The role of the Alaska seafood industry: a social accounting matrix (SAM) model approach to economic base analysis

Abstract

A social accounting matrix (SAM) model for Alaska is constructed to investigate the role of the state’s seafood processing industry. The SAM model enables incorporation of the unique features of Alaska economy such as (a) the existence of a large nontraditional economic base, (b) a large leakage of labor income, and (c) a very large share of intermediate inputs imported from outside the state. The role of an industry in an economy with these features cannot be examined correctly within an input–output framework, which is the method most often used for examining the importance of an industry to a region. Taking an export base view of the economy, we found seafood processing to be an important industry, generating 4.5% of the state’s total employment. While an important driver of the state’s economy, the industry has the smallest SAM multiplier mainly due to a large leakage of labor earnings and a large share of imported intermediate inputs. We also found that nontraditional economic base components such as (a) federal transfers to state and local governments, and (b) federal transfers, permanent fund dividend (PFD) payments, and other extra-regional income received by households generate about 26% of the state’s total employment and earnings.

1 Introduction

Magnuson–Stevens Fishery Conservation and Management Act National Standard 8 (MSFCMA Section 301[a][8]) requires that fishery conservation and management actions consider the importance of fishery resources to fishing communities. Therefore, fishery managers are interested in analyses that estimate the importance of the seafood industry to a region.

This study examines the role of the seafood industry in the Alaska economy using a social accounting matrix (SAM) model. This study has several contributions. First, unlike most previous input–output (IO) studies measuring the importance of an industry to a regional economy, this study uses a SAM model instead of an IO model. The SAM model enables one to incorporate the unique features of Alaska’s economy such as (a) the existence of a large nontraditional economic base, including exogenous income transfers to households (e.g., permanent fund dividend (PFD), federal government transfer payments, and other exogenous income) and exogenous revenue sources for state and local government (e.g., federal government transfers); (b) a large leakage of labor income, especially from the resource-based industries including seafood industry; and (c) the fact that much of the intermediate inputs used by Alaska industries are imported from outside of the state. The role of an industry in an economy with these features cannot be examined effectively (correctly) within an IO framework. Furthermore, within a SAM framework it is possible to address distributional effects across different types of households and institutions. Finally, this study integrates various important data (such as nonresident employment and earnings data, PFD data, and Census data) into the Alaska IMPLAN¹ (IMpact analysis for PLANning, Minnesota IMPLAN Group) data set.

Seafood is an important industry in Alaska. In 2002, about 5.1 billion pounds of fish and shellfish were harvested in waters of Alaska with an ex-vessel value of about $812 million (National Marine Fisheries Service (NMFS) 2003). In the same year, groundfish accounted for 58% of the ex-vessel value; shellfish, 15%; salmon and halibut, 13% each; and herring, 1% (North Pacific Fishery Management Council (NPFMC) 2003). Over the past 5 years, about 23.9 billion pounds of seafood have been harvested in waters off Alaska (NMFS, various years). In 2002, 54%, by weight, of the total U.S. commercial fishing harvest came from Alaska (NMFS 2003).

According to the Alaska Department of Fish and Game (ADFG 2001), Alaska’s seafood industry, including both harvesting and processing sectors, translates into more than 16% of the state’s basic sector employment, and more than 47% of private basic sector employment (ahead of oil and gas, mining, forest products, and tourism). Seafood is Alaska’s top international export with fish products representing approximately 40% of these exports. In 2001, the Alaska seafood industry directly accounted for about 3.1% of total state employment of 287,941 jobs, and about 2.2% of $10.3 billion total state earnings (Alaska Department of Labor and Workforce Development (ADOL 2005)).²

2 Innovative features of this study

In this study, we use economic base (or export base) concepts to measure the importance of the seafood industry within a SAM model framework. Economic base theory states that the economic base determines a region’s total level of economic activity. Sectors that export a large share of their production are called basic sectors. These sectors bring in revenue from outside the region, which is spent and re-spent on goods and services produced by other nonbasic sectors within the economy, generating multiplier effects. Basic sectors in Alaska include the seafood industry, oil and gas mining, other mining, forest products, tourism, and the Federal government. More detailed discussion of economic base theory can be found in North (1955), Tiebout (1956), and Richardson (1973). Excellent critical reviews of economic base models may also be found in Krikelas (1992).

