North American Journal of Fisheries Management 10:269-278, 1990© Copyright by the American Fisheries Society 1990

Resource Scarcity and Inequality in the Distribution of CatchCOURTLAND L. SMITH

Department of Anthropology. Oregon State UniversityWaldo 238. Corvallis. Oregon 97331-6403, USA

Abstract.—What happens to the distribution of catch when resources become scarce? Catchdistributions for most fisheries where the stock is limiting are skewed such that the largest numberof fishers take the smallest catches. Economic efficiency goals result in suggestions to eliminatesmall, inefficient producers. Oregon salmon fishing and bottom trawling, plus two nonfisheryexamples, illustrate that catch distributions become even more skewed with increasing resourcescarcity. A simple model shows how increases in the number fishing and improvements in technicalefficiency contribute to the creation of a distributional pattern in which a very few fishermen arevery successful, whereas most other fishermen take increasingly smaller shares of the resource.Fishery management rules that apply equally to each participant in the fishery result in greaterrelative advantage for some producers and increase inequality.

In fisheries requiring management, fishing effortneeds to be controlled to protect the stock fromoverexploitation. Economic analysis of fisheriesmanagement points to the greater inefficiency thatusually results from restrictions on fishing effort.Management rules such as those limiting effort onthe basis of seasons, open and closed periods, gearrestrictions, and area limitations make fishermenless efficient because the rules cause equipment tobe left idle or not used to full capacity. In additionto efficiency effects, management also affects thedistribution of catch.

The general view is that the people hurt mostby limitations on effort are the large, efficient pro-ducers. Because of their larger number, small pro-ducers are said to use the political process to gainan advantage. For example, Karpoff (1987) said,"It appears that regulations, particularly capitalconstraints and season closures, convey distribu-tional advantages to politically dominant fisher-men at the expense of their more efficient com-petitors." KarpofF cited studies by the AlaskaCommercial Fisheries Entry Commission (1982)and Knapp (1984), which showed that regulationsgive advantages to smaller, resident fishing oper-ations.

The catch distribution effects between large andsmall producers are complex. For example, the top10% of the fleet in the Pacific groundfish fisherycaught 43% of total fleet landings in 1981 and 44%in 1982 (Huppert and Korson 1987); these valuesrepresent common observations. Most catch dis-tributions are positively skewed. As a resource be-comes scarcer and management more intensive,how does the distribution of catch between vesselschange? Management rules that reduce the effi-ciency of larger, more efficient fishing vessels im-

ply that more catching opportunities go to small,less-efficient vessels. When the fish stock has to bedivided among a larger number of fishermen oramong fishermen who are becoming more tech-nically efficient, how does the catch distributionchange to reflect gainers and losers? What is theeffect of fishery management regulations to restricteffort? The first step in answering these questionsis to investigate the hypothesis that effort restric-tions create a disadvantage for larger, more effi-cient producers.

Measuring InequalityOne way to represent a catch distribution is with

a histogram. Figure 1 shows the income from catchby Columbia River gill-netters in 1899. Figure 2gives the distribution of income to Oregon salmonfishermen in 1971 (Lewis 1973). The two histo-grams differ considerably in shape. The 1899 his-togram shows a smaller range and is only slightlyskewed. Note the much greater range and far great-er skewing for the 1971 catch distribution. The1971 histogram, because of a much larger numberof people being at the low end and a much greatergap between the top and bottom of the distribu-tion, shows greater relative inequality.

When the variation in distributions becomesgreat, histograms, like those in Figures 1 and 2,are impractical to plot. More common is the useof cumulative frequency curves, Lorenz curves,logarithmic plots, or percentiles. To measure in-equality in resource catch distributions, one needsto represent frequency curves with an index thatallows comparison between distributions. Manyinequality measures are available that show changesin patterns of inequality. The most commonly usedare percentiles and Gini coefficients calculated from

269

270 SMITH

35-

25-

ID-

25 225llM

425 625 825 1025 1225Value of Catch ($)

FIGURE 1. —Distribution of income (in US$) for Co-lumbia River gill-netters, 1899.

Lorenz curves. Many other measures of equalityhave been proposed and used (Bronfenbrenner1971;Champernowne 1973; Atkinson 1975;Cow-ell 1977). Hibbs and Dennis (1988) suggested cal-culating the ratio of the amounts held by the top20% versus those held by the bottom 40%. For theanalysis in the present study, the Gini coefficientshows changes in relative inequality. A Gini co-efficient gives a numerical measure of the shapeof a catch distribution plotted as a Lorenz curve.

