The Shipbuilding Composite Index and Its Rates of Change Compared With Economywide Inflation Rates

Summary

In most years, the Department of Defense provides a 30-year shipbuilding plan to the Congress. The plan details the ships that the Navy expects to procure, along with an estimate of the costs of procuring those ships. The Congressional Budget Office is directed by law to provide an independent assessment of the costs of the Navy’s plan.1

CBO’s estimate of the costs of the Navy’s shipbuilding plan typically exceeds the Navy’s estimate. One reason is that when the Navy estimates the cost to build a ship in the future that is identical to one already built, it assumes that construction costs—in constant dollars—do not change. But the costs of labor and materials used in building naval ships have tended to grow faster than the costs of other goods and services in the economy. CBO accounts for that difference by comparing past increases in costs of building naval ships (that is, shipbuilding inflation)—using the Navy’s shipbuilding composite index (SCI), which has existed in its current form since 1980—with increases in costs of other goods and services in the economy (that is, economywide inflation), as measured by the gross domestic product (GDP) deflator. The agency regards the difference between shipbuilding inflation and economywide inflation as growth in the constant-dollar cost of building naval ships.2

The SCI is divided into numerous categories and sub­categories. Those individual components can have different annual rates of change. In this report, CBO describes how the differences between the largest categories of the SCI and the GDP deflator have changed over time. On average, all the major components have grown more quickly than the GDP deflator has. All told, between 1980 and 2022, the SCI grew an average of 1.2 percentage points faster per year than the GDP deflator did, although the average difference between 2015 and 2022 was 0.4 percentage points. Looking ahead, a gap between the SCI and the GDP deflator of roughly 1 percentage point would be consistent with historical experience, in CBO’s assessment.

Measuring Economywide Inflation

In theory, measuring inflation is straightforward. A representative bundle of goods is defined, and then the prices of those goods are compared at two or more points in time. A five-pound bag of flour is the same product in 2022 as it was in 2021 (or 2011 or 1981), so calculating the percentage change from the 2021 price to the 2022 price provides a measure of inflation over that period.

In practice, however, measuring economywide inflation is more challenging, in large part because many products are not identical from year to year. For instance, although the average price of cars sold in the United States has increased considerably, some of that price increase reflects the fact that newer cars have enhanced capabilities, such as backup cameras, that were not previously available. The Bureau of Labor Statistics (BLS) attempts to account for such improvements when it measures inflation.3 Because products are often improved over time, inflation rates are typically lower when a quality adjustment is undertaken.

Measuring Changes in the Costs of Building Navy Ships

The challenges in measuring rates of change in the prices of Navy ships are even greater. Though adjusting for the increase in the quality of modern cars is difficult, it is even harder to adjust for, say, the enhanced capability of a Ford class nuclear aircraft carrier relative to a Nimitz class nuclear aircraft carrier. Navy ships are more complex than cars, and new ships could be enhanced in many more ways. In addition, the Navy does not procure each type of ship every year, which further complicates estimating annual rates of change in prices.

Rather than measuring the prices the Navy pays for ships, the SCI tracks the costs of building ships. The SCI has no direct relationship with shipbuilders’ productivity; that is, it measures the costs that shipbuilders incur, not what or how much the shipbuilders produce. It does not adjust for changes in ships’ capabilities over time.

Estimates of the annual rate of change of the SCI are derived from an aggregation of annual rates of change in several categories associated with ship construction. CBO breaks the SCI into five categories. (The Navy breaks it into more granular subcategories; those are described later in this report.) CBO’s categories are as follows:

• Shipbuilder labor. The hourly costs of labor at the seven shipbuilders with which the Navy contracts.4
• Nonnuclear industrial machinery. The costs of materials used in the construction of conventionally propelled ships for which, according to the Navy, BLS’s general industrial machinery and equipment producer price index provides a reasonable proxy.
• Nuclear material. The costs of materials used in the construction of nuclear-propelled ships, as well as subcontractor labor purchased by shipbuilders that build them.
• Nonnuclear aeronautical instruments. The costs of materials used in the construction of conventionally propelled ships for which, according to the Navy, BLS’s aeronautical, nautical, and navigational instruments producer price index provides a reasonable proxy.
• Other nonnuclear material. The costs of all other material used in the construction of conventionally propelled ships.

