In this chapter, we will explore why the various primary contracting strategies perform in characteristically different ways. We will ask if there are ways to mitigate some of the problems we see in certain important contracting approaches, EPC‐LS and EPCM in particular. Given the popularity of EPCM around the world, any approaches to improving the outcomes of that particular contract form would be very useful. We will also discuss one of the great paradoxes in contracting strategy: why the best contracting strategies (for owners at least) are less popular than strategies that are more damaging to owner interests in cost and schedule effective projects.

Traditional EPC‐LS (Design‐Build Fixed Price)

EPC‐LS is the contracting strategy that minimizes the joint product nature of capital project delivery. The goal is to divide the project into three clearly delimited phases: owner‐led project preparation, contractor execution, and owner operation. Sometimes, contractor execution extends to starting up the facilities, in which case the contract form is “turnkey.” (We will address whether that is a good idea later.) The clear demarcation of responsibilities is one of the key features of EPC‐LS. It simplifies the management of the approach while creating important limitations as well.

The EPC‐LS form is used in some regions as the primary contracting strategy because government rules make other forms impossible or impractical. Government‐owned industrial companies (so called national companies) are most likely to follow EPC‐LS contracting because it is required or preferred by their government owner. It is the dominant form in the Gulf Cooperation Council countries in the Middle East for national companies and is common for national companies elsewhere in the world as well. Government preference for EPC‐LS is usually ascribed to the desire to make corruption in capital projects more difficult, although I am unaware of any evidence that it does so.¹ Sometimes governments make EPC‐LS the only feasible contractual form by requiring government review of contracting decisions down to a very low level—in one case every spending decision over $10,000 USD has to be reviewed. That forces a minimum of contracting decisions and, therefore, EPC‐LS.

When projects are externally funded by financial institutions, the lending syndicate often mandates EPC‐LS. Many lenders go further and prefer a “wrap contract” with a single EPC‐LS turnkey contractor with operational performance guarantees and liquidated damages for late completion. As it is indelicately put, “It is good to have one throat to choke!” In reality, EPC‐LS offers little protection against overruns. The variation in actual‐to‐estimated cost is almost exactly the same for EPC‐LS and other contract strategies. The belief that the contract price is a ceiling on what the project will cost is a form of magical thinking or perhaps just confused vertical orientation. The EPC‐LS contract price is the floor on what the project will cost but certainly not the ceiling! The efficacy of liquidated damages and other penalties will be discussed in Chapter 8. Of course, a lending syndicate's desire to have a single contractor execute a project is not feasible if the project is too large for a single contractor to handle. In fact, about 40 percent of EPC‐LS projects overrun the contractor's bid, and when they do, the average overrun is 17 percent of total project cost with the largest overrun being 150 percent.

When EPC‐LS projects overrun, changes are the culprit (Pr.|t|<.001). The changes are in turn driven by the defects in the front‐end loading (Pr.|t|<.001), especially the FEED quality and the completeness of the project execution planning. In traditional EPC‐LS projects, the first task of the winning contractor is to thoroughly review every aspect of the FEED package. Usually, the contract will require the new contractor to identify and request change orders for all of the deficiencies in the FEED package within 60 to 90 days of contract award. After that point, FEED deficiencies will not generally be accepted as legitimate change order items.² Of course, what is a “deficiency” is to some extent in the eye of the beholder. The two‐ to three‐month review period is an inevitable feature of the traditional EPC‐LS structure. It is thoroughly disliked by most owners as it is often quite contentious, delays the full start of execution, and pushes up the price if FEED errors and omissions are uncovered.

