Prepared by

Gunnar Knapp

Professor of Economics

University of Alaska Anchorage

3211 Providence Drive

Anchorage, Alaska 99508

907 786-7717 (telephone)

907-786-7739 (fax)

afgpk@uaa.alaska.edu (e-mail)

March 2000

This paper presents estimates of United States salmon production and consumption of salmon for the years 1990-1999. The estimates are for two kinds of products ("canned" and "fresh & frozen) and for six species of salmon (Chinook, Sockeye, Coho, Pink, Chum, and Atlantic).

Estimated U.S. consumption of fresh & frozen salmon increased substantially during the 1990s, reflecting growth in imports as well as consumption of domestic fresh and frozen production. In contrast, canned salmon consumption has varied from year to year reflecting changes in harvests, with no clear upward or downward trend.

The estimates of per capita salmon consumption show similar trends to National Marine Fisheries Service estimates for canned salmon and National Fisheries Institute estimates for all salmon. However, the estimates show much greater detail about the components of U.S. salmon consumption than is available from those other sources.

This paper presents estimates of United States salmon production and consumption of salmon for the years 1990-1999. The estimates are for two kinds of products ("canned" and "fresh & frozen) and for six species of salmon (Chinook, Sockeye, Coho, Pink, Chum, and Atlantic).

I estimated U.S. salmon consumption using a "disappearance" method. This method involved the following seven steps, for each species:

1. Estimation of U.S. salmon harvests. I estimated total U.S. salmon harvests (live or round weight) by adding data for Alaska wild salmon harvests, Lower 48 wild salmon harvests, and U.S. farmed salmon harvests.

2. Estimation of U.S. salmon production. I estimated canned salmon production weight (canned weight basis) from pack data reported by the National Food Processors Association. I estimated fresh & frozen production weight (H&G basis) from production data reported by the Alaska Department of Fish and Game and assumed yields from Lower 48 wild salmon harvests and U.S. farmed salmon harvests.

3. Estimation of Net Inventory Accumulation. I estimated canned salmon net inventory accumulation, or the change in inventories over the year, from canned salmon inventory data reported by the National Food Processors Association (NFPA). Because NFPA has not reported canned inventory data since 1997, I estimated canned salmon net inventory accumulation for the years since 1997 from pack data, based on the previous historical relationship between pack size and net inventory accumulation. I estimated frozen salmon net inventory accumulation from National Marine Fisheries Service data for frozen salmon holdings.

4. Estimation of U.S. consumption of U.S. salmon production. I assumed that U.S. salmon production which was not exported or accumulated in inventory was consumed in the United States. I calculated U.S. consumption of U.S. production by subtracting exports (as reported by NMFS) and net inventory accumulation from U.S. production, and then multiplying by assumptions about the yield of edible weight from production weight:

5. Estimation of U.S. consumption of imports. I estimated U.S. consumption of imports by multiplying U.S. imports (as reported by NMFS) by assumptions about the yield of edible weight from import weight.

6. Estimation of total U.S. salmon consumption. I estimated total U.S. salmon consumption by adding estimated U.S. consumption of U.S. production and estimated U.S. consumption of imports.

7. Estimation of U.S. per capita salmon consumption. I estimated per capita consumption by dividing estimated total U.S. consumption by United States resident population.

Keep in mind that the estimates are only as good as the data and assumptions that they are based on. For numerous reasons, actual U.S. salmon production and consumption may differ from these estimates. These reasons include:

1. The data sources on which the estimates are based are not necessarily accurate. Not all harvests, production, inventories, exports and imports are necessarily reported.
2. Some data were missing and had to be estimated. Canned pack data for 1997 were missing and were estimated based on Alaska production data. Canned salmon net inventory accumulation was estimated from pack data for years since 1997--which makes the canned consumption estimates less reliable for years since 1997. Alaska production data for 1999 were not yet available and were estimated by assuming the same product mix and yields as in 1998.
3. Processing and edible yield assumptions are not necessarily correct. The estimates assume standard yields for what are in fact a very wide variety of salmon products produced and consumed in the United States.
4. The estimates ignore exports and imports of salmon other than canned, fresh or frozen products. The estimates also ignore Alaska production other than canned, fresh or frozen salmon. However, this "other" production is still a very small share of total production.

In the remainder of this paper I briefly review what I consider to be some of the more interesting trends in United States salmon consumption suggested by these estimates.

