It’s Time for the Defense Production Act’s Digital Transformation
President Donald Trump’s invocation of the Defense Production Act (DPA) marked an important step in the fight against COVID-19. Doing so allowed the president to order 3M to make millions of N95 masks, initiate a 30,000 ventilator contract with General Motors, and force meat processing plants to remain open. Despite these efforts, the administration’s use of the DPA has received mixed reviews.
Behind the scenes, America’s capacity to leverage the DPA is running up against a modernization problem. American leaders are flying blind as they attempt to manage this pandemic equipped with antiquated digital technology, spotty datasets, and inaccurate models. Even more challenging, the tools required for a digitally driven federal response strategy do not exist, nor do regulations to enforce compliance with emergency data management. That gap exists because the 1950s authors of the DPA could not have foreseen today’s digital reality.
The United States should modernize the DPA’s authorities, invest in digital infrastructure, and enhance private sector collaboration. Specifically, Washington should amend the DPA to reflect the critical role data plays in proactive government response efforts, expand bilateral data sharing with the private sector, and challenge norms around digital talent in government.
The Defense Production Act Is a Relic of the Past Century
The DPA is a product of 20th-century problem solving. When drafted in the 1950s, it expanded the president’s emergency authorities over the nation’s lifeblood — industrial production. At the time, nearly 50 percent of America’s private sector was employed in manufacturing. It was America’s manufacturing resources that helped shape the outcome of World War II, starting with the Lend-Lease program. Manufacturing’s prominence in American strategy led President Franklin Roosevelt to demand “we must out-produce them overwhelmingly, so that there can be no question of our ability to provide a crushing superiority of equipment in any theatre of the world war.”
In the decades since, America’s economy has changed a great deal. As of 2018, the manufacturing sector employed just over 8 percent of America’s workforce. This has decreased the comparative impact of mobilizing private sector manufacturing resources during emergency response efforts. Accordingly, even when industrial leaders step forward to help fight COVID-19, like Tesla’s offer to build ventilators, efforts have run into challenges.
The problem, highlighted by Gov. Andrew Cuomo, is that even General Motors or Ford “can’t make ventilators that fast because there are parts that have to come from other countries.” International supply chains, critical for decreasing the price of consumer goods, have displaced America’s 1950s industrial base. While COVID-19 has rekindled conversations about reshoring American manufacturing, this would take years or even decades — too long to have a meaningful impact on the pandemic response.
Data and the Pandemic Response
While America’s industrial base commands a smaller share of the economy than it used to, there is another critical resource the government could use to enhance response efforts aligned to the DPA’s intent: data. Much as manufacturing dominated America’s strategic options in the 20th century, the power of big data and artificial intelligence to leverage the 2.5 quintillion bytes of data created each day can fundamentally change 21st-century response efforts.
A data-centric emergency response strategy can achieve goals that a manufacturing-oriented DPA cannot. For example, digital platforms capable of real time tracking for COVID-19 caseloads, medical equipment burn rates, and available inventories of medical supplies would provide policymakers immediate situational awareness about resource shortages. In turn, data from healthcare systems could be used to inform predictive models that would improve resource allocation and accelerate resource matching, enabling an iterative approach to policy development and implementation. If this data was made publicly accessible in an open-source format, non-profits and private individuals could leverage these datasets to generate innovative solutions to emerging challenges. This granular understanding of the challenges facing American communities could allow the White House to best allocate America’s manufacturing resources, achieving optimization levels that are impossible in the current system of fractured data systems. While individual efforts tackling this problem exist across the country, many are ad hoc and lack the robust digital infrastructure or data sharing networks needed for optimal data utilization.
A competent, data-driven approach could have helped mitigate the personal protective equipment shortage confronting healthcare workers. While much public attention focused on the need to surge manufacturing efforts, far less attention went to the inability of healthcare and government teams to purchase millions of masks and other medical equipment from American warehouses. The biggest obstacle confronting their efforts was speed, as traditional medical supply purchasing systems could not keep pace with the frenzied pace of pandemic demand. These challenges were compounded when the Federal Emergency Management Agency began purchasing masks and struggled to maintain objective and efficient systems for distribution. Accordingly, medical personnel were exposed to enormous risks and the federal government struggled to understand how to best use resources like the national stockpile.
The ineffective use of data when making federal decisions is due in large part to antiquated technology and data illiteracy. The term “data science” was first coined in 1996, yet government agencies still often struggle to recruit and retain digital talent, and lack big data and cloud infrastructure. The government should work to keep pace with technological change.
The government has not fully embraced the power of digital transformation in the DPA’s current structure. This can be seen in how little it has been amended to account for advances in digital technology. Even when the DPA was updated to include “cyber” assets within the definition for “critical infrastructure,” there was no accompanying definition of data. Further, current references to data within the DPA are primarily relegated to internal reporting concerns. While the DPA has been updated over time, many existing amendments have emphasized tangible objects, like the 1980s redefinition of “strategic and critical material” to include energy. Even the DPA’s section about “modernization of equipment” focuses on manufacturing resources.
Recent Digital Failures
At present, government digital infrastructure is antiquated and disorganized. Already this has had grave consequences for the American public. This spring, as our economy contracted in the face of an international recession whose scale has not been seen since World War II, government platforms struggled to provide public services. States were buried beneath a surge of benefit seekers that overwhelmed Florida’s digital processing capabilities, left New York lagging by hundreds of thousands of unemployment claims, and led New Jersey to seek coders fluent in a 60-year old programming language.
We saw a similar pattern of digital unpreparedness hurt small businesses despite Congress allocating $377 billion of the $2.2 trillion stimulus plan towards small business owners. Delivery efforts struggled to issue small business loans, pay the intended companies, and provide clear guidance for using funds. As a result, one in four small businesses is at risk of closing permanently. A contributor to these challenges was the collapse of the Small Business Administration’s systems for processing these loans.
The medical sector also suffered due to digital infrastructure gaps, which contributed to delayed response efforts and resource misallocations. The visible consequences included difficulties in understanding pandemic growth rates, difficulty adapting policy due to modeling challenges, and establishing military field hospitals that went largely unused. Even the government’s capacity to engage America’s entrepreneurs to create new medical solutions was bogged down by arcane grant award processes that should have been automated.
This cascade of public sector digital incompetence was as tragic as it was preventable. For years our technologists, and satirists like John Oliver, have argued that the country’s digital infrastructure is broken. The problem is the lack of federal regulations or compliance requirements to mandate digital upgrades across the public sector. While the need to develop strict guidelines at the federal level is most often associated with cybersecurity, America’s collective failure to respond to COVID-19 demands seeing these digital upgrades as a critical government investment. Failure to make these changes will deny the United States the robust data infrastructure needed to protect its communities from the economic and public health consequences of the “inevitable” second wave.
Updating the Defense Production Act for the Digital Age
The magnitude of the challenge America faces requires aggressive and proactive legislation to update the DPA that addresses a two-fold challenge. First, legislators should expand the DPA to provide the executive branch with data access to enable more effective response, while simultaneously safeguarding constitutional rights. The legislative language be flexible enough to adapt to different national challenges and relevant data providers, while being consistent enough to normalize data sharing frameworks. Second, Congress should invest in the digital infrastructure needed to leverage that data once in government possession, from acquisition efforts like the Pentagon’s contested JEDI contract, to expanding National Science Foundation and Small Business Innovation Research grants, and prioritizing talent reform efforts. This second challenge requires developing the appropriate human capital, cloud infrastructure, and data platforms needed to transform data access into data-driven decisions and insights. Failure to accompany legal authorization for data access with modernization of the federal government’s digital infrastructure will blunt the value derived from that data once in public hands.
From a legislative perspective, the DPA provides grounds for the President to offset the risk posed by antiquated systems by recognizing the need “to ensure the vitality of the domestic industrial base . . . [and] support continuing improvements in industrial efficiency and responsiveness.” Further, the president is empowered to enforce the “domestic productive capacity to meet essential national defense needs that can result in emergency conditions and unique technological requirements.” These permissions need to be extended to data rights and relevant digital infrastructure.
The legislative challenges associated with maintaining constitutional protections while expanding data access would be immense. The government’s power to “seize” data is in many ways constitutionally limited, but its power to incentivize and facilitate cooperation between the government and private sector is much broader. Contemporary legislation involving data access, like the Stored Communications Act, has already recognized the potential for “emergency” circumstances that may warrant a provider’s otherwise impermissible disclosure of customer information to government entities.
Much of the data needed to improve response efforts will not require personally identifiable information. Sanitized data and metadata sources would be used to understand trends within populations to make informed policy recommendations and predictive models. Accordingly, these data sets are not used for prosecution, nor to violate intellectual property protections, and would have to protect individual identities and privacy rights.
The complexity and nuance associated with writing legislation that would empower government response efforts, while protecting constitutional Fourth Amendment rights will require leveraging both public- and private-sector expertise. These conversations must explore the most efficient ways to create standard data types and sharing frameworks that bridge the public-private sector divide, protecting commercial secrets while overcoming proprietary data types and siloed systems. A strategy for confronting this challenge could include establishing a board of trusted civilian experts, like the Pentagon’s Defense Innovation Board. The Defense Innovation Board, chaired by Eric Schmidt, applies Silicon Valley strategies to help the Pentagon overcome being “stuck in software in the 1980s.” Their insights have created policy recommendations and programs that aided the Pentagon while saving millions of dollars.
Josh Marcuse, the former executive director of the Defense Innovation Board, noted that
Congressional action would be needed to formulate a collective strategy and associated legal authorities to leverage data as an emerging class of resources that the government needs to be able to access to serve the American people in times of an emergency. It’s vaguely analogous to eminent domain, DPA, and the 3rd and 4th Amendments in terms of an overriding public need, but now it must be updated from raw materials, factories, or land to encompass the 21st Century’s most precious resource: data. The civil liberties concerns and practical concerns are massive, but the COVID crisis shows we need to wade into these deep waters.
After tackling constitutional data sharing challenges and writing a digitally transformed DPA, legislators will need to invest in the digital infrastructure needed to leverage this data. This is a long overdue national priority, and failure to own this responsibility only makes it more expensive to catch up with technology standards due to mounting technical debt.
While many of these digital investments will likely mirror private-sector practices, Congress could support the American people through COVID-19 specific resource allocations. A framework for this style of investing was highlighted in a recent letter to congressional leaders about investment priorities from technologists across the country. In it, technologists advocated for focused investment in teleworking tools, state and local technology implementation, cybersecurity, and modernizing government technology. Legislators would also benefit from recommendations made by Nicholas Thompson, who urged political leaders take strong stances around “broadband for all, a national strategy in AI and quantum, regulatory reform on issues like telemedicine where the US should lead the world, a federal job retraining strategy for the era of AI, [and] easing immigration policies for high-tech workers.”
The expertise required to guide digital infrastructure investment decisions could be supplied by reinvigorating programs that hire technologists as government policy advisers. The government should reestablish an Office of Technology Assessment. During the Carter administration, congressional technologists in the Office of Technology Assessment conducted policy reviews that saved over $60 billion and boosted America’s competitive advantages by applying experts against gaps in legislator experience. Unfortunately, it was eliminated in 1995, leaving legislators without impartial experts to guide technology policy formation. Now, when staffers require technical expertise, they are forced to turn to “industry lobbyists, or think tanks, or advocacy groups”, some of whom may be influenced by institutional agendas.
The progress towards closing the federal government’s technical expertise gap gained by refunding the OTA could be complemented by expanding investment in programs like the Congressional Innovation Fellowship and the Presidential Innovation Fellowship While these programs have a broad mandate, emphasizing digital transformation efforts could provide immediate value for legislative agendas. These investments could improve the quality of legislative decision making and create opportunities for government service, both of which would stave off the risk of future response shortfalls. Further, ongoing discussions about future stimulus packages could set the condition for long overdue investment in digital infrastructure.
Federal investment in digital infrastructure would require equivalent investment in workforce development, or federal agencies would risk being buried beneath data wrangling problems. Thankfully, the Pentagon again has relevant experience with these digital transformation challenges. When the Air Force pursued digital transformation, service leaders recognized that maintaining modern military dominance required treating software development as a critical organic capability, instead of relying solely on external contractors.
This strategy gave rise to software initiatives like Kessel Run, which provided the Pentagon with more agility than traditional acquisitions models. This same approach was replicated when Space Force created its software factory, Section 31. Software factories have tackled challenges like creating new software applications, replacing legacy platforms, automating workflows, or optimizing database structures. Further, service-specific software factories are complemented by Pentagon-level efforts that bring in skilled civilian technologists to create bespoke solutions, like the Defense Digital Service.
A similar strategy could be undertaken by critical federal agencies, like the Federal Emergency Management Agency. An inward-facing software development strategy upskills and empowers existing personnel, while providing organizations with developers who have the relevant context to solve the right problems. This unique blend of insights and direct experience cannot be easily replicated by external contractors, thereby minimizing challenges with product-mission fit. Further, internal developers understand the complex technical interdependencies and value chains that solutions must nest within for maximum impact.
Perhaps just as critically, Congress should incentivize greater collaboration and data sharing between the government and the technology sector. This includes the creation of cooperative research and development agreements and awarding grants to support non-profit research efforts. An example of this approach is the COVID-19 High Performance Computing Consortium, which offers researchers access to America’s best computing labs, with the support of U.S. Chief Technology Officer Michael Kratsios. This ad hoc effort aligns with other ongoing philanthropic programs to improve information sharing that provide ready partners for government efforts, like Microsoft’s Open Data Initiative and the Open Data Institute.
These philanthropic projects allow non-profits and other public-sector partners to share data and collaborate on critical social challenges, using vast and underleveraged private-sector data sets. The leaders within these and related initiatives could also help populate the recommended government advisory board, and could provide the initial surge of datasets needed to mobilize effective response.
The United States has seen the consequences of responding to COVID-19 without the appropriate digital infrastructure and data-enabled frameworks. Since many healthcare organizations project a second wave, failure to take aggressive action would needlessly subject the American people to suffer the consequences a second time.
Central to preventing future tragedies is the government transforming the DPA to fit the digital age. If successful, federal leaders can rebuild America’s digital infrastructure and capabilities with the strength needed to protect our communities for COVID-19 and future crises. This would enable unprecedented data sharing and government modernization, while enhancing economic and public health resilience — all while regaining critical competitive advantages.
It’s time to invest in America’s digital and data resilience, starting with a new DPA.