Many of the alternative EPC‐LS forms, such as design competitions, functional specification contracts, and covert‐to‐lump‐sum contracts, are attempts to avoid the “hard hand‐off” between the FEED contractor and the execution contractor while still maintaining the EPC‐LS format, at least in part. Occasionally, the owner seeks to avoid the hand‐over problem by allowing the FEED contractor to bid and surreptitiously giving that contractor an advantage in the bidding competition. This practice is quite unethical and often deters the best EPCs from entering the competition as they fear they are wasting their time and money.³

When projects are well defined, EPC‐LS does provide more protection against overruns than any other contractual form, but with some important caveats. If the project was too large for the contractor's balance sheet to absorb anything but a smallish overrun, the EPC‐LS form provides little protection to the owner. EPC‐LS projects are a major source of contractor bankruptcies, and in my experience that never works out well for the owner. Also, when a lump‐sum contractor is struggling with the job, quality will suffer unless the owner has a very strong inspection organization.

Factors Militating Against Use of EPC‐LS

A number of factors make the use of EPC‐LS difficult or even impossible. The most important limitation by far is the state of the contractor market when a project is coming up for a contracting strategy decision. Busy sellers' markets make EPC‐LS either much more expensive or simply impracticable. When the market is busy, a contractor's need to win any particular bid declines. That enables contractors to more fully price any risk they see in the work.

The problem for owners is that contractors will price risk at a higher level than most owners would. This conclusion simply accords with the first principle of risk pricing: the pricing of the risk is a function of one's wealth position relative to the downside of the bet. Richer people can afford to be less risk‐averse than poorer people. In the industrial world, owners are almost always wealthier than contractors. The financial implications of absorbing a large overrun on a project are profoundly different for a contractor and an owner. On a lump‐sum contract, every dollar of overrun that cannot be avoided is deducted from the contractor's balance sheet. By contrast, every dollar of overrun on a reimbursable project is added to the owner's balance sheet as an asset. The owner is not happy about having to earn against an asset that is larger than expected, but even a large overrun is very unlikely to do serious financial damage to an owner company. A large overrun on even a medium‐sized project could be debilitating for a contractor.

The other factor that becomes important in hot markets is that one may not be able to generate a bid list of three to five genuinely interested contractors. Without at least three bidders, the competition is likely to end up with a high‐cost project, and three bidders is far from optimal.⁴

The present market is instructive about the differences between owners and contractors. The current market is not overheated in most areas of the world. Even so, contractors are highly risk‐averse. Many major EPC firms are declining to entertain any industrial EPC‐LS work because their financial condition is weak and they feel insecure with fixed‐price work. Many contractors are declining EPC‐LS for infrastructure work as well.

Too Much Time‐Pressure. A second consideration that must be weighed carefully is whether the time required for the EPC‐LS cycle time is going to be palatable to the business sponsors of the project. As shown in Figure 4.1 EPC‐LS projects average more than 130 days from issuance of the invitations to bid (ITB) to bid receipt.⁵ Larger (more than $500 million) greenfield projects require 215 days on average with a P25 to P75 range from 135 to 267 days. Complex process technology adds to the cycle as well. The bidding cycle is only part of the process. After a winner is selected, the contract terms must be negotiated, and the FEED review process delays true project start even more. If the business case cannot stand the added time, EPC‐LS may be a nonstarter.

Note that the bidding cycles for construction only competitions are much shorter and that there is no difference between lump‐sum construction and unit rate contracts. This becomes important as we discuss other contracting forms.

Too Much Residual Uncertainty After FEL. Another significant issue in the EPC‐LS decision is whether uncertainties in the particular project can be reduced very substantially before we need to issue the ITB. The timing of uncertainty reduction should look like the pattern shown in Project A in Figure 4.2. The owner's goal should always be to present a low‐risk project to the contractor community—a project that can be executed quickly and without significant upsets. Smart contractors do not like to enter EPC‐LS bidding competitions for messy projects.

If uncertainty is high at the end of FEL 3 (FEED and execution planning), such as in projects B and C, contractors will not bid, or will bid high, or the resulting contract will not be sustainable as an EPC‐LS.

There are a number of factors that can drive high levels of uncertainty into execution.

• Project unfriendly locations are places in which regulatory approvals are problematic and government interventions in projects during execution are likely and politics are unstable.
• Remote sites and other sites with inherently difficult logistics are probably poor candidates for EPC‐LS.
• Projects using substantially new technology in which significant design changes are likely through much of detailed engineering are not conducive to EPC‐LS.⁶
• Projects with significant shaping problems such as unhappy partners and other stakeholders, where disruption of execution is likely, should not be attempted EPC‐LS.⁷
• Finally, projects with poor quality or very incomplete FEED are going to be a disaster in the unlikely case that an EPC‐LS contract can even be had.