To examine the role of an industry in a regional economy, IO models are often used. Although IO models have well known shortcomings such as fixity of prices and non-substitutability of inputs in production and goods in consumption,³ IO models are useful for tracing the impact of a basic sector on nonbasic sectors of the economy via interindustry linkages. Leones et al. (1994) identified 27 state-level studies conducted from 1987 to 1994 that examined the role of agriculture in state economies. About half of those studies used IO models. More recently, Sharma et al. (1999) investigated the importance of agriculture in Hawaii using an IO framework.⁴ However, there are significant disadvantages to using IO models to examine Alaska’s economy. First, in Alaska, Federal transfer payments and PFD payments to residents are important components of the economic base. In 1998, about 13.5 and 5.1% of the total personal income of Alaska residents was from Federal government transfers and the PFD, respectively (U.S. Department of Commerce). These transfers serve as a major source of personal income. However the traditional IO framework is not able to account for these nontraditional elements of the economic base. A SAM model does allow incorporation of these nontraditional components of Alaska’s economic base.

A second major problem with the IO accounts is that they do not trace factor payments or indirect business taxes to their respective institutional accounts. As a result, it is impossible to close an IO model with either household spending or government spending as endogenous variables.⁵ The Alaska economy is characterized by a large leakage of labor income since a large proportion of workers in some Alaska industries are nonresidents. For example, in 1998, nonresidents accounted for about 19.5% of total private and state and local government employment in Alaska. Consequently about 11.3% of the total labor income earned in private and state and local government sectors leaked out of the state. Leakages of labor earnings are highest in seafood processing (59.5%), commercial fishing (32.0%), oil and gas mining (26.4%), air transportation (24.7%), water transportation (22.2%), and lumber and wood products sectors (22.1%) (ADOL 2000). The SAM model developed for this study is able to incorporate the effects of labor income leakage by sector.

The third advantage of using a SAM model over an IO model is the SAM’s ability to address distributional effects. Within an IO framework, it is not possible to address distributional impacts across different types of households and institutions since these accounts are usually exogenous. The accounts in the Alaska SAM model help trace factor payments to institutional spending accounts by place of residence.

In sum, the SAM model developed for this study (1) includes the nontraditional sectors of the economic base, (2) incorporates leakage of labor income, and (3) addresses the distributional effects of the basic sectors on regional households and government. In the SAM model, expenditures by regional households and state and local government are treated as endogenous variables. Household income is assumed to drive household consumption. State and local government revenues are assumed to drive state and local government expenditures. These features of the Alaska SAM model represent a methodological improvement over previous IO studies such as those reviewed in Sharma et al. (1999) and many of the studies reviewed in Leones et al. (1994). The SAM approach used in this study is similar to Waters et al. (1999), which used a SAM to analyze the role of agriculture in Oregon. However, the present study differs from that study in that the Alaska SAM model was designed to address an economy that is characterized by (a) the largest seafood industry of any state, (b) very large exogenous income payments to households, (c) large leakage of labor income, and (d) large imports of key inputs used by industries.

Another noteworthy feature of this study is that data from many different sources were combined with the Alaska IMPLAN SAM data set. For example, this study incorporated (a) ADOL’s nonresident employment and earnings data to capture the leakage of labor income for each industry, (b) Alaska Department of Revenue’s (ADOR 2003) PFD data to estimate exogenous income payments to households, and (c) Census data on the number of households by income level and average household size. In addition, this study employed a realistic assumption about the source of other property income, following the treatment by Waters et al. (1999).

Section 3 below discusses how the 1998 Alaska SAM model was constructed. Section 4 explains the data methodology. Section 5 presents and discusses the results. Section 6 offers some conclusions.

3 The Alaska SAM model

A standard IO model includes intersectoral flows of intermediate inputs, and so captures a major source of economic linkage in an economy. However, the standard IO model ignores the income that flows from producing sectors to factors of production (value added), and then on to entities such as government and households, which generates endogenous demand for goods and services. A SAM model captures these flows in detail. A SAM can also investigate the distribution of factor income to various types of institutions and households. For a more detailed discussion of a SAM and SAM models, see King (1985), Holland and Wyeth (1993), and Adelman and Robinson (1986).