A Lorenz curve gives a picture of all the datain a compact space. The measure of equality in a

Lorenz curve is a frequency curve in which ev-eryone is the same: perfect equality. A Lorenz curveplots the cumulative percent of people along theA'-axis. The cumulative percent of the fish caught,dollars earned, or other resource allocation is plot-ted along the y-axis. Because the percentage of anitem controlled is plotted against the percentageof people, a Lorenz curve is a relative measure. A45° line running from the origin to the upper rightcorner of the graph is the line of perfect equality;it is when everyone has the same amount of theitem being measured.

Numerical representation of equality in a Lo-renz curve is the Gini coefficient. Taking the ratioof the area between the actual distribution and the45° line divided by the total area under the 45°line gives the Gini coefficient. The more the catchcurve departs from perfect equality, the greater theGini coefficient. Complete inequality is 1.0. In Fig-ure 3, perfect equality with a Gini coefficient ofzero is represented by the 45° line from the originto 100%. Dividing the area between the line ofperfect equality and the catch curve for ColumbiaRiver gill-netters in 1899 by the total area underthe line of perfect equality (Figure 3) gives a Ginicoefficient of 0.31. For Oregon salmon fishermenin 1971, the Gini coefficient is much higher, 0.74.Therefore, as the Gini coefficient gets smaller itapproaches perfect equality, which is zero.

A Gini coefficient is a relative measure of in-

4000

3500

3000-

2500-

E 2000-13

1500-

1000-

500-

2500 22500 42500 62500 82500 102500122500142500Dollars of Income

FIGURE 2.—Distribution of income (in US$) for Oregon salmon fishermen, 1971.

CATCH DISTRIBUTION AND RESOURCE SCARCITY 271

10 20 30 40 50 60 70 80 90 100Percent of Rshermen

FIGURE 3.—Lorenz curves for perfect equality in re-source distribution among Columbia River gill-netters,1899, and Oregon salmon fishermen, 1971.

equality. It shows how far the distribution departsfrom perfect equality. Depending on the value ofeach species in the catch, the catch distributionand the value of the catch may be quite different.The analysis in the present study is limited by thedata available. The objective is to look at changein the pattern of distributions in which there isincreased resource scarcity.

Lorenz curves and Gini coefficients can be am-biguous, as Atkinson (1973) shows in comparingthe overall income distributions for West Ger-many and Great Britain. The Great Britain curvestarts out below that for West Germany and thencrosses the West German curve. Both have com-parable Gini coefficients, but the actual curves showthat income in West Germany is less concentratedamong the lowest 40% of the population, but ismuch more concentrated at the high end of theincome distribution than in Great Britain. GreatBritain shows less equality at the low end but ismore equitable at the high end of the distribution.

A more important criticism from the point ofview of fisheries is the goal of perfect equality im-plied by a Lorenz curve. People enjoy fishing be-cause of the chance to excel. Highliners are re-spected for their skill. In using Lorenz curves andGini coefficients, the intent is not to suggest a goalfor perfect equality in catch distributions. Randomfactors of weather, oceanographic conditions, andbreakdowns will always make for catch differ-ences. Aitchison and Brown (1957) showed thatthe general pattern of income distributions is log-normality. Lognormal distributions multiply thedifferences between people, and the mode is much

less than the mean. Catch distributions with a log-normal shape have higher Gini coefficients.

Some inequality is a fact of life in fishing, butwhat is the pattern observed for change in catchdistributions? If management rules discriminateagainst large, efficient producers and benefit small,inefficient ones, does this result in distributionsthat sacrifice efficiency for equality? Such a dis-tribution would show less skewing and a lowerGini coefficient.

Despite criticisms, Lorenz curves continue to bewidely used to measure inequality. For consisten-cy, the Gini coefficient will be used to comparechanges in the pattern of catch distributions dis-cussed in the examples that follow and in the mod-el used to explain the patterns observed.

