relevant to private projects),
some of which could be validated (or not) or quantified by the types of techniques described
in the previous paragraph. These include

loss of productivity claims; Zack questions the basis of numbers used to calculated
productivity losses claimed this forms an important part of SD models, which seek to
explicate the chains of causality and explicitly quantify the different elements of
productivity losses

cardinal changes; this is a US concept, in which a change to a public contract which

substantially changes the nature of the agreement is called a cardinal change and
should have a separate procurement activity: but contractors can say the accumulation
of many small changes eventually led to a cardinal change, effectively giving a costplus situation; again, in Eden et al [23] the cumulating effect of many small Change
Orders is singled out as one of the effects for the modelling of which SD models are
particularly appropriate.

project float claims; [9] includes the principle that float belongs to the contractor so
you can claim a compensable delay even when the project isnt delayed; Zack disgrees
with this. SD models allow a transparent analysis of the effect of using up float (which
of course is not a zero effect as assumed in simple network models).

Other games include bidding games, taking claims into account in the bid, reservation of
rights (when agreeing change orders), delayed early completion claims, accelerated delay
claims (merging the delay and acceleration claims), total cost claims, DBE claims, and
Hail Mary change orders (throwing everything into one change order at the end of the

14

contract. Zack also lists 11 claims games starting to be played by public owners
(generally new and untested), some of them simply opposites of the above, including
interestingly time-impact-analysis requirements (e.g. using fragnets) and concurrency of
delay (other games include claims audit, project partnering, escrow of bid documents,
waiver-of-claim language, ownership of float clauses, keeper-of-schedule specification,
liquidated-damages-plus-actual damages clauses, ADR.

However, SD is not a panacea to answer all problems. SD models will not normally be at
the operational level of a network, and practically will need to consolidate activities into a
smaller number of units. SD also assumes continuous flows of homogenous entities through
the models, and some approximation is often needed compared to Discrete Event
Simulation models both in modelling discrete units with a continuous flows (e.g. when a
short-length production run is involved) and in the assumption of homogeneity (e.g. when a
small number of different types of product is being manufactured in a single line).
However, it has been found to be a very powerful technique for capturing the counterintuitive temporal effects which are seen in projects and in particular which form an
essential part of most EOT claims where small causes compound to give significant
delays, or where management actions have reduced or even opposite effects to those
expected.

PROPOSED PROCEDURE AND CONCLUSION

So, given that conventional network-based techniques have some clear disadvantages, but
SD also has some clear disadvantages as well as clearly lacking the court-room credibility
of network-based techniques, what should we do?

The first step is clearly to find out what actually was the outturn in the project. For this,
Gantt charts or some sort of time-profile showing key dates (planned and actual) and the
dates of key disruptive events is clearly vital to illustrate the overall project timeperformance. This doesnt explain why the out-turn was as it was, or whose fault it was, but
it does give an overview of what happened and, surprisingly, management withing
projects managed in a distributed way with no clear project manager (e.g. under a coordination-matrix management approach) can have only a hazy view of how late key
interim dates were.

15

Then, if there are a limited number of delays or disruptive events (and, if change orders are
an issue, the number of change orders is not sufficient to consitute a cardinal change),
then the standard network-based methods described in this paper are applicable and can be
used.

Otherwise, if there are many effects on the project, or the project was so complex that the
out-turn cannot be intuitively expected from the effects known to have triggered the
behaviour (see Eden et al [23] and Williams [24]), it is necessary to seek to understand why
the out-turn occurred and to trace the causality from the triggering effects. For this, causal
maps (to understand the causality) and System Dynamics (to quantify these effects) are
necessary, as described in the section entitled problems with current CPM methods
above. Networks can give the answer what would have happened (had these effects
not compounded together; had there not been systemicity within the effects; in particular,
had management not acted in response to the triggering events and (for example)
accelerated the project). But even an analysis that says this project would have been 1 year
late due to the actions of the client but in fact was only 3 months late because the
contractor accelerated, while having some usefulness (it can be at least a defence against a
claim for Liquidated Damages if the contractor is completely blameless and caused no
parallel delays), in general does not help to credibly explain delays and replicate what
actually happened in the project.

In practice, a combination of methods for different parts of the project has been found to be
often most appropriate, as the effects on different phases of the project may be different. For
example, it may be that design and manufacturing were heavily accelerated in the face of a
high level of delay and disruption (which would imply a very high level of over-spend); in
this case, a cause-map/SD analysis would be needed to explicate the factors and effects and
replicate the out-turn, showing why this limited over-run (and high over-spend) occurred;
but it may be that the commissioning phase was delayed as discrete events affected discrete
activities (albeit the causes of those events might have arisen from complex interactions of
triggers which can be shown by the causal-map analysis), which can be modelled using
networks.

Further work in this area is looking at closer combination of these methods bringing the
benefits of systemic modelling to the operational models provided by networks - but this is

16

still under development. For now we must use judgement to apply the appropriate method
to the appropriate project phase to provide effective understanding of the effect of delays
and their claimability; other further work will formalise and systemise such judgements.