In this section, a 1998 Alaska SAM model is presented. This section draws on Holland and Wyeth (1993), Adelman and Robinson (1986), and Waters et al. (1999). Table 1 shows the structure of the Alaska SAM. In the Alaska SAM, there are a total of 46 accounts—42 endogenous accounts and four exogenous accounts. The 42 endogenous accounts include 29 industry accounts, three value added accounts, nine household accounts (nine different income categories of households), and a state and local government account. The four exogenous accounts are PFD, Federal government, capital, and the rest of the world (ROW).

Table 1 Structure of the 1998 Alaska SAM

Following procedures used in Holland and Wyeth (1993) and Adelman and Robinson (1986), the matrix of direct coefficients in the Alaska SAM model, denoted S, is derived as follows:

$$ {\text{S}} = {\left[ {\begin{array}{*{20}l} {{\text{A}} \hfill} & {0 \hfill} & {0 \hfill} & {C \hfill} & {{{\text{GD}}} \hfill} \\ {{\text{V}} \hfill} & {0 \hfill} & {0 \hfill} & {0 \hfill} & {0 \hfill} \\ {{{\text{IBT}}} \hfill} & {0 \hfill} & {0 \hfill} & {0 \hfill} & {0 \hfill} \\ {0 \hfill} & {{\text{F}} \hfill} & {0 \hfill} & {{{\text{IHT}}} \hfill} & {{{\text{STR}}} \hfill} \\ {0 \hfill} & {{{\text{SF}}} \hfill} & {{{\text{BTS}}} \hfill} & {{{\text{HTX}}} \hfill} & {{{\text{IGT}}} \hfill} \\ \end{array} } \right]} $$

(1)

where:

S:

matrix of SAM direct coefficients

A:

matrix of technical coefficients

V:

matrix of primary factor payments coefficients

IBT:

matrix of indirect business tax coefficients

F:

matrix of factor payment to household coefficients

SF:

matrix of state and local factor tax coefficients

BTS:

matrix of state and local indirect business tax coefficients

C:

matrix of household consumption coefficients

IHT:

matrix of interhousehold transfer coefficients

HTX:

matrix of state and local direct household tax coefficients

GD:

matrix of state and local government demand coefficients

STR:

matrix of state and local government transfer coefficients

IGT:

matrix of intergovernmental transfers

Then the SAM model can be represented as:

$$ {\left[ {\begin{array}{*{20}l} {X \hfill} \\ {V \hfill} \\ {{IBT} \hfill} \\ {H \hfill} \\ {{SG} \hfill} \\ \end{array} } \right]} = {\left( {I - S} \right)}^{{ - 1}} {\left[ {\begin{array}{*{20}l} {{ex} \hfill} \\ {{ev} \hfill} \\ {{et} \hfill} \\ {{eh} \hfill} \\ {{eg} \hfill} \\ \end{array} } \right]} $$

(2)

where:

X:

vector of industry regional output

V:

vector of total primary factor payments

IBT:

indirect business tax payments

H:

vector of total household income

SG:

total state and local government income (revenue)

ex:

vector of exogenous demand for regional output

ev:

vector of exogenous factor payments

et:

exogenous indirect business tax payment

eh:

vector of exogenous federal transfers and PFD payments to households

eg:

federal transfers to state and local government

Here (I–S)⁻¹ is called the SAM multiplier matrix or matrix of SAM inverse coefficients.

In the SAM model represented by Eq. 2 above, the endogenous variables are X, V, IBT, H, and SG. The exogenous variables are ex, ev, et, eh, and eg. There are three non-zero exogenous demand vectors: ex, eh and eg. The components of ex are final demand generated by investment, federal government and exports. The components of eh include Federal transfers to households, PFD payments to households, and financial returns from capital holdings outside Alaska. The components of eg include Federal transfers to state and local government; income from leases, trusts, and investments; and taxes paid by non-residents. Injections of income into the region occur through final demand components in ex and extra-regional payment components in eh and eg. Leakages include factor income payments to nonresident factor owners, federal taxes, savings and payments for imports.