United States harvests are dominated by Alaska harvests. United States harvests of Lower 48 wild salmon and farmed Atlantic salmon are very small in comparison with wild Alaska salmon harvests. (Figure 1)

United States harvests of all species vary significantly from year to year. (This contributes to significant year-to-year variation in U.S. consumption of U.S. production.) U.S. salmon harvests are dominated by sockeye salmon and pink salmon. United States sockeye salmon harvests declined significantly from 1996 through 1998, but rebounded in 1999. United States harvests of chum salmon rose during the 1990s, reflecting an increase in Alaska harvests of chum salmon released by Alaska hatcheries. (Figure 2)

Canned salmon continues to account for an important share of U.S. salmon production. United States canned salmon production varies substantially from year to year, reflecting variations in harvests of pink and sockeye salmon. (Figure 3)

Canned salmon accounts for the largest share of pink salmon production. A significant share of sockeye salmon harvests is also canned, although the largest share is frozen (and exported to Japan). Relatively little chum, coho, chinook or Atlantic salmon is canned. (Figure 4)

Most U.S. production of canned pink salmon is consumed in the United States, although significant volumes are exported. (Figure 5)

Most U.S. production of canned sockeye salmon is exported (primarily to the United Kingdom and former Commonwealth countries). (Figure 6)

Note that my estimates probably understate U.S. canned sockeye consumption. In some years reported exports exceed reported total canned pack and inventory reductions. For those years, estimated U.S. consumption of canned sockeye salmon is zero--although that is obviously not the case. This is a good illustration of why estimates such as these should be taken with a grain of salt--they show general trends but they are limited by imperfect data!

Canned pink salmon accounts for most of U.S. salmon consumption. U.S. canned salmon consumption varies substantially from year to year, reflecting variations in harvests and pack of pink and sockeye salmon. (Recall that the consumption estimates since 1997 are less accurate because they do not account for changes in inventories from year to year.) (Figure 7)

U.S. production accounts for almost all United States canned salmon consumption: U.S. imports of canned salmon are very small.

United States fresh & frozen salmon production varies substantially from year to year, reflecting variations in harvests. Sockeye salmon accounts for by far the largest share of United States fresh & frozen salmon production. (Figure 8) However, most of this fresh & frozen sockeye production is exported, and relatively little is consumed in the United States. As a result, other salmon species account for most United States consumption of United States production. (Figure 9)

The United States consumes significant volumes of U.S. production of fresh and frozen chum, pink, chinook, Atlantic, and coho salmon--in approximately this order of importance. The relative importance of these different species varies from year to year, reflecting variation in U.S. harvests. Total U.S. consumption of U.S. production of fresh and frozen salmon has risen in the 1990s, reflecting larger U.S. harvests--in particular of chum salmon. (Figure 10)

United States Consumption of Imported Fresh & Frozen Salmon

United States consumption of imported fresh & frozen salmon has risen dramatically in the 1990s, and continues to increase at a rapid rate. Throughout the 1990s, estimated U.S. consumption of imported fresh & frozen salmon has exceeded estimated U.S. consumption of fresh & frozen salmon from U.S. harvests. (Figure 11) United States consumption of imported fresh & frozen salmon is dominated by Atlantic salmon (almost all of which is farmed).

Estimated U.S. per capita canned salmon consumption has varied during the 1990s, reflecting variation in Alaska harvests and pack of pink and sockeye salmon. There is no obvious upward or downward trend over time. (Figure 12) (Recall that the per capita canned salmon consumption estimates since 1997 are less accurate because they do not account for changes in inventories. As a result, they may overstate per capita consumption in some years and understate it in others.)

Estimated U.S. per capita consumption of fresh & frozen salmon increased substantially during the 1990s. Most of this increase was due to growth in farmed salmon imports.

Estimated total U.S. per capita consumption of salmon increased substantially during the 1990s, primarily due to growth in farmed salmon imports.

My estimates of per capita salmon consumption (labeled as "UA" for "University of Alaska") show similar trends to National Marine Fisheries Service (NMFS) estimates for canned salmon and National Fisheries Institute (NFI) estimates for all salmon. Differences between the estimates are attributable to differences in data sources and assumptions. (Figure 13)

As time and resources permit, I hope to update and expand these estimates in the future. I would welcome any comments or suggestions as to how to improve these estimates or to make them more useful. Please feel free to contact me at afgpk@uaa.alaska.edu.

I have posted more detailed information about this analysis on the world wide web. A slightly different version of this paper may be found at:

www.iser.uaa.alaska.edu/iser/people/knapp

Linked to this paper are graphs providing additional details of the analysis, tables with the data shown in the graphs, and a discussion of data sources used for the analysis.

Back