For each fiscal year, each SCI category has both a weight and a rate of change (see Table 1). A category’s weight is its share of the total annual costs of shipbuilding. The combined weight of the five categories equals 100 percent each year, even though the weights of individual categories may change from year to year. That is, an increase in the percentage share of one or more categories must be offset by a decrease in one or more of the remaining categories. A category’s rate of change measures how much costs within a category have increased or decreased from the previous year. The annual SCI is a weighted average of all the composite categories: First, the weight and rate of change for each category are multiplied, and then those products are summed. The SCI increased by 7.17 percent in 2022 from the previous year, though the rates of increase in the costs associated with the other nonnuclear material and nonnuclear industrial machinery categories were much greater.

The weights of the five categories have varied from year to year, reflecting changes in the types of costs incurred in shipbuilding (see Figure 1). The relative weight of non­nuclear industrial machinery roughly doubled, from 17 percent to 35 percent, between 2007 and 2013. That increase came at the expense of shipbuilder labor, which declined from 48 percent to 39 percent, and nonnuclear aeronautical instruments, which decreased from 18 percent to 6 percent. Since 2013, the weight of the nonnuclear machinery category has declined to 26 percent, whereas the weights of shipbuilder labor and nonnuclear aeronautical instruments have increased, to 44 percent and 9 percent, respectively. The evolution of those weights reflects the Navy’s estimates of the proportional importance of each category in the total costs shipbuilders incur.

The nuclear material weight fluctuated in a fairly narrow range between 13 percent and 19 percent over the 2006–2022 period, even though about half of appropriations for new-ship construction went to purchase nuclear-propelled vessels over that period. In the SCI’s structure, any government-furnished material (GFM) is tied to a nonnuclear price index, even if it is used on a nuclear-propelled ship. Nuclear material weights correspond solely to material directly purchased by the two shipbuilders that build nuclear-propelled vessels—Electric Boat Corporation and Newport News Shipbuilding.

The Difference Between the SCI and the GDP Deflator

Between 1980 and 2022, the SCI increased an average of 1.2 percentage points faster than the GDP deflator did; the SCI’s rate of change was lower than that of the GDP deflator in only three years over that period (see Figure 2). The difference has shrunk in recent years, however: Between 2015 and 2022, the SCI increased an average of 0.4 percentage points per year faster than the GDP deflator did.

The compounded effect of the difference between the SCI and the GDP deflator is sizable. For example, $100 at the beginning of 1980 growing at the rate of the GDP deflator would be $328 at the end of 2022. The same $100 growing at the rate of the SCI would be $546 at the end of 2022, or 66 percent more.

CBO obtained category-level weights and rates for the SCI for 2006 through 2022 and therefore focused its analysis on that period. During those years, the SCI increased, on average, 1.1 percentage points faster than the GDP deflator—very similar to the average for 1980 through 2022. During the 2006–2022 period, each of the CBO-defined categories of the SCI, on average, grew more quickly than the GDP deflator (see Table 2).

• The hourly costs of shipbuilder labor increased 1.0 percentage points faster than the GDP deflator, on average.
• Costs of nonnuclear aeronautical instruments increased, on average, 0.9 percentage points faster than the GDP deflator, the lowest average rate of increase across the five categories.
• Other nonnuclear material had the greatest difference, increasing 1.5 percentage points faster than the GDP deflator, on average.
• Each of the five categories had at least one year with slower growth than the GDP deflator. The category with the highest variance, other nonnuclear material, had six years in which its rate of change was below that of the GDP deflator.

  Table 2.

  Trends in the Shipbuilding Composite Index Relative to the Gross Domestic Product Deflator, by Category

  Percent

  

  Notes

  Data source: Congressional Budget Office, using data from the Bureau of Economic Analysis and the Department of the Navy. See www.cbo.gov/publication/59026#data.

  The average difference is the observed rate of change in a category’s costs minus the annual change in the gross domestic product deflator

Notably, the SCI increased less than the GDP deflator did in 2019, driven by a decline in shipbuilders’ average labor costs per hour. (That labor rate is fully burdened, meaning that it includes shipbuilders’ overhead costs.) The change in labor costs refers to the rate of change in cost per worker, not the rate of change in total labor costs. That rate could decline for a variety of reasons, even as total spending on labor increased. For example, if a shipbuilder had a larger workload requiring more workers, fixed overhead costs would be spread over more employees, leading to a decline in average labor rates. If that shipbuilder hired additional junior employees who were paid less than more senior workers, the average labor rate would also be lower. In both cases, the labor weight would increase because total spending on labor would rise.