We define the employment dependency index for a sector as the share of total state employment that depends on the sector’s economic activity. The employment dependency index measures the relative contribution of each sector or industry to total state (regional) employment. Specifically, the employment dependency index is derived as follows. By multiplying the SAM multiplier matrix, (I–S)⁻¹ by the diagonalized vector of exogenous demand, ED, the matrix of total impacts on output, Z, is obtained:

$$ {\text{Z}} = {\left( {{\text{I}} - {\text{S}}} \right)}^{{ - 1}} {\text{ED}} $$

(3)

Next, each row of Z is multiplied by the corresponding sectoral employment-to-output ratio to generate a matrix E of impacts on total employment attributable to each sector. The column sums of E represent total employment associated with a given sector’s exports or receipts of exogenous payments. These sums were divided by total state employment to obtain employment dependency indices, which measure the relative contribution of each component (sector or industry) of the economic base (Waters et al. 1999). The labor earnings dependency index is derived in a similar way. However Z is multiplied by the corresponding sectoral earnings-to-output ratios. The earnings dependency index for a sector measures the relative contribution of each sector or industry to the total labor earnings in the state.

4 Data methodology

For this analysis, 1998 IMPLAN data were used.⁶ The 528 IMPLAN sectors were aggregated into 29 industrial sectors. The sectoral aggregation scheme used is available on request. Information on nonresident earnings was obtained from ADOL (ADOL, 2000). The ADOL data include information on nonresident employment and earnings for each of 77 SIC-based industries (75 private industries and state and local government industries). Leakage of Federal employees’ income was estimated residually by subtracting the total nonresident labor income in the ADOL data ($927.2 million) from the total leakage of labor income shown in the IMPLAN data.

In general, SAMs derived from IMPLAN data tend to overestimate respending of other property-type income generated in a regional economy. It seems likely that payments by Alaska corporations to their creditors and shareholders would tend to leave the region, especially for a sparsely populated state like Alaska, and that the dividends, interest and rent received by Alaska residents would be more likely to come from their holdings of worldwide assets than exclusively from local companies. But the initial SAM derived from IMPLAN shows all of the other property income and enterprise payments to households originating in the state. Assigning so much local property income to resident households will overestimate economic multipliers. Therefore, following the treatment in Waters et al. (1999), in industries where the proprietors’ income (assumed paid to residents) is a relatively large component of value-added, we assumed that other property income also accrues state residents. In other words, if the capital stock in a given industry seems likely to be owned by Alaska residents (as evidenced by a large proportion of proprietor earnings), then that entire sector’s other property income, net of depreciation allowances and retained earnings, was allocated to Alaska households. Otherwise, other property income was assumed to be leaked. This is in contrast to a Type II IO model closure, where only returns to labor and proprietors are considered endogenous. In the Alaska SAM it was also assumed that there was an inflow of other property income just sufficient to balance total expenditures by the institutional accounts, minus any regional other property income received.

Data on the PFD payments was obtained from the ADOR (ADOR 2003). In 1998, total PFD payments to Alaska residents were about $870.7 million. To allocate the total PFD payment to nine different types of households, we used Census information on the number of households by income level and average household size for the state of Alaska (U.S. Census Bureau 2001).⁷ Steps followed to balance the Alaska SAM accounts are available on request.

5 Model results

Tables 2, 3, and 4 present the results from the Alaska SAM model. In presenting the results, 21 of the original 29 industries are aggregated into six larger industries so that the total number of industries shown in the tables is 14 (six aggregated industries plus the remaining eight industries). Also, in Tables 2, 3, and 4, households are aggregated into a single household sector to summarize the results.