17

References
1
2

3
4

5
6
7

8
9
10
11

12
13
14

15
16
17
18
19
20
21
22

Clarke, L (1994) The essence of change Prentice-Hall, Herts, UK

Cushman, KM, Hoolyday, JD, Coppi, DF and Fertitta,TD (1996) Delay claims. In,
Cushman, RF, Jacobsen, CM and Trimble, PJ (Eds) Proving and pricing construction
claims. Aspen Law and Business, Frederick, MD
Jergeas, GF and Hartman, FT (1994) Contractors construction-claims avoidance
Journal of Construction Engineering and Management 120 (3) 553-560
Semple, C, Hartman, FT and Jergeas, G (1994) Construction claims and disputes:
causes and cost/time overruns. Journal of Construction Engineering and Management
120 (4) 785-7947
Scott, S (1993), Dealing with delay claims: a survey. International Journal of Project
Management, 11, 143-154
Arditi, D and Patel, BK (1989) Impact analysis of owner-directed acceleration
Journal of Construction Engineering and Management 115, 144-157
Alkass, S, Mazerolle, M, Tribaldos, E and Harris F (1995) Computer aided
construction delay analysis and claims preparation. Construction Management and
Economics 13, 335-352
Wickwire, JM and Smith, RF (1974) The use of critical path method techniques in
contract claims. Public Contract Law Journal, USA, 7 (1) 1-45
Rubin, RA et al (1983) Construction claims: analysis, presentation, defense. Van
Nostrand, New York
Reams, JR (1990) Substantiation and use of the planned schedule in a delay analysis.\
Cost Engineering 32 (2) 12-16
King, RA and Brooks, PL (1996) Types of claim In, Cushman, RF, Jacobsen, CM and
Trimble, PJ (Eds) Proving and pricing construction claims. Aspen Law and Business,
Frederick, MD
McCullough, RB (1989) CPM Schedules in construction claims. Cost Engineering 31
(5) 18-21
Kraiem, ZM, and Diekmann, JE (1987) Concurrent delays in Construction projects.
Journal of Construction Engineering and Management 113, 4
Yogeswaran, K, Kumaraswamy, MM and Miller, DRA (1998) Claims for extensions
of time in civil engineering projects. Construction Management and Economics 16,
283-293
Hohns, HM et al (1981) The law behind construction schedules, deskbook of
construction contract law. Prentice-Hall, Englewood Cliffs, NJ
Institution of Civil Engineers (1974) Conditions of Contract 5th Ed. Stanhope Press,
UK
Bordoli, DW and Baldwin, AN (1998) A methodology for assessing construction
project delays. Construction Management and Economics 16, 327-337
Galloway, PD and Nielsen, KR (1990) Concurrent schedule delay in international
contracts. International Construction Law Review 7th October 1990
Bartholomew, SH (1988) Discussion of Kraiem and Diekmann (1987), Journal of
Construction Engineering and Management 114, 333-335
Logcher, RD (1988) Discussion of Kraiem and Diekmann (1987), Journal of
Construction Engineering and Management 114, 335-337
Kraiem, ZM, and Diekmann, JE (1988) Discussion of Kraiem and Diekmann (1987),
Journal of Construction Engineering and Management 114, 337-338
Alkass, S, Mazerolle, M and Harris F (1996) Construction delay analysis techniques.
Construction Management and Economics 14, 375-394

18

23

24
25
26

27
28

29

30
31

32
33
34

35
36
37
38

39
40
41

Eden, CE, Williams, TM, Ackermann, FA and Howick, S (2000) On the Nature of
Disruption and Delay (D&D) in major projects. Journal of the Operational Research
Society 51 (3) 291 300
Williams, TM (1999) The need for new paradigms for complex projects.
International Journal of Project Management 17 (5) 269-273
Baccarini, D (1996) The concept of project complexity - a review. International
Journal of Project Management 14, 201-204
Turner, JR and Cochrane, RA (1993) Goals-and-methods matrix: coping with projects
with ill defined goals and/or methods of achieving them. International Journal of
Project Management 11, 93-102
Williams, T M (1992), Criticality in probabilistic network analysis. Journal of the
Operational Research Society, 43, 353-357
Bowers, JA (1995) Criticality in resource constrained networks. Journal of the
Operational Research Society, 46, 80-91
Williams, TM (1997) (Editor) Managing and Modelling Complex Projects NATO
ASI Series, Kluwer Academic Publishers, Dordrecht, Netherlands ISBN 0-79234844-3
Williams, T M (2000), Systemic project risk management - the way ahead.
International Journal of Risk Assessment and Management 1 (1) 149-159
Jorgensen, T and Wallace, SW (2000) Improving project cost estimation by taking
into account managerial flexibility. European Journal of Operational Research 127,
239-251
Williams, TM (1999) Towards realism in network simulation. Omega 27 (3) 305-314
Cooper, KG (1994), The $2,000 hour: how managers influence project performance
through the rework cycle. Project Management Journal, 25, 11-24
Howick, S and Eden, C (2001) Using System Dynamics to Explore the Role of
Disruption and Delay in the Cost of Compressing Large Projects. Journal of the
Operational Research Society 51 (1) 26-34
Bowers, JA (2000) Multiple schedules and measures of resource constrained float.
Journal of the Operational Research Society 51 (7) 855 - 862
Ackermann, F, Eden, C and Williams, T (1997) Modelling for litigation: mixing
qualitative and quantitative approaches. Interfaces 27 (2) 48-65
Williams, TM (2000) The risk of safety regulation changes in transport development
projects. International Journal of Project Management 18, 23-31
Williams, TM, Ackermann, FR and Eden CL (2000) Structuring a Disruption and
Delay Claim. Working Paper 2000/1 University of Strathclyde Dept Management
Science
Rodrigues, AGR and Bowers, JA (1996) The role of system dynamics in project
management. International Journal of Project Management 14, 213 - 220
Cooper, KG (1980) Naval ship production: a claim settled and a framework built.
Interfaces, 10 (6) 20-36
Zack, JG (1993) Claimsmanship: current perspective Journal of Construction
Engineering and Management 119 (3) 480-49

19
View publication stats