What Structural Factors Inhibit the U.S. Government Efficiently Acquiring Ships?

It's déjà vu all over again

(Note: The next Green Fuels update is almost ready and will be out in mid-July. I expect to alternate between a decarbonization post and a U.S. domestic issue post for the next few months.)

The U.S. Government’s ship acquisition crisis appears to be in a (at least local) maxima. Since the start of 2024, reports of struggling programs and approaches have appeared one after the other like the surf breaking against a beached vessel. In January, initial reports that FFG-62 would be up to a year late surfaced, along with the start of the SECNAV’s 45-day shipbuilding review. Ongoing delays to Columbia’s completion appeared in March. The 45-day review appeared in April (see page 11 of this CRS report for the one-page output), which concluded almost all building programs (SSN, CVN, etc.) are running 12-36 months behind schedule, and FFG-62 is on the 36-month, not 12-month, end of the scale. May opened with the GAO launching a broadside chronicling significant shortcomings in the Navy ship acquisition approach compared to commercial practice. The FFG-62 bad news continued, with a May report indicating that weight growth was also complicating the design, while concerns about the U.S. submarine industrial base’s capability to support both domestic production and AUKUS continue to simmer in the background. Congressional frustration resulted in proposed legislation to turn the GAO recommendations on design completion into law. The USCG was probably thankful that the Navy’s rough six months helped distract attention from the 4(?) year delay to their new heavy icebreaker.

Reading the recommendations from all of these reports, I was struck with a profound sense of déjà vu from past reports. For example, the following quote from a 2014 benchmarking study dovetails perfectly with current discussions:

“International shipyards have a standard approach to ship design for all vessel types with well-defined design stages and clearly specified outputs for each stage, including design for production elements. In addition, long intervals between design cycles in the US reduce the opportunities for applying design for production principles, introducing new methods and maintaining core design capability. Maintaining the technical workforce is of particular concern with regard to the future capability and performance of the industry…”

So why hasn’t more progress been made? I do not believe that the U.S. Government employees running Navy and Coast Guard acquisition programs are unaware of these findings, I also believe they are highly skilled at their jobs. Lack of knowledge of commercial practice isn’t at the root of this struggle. I propose an alternative thesis for discussion - that best practices in modern commercial ship design and construction rely heavily on two structural features of the commercial world:

1. a deep collective knowledge structure built up in the industry, with many organizations (class, IMO, vendors, yards, owners) actively investing in extending and maintaining this structure.

2. A market view that provides both a clear basis for design judgment and the appearance of market elasticity that places efficiency of production as a key to achieving higher profits.

I’d argue that we’ve lost the equivalent structural features for government ship design domestically, and missing these features makes it impossible to simply copy commercial ship design approaches.

Benchmarking Against Commerical Design

Contrasting U.S. government practice with commercial practice helps develop a framework for looking at these different structural features. Fortunately, the recently produced GAO report provides an excellent framework for this comparison. Through interviews with numerous commercial shipowners and shipyards, the GAO has produced a structured list of themes, with 13 themes under 4 higher-level headings. I’ll adopt the GAO’s framework below to explore the structural factors in more detail.

GAO Theme Area #1: Establish business cases and requirements that support predictable design outcomes

Prioritize timeliness of ship design and delivery

Here, the large size of the commercial shipbuilding market compared to the output of any one shipyard, giving the appearance of market elasticity, is a central difference compared to government shipbuilding. For all yards, faster throughput lowers the yard overhead charge (yards are highly capital intensive) for each vessel delivered, reducing the vessel cost. With a market that looks elastic to each yard, faster throughput means order books can be kept near optimal levels for successful yards, and in “boom” times where orders come in faster than yards can produce vessels, prices can rise far above cost, giving they yard a (rare) opportunities to make large profits.

For shipbuilding under a U.S. Government contract, the same market forces do not apply. Most government contracting is designed to prevent companies from making excessive profit on government work, and imposes a strong coupling between cost and price, with an implicit assumption that getting the ship “correct” matters most. A side effect of this is smaller opportunities to make large profits from decreased cycle time. Additionally, the number of vessels to be built is set by operational requirements and Congress together. Therefore, a yard cutting the price and time to build a vessel by 15% won’t always result in 15% more orders or a higher profit. Additionally, the slower pace of government decision making could mean that finishing early results in a gap in production that is very destabilizing for a shipyard’s viability.