Inequality and Resource ScarcityIf management discriminates against large, ef-

ficient producers, an hypothesis would be that catchdistributions should narrow in their range and showgreater equality. Fishermen would be forced to-ward lower and more similar catch levels. Fourexamples test this hypothesis. One is the catchdistributions for Oregon salmon Oncorhynchus spp.fishermen in 1899, 1921, and 1971 (Table 1). The1899 and 1921 income distributions are for sam-ples of gill-netters fishing in the Columbia River.At those times, gillnetting in the Columbia Riverwas the dominant mode of salmon fishing. If thesesamples are representative of all Oregon salmonfishermen, they show the pattern of change in catchdistribution. The total population of gill-netterswas about 2,400 in 1899. In 1899, each gill-netboat was sail-powered and required a puller andsomeone to lay out the net. In the early 1900s,motorized boats enabled one person to fish a gill-net boat. By 1921, the total number of gill-netterswas about 2,000 (Smith 1979). By 1971, mostsalmon-fishing effort shifted to trolling in the ocean.The 1971 distribution represents 4,358 commer-cial fishermen catching salmon in the ocean andgillnetting in the Columbia River.

Table 1 gives the Gini coefficients for 1899,1921,and 1971. Each successive value shows progres-sive inequality. In addition, the earnings range be-tween the top and bottom increased from $1,375in 1899 to over $137,500 in 1971. This is a 100-fold increase. During the same time period, thevalue of the dollar decreased by a factor of 10.These are gross earnings; cost data were not avail-able to calculate net earnings. The gross earningsdata are directly proportional to the number ofpounds landed.

272 SMITH

TABLE 1.—Measures of the patterns of inequality indistribution of resources among Oregon salmon fisher-men, Skolt reindeer herders, and Rampur landowners.

Year

189919211971

19581971

19101953

NGini

coefficientOregon salmon fishermen

2,400* 0.312,000* 0.424,358 0.74

Skolt reindeer herders35 0.3735 0.59

Rampur landowners38 0.2678 0.37

Maximumresource

allocationper individual

US$1,375US$4,500US$137,500

293 reindeer162 reindeer

375 bighasb

240 bighasb

a Estimated from Smith (1979).b A bigha is about one-tenth of a hectare.

Discussion of the salmon fishery lumps togetherseveral processes that are difficult to separate. Firstis the trend in stock size. Natural runs of salmonhave declined substantially. Some of this loss hasbeen compensated for with hatchery stocks. Al-though the overall stock size has decreased, thenumber of fishermen and their technical efficiencyhave both increased. The salmon fishery has beensubjected to increasingly stringent management.Rules designed to protect Columbia River salmonstocks date from the 1870s. In any specific fishery,separating the effects of resource scarcity, numbersof fishermen, technical efficiency, and manage-ment impacts is very difficult.

A second case is not from a fishery, but is Pelto'sanalysis of the Skolt Lapp community of Sevet-tijarvi in Finland. This community of 50 families"up to 1960 was basically that of an egalitariansociety" (Pelto 1973). The Skolt are an exampleof how improved technical efficiency from thesnowmobile, which made reindeer herding easier,affected the ownership distribution. Introductionof the snowmobile along with other technologicalinputs like chain saws, which were designed toincrease the productivity of the Skolt, broughtabout "visible economic stratification."

Fishing and reindeer-herding were the primaryeconomic activities. Some Sevettijarvi families alsoengaged in wage labor, but the Skolt Lapps con-sidered themselves reindeer herders above all.Reindeer ownership was compared by Pelto (1973)between 1958, before the snowmobile, and 1971,when Sevettijarvi had 70 snowmobiles (Table 1).For 35 reindeer-herding households, Pelto fol-lowed the changes that occurred and found thatreindeer herds diminished sharply during the years

of the "snowmobile revolution." The concentra-tion of ownership increased such that one house-hold ended up owning 31% of the total herd in1971. In 1958, the household with the most rein-deer owned 7% of the total herd.

The noise of the snowmobiles scared the rein-deer, and they moved to the remotest portions oftheir range. The noise also reduced the reproduc-tion rate. The herd size decreased from 2,179 in1958 to 947 in 1971. The Gini coefficient jumpedsubstantially in this 13-year period (Table 1).

In the north Indian village of Rampur, Lewis(1958) compared the distribution of land amonglandowning families in 1910 and 1953 (Table 1).In the Skolt case, new technology had a negativeimpact on herd size, whereas the population ofherders remained relatively constant. In Rampur,technology did not change, but the population in-creased. Because of inheritance rules that dividedland equally among heirs, Rampur tended to spreadpoverty broadly. This situation is much like theone described for fisheries. During the 43-year pe-riod, Lewis found that the number of Rampurhouseholds doubled from 38 to 78. According toLewis, none of the landowning families in 1910had less than 6.25 acres. By 1953, nearly half thefamilies owned less than 6.25 acres. The amountof land held by the top 5% increased in Rampurslightly from 16% to 18%. The big change accord-ing to Lewis was in the bottom 50%, for whomthe share of land held declined from 30 to 19%.Lewis summarized the problem: "Despite the dif-ferences in landholding, wealth, and power withinthe village, the main problem in Rampur is notone of excessive concentration of land in the handsof a few.. . . There is no absentee landownershipin Rampur, and there are no large landholdings tothe extent found elsewhere. The main problem inthe village seems to be simply an inadequacy ofland resources" (Lewis 1958). In Rampur, the in-creased population and the inheritance pattern ofdistributing land equally among heirs merely spreadpoverty more broadly and lead to greater inequal-ity. The Rampur data suggest that rather than thelargest and most productive being most adverselyaffected, it is the greater number and reduced suc-cess of small producers that change the most.