Lack of the Proper Skills. EPC‐LS requires the skill to select and manage FEED contractors. That requires a substantial amount of engineering skill. The ability to write a solid ITB and solicit a good number of bids is an under‐appreciated skill set. Finally, as mentioned, a strong inspector cadre is essential to mitigating the quality concerns that always accompany EPC‐LS contracts.

Insufficient Internal Discipline. Some companies should not attempt EPC‐LS simply because they lack the internal disciple essential to keeping EPC‐LS projects under control. By this I do not mean the controls function, but the willingness of all owner functions to get their project scope items addressed in FEL‐2 and not revisit the scope throughout the project execution process. As the old expression goes, the EPC lump‐sum with lots of owner changes is the contract from hell for the owner and the contract from heaven for the contractor.

The contract form depends on allowing the contractor to execute the project without repeated owner intervention. Owners who cannot stomach that should find another contracting strategy. Often changes and interventions lead to conflict with the contractor as well as unending cost growth and schedule slip. We even had one project in which the very reputable Japanese contractor quit the project in the middle of execution, arguing that the endlessly dabbling owner was “not an honorable owner!”

The Operability Penalty. The EPC‐LS form carries a small but consistent operability penalty of about 4 percent of nameplate capacity when compared to equivalent projects executed on different contract strategies. This penalty is based on the second six months production after startup, after startup upsets have been resolved. The hit to operability is not surprising as it is known that cutting corners is an ever‐present danger with the EPC‐LS forms. This result underscores the importance of quality inspectors on the owner team during execution.

Understanding EPC‐Reimbursable

Reimbursable EPC is a relatively rare contracting form these days for major projects. It is sometimes an owner's contracting strategy of last resort. When no contractor will bid lump‐sum on a project or when EPC‐LS bids come in unacceptably high, EPC‐R is sometimes the fallback strategy. It is often used when high levels of uncertainty persist in a project into execution and an EPCM approach is not feasible. During the hot market periods in the U.S. Gulf Coast and Western Canada regions, EPC‐R was used when other forms, including lump‐sum construction, were simply not available in the market. EPC‐R is used more often when new technology is in play—about 10 percent of EPC‐R projects had some technology that was new in commercial use versus about 6 percent of other projects.⁸

If EPC‐LS typically has the highest of risk transfer to contractors, EPC‐R surely has the lowest. EPC‐R places more burden on owners to control the project than any other form. Most owners fear the possibility of the contractor “cranking hours” endlessly on EPC‐R projects. Unlike other contractual forms, there is seemingly no built‐in stop to the contractor spending more hours due to low productivity in both engineering and the field. The other EPC reimbursable form, EPCM, at least has the change of contractors for construction even if not for any other activity.

Is the owner fear of “cranking hours” a justified fear? We find it is only to a very limited degree. Most of the time when hours are seemingly being “cranked,” it is because the front end was not completed. EPC‐R projects have more increase in material quantities than any other contract form from authorization to completion. It is true that the EPC‐R contractor almost always executed FEED as well as execution. But one of the major downsides of not having a break between contractors at the end of FEED is that owners are tempted (often with contractor encouragement) to “slide” into execution, blurring the line between FEED activities and execution activities. That tendency is no doubt exacerbated by EPC‐R projects being schedule‐driven more often than any other contractual form except guaranteed maximum price.⁹

In recent years, even EPCs appear to have shied away from EPC‐R in preference of EPCM. EPC‐R usually involves direct‐hire construction by the EPC. EPC contractors have experienced a good deal of difficulty with direct‐hire projects because the deep level of construction management that is required is now beyond what many can comfortably deliver. Depending on the terms of the contract, EPC‐R often involves significantly more liability for the contractor than an EPCM assignment does. This is especially true if the EPCM has the constructors prime to the owner rather than subcontractors of the EPCM. The EPC‐R contractor may be liable for operational performance problems as well, whereas the EPCM contractor rarely is.

