In a recently published article, my co-author and I focus on the issue of contracting by US federal agencies in response to disasters. Specifically, we investigate the choice of procurement contract type (cost reimbursement vs. fixed price) by US federal agencies during disaster management operations. The research relies on 47,560 contracts issued by the US federal government in response to 14 major disasters between 2005 and 2016. While earthquakes represent the majority of disasters in our dataset (10 out of 14 disasters), we also consider other disasters (such as earthquakes, tornado, and oil spill) for which federal agencies procure products and services for relief operations. The dataset includes events that mainly affected the US; however, two events outside this geographical scope were included (Haiti earthquake in 2010 and Pacific earthquake in 2011) because they represented a significant expenditure by US Federal Government agencies. We build on agency theory to investigate the choice of the contract type made by federal agencies at the different stages of a relief operation. This research provides empirical evidence of the key factors underpinning the choice of contract in the context of disaster management, namely the amount of spend per contract and the type of acquisition (product or service), and reveals the moderating role of the stage of the relief operation.
Our finding shows that in disaster-relief operations, as spend increases, the likelihood of choosing cost reimbursement also increases. Our results indicate that government agencies tend to be more guarded in this respect by preferring cost reimbursement deals as the contract value increases. Our findings show that service contracts are more likely to be cost reimbursement. We found no significance in the direct relationship between stage of disaster (preparation, response, and recovery) and contract choice, when the response stage was operationalized as 100 days after a disaster strikes. However, the relationship was positive and significant for 200-day model. Interestingly, we expected more cost-reimbursement contracts in the Response stage. Similar logic would lead us to expect that more cost-reimbursement contracts relative to fixed-price contracts during all disaster stages relative to “normal times.” However, a report by the Government Accountability Office (GAO, 2017) reveals that the split of contract types for all government contracts across all agencies between 2011 and 2015 was 63% fixed-price; 37% cost-reimbursement. This contrasts with our dataset in that fixed-price is much more prevalent during disaster contracting, accounting for 92% of all contracts, notably driven by Hurricane Katrina contracts. It appears that the preference for fixed-price arrangements is emphasized during disasters, particularly during the response stage. One explanation may be that fixed-price contracts might be quicker to construct (and much easier to monitor) than cost-reimbursement contracts; give the hectic nature of disasters, this may dominate. This is in line with prior arguments about the ease of execution of fixed-price contracts. This also helps explain that as we expand the response-stage window to 200 days, that is, we include more contracts that are not subject to time pressure, the relationship between stage and contract choice becomes significant.
The split-sample analysis of the Katrina/non-Katrina contracts revealed that in the non-Katrina sample there is a higher likelihood of using cost-reimbursement contracts during the response stage than in the preparation and recovery stages. However, the finding was opposite and significant for the Katrina only sample. Katrina is the first and most severe event in our dataset, and the response effort was heavily criticized for handing out large contracts without competition. These criticisms resulted in an investigation by the White House and a legislation reform with extensive revisions to emergency response provisions (Post-Katrina Emergency Management Reform Act of 2006, 2006). While these recommendations and legislation do not specifically refer to fixed-price contracts, they stipulate changes to staffing, organization structure, and contract competition. In 2015, 10 years after Katrina, the GAO reported (GAO-15-783, 2015) on the changes to emergency response provisions triggered by Hurricane Katrina reported a tripling in the number of contract officers and the establishment of the Disaster Acquisition Response Team (DART) and a new rule that prevents contracting officers from issuing non-compete contracts exceeding 150 days without proper authorization. These changes to staffing, legislation, and procedures are likely to have affected contract officers' behavior post-Katrina, particularly regarding the award of service contracts, which tend to extend over time.
Interestingly, while we hypothesized that larger contracts during the response stage would lead to more cost-reimbursement contracts, the opposite was true in the whole sample. The result is driven by extremely large contracts. To explore this further, we inspected the largest contracts during the response stage, specifically those over $75 million, which is the threshold at which the senior procurement executive of the agency is required to approve the contract according to FAR 6.304. We found 39 contracts, of which 31 (79.5%) are fixed-price and eight (20.5%) are cost-reimbursement. However, this split is not homogenous across government departments. Of the eight cost-reimbursement contracts in this sub-set, seven were issued by one Agency, FEMA; the agency tasked with “helping people before, during and after disasters.” Of these seven contracts, five relate to architectural/ construction services and appear to be associated with providing accommodation for people in the aftermath of a disaster, in line with FEMA's mission. The remaining two are technical assistance contracts.
In our sample, as expected, FEMA has the largest number of contracts 13,742 (29% of all contracts), because of its mission. Even though FEMA has a larger proportion of the extremely large cost-reimbursement contracts, they predominantly award fixed-price contracts during the response stage due to the speed of awarding and monitoring them. It seems that FEMA follows a dual strategy, relying on fixed-price contracts for speed, and also using extremely large cost-reimbursement contracts for major projects. While practitioners have long recognized the benefits of fixed-price contracts in terms of speed, this is not broadly recognized in the extant literature.
We ran a sub-sample analysis for the two major agencies: FEMA and the Department of the Army, which represent over 43% of all contracts in the dataset. This analysis indicates that the findings regarding FEMA are significant and align with the results in our main model results. Even the outcome pertaining to the disaster response stage exhibits a higher likelihood of using cost-reimbursement contracts during the response stage than in the preparation and recovery stages. Conversely, in the case of the Army, we find that in disaster-relief operations, the amount of spend associated with a contract increases the likelihood of using cost-reimbursement contracts over fixed-price contracts. We also find that in disaster-relief operations, the Department of Army was more likely to consider cost reimbursement contracts for procuring services than for procuring products. In the case of the Department of Army we couldn't find evidence of higher likelihood of using cost-reimbursement contracts during the response stage than in the preparation and recovery stages. Neither did we find evidence of higher use of cost reimbursement contracts during the response stage (as compared to other stages) to be related to (a) the amount of spend and (b) the procurement of services as compared to products.
This indicates that while both government departments issue many contracts during a disaster, only FEMA appears to be under pressure during the response stage. This outcome corresponds to FEMA's primary mission of addressing urgent needs in the immediate aftermath of a disaster. An alternative explanation is that because agencies might have already negotiated per-unit costs with selected suppliers for services commonly needed during emergencies, such as medical care, provision of meals, and so forth, pre-positioned contracts could thus help mitigate opportunistic behavior. Policy guidelines under FAR make pre-established arrangements available to government officials, and researchers have reported on their effectiveness. For example, in 2019, the Army Corp of Engineers established a $999 million advance joint venture agreement with six contractors to provide time sensitive disaster support to repair critical infrastructure such as roads, bridges, and electric distribution systems.
Our robustness checks reveal that results regarding the amount of spend during the response stage for the two largest sub-samples regarding the type of competition, “Not Competed” and “Full and Open Competition,” (which together represent about 47% of all the contracts in the dataset) are significant and negative, in line with our main model results. These two competition categories, which are at opposite ends of the spectrum (from no competition to full and open competition), indicate that the negative interactions between the stage of the disaster and the amount remains under vastly different types of competitive environment. Finally, our results regarding the interaction between the service variable and disaster stage, show that services contracted during the response stage are more likely to use cost-reimbursement contracts than those awarded during other disaster stages, in line with agency theory predictions and our logic presented there.
Source: Mena, C., & Nair, A. (2024). US federal government contracting for disaster management. Journal of Operations Management, 70(4), 523-547.