The Columbia River, Skolt, and Rampur ex-amples show greater inequality with resource scar-city. Is there an example for which the processreverses when relative abundance returns? TheOregon bottom-trawl fishery from 1976 to 1985is such an example. Changing patterns in catchdistribution are shown in Table 2. Targets of thefishery were groundfish—roundfish (principally

CATCH DISTRIBUTION AND RESOURCE SCARCITY 273

TABLE 2.—Oregon trawl-fleet characteristics, 1976-1985.

Year

1976197719781979198019811982198319841985

Number ofvessels

7679

104152168152186182162133

Average netvessel displace-

ment8

35364348474750474646

Average catch(tonnes)

149120139136101135154162130158

Total catch(tonnes)11,3549,498

14,43920,66416,97520,55028,68529,45321,13721,069

Gini coefficient0.520.580.550.590.630.590.550.540.530.54

a Net registered tons (NRT); 1 NRT = 2.8 m3 of cargo space.

sablefish Anoplopoma fimbria, Pacific cod Gadusmacrocephalus, and lingcod Ophiodon elongatus),rockfish (Sebastes spp. and shortspine thornyheadSebastolobus alascanus), and flatfish (Dover soleMicrostomus pacificus, English sole Parophrys ve-tulus, petrale sole Eopsetta jordani, arrowtoothflounder Atheresthes stomias, and rex sole Glyp-tocephalus zachirus). During the 10 years covered,the number of vessels varied between 76 and 186(Table 2). The fishery experienced growth in bothnumber and size of vessels. It also experiencedmore intensive management. Bottom-trawl fish-ermen were very innovative during this period andadopted several new methods to increase theirtechnical efficiency (Dewees and Hawkes 1988).Those operating in the groundfish fishery alsoswitched to other fisheries (Carter 1981). Sometrawlers caught Dungeness crab Cancer magisterin the winter and ocean shrimp Pandalus jordaniin the spring, or participated in joint-venture fish-eries in the summer or fall. Vessels fished in Alas-kan waters when opportunities were better there.

The Oregon bottom-trawl fleet grew very rap-idly in number and size from 1976 to 1982 (Table2). The number of vessels increased 32% between1977 and 1978, and the average vessel size in-creased by 19%. Then between 1978 and 1979,numbers grew 46% and size increased 12% more.Numbers jumped 22% between 1981 and 1982,but size only went up 6%.

Scarcity of the resource and the recession of theearly 1980s caught the fishery in a cost-pricesqueeze—reduced catches, high interest rates, highfuel prices, and escalating insurance rates com-bined with low ex-vessel prices for groundfish.1

The number of vessels declined 11% from 1983to 1984 and dropped 18% more in 1985 as theeconomic readjustment throughout the economyproceeded. Other options such as midwater trawl-ing, joint-venture opportunities, and moving toother areas also reduced the number of vesselsparticipating in the Oregon bottom-trawl fishery.

Total catch nearly doubled between 1976 and1979. From 1979 to 1980, catch declined by 18%,whereas the number of vessels grew 11%. The larg-est catch for a single vessel increased every year.It was 467 tonnes in 1976 and more than doubledto 984 tonnes in 1980. The number of vesselsdropped in 1981, the fleet catch grew, and themaximum single-vessel catch dropped to 806tonnes. In 1982 and 1983, both catch and numberof vessels grew. Total catches dropped back to1981 levels in 1984 and 1985. The number ofvessels, too, dropped in these two years. The totalfishery landings in 1983 were nearly 75% greaterthan in 1980. For 1984 and 1985, total catchdropped more than 25% below the 1983 level.

The average catch was lowest and the Gini co-efficient was highest in 1980 (Table 2). Lower Ginicoefficients correlate with higher average catches.The correlation between Gini coefficient and av-erage catch per vessel was -0.76 (P < 0.006).