Table 2 Sectoral output and exogenous demand

Table 3 Sectoral and export-dependent employment

Table 4 Sectoral labor earnings and export-dependent labor earnings

Table 2 shows base-year sectoral output and exogenous demand. The second column in Table 2 shows that in 1998 services accounted for 16.3% of the total state output of $34.7 billion, followed by TCPU (Transportation, Communication, and Public Utilities, 13.5%) and FIRE (12.5%). Federal government and state and local government accounted for 8.7 and 7.7% of total state output, respectively. The seafood processing industry accounted for 3.8% of total output. However examining the composition of Alaska’s export base, as shown in the last two columns of Table 2, tells a somewhat different story about the importance of individual sectors. The third column in the table shows the export base for the state of Alaska—exogenous demand for the output of the 14 industries, and extra-regional payments to households and state and local government. According to the last column in Table 2, extra-regional payments (such as federal transfers, PFD payments, and capital income) to households constitute the largest share (18%) of Alaska’s export base. Federal government payments account for an additional 12.5%, TCPU accounts for 12.1%, and exogenous state and local government revenue accounts for 11.2% of the state’s total economic base. Table 2 also shows that the next most important sectors (industries) in driving the state’s economy are mining (10.4%) and construction (8.9%). The seafood processing industry constitutes 5.4% of the state’s economic base while exporting about 98% of processed seafood. The two non-industrial sectors—the combined households sector and state and local government—are two of the most important economic base sectors.

The contribution of an industry is often measured by the number of workers it employs. Table 3 shows base-year sectoral and export-dependent⁸ employment. The second column in Table 3 presents base-year employment for 14 industries in terms of their percentage of the total state employment of 392,193 jobs. Services ranks first, employing 27.0% of total employment, followed by trade (17.8%), state and local government (12.4%), Federal government (11.2%), and TCPU (7.1%). The seafood processing industry employs 2.3% of total workers in the state.

The last four columns in Table 3 show export-dependent employment for the 14 industries and other endogenous sectors in the model. The Federal government, which ranks fourth if measured by number of employees, ranks first in terms of export-dependent employment as shown in the last column in the table. This means that the largest portion (15.7%) of total employment in the state is dependent on the Federal government. Exogenous state and local government revenue is the next most important sector, accounting for 14.1% of total state employment. Household spending of exogenous income accounts for 12.3% of total state employment. Thus the combined household sector is the third most important employment-generating sector in the state, after the Federal government and state and local government. The next two most important sectors are TCPU (10.7%) and construction (9.8%). The seafood processing industry, which employs 2.3% of total workers, becomes much more important if measured by export-dependent employment: 4.5% of total state employment is attributable to the seafood processing industry. Table 3 shows that over all, more than 26% of total state employment is dependent on nontraditional economic base—exogenous income received by households and state and local government.

Table 4 presents base-year labor earnings by industry and export-dependent earnings by sector. The first column in Table 4 shows labor earnings for the 14 industries as a percentage of total state labor earnings of $12.6 billion. The highest share of earnings is from state and local government (18.7%), followed by services (17.7%), Federal government (16.2%), and trade (12.2%). Earnings in the seafood processing industry constitute 2.1% of total earnings in the state.

From an export-base view, the four columns on the right of Table 4 show the largest share (19.7%) of total export-dependent labor earnings in the state is attributable to the Federal government, followed by the state and local government (16.5%), TCPU (11.4%), and construction (9.9%). Exogenous income received by households generates 9.4% of the total state labor earnings; households are the fifth most important sector in generating labor earnings. Table 4 shows that about 26% of total state earnings are attributable to exogenous income received by households and state and local government.

Labor earnings dependent on the seafood processing industry account for 3.1% of total state earnings; this is slightly greater than the sector’s base-year earnings share of 2.1%. The earnings dependency index for the seafood processing industry (3.1%) is smaller than its employment dependency index (4.5%, Table 3). This implies that the seafood processing industry has somewhat larger employment-generating capability than earnings-generating capability, as evidenced by a higher ratio of indirect and induced effects to direct effects for employment than for earnings (Tables 3 and 4).