A Deeper Dive into EPCM

As industrial landscape changes go, the development of EPCM as the dominant contracting form has been quite dramatic. That begs the question of why this form has become so popular so quickly. A number of forces are at work.

First, a reminder: the form is actually FEED‐EPCM where FEED stands for “front‐end engineering design.” FEED is the engineering side of the final stage of the front‐end loading process. When FEED is complete, the project moves to full‐funds authorization, aka final investment decision (FID). For a typical project in the process industries, FEED is relatively expensive, typically accounting for 2 to 4 percent of total capital cost. When the EPCM form is selected, the FEED contractor is the EPCM contractor almost all the time (>95 percent). When the FEED contractor is not selected for the EPCM work, it is usually because the FEED contractor and the owner are not getting along.

The cases in which the owner is so unhappy with the FEED contractor as to end the relationship at FEED tend to be highly problematic projects. The core problem is that as soon as the FEED contractor knows it is not going forward into execution, the already problematic FEED process slows down, and the contractor often withdraws his best people. Often the actual contract is of almost no help at all here because the contract did not lay out with enough specificity what all the FEED deliverables were required to be and did not include any consequences for failing to get the work done in a timely way with quality. Every FEED contract should be written so that it is enforceable as a stand‐alone contract. “Off‐ramps” are difficult to write well but are important for a number of contract types.

Because it is an FEED to EPC form, EPCM avoids one of the most difficult transitions for capital projects: changing contractors at the FEED‐to‐execution point. Although FEED is considered part of the “front end,” it is actually more akin to execution engineering than it is to planning. FEED entails completing all of the engineering drawings that will be picked up and detailed in the next phase. The plot plans, the equipment layouts, the piping and instrumentation diagrams, line sizes, the electrical (single) lines, and the heat and material balances are all part of this work. The break between FEED and execution is largely artificial when considered only from the engineering perspective. When contractors are changed at the FEED‐to‐execution gate, for whatever reason, the new contractor needs to thoroughly review the work of the FEED contractor and almost always finds it deficient in a large number of respects. This finding seems to be completely independent of the actual quality of that FEED package! The result is lots of changes, sometimes important and sometimes cosmetic, and a slowing of the project. This transition costs the owner money and time and is thoroughly disliked by most project teams.

If the owner has not planned to change contractors after FEED, the change is very disruptive because new contractors have to be prequalified, and a tendering process has to be initiated.¹⁰ Alternatively, a contractor is selected as a sole source, but that is not a good solution either.¹¹ So, the most important advantage of F‐EPCM is that it avoids the FEED‐to‐execution transition.

There are reasons other than the avoidance of transition for owners to like EPCM. As an EPC contractual form, EPCM is a “one‐stop shopping” exercise. Only one contracting decision has to be made; all the others are usually the responsibility of the EPCM, including the selection of the construction contractors. In practice, this means that owners feel like they require little or no construction expertise. Decisions such as when to start work in the field, selection of subcontractors, interface management, controls, and so forth are all considered responsibilities of the EPCM, with the owner providing guidance and consultation but not active management. The owner avoids all of the inspection challenges associated with EPC‐LS. EPCM reflects the reality of today's industrial company projects organization: profound weakness in knowledge of construction and field management practices. The out‐sourcing of construction management that started in Western countries in the late 1980s made contractual approaches such as EPCM inevitable in the 21st century.

The other positive of EPCM is that contractors really like the form, even more than EPC‐reimbursable. Contractors generally tout the form to owners as the best contracting strategy for any and all projects. Contractors' preference for EPCM is understandable: it provides them work with essentially no risk but with a number of avenues for profit, which we will discuss a bit later. The EPCM form also relieves the contractor of having to do direct‐hire of construction labor¹² and the full and complete construction management that goes with it. EPCM has significant pluses for both owners and EPCM contractors. But when owners are saying “EPCM is the best way to go!” and contractors are saying “EPCM is the best way to go!” somebody ought to be asking “What's wrong with this picture?” So in keeping with the mantra of “there's no free lunch,” what are the downsides of EPCM?