Many factors affect the availability of and mar-kets for fish. For the Oregon bottom-trawl fishery,catches were most constrained and there was alsothe greatest inequality in the distribution of catchin 1980. The 1980 conditions did not continueand the fishery improved.2 As they did, catch in-equality declined.

1 Because of the large variety of species landed, vari-ability in prices between time and place of landing, andthe absence of consistent records on the price offish, itis not possible to trace trawl incomes during this period.

2 Oregon bottom-trawl fisheries show a better fit to aBeverton-Holt (1957) model that explains change in stocksize. A Beverton-Holt model does not have a maximumsustainable yield. Catch rates are most influenced by age-class strength.

274 SMITH

These are but four purposely selected examples.The Gini coefficient shows greater relative in-equality with increased resource scarcity. In eachexample, the highest Gini coefficient is associatedwith the greatest resource scarcity. The model thatfollows explains why this may be so.

A Model of the ProcessA simplified example illustrates the process of

greater inequality coining from an increased num-ber of vessels and technological innovation thatputs greater pressure on the fish stock. Commercialfishing is a competitive activity in which techno-logical innovation enables some fishermen, likethe Skolt reindeer herders, to achieve greater suc-cess than others. Further, the number of fishermenoften grows like the population of Rampur, so thatmore people are competing for the same resource,but unlike the situation in Rampur, cultural rulesdo not require sharing the commercial fishing catchequally. In addition, commercial fishermen im-prove their technology, as shown for the Oregonbottom-trawl fishery; they continually work at in-creasing their ability to catch fish. As fishermenbecome more skillful, management rules restricttheir effort and make them less efficient. The con-tinued growth of numbers and technological effi-ciency makes fishing much more competitive anda few people outcompete others. These examplesindicate that, despite being less efficient, largerproducers gain a relative advantage over smallerones.

To illustrate this process, begin with an unusedfish stock at the environmental carrying capacity.Assume the stock-recruitment curve is of the Rick-er (1975) type. Suppose the community has anunmet need for food, and an innovative individualinvents a technology that allows catching exactly100 fish. This technology spreads and each entrantto the fishery catches 100 fish their first year.

Assume that catching 100 fish produces greaterincome than economic opportunities elsewhere.This is the traditional full-employment assump-tion common to economic analysis. Incentives arefor people to switch to fishing because they earnmore income. Assume the price for fish is inelasticwith respect to supply, so that as more fish arecaught earnings remain directly proportional tothe amount caught. This assumption is for sim-plicity in illustrating the model. Without this as-sumption, the increased supply of fish could beexpected to force the price to drop. The inelasticityassumption could be relaxed and would not changethe overall result. Decreasing the economic incen-

tives to fish should help reduce the number offishermen, as was the case in the Oregon bottom-trawl fishery.

Table 3 shows the growth in number of fisher-men entering the hypothetical fishery. The in-crease in the number fishing is designed to followa logistic growth curve. In the first year, two newpeople enter the fishery. This number grows at anincreasing rate until the seventh year, when theincrease in number begins to slow down. Thegrowth in number slows because, with more andmore people fishing, the cost of fishing is reducingthe net benefit, and not entering the fishery looksbetter for most people. In year 10, growth is only5 new units, down from the maximum increase of60 in year 6.

The increase in numbers is analogous to thepopulation growth experienced by a village likeRampur, whose agricultural land was limited. Whatabout the effects of technological advances as ex-emplified by snowmobiles among the Skolt andthe size of Oregon bottom-trawl vessels? Assumewith experience that fishers improve their tech-nical efficiency. To include the process of becom-ing better fishers, the hypothetical fishery com-pounds an 8% improvement in technical efficiencyeach year. This means that a person has the abilityto catch 108 fish in the 2nd year and 117 fish inthe 3rd year, and by the 10th year the individual'scatching capacity doubles to 200 fish (Table 3).The assumption is that people improve technicalefficiency only by fishing. Research on the Oregonbottom-trawl fishery (Stephenson 1980; Standeret al. 1982) showed that this is generally true dur-ing the early years of a trawler's career. Some bot-tom-trawlers also gain experience in fishing forshrimp before "graduating" to the trawl fishery.

Table 3 shows that by the fifth year the catchingcapacity is 7,462 fish, and it nearly doubles to14,064 in year 6. The number fishing increases by86% and their technical efficiency grows 8%. Table3 reflects the dual process of increasing numberand technological innovation on effort growth. Oneis from more people fishing, and the other is theinnovative