Avoid overly prescriptive requirements

Here, the GAO noted that commercial owners focus on high-level requirements and leave the shipyards to develop the details - the GAO report gives a good example of Chevron providing a 5-10 page summary of what is needed, and the shipyard driving the detailed design process, responding with 50-100 page and then 500-1000 page design specifications. In the commercial world, this division of labor makes sense- a major shipyard could easily be delivering 50 vessels a year, while a shipyard group like DAMEN might have 150 vessels delivered per year. No owner is developing anywhere close to that many designs per year. Commercial shipyards have invested and owned the knowledge structure necessary to handle this level of design activity, often, but not exclusively, using standardized designs optimized for rapid construction in their shipyard (see point above). This knowledge is a source of competitive advantage; it is the expectation of the market that they will perform this task, and their cost is recouped from their high output.

In the U.S. Government market, there may be a total of only one to two new designs per year between all the Navy and Coast Guard building programs, which are split up over multiple yards. It’s not possible for individual yards to maintain the same type of design knowledge, especially across all ship types desired by the Government. Thus, design invovles both the government and the shipyard. In the government space, pure requirements-driven approaches for complex government projects have historically failed - for example, USCG Deepwater. I’d argue that a key factor in these failures is that no one- not the government nor the shipyard/LSI (in system-of-systems speak) has the equivalent of a commercial shipyard’s knowledge structure around the product to be built. Without this knowledge structure, it is impossible to translate high-level requirements into a complete engineering product efficiently - at some point, either one has created the necessary knowledge on the way (which often looks like inefficiency to external observers) or one has “discovered” the missing knowledge the hard way via expensive product failures/late-stage emergent issues.

Maintain a sound business case through continued reevaluation

Here, there is a clear market role - both the shipyard and commercial shipowner have the same metric for success at the end of the day - money. Monetary profit is a single objective, and, in hindsight, success or failure is clear to all parties regardless of vantage point. Removing or modifying features for naval vessels is much harder as all the roles they will fulfill and metrics for success are debatable, as the conclusion often depends on one’s perspective. For example, the UK invested heavily in moving their Navy to perform ASW for NATO in the 1960s-1970s, then had to use ships optimized for such missions in the Falklands War - an entirely different “business case”. Likewise, the USN successfully constructed the limited, design-to-cost FFG-7 class as a deep ocean escort but then used the class intensely in the Persian Gulf. Both navies had vessels damaged (in the UK case, several sunk) in these operations that were beyond the initial “business case” for the vessels. In hindsight, should more multi-mission capability have been added? What would have to be sacrificed to do so? Could that decision be made logically and defensibly in the 1970s? The lack of a clear market measure of success complicates answering such questions.

GAO Theme Area #2: Use iterative design to accelerate design maturity

Prioritized user involvement in the ship design process

This one I just completely agree with!

Leverage existing ship design and systems in digital libraries

This ties in very closely with my response to the “prescriptive requirements” above - this is a good idea, but it assumes someone, somewhere, is developing and maintaining this knowledge base. As NAVSEA downsized in the 1990s, it took a market-oriented model, where programs with active funding and requirements became the focus of the organization, and charges against these programs were how units justified their existence. With such a worldview, maintaining a broader knowledge structure that might have future re-use benefits inside NAVSEA appears to be wasteful inefficiency. As subsequent programs encountered schedule and budget delays, spending time and money to rebuild a wider knowledge base for future re-use becomes less and less attractive to underfire program managers.

Prioritize timely vendor decisions and information

The GAO report points out that vendor information is normally available far sooner for commercial designs than for Navy designs. This is an important point, as vendor influence on shipbuilding is probably the most overlooked aspect of ship design. Roughly 70%-80% of the value chain of a ship contract will go to vendors.

In looking at vendor information, I feel the role of classification societies - independent regulators of risk (overview) is essential to understand how the commercial market functions. A marine vendor wanting to introduce a piece of equipment will typically seek class approval for the equipment, starting with an initial approval-in-principal, providing an outside assessment of the risk and maturity of the equipment. This is relatively cost-effective as the class societies have built up a large knowledge structure around this process through classing thousands of pieces of machinery. Access to this large, shared knowledge structure provided by a third party at an affordable rate helps all parties - the vendor gains feedback on their design and, with class approval, makes it product more attractive, and the risk to the shipowner and shipyard is reduced.