The Downsides of EPCM

As discussed earlier, EPCM is an expensive contractual approach for owners. At any project size, EPCM is less cost‐competitive than EPC lump‐sum by 5 to 10 percent, depending on project size.¹³ And this is despite that, unlike EPC‐LS, EPCM is transferring little or no risk to the contractor! The EPCM form is much less cost effective than the split forms—more than 13 percent on average. The EPCM form has more cost growth than EPC‐LS (Pr.|t|<.006), and there is no difference in schedule slip. Therefore, the desire for predictability is not a reason for owners to prefer EPCM over EPC‐LS. Cycle time¹⁴ is about 9 percent shorter for EPCM than EPC‐LS (Pr.|t|<.01) due to the long bid cycle associated with traditional EPC‐LS between the completion of FEED and start of execution. There is no difference in average execution time.¹⁵

The biggest problem with EPCM from an owner's perspective is that cost performance degrades very rapidly with project size. Figure 4.3 provides the relationship between cost effectiveness and project cost for three types of contract approaches: EPCM, EPC‐LS, and the split forms. Both EPC‐LS and EPCM degrade with project size, but EPCM degrades at a much faster rate. The split forms, however, show no relationship between cost effectiveness and size at all; they maintain their superior cost effectiveness as well for megaprojects as for smaller projects. Ironically, EPCM becomes more common as projects get larger, while split forms become less common.

Why EPCM Contracting Is Problematic

When I say that EPCM contracting is problematic, it is only problematic for owners. It has probably been instrumental in keeping EPC contractors afloat financially over the past decade. But why is EPCM so problematic for owners, and what, if anything, can be done about it?

Because FEED is almost always included in the remit of the EPCM, the project's cost estimate and schedule are developed by the EPCM. It is the cost estimate that is key. When contractors who are going to execute develop the cost estimate, they are strongly incentivized to estimate high rather than 50/50. This is not a matter of evil intent; this is just a matter of common sense. As a contractor, would I rather underrun or overrun? Would I rather have more money in the engineering and construction management and construction accounts or less? Figure 4.4 provides the obvious answer to this obvious question.

When contractors make the estimates for projects in which they are going to execute engineering and construction, the estimate is 13 percent higher than the expected value of the estimate or engineering and construction management, and the estimates contain 8 percent more money for the total project than cases in which the contractor is going to execute engineering only.

Why doesn't owner estimate validation remove the excess money in the contractor's estimate? When owners went through a validation process on EPC‐R and EPCM contractor estimates, they removed little or none or the overestimated amounts. When the FEED contractor was going to execute engineering only, owner validation resulted in systematic improvement of the estimate (Pr.|t|<.02). What this tells us is pretty simple: when contractors have a very strong incentive to fatten up cost estimates, they will likely to do so and will do so in ways that are difficult for owner validation to find. This phenomenon is merely the principal‐agent problem at work. It is also worth mentioning that the desire to plump up the estimates will occur regardless of whether additional incentive payments are employed. It is just a fact of the situation, and it erodes the cost effectiveness of EPC‐R and EPCM strategies. The owner weakness that makes this process possible is the inability of most owner estimating functions to generate a bottom‐up detailed material takeoff‐based estimate. Because they do not know how to make such an estimate, they do not know how to find fat and errors in such estimates whether they are “open book” or not.

Owners' interests and contractors' interests do not, cannot, and never will fully align. As discussed in Chapter 3, even owner functions are often not fully aligned. There is another fact that really seals this issue. When owner personnel do EPCM, the projects results do not look like EPCM projects. Instead, the results look just like Re/Re or Re/LS projects, depending on how construction is contracted. A few owners maintain detailed engineering capability in‐house, and those owners usually execute F‐EPCM on their projects. They do not overestimate the projects. The projects are much more cost effective than EPCM, but the work structure is otherwise identical to an EPCM. The only thing missing is the principal‐agent problem.