By scaling the differences for each category relative to the GDP deflator by the relative weight of each category in each year, CBO found that the shipbuilder labor category accounted for about 40 percent of the overall 1.1 percentage point average difference between the SCI and the GDP deflator observed between 2006 and 2022. Nonnuclear industrial machinery accounted for about 30 percent of the difference, nuclear material accounted for about 20 percent, and nonnuclear aeronautical instruments accounted for most of the rest. The other nonnuclear material category had a low average weight (about 2 percent), so its contribution to the SCI’s difference relative to economywide inflation was minor, despite its volatility.

Projecting the Difference Between the SCI and the GDP Deflator

Because the difference between the SCI and the GDP deflator has been relatively small in the recent past, CBO also explored whether that trend was more likely to persist or to revert to the greater historical difference.

To gain insight into future trends, CBO analyzed how the categories of the SCI have evolved over time relative to the GDP deflator. Although CBO only had a weight-and-rate category breakdown of the SCI back to 2006, the Navy provided CBO with estimates of the categories’ annual rates of change back to 1980.

Since 1980, all of the categories of the SCI have, on average, grown more quickly than the GDP deflator has. The SCI for shipbuilder labor grew more slowly than the GDP deflator in 2018, 2019, 2021, and 2022. The negative difference in 2019 was caused by a reduction in the average shipbuilder labor rate. In 2018, 2021, and 2022, average rates increased, but not as rapidly as the GDP deflator. Those differences resulted in a downward trend in the costs of shipbuilder labor relative to the GDP deflator. However, CBO anticipates that the difference will return to close to its historical average in the future, in part because of adjustments for economywide inflation in union labor contracts.

The other categories have not had statistically significant trends in their differences over time. That is, there is no evidence that their rates of change are converging to the GDP deflator’s rate of growth. In CBO’s assessment, therefore, it would be consistent with historical experience for a roughly 1 percentage point gap between the SCI and the GDP deflator to continue.

Components of the Shipbuilding Composite Index

The Navy breaks the SCI into 25 subcategories. Those subcategories are more granular than the categories CBO uses; for details on how they correspond, see Table 3.

This section provides further details about the Navy’s subcategories. First, the SCI divides costs into two broad categories: those borne by shipbuilders and those classified as government-furnished material. GFM includes all the items the federal government provides to shipbuilders in the process of building a ship.

Shipbuilders’ costs are then broken into three components: labor, nonnuclear material, and nuclear material. Labor costs are composed of direct costs, indirect costs, and a subcategory called plan costs (the costs of naval architects and other professional personnel involved in ship design). Although the labor subcategories cover different types of laborers, the SCI’s annual rates of change have been the same for all three subcategories.

Material costs represent the costs of goods and services that the seven shipbuilders buy from their suppliers and subcontractors. (Material costs include the suppliers’ and subcontractors’ labor costs.) Nonnuclear materials are those used to build conventionally propelled ships, and nuclear materials are those used to build nuclear- propelled ships. The same item could be both nonnuclear and nuclear material if it was used on both conventionally and nuclear-propelled ships. For example, Electric Boat Corporation, which manufactures nuclear- propelled submarines with Newport News Shipbuilding, purchases nuclear material because it builds only nuclear- propelled ships. By contrast, Ingalls Shipbuilding builds only conventionally propelled surface ships, so even if it purchased the same material as Electric Boat Corporation, Ingalls Shipbuilding’s purchases would fall into the nonnuclear material subcategories. Nonnuclear and nuclear material costs are each broken into nine ship work breakdown structure (SWBS) subcategories that correspond to different parts of ships (see Table 4).

The GFM category is also broken into SWBS subcategories, but only for subcategories 200, 400, 500, and 700. Taken together, the 25 subcategories aggregate to form the SCI (see Table 5). However, some rates of changes in costs typically apply to more than one subcategory:

• The plan costs, shipbuilder direct labor, and shipbuilder indirect labor subcategories shared one annual rate of cost change for each year between 2006 and 2022.
• In almost all years and subcategories, the GFM annual rate of cost change has been the same as the corresponding nonnuclear material rate of cost change. (Although GFM might be used on nuclear-propelled ships, in the calculation of the SCI, nonnuclear annual rates of change are used for all GFM.)
• Over the 2006–2019 period, the SCI incorporated the same annual rates of change from the BLS for nonnuclear materials in SWBS subcategories 200, 300, 500, 600, and 700.
• The annual rate of change for nuclear subcategory 900 (ship assembly and support services) material has generally been the same as that for nonnuclear subcategory 900 and nonnuclear subcategory 800 material.

In effect, 13 unique annual rates of change generate the SCI—1 for labor, 8 for nuclear material, and 4 for nonnuclear material (see Table 6).

---