Utilization of Learning Curves in Damage for Delay Claims
University of Colorado
Court actions and claims for increased costs due to delays in construction are extremely common , . From a contractor’s viewpoint, construction delays can be categorized into three partitions:
1) delays caused by acts of the owner or his agents (compensible)
2) delays caused by the contractor or his agents (non-compensible)
3) delays chargeable to neither the owner nor the contractor (excusable)
This paper will concentrate on the first class of delays with particular emphasis on proving damages for delay to repetitive construction operations.
Delay claims brought by a contractor against an owner are one of the most difficult of all areas of contracts administration for several reasons. First, delay claims are tied closely to the specific contractual language (no-pay-for delay clauses, time is of the essence clause, etc.). Second, delay claims are often presented with other contemporaneous legal issues such as acceleration, maladministration, defective design, changes to the work, etc. A contractor seeking to recover delay damages must establish contractual justification for his claim, typically in the form of a breach of contract argument or breach of an implied warrantee by the owner that the contract can be completed on time. Then the contractor must prove that the owner has failed in his obligations, and finally the contractor must prove the extent of the delay damage.
This last step, the proof of delay damage, is a most difficult task. To illustrate judicial attitude toward proof of delay damages several legal citations follow:
1. Uncertainty as to the amount of damages does not prevent recovery if evidence affords a sufficient basis for estimating their amount in monetary terms. Prejean vs Delaware-Louisiana Fur Trapping Co. 5 Cir, 13 F 2nd 71-2.
2. Plaintiff (Sub) brought action against defendant (GC) for damage due to breach of contract including loss of profits. Court disallowed the claim because it was established by plaintiff’s opinion of his loss without supportive facts. Hadden vs Western Contracting Co. 76 F Sup 987.
3. Court ruled that if the loss occasioned by a breach of contract is pecuniary and susceptible to proof with approximate accuracy, then proof is necessary. Belcher vs King 96 WVa 362, 123 SE 398.
4. Supreme Court disallowed claim when plaintiff refused to break down the amount in detail. Also stated, “We admit the admissability of opinion evidence but insist upon its weakness and upon the necessary data to enable the court to test its admissability, weight and value.” Doman vs Baltimore & Ohio R.R. Co. 1942, 125 WVa 8, 22 SE 2nd 703-5.
5. The court stated, “We are not unmindful of the legal generalizations which favor upholding the verdict. We are reluctant to disturb a verdict of juries. But reluctance must yield to duty whenever the record affords no substantial evidence to support the quantum of the verdict.” Chesapeake & Ohio R.R. vs Allen, 113 WVa 691, 169 SE 610-12.
6. The court stated that absolute certainty as to the amount of damages is not essential, this being a matter of determination of circumstances in each case. There is no objection to damages that are difficult to ascertain for the underlying principle is full compensation for wrong done. Needles vs U.S. 101 Ct Cl 535.
It is obvious, from these examples, that courts seem to recognize the need to award certain damages to contractors for delay. However, they require some degree of proof which goes beyond opinion and expert testimony.
Sweet , in recounting a U.S. Court of Claims case, states,
It is a rare case where loss of productivity can be proven by books and records; almost always it has to be proven by the opinions of expert witnesses. However, the mere expression of an estimate as to the amount of productivity loss by an expert witness with nothing to support it will not establish the fundamental fact of resultant injury nor provide a sufficient basis for making a reasonably correct approximation of damages (Emphasis added).
This quest for more certainty in establishing delay damages has led the authors to see if learning curve theory could be applied to delay claims.
Learning Curves in the Construction Industry
Learning curves are used in many industries to depict the increase in efficiency from performing a task a number of times. The basic form of the learning curve is:
Y = Axb
A = manhours (or cost) to produce the first unit in a series
Y = manhours (or cost) to produce the xth unit in a series
b = a parameter which describes the rate of decrease in manhours with successive units
A typical value for b is -0.322, which is to say that the accumulated manhours (or costs) will be reduced to 80% when doubling the number of units.