Another disturbing aspect of EPCM from an owner's viewpoint is that the form becomes progressively more expensive with project size (Pr.|t|.004). This is a tendency more pronounced in EPCM than any other contract form.¹⁶ Part, but only part, of the explanation is that the estimates are fatter with size. Given the incentive structures at work and the decreasing efficacy of validation as a function of size and complexity (there are more places to hide money subtly), this result is not surprising.

The big difficulty with EPCM is not the F‐EP portion of the work; that is clear enough. However, the CM part frequently is not. Construction management is a general term that applies to a broad array of field activities, ranging from very high‐level progress reporting to very detailed hands‐on management responsibilities. The following is a short list of activities and functions that fall under the rubric of CM:

• Planning and sequencing of construction
• Ensuring timely completion systems for turnovers to ops
• Interface management (mostly between construction contractors on‐site)
• Work package planning: primary and backup tasks planning and tools
• Quantity surveying
• Progress measurement and reporting
• Change management
• Review and approval of submittals to the engineer
• Taking of questions for design and tracking responses
• Safety program—planning and execution
• QC
• Inspection and testing, such as weld tests, hydro testing, etc.
• Construction equipment leasing, maintenance shops, sometimes provision of certain equipment operators as well
• First‐line gang supervision
• General foreperson
• Materials management
• Receipt of equipment and materials inspection
• Yard management
• Lift planning
• Specialty contractor management
  • Scaffold
  • Paint and insulation
  • Dining areas
  • Labor camp

This list is not exhaustive, but the point should be clear. CM can range from just a few of these at a high level to every one of these activities. How much CM is required depends very much on the strength of the general construction contractor or the disciplinary contractors who are hired to construct the facilities. When the EPCM contract is signed, generally prior to the start of FEED, none of the actual constructors have usually been identified and qualified. If they have been identified and qualified, they may turn out to be unavailable or uninterested in bidding when the time comes later in the project. The contract may list and the contractor may promise every CM service imaginable. But that doesn't govern the contractor's assumptions about what will really have to be staffed. If the EPCM underestimates the range and depth of CM that will be required, the contractor will need very good bench strength to be able to meet the unexpected challenge. This is where the type of economic entity contractors must be remembered: carrying a strong bench in a professional services organization is a recipe for bankruptcy.

If the EPCM that was selected at the beginning of FEED assumed that the level of CM required would be relatively shallow, the contractor may be woefully unprepared if the construction contractors selected are not able to carry most of their own CM. This sort of mismatch of expectations about the depth of CM that will be required is the single most intractable problem that we see in this contracting form. As projects get larger and more complex, the probability of a mismatch of expectations and reality seems to become almost a certainty. As the number of construction contractors increases on the site, the interface management challenges increase exponentially, and it always seems that at least one of a large number of constructors will turn out to be less competent than expected. The EPCM firm is often not equipped to rescue the situation.

Making EPCM Work Better

After reviewing the detailed histories of more than 300 EPCM projects, I see nine steps that owners can take that will systematically improve the outcomes of these projects. These steps will not eliminate the problems but at least will mitigate them.