Although not reported here, the results from the Alaska SAM model showed that the seafood processing industry’s output multiplier of 1.73 is among the smallest multipliers in the Alaska SAM model. This is because much of the labor income earned in, and expenditures made by, the industry leak out of the state. As was mentioned earlier, 59.5% of labor earnings from seafood processing flow out of the state. In addition, a large portion of intermediate inputs used by the industry is imported; according to IMPLAN data, 69.3% (by value) of intermediate inputs used in 1998 by the seafood processing industry were imported (compared to an average of 43.3% (by value) for the whole state economy). The main commodity imported by the seafood processing industry was raw fish—those fish caught by catcher vessels owned by nonresidents but landed for processing in Alaska.⁹

Summarizing the role of seafood industry in Alaska, in 1998, seafood processing accounted for $1,312 million in output (3.8% of total output, Table 2), 9,073 jobs (2.3% of total employment, Table 3), and $269 million in labor income (2.1% of total labor income, Table 4). In the same year, total export demand for seafood was $1,285 million (5.4% of total exogenous demand, Table 2), which contributed to 17,820 jobs (4.5% of total employment, Table 3), and $396 million in labor income (3.1% of total labor income, Table 4). About half of employment (Table 3) and a third of labor income (Table 4) generated by export demand for seafood were indirect and induced effects. These results indicate that the seafood industry is heavily linked with other sectors in Alaska.

6 Conclusions

This study employed a SAM model to examine Alaska’s seafood processing industry. The SAM model took into account leakage of labor income from each aggregated industry, and endogenized household and state and local government accounts. We found that seafood processing accounts for 5.4% of the state’s total economic base. Although the seafood processing industry employed only 2.3% of total state workers in the base year, 4.5% of the state’s total employment depends on the industry. While seafood processing is an important driver of the state economy, the industry has a relatively small SAM multiplier since much of the labor income earned in the industry leaks out of the state, and much of the intermediate inputs used in the industry are imported.

We also discovered that payments from the Federal government are very important in generating employment and earnings in the state. Private industries found to be major drivers of the state economy include TCPU, construction, trade, and mining. Non-traditional economic base components such as (1) Federal transfers to state and local government, and (2) Federal transfers, PFD payments, and other extra-regional income payments to households, are two of the most important sectors in generating employment and earnings in the state. Exogenous revenue of state and local government generates more jobs and labor earnings than does the exports of any private industry in the state. Overall, more than 26% of total state employment and earnings are dependent on nontraditional economic base.

This study also demonstrates the use of a SAM to model the unique features of the Alaska economy. These features cannot be effectively modeled using IO models. This study helps fill a gap caused by the absence of a systematic evaluation of the role of the seafood industry in an economy as unique as those Alaska’s, and can serve as a valuable tool for Alaska fisheries policymakers. This model can also be used for economic impact analysis. Most fishery management policies result in a change in landings, which may lead to a change in the amount of fish processed in a region. The type of model used in this study can be utilized to calculate the economic impacts of those policies on a region, providing more complete results than an IO model constructed using unrevised IMPLAN data.

A few caveats are in order concerning use of IMPLAN data. First, IMPLAN underestimates employment in the commercial fishing and seafood processing industries. IMPLAN data exclude many crew members and fishermen because they are self-employed or are casual or part-time workers and therefore are not covered by state unemployment insurance. To the extent that this employment is underestimated, the results from this study may also be underestimated. To improve the results, it would be necessary to obtain primary data on employment, earnings and costs for harvesting and processing industries through detailed industry surveys. Future planned research efforts will conduct such studies for subregions in Alaska that are heavily dependent on fisheries. Second, IMPLAN data assumes a national-level production function for regional industries, including fishery industries. But production technologies used by Alaska industries may differ from the national average. To the extent that the national technical coefficients do not correctly describe the Alaska economy or Alaska seafood industry, the results presented in this study may be biased. However since the Alaskan harvest comprises about half (by weight) of total U.S. harvest, any bias with respect to technologies used in seafood harvesting and processing sectors may be less for Alaska than for many other states. Third, IMPLAN uses national average household demand functions. IMPLAN regional personal consumption expenditure (PCE) estimates are based on the consumer expenditure survey (CES) and national income and product accounts (NIPA). Since many communities in Alaska are very remote, and subsistence activities are not generally included in the data, estimates of national average consumption behavior may not be very representative of Alaska households. Finally, IMPLAN regional purchase coefficients (RPCs), which determine the proportions of regional use of regionally produced output and imports, are estimated econometrically using a set of standard equations and regional data. Since the Alaskan economy is very different than the “average” U.S. regional economy, the RPCs and resulting trade flows may not be entirely representative of the Alaskan economy. However, short of doing an extensive primary data survey of Alaskan industry, there does not seem to be a practicable solution to this issue.