Applications of learning curve theory to construction are discussed by Paulson  and Parker and Oglesby  and well documented in a United Nations report on repetitive building operations , The United Nations report states that for learning curve theory to be applicable to construction, two prerequisites must be met. First, the work elements must be repetitive and, second, the work must be continuous. There are many examples of repetitive work in construction; high-rise office building, condominiums, tract housing to name a few. But what is the effect on the learning curve improvement if the repetitive work operations are not performed in a continuous operation?
The United Nations report suggests that considerable unlearning can take place if a repetitive operation is delayed for any length of time. A study conducted in Finland as part of this report shows the effect of delays during construction (See Figure 1). The data refer to the construction of 45 single family houses. Construction was halted between the 29th and 30th houses, with obvious effects on productivity. In this study, manhours decrease quickly with the first five houses and efficiency appears to level off. Although the complicated tasks show more improvement, they are also strongly influenced by an interruption to the learning curve.
Figure 1 Unlearning Behavior
A Swedish study, which is also part of the United Nations Report, attempts to quantify the unlearning effect (See Figure 2). We quote from the original,
If an interruption occurs in the course of the execution of a sequence of identical operations, owning either to a time break or a change — whether complete or partial…the improvement curve will show a discontinuity (Emphasis added).
In this study the length of interruptions, length of time operation proceeded prior to interruption, and the type and complexity of work all have an effect on productivity.
a = break between 9th and 10th units
b = break between 19th and 20th units
a1 = a 12-week interruption
a2 = an 8-week interruption
a3 = a 4-week interruption
b1 = a 20-week interruption
b2 = a 16-week interruption
b3 = a 12-week interruption
b4 = an 8-week interruption
b5 = a 4-week interruption
Figure 2 Effect of Breaks on Repetitive Tasks
A final example, provided by Parker and Oglesby , shows the results of a detailed study of sequential installation of turbines (See Figure 3). The manhours expended after installation of the eighth turbine begins an upward trend. Prior to this period, material delivery was constant and an identical crew was available with an ever-ready work site. Units 9-14 show the results of loss of momentum due to forced crew changes, variable delivery time, and conflicts in work site availability. These interruptions reversed the increasing efficiency trend and resulted in increased costs.
Figure 3 Learning Curve Effect
E.W. Bliss Company vs U.S.
The unanswered question is if the courts will accept learning curve data as a basis for a quantum judgement in a delay case. The only case uncovered claiming dollar amounts from a loss in the learning curve effect due to delay was E.W. Bliss Company vs U.S. in a case before the Armed Service Board of Contract Appeal (ASBCA) on 15 February 1968 ,
E.W. Bliss Company was awarded a contract by the U.S. Navy on 12 December, 1961, as a negotiated fixed price contract to provide torpedo tube assemblies and ejection pumps for consignment to Mare Island Naval Shipyard, Mare Island, California. This equipment was for two Polaris submarines, SSB (N) 629 and SSB (N) 634. The contract price was $636,010.07 but Bliss’ actual total cost was $1,170,300. A summary of the total claims is listed in Table 1. Of the $500,338.04 claimed for production interference, a specific claim of $169,035.94 was for learning curve effect loss.
Of the eight ship sets of equipment produced under various contracts, the two Mare Island submarines were the fourth and fifth sets. Bliss projected, at an 85% learning curve, the manhours required should have been 40,662 and 38,000, respectively. Actual manhours were 50,771 and 50,779 for an excess of 22,026 total manhours at a cost of $7.65 per hour.