1. As we will discuss in more detail in Chapter 7, it is important to prequalify potential F‐EPCM contractors. EPCM contractors are selected without prequalification more than in any other contractual form. Be clear that this is not a recommendation to have a formal competitive bidding process for F‐EPCM work; it is rather to fully review the field before you make a selection of the F‐EPCM contractor. Owners learn important things during the prequalification process that may cause them to look beyond the contractor they were intending to hire sole‐source.
2. Finish FEED completely before mobilizing design. That means completing all of the P&IDs, not just the inside battery limits portion. It means all the line sizes have been checked, the heat and material balances are fully closed, and all utility requirements are fully known and specified. The failure to complete FEED, which occurs in nearly 40 percent of EPCM projects, adds 10 percent to total installed cost (Pr.|t|<.0001) and is completely unnecessary. Because the FEED contractor will carry on into execution, it is too easy not to complete the FEED; it seems unnecessary. But without completed FEED, there is no quantities baseline for the project, maintaining control of engineering hours becomes more difficult, and work will be done out of order and therefore done more than once.
3. Validate the estimate made by your FEED‐EPCM thoroughly. Challenge the bulk material quantities—it is the most common and insidious place to hide money. Benchmark the engineering hours by discipline. Carefully look at all the key ratios by material: bulks‐to‐equipment, engineering hours to bulks, engineering hours per piece of major equipment, labor hours to bulks, CM hours and cost to field hours, etc. If your EPCM balks at giving you complete access to the estimating information, you have selected a predatory EPCM.¹⁷
4. Work with the contractor at the start of FEED (or even before) to assess the construction market and the capabilities of the general contractors (GCs) and disciplinary construction contractors to understand what depth of construction management will be required for the project. Make the decision about whether a GC will be needed early and start the prequalification process for GCs. If deep CM will be needed by your EPCM contractor, start discussing the staffing requirements during FEED, not later.
5. Again work with your assessment of the construction contractor market, and decide what construction contracting strategy will be used if a GC is not going to be employed. For EPCMs, unit rate contracts are the most cost effective—about 8 percent more cost effective on a whole project basis (Pr.|t|<.04). Lump‐sums are directionally cheaper as well. Straight reimbursable and T&M contracts are the most expensive. Recall the principle that says that risks should be assigned to the party most capable of managing those risks. The construction firms doing the work are more capable of handling field productivity risks than any owner as long as the engineering and materials are available.
6. Do not attempt to do EPCM by directly contracting disciplinary subs if the disciplinary construction contractor market is so thin that many subs will be required to get the work done on any one EPCM contract. The EPCM will be overwhelmed with interface management problems. When you have a large number of contractors working side by side doing construction, it is almost certain something will go badly wrong. Go with a GC instead, preferably with a unit rate or lump‐sum contract where available. Then allow the GC to manage and coordinate any disciplinary constructors as needed.
7. As discussed earlier in this chapter, when smart owners use EPC‐LS strategies, they do very detailed and complete reviews of risk and risk assignments with the EPC during the negotiation process. Unfortunately, we rarely go through the same exercise with EPCM contractors. Part of the reason we do not do so is because the EPCM contract is usually negotiated before the start of FEED when project execution risks seem far in the future. Go through the same risk assignment exercise for EPCMs. Owners should not automatically be agreeing to a zero‐risk‐to‐the‐contractor EPCM form. At a minimum there should be target hours for engineering and CM (separately) that result in a loss of overhead contributions if those hours are exceeded without owner changes.
8. If the owner chooses to be schedule‐driven when using an EPCM form, be clear that there is a very sizeable penalty for being so—an average of 8.5 percent of total installed cost. The penalty for being schedule‐driven is higher for EPCM than any other form except EPC‐R.
9. Finally, owners have to exercise more discipline about allowing work to go to the field or fabrication yard too early. My advice would be never go to the field until after the final HAZOP is complete and changes incorporated and second model review is completed and changes incorporated. Then do a full review of procurement status. Understand exactly what is going on in the shops fabricating steel and pipe for the project. Then and only then, if everything looks sound, trigger field start. Owner project teams need to explain to the business sponsor that if the contractor attempts to go around the team to the business sponsor to suggest getting the field start sanctioned, they are doing so for their benefit, not the owner's.

EPCM is the single most popular industrial project contracting strategy in the world. Despite its many drawbacks, that is probably not going to change anytime soon. Therefore, it behooves all of us to do the things necessary to make EPCM as effective a strategy as possible.

Understanding the Split Strategies

Just as a reminder, there are four split strategies: Re/LS, Re/Re, LS/LS, and LS/Re. The last of the set is not used very often (about 2 percent of projects) and sometimes came about when the LS/LS strategy unraveled because construction contractors would not work lump‐sum. However, the LS/Re was nonetheless more cost effective (97 percent of industry average) than any of the EPC strategies. Because LS/Re is employed so rarely, we will not bother discussing it as a separate strategy.