There is no such equivalent for many government vessels, especially where different standards are required from commercial class approval. There isn’t the equivalent of an approval-in-prinicple stage and the market sizes are vastly different - Europe alone has 22,000 marine equipment manufacturers, with an annual output of ~$75 billion, roughly twice the entire proposed USN shipbuilding budget. This makes investing in new equipment higher risk for government-focused vendors, and prevents a clear route from doing so without an active aquistion program. The result is a less robust market when VFI is needed in ship design.

Make risk-based decisions to off-ramp design features

Minimize and isolate changes to existing designs

Carefully manage design innovation

I’ve grouped these three together, as many of the points that apply here have already been mentioned. Risk-based decisions to off-ramp features is difficult to do without a single (money) metric for success - the same problem as presented in “Maintain a sound business case through continued reevaluation”. As design programs have become bigger, with higher levels of innovation, they have also taken longer to complete. The extended timeframe has reduced the collective knowledge structure around existing designs significantly. The extended timeframe is also counter to minimizing changes to existing designs and limiting high-risk innovation as every new program attempts to incorporate decade(s) of progress in shipboard systems. There is a clear contrast with past acquisition practices, such as the late 1950s and 1960s when the USN designed and built several classes of frigates in short order, gaining experience with the design of this type of vessel.

GAO Theme Area #3 Use efficient ship design collaboration and decision-making practices

Use processes that support timely design decisions

Align decision-making with design maturity measures

The examples the GAO report provides here overlook significant portions of the knowledge structure accessible to commercial owners and shipbuilders. In stating that owners routinely return documents after a 10-21 day review period, it overlooks the role of classification societies in providing separate safety-oriented reviews beyond that of the owner. Class societies have built up an extensive knowledge structure around hull, machinery, and materials, promoted through comprehensive rule sets and international collaboration in setting standards through IACS and the IMO move to goal-base standards. They maintain large research staff to support this knowledge structure, actively leading efforts at IACS, ISSC, etc. Class societies, especially when looking into alternative compliance approaches, are going to take more than 10-21 days of working with the shipyard to approve a design. However, here, the commercial industry benefits from economies of scale - the class rules, procedures, and standards are clearly set, frequently applied, and supported through class fees paid by thousands of vessels. Class’s central role in approval also promotes design maturity before construction.

This contrasts with the situation for federally procured vessels. The government is normally both the customer and the final regulatory approver. Early-stage (pre-contract) interaction with shipyards is typically more difficult owing to competition and contracting regulations. Additionally, with a far slower throughput of vessels, U.S. yards often feel compelled to start construction to maintain work for the skilled trade shop floor workforce, even if the design is still not fully stable. Losing a large portion of the skilled trade employees from a production gap would be fatal to most shipyards - probably a higher risk on average than proceeding with an unstable design.

GAO Area #4: Employ robust in-house ship design capabilities and tools

Maintain strong in-house design workforce capabilities

NAVSEA’s 1990s move to a market-oriented system, including switching large parts of the organization to a Defense Working Capital Fund (DWCF) model, caused unintended damage here. Under this model, R&D and other services charge programs for their products to fund their employees. However, coupled with a low rate of new programs, the DWCF made it difficult to maintain a complete knowledge structure around naval construction. Much of the safety-critical aspects of this work have fallen to overburdened technical warrant holders in NAVSEA, while other design processes have been built up for specific programs but then faced difficult transition when programs end. The capital-funded model assumes that this in-house engineering force can pivot to support whatever is needed next without any direct funding provided for the maintenance of existing knowledge. Unfortunately, in reality, specialist knowledge is highly perishable if not frequently used, and once lost, it is difficult and expensive to re-create. There is also a clear shortage of qualified engineers domestically, as highlighted in previous posts.