The ASBCA ruling on the claim stated, “It is not established by the record that Bliss used a reasonable and realistic number of manhours per ship set as a starting basis for the plotting on its learning curve; hence, we are unable to verify the correctness of its calculation of excess production hours. The computation of manhours to be expected under an 85% learning curve appears to be based on the work called for by the original specifications, while the computation of actual manhours incurred in the production of the two Mare Island ship sets reflects the additional manhours required by the changes for which the parties have already agreed on price increases totaling some $20,000 and for which the Board has granted price increases totaling $92,007.97. Hence, the excess manhours shown as “learning curve effect loss” include the manhours for additional work for which appellant has already been granted price increases. The learning curve effect loss computation serves to verify to some extent that the contractor actually experienced the increased costs of performance (emphasis added) for which price adjustments have been granted, but there is no evidentiary basis for holding that the entire amount shown as learning curve effect loss was caused by change orders, actual and constructive. The causal connection between the changes and the amount claimed as learning curve effect loss has not been established. The claim based on learning curve effect loss must be denied.”
Table 1 Total Claim for Production Interference E.W. Bliss Company vs U.S.
In this case the ASBCA recognized damages caused by the learning curve effect loss. However, ASBCA objected, not to the reason for the claim, but the insufficient supporting evidence. Bliss was unable to fulfill the requirement for adequate evidence in order to allow the Board to consider its value and make an award. The Board felt that the learning curve effect loss computation only further justified a previous award and was unclear as to any additional damages. Since the learning curve effect included a previously granted price increase, the Board disallowed the entire claim as redundant.
Bliss should have validated its original learning curve with historical data from previous contracts rather than relying on an estimated projection. Additionally, a new learning curve should have been developed to reflect the change in scope for which it was compensated by the U.S. Navy. This curve should have reflected additional damages due to delays and changes not covered by the original compensation.
Summary and Conclusions
Reviewing the existence and application of the learning curve effect in construction, judicial statements on delay claims, and the Bliss vs U.S. case, we provide the following observations:
- Interruption in repetitive work causes an increase in time and labor.
- These increases can be defined in monetary terms by accurate recordkeeping and job analysis.
- Claimants in general are reluctant to base damage claims in only one area. Rather, a claim will contain several reasons for compensation.
- Courts recognize the need to compensate contractors for provable delay damages.
This would seem to support the view that damages due to learning curve effect loss, if accompanied by full supporting evidence, would be recognized by the courts. However, the implication in observation 3 above (claiming many areas) is significant. The supporting evidence required for learning curve effect loss justification includes many areas; material delays, work stoppages, scope changes, etc. Each of these areas by itself may be justifiable cause for compensation and have been previously recognized in the courts. In addition, learning curve analysis requires meticulous recordkeeping in order to accurately pinpoint the claim cause. It seems likely that separate claims for learning curve effect loss may be accompanied by additional redundant claims and would therefore be disallowed by the courts. Also, if all or the majority of a claim may be justified on areas of previous precedence, there will be a tendency to follow the previously established route. Finally, construction operations are often subject to numerous non-owner caused delays, such as weather, labor disputes, etc. A constructor, if he wishes to use learning curves as a basis for claim, must be able to segregate the effects of non-owner delays from owner delays.
Considering these facts, the use of learning curve effect loss seems destined to serve a supporting role rather than a primary role in justification for total damages due to various owner delays.
1. Baldwin, J.R., et al., “Causes of Delay in the Construction Industry”, ASCE Journal of the Construction Division, November, 1971.
2. Board of Contract Appeals Decision, 68-1, E. W. Bliss Company, 6906.
3. Koehn, E., et al., “Cost of Delays in Construction”, ASCE Journal of the Construction Division, September, 1978.
4. Parker, H.W. and C. H. Oglesby, Methods Improvement for Construction Managers, McGraw-Hill, 1972, p. 59.
5. Paulson, B., “Estimation and Control of Construction Labor Costs”, ASCE Journal of the Construction Division, September, 1975.
6. Sweet, J., Legal Aspects of Architecture, Engineering and the Construction Process, West Publishing, 1977.
7. U.N. Committee on Housing, Building and Planning; “Effect of Repetition on Building Operations and Processes on Site”, United Nations, New York, 1965.
Have you completed and returned the Project Management Certification Survey on pages 57-58?