The common feature of all split form strategies is the contractual break between engineering and procurement on the one hand and construction on the other. That break has a profound effect on the cost effectiveness of projects. For industrial companies making commodity products, the cost advantage of the split forms can over time provide a substantial boost to their return on capital employed (ROCE), which is one of the core financial measures of company success. Furthermore, as we saw in Chapter 3, the split forms can deliver on schedule‐driven projects as fast as or faster than the EPC strategies. In this section, we will explore the advantages, disadvantages, and challenges of using these generally superior contract strategies.

Differences Among Re/Re, Re/LS, and LS/LS

There are some important differences in how the various split forms work. Re/Re is the most robust of the three forms. It provides very good cost and schedule performance, very little cost growth (4 percent), and less schedule slip than average (15 percent, which is a lot but still less than average). Re/Re is especially cost effective when unit rate construction contracts are used but are cost effective on average in any case. The overall performance is dependent only on completing front‐end loading with good quality. When speed in execution is important, Re/Re becomes the fastest of all contracting strategies (82 percent of industry average) with no degradation of cost performance. By contrast, the Re/LS and LS/LS forms are more sensitive to site knowledge and to team continuity than Re/Re. Because it does not use lump‐sum construction, there are fewer “gotcha” items.

Under schedule pressure, the Re/LS responds just as well as the Re/Re, but LS/LS does not. When schedule‐driven, the LS/LS form gathers very little speed, and the cost performance degrades from .92 in the non‐schedule driven case to .98. Re/Re and Re/LS show no cost degradation at all. The culprit is the poor flexibility under schedule‐driven conditions of the lump‐sum engineering. In case after case, we saw that the engineering firm was unwilling to accelerate because they were working lump‐sum and were seeking to maximize profits. Lump‐sum engineering tends to experience more slip than reimbursable engineering and not respond well to any attempt to increase speed. This should not be a surprise. Engineering tends to slip because the engineer does not have readily available the engineering disciplines needed in the right order. On a reimbursable, he can hire from the street to fill gaps. On lump‐sum engineering, that is a recipe to lose money unless the fee formula is altered, which we discuss in Chapter 8.

Why So Much Reimbursable Construction?

Outside EPC‐LS contracts, only about 40 percent of projects in the process industries use lump‐sum contracts for construction. Straight reimbursable (fixed‐fee) and T&M together are more common despite generating inferior results. The amount of reimbursable and T&M construction goes up as projects get larger. Fewer than a third of projects more than $200 million have predominantly lump‐sum construction.

Some owners tell us that the construction companies resist lump‐sum construction, and some tell us they resist unit rates as well. But that is not really my experience in talking to construction firms. A friend of mine was CEO of a first‐rate construction company. He was adamant that lump‐sum construction was better for his firm because it kept them sharp and focused. “Reimbursable makes you sloppy.” I suspect what is actually going on is the construction firms resist lump‐sum with owners who don't understand construction management and with engineers that will not sequence their work to ensure that design and materials are available when needed.

When constructors work on a reimbursable or T&M basis rather than lump‐sum or unit rates, the owner is taking responsibility for risk associated with craft labor productivity. That violates a basic principle of contracting: when a risk is assigned to a party who does not control the risk, the risk goes unmanaged. Owners do not have the skills needed to control labor productivity. Therefore, they need to ensure the project is set up properly in front‐end loading and that the detailed engineering will support transfer of that risk to those who do.

Summary

The contracting strategies that we have reviewed in Chapters 3 and 4 are the standard approaches for industrial projects. Each strategy is appropriate in some circumstances, depending on the project, the strength of the owner, and the contractor market. Most industrial owners are able to use only one or two of the standard strategies effectively, and most often those are the EPC strategies. The reluctance of owners to employ the split strategies costs industrial owners massive amounts of money. Developing the capability needed to employ split strategies on a wide variety of projects would be a very profitable investment.

Many owners, however, are not satisfied that the standard approaches produce the best possible outcomes and look to other strategies to deliver better results. We discuss a number of those less common strategies in the next two chapters.