Use ship design tools to shorten cycle time

Similar to people, tools have lagged under the market-based approach to running NAVSEA. R&D funding for new tools to address physics more accurately and to integrate it into the design process has excelled, and the U.S. R&D community remains strong. However, specialist knowledge is not a linear process where newly-created knowledge is appended to existing knowledge, which remains forever accessible and available. Investments in more mundane tools that do not promise S&T breakthroughs have lagged, as has training and documentation on such tools. Shipyards, typically building 1-2 new designs over several years (HII Pascagoula being a notable exception), do not have the same incentive as a commercial yard with 30 deliveries a year to invest in making these tool chains as efficient as possible.

Structural Issues Summary

While shortcomings in knowledge structure maturity and market forces cannot explain every difference between commercial and government shipbuilding noted in the GAO report, I believe the analysis above points to their central role in the current shipbuilding struggles. Commerical ship owners and builders exist in a much larger knowledge structure than U.S. government shipbuilding, large commercial shipyards typically turn through tens of a designs a year. Furthermore, they are supported by a large ecosystem of vendors, class societies, IACS, and IMO who perform vital knowledge management roles. These structures allow commercial designs to be developed rapidly and at low risk. In comparison, the U.S. government design throughput is much lower, and the surrounding knowledge structures are simply not as robust.

Commercial ships are also designed in a marketplace that provides a single, clear, metric for success, and rewards production efficiency. U.S. Government ship design and shipbuilding does not have this luxury, the required number of hulls is set by operational and budgetary needs of one customer, thus reducing the time and cost to build a ship will not always lead to more orders or significantly higher profits. The metric for success is also unclear, which makes it harder to determine when to reduce complexity or which features are really needed. Past design-to-cost ships have struggled to deal with real-world evolution in operational needs. Where market-simulating mechanisms have been applied, such as contracting and working capital funding models, they have unintentionally complicated the building and sharing of the knowledge structures discussed above.

Beyond the GAO report, there is more evidence to support this thesis. The past 20 years have shown signs of the struggle around knowledge structures - for example, the U.S. Navy’s move in the 2000s to work with a class society on developing naval vessel rules (along with the ASLP ), though the dynamics of the small market have made such tie-ups difficult. Attempts have also been made to access existing product structures (e.g. design artifacts like ship drawings) for new designs by requiring a “parent hull” design for the icebreaker, FFG-62, and other classes. However, as has been painfully learned, acquiring the product structure does not automatically re-create the associated knowledge structure used to design it. Furthermore, the highly sucessful NMSV program shows that domestic shipbuilding can deliver on-time and on-budget when the design allows access to the broader commercial knowledge structures and markets. In the NMSV case, this access was provided by using a commercial design manager, class society, and allowing the Korean design firm DSEC to complete the detailed design.

What can be done?

The differences between commercial and Government design highlighted in the GAO report are all still valid, and important topics to work on. Directly implementing commercial standards as an end-run around this problem is likely unworkable - government ships have unique operational needs, and most Navy designs (CG-47, FFG-7, DDG-51, etc.) have had at least one member of the class suffer combat damage in service. Some topics for discussion in addressing these structural factors include:

• Accepting that building up a knowledge structure takes time, and can look inefficient from a single-program viewpoint. Howevre, compared to the cost of delays, overruns, and mistakes, it makes sense to invest in such knowledge structures. NAVSEA and shipbulding personel will know what to prioritize here.

• Accepting that shortcuts around building up knowledge structures, such as parent hulls to skip directly to product structure (the topic of an upcoming post!) have a poor record of success.

• Think about risk assessment for design, both in terms of initial risk appetite and when to cut features. Explore new methods to address the shortcomings of system engineering and market measures for such trade studies.

• Given the small size of the raw number of designs likely to be produced, carefully review the market and contracting assumptions used today:

  • Does it make sense to contract for ongoing design capability in shipyards separately from programs of record?

  • How can we reward early deliveries with more work, not just better profit rates?

  • How can we avoid starting construction early to keep skilled trades employed?

  • Can we expand our work (which has occurred at a technical level since WWII) with allies to build these knowledge structures?

  • How do we keep contracts fair to all parties while maintaining a level workload going forward? How should industrial base stability factor into 30-year shipbuilding plans?

  • How do we better support vendors and get vendor information available earlier? They are a critical part of the shipbuilding puzzle, how do we make it easier for them to innovate and stay alive?

If we work on these issues, in another 10 years, hopefully, the GAO can write about something new related to Government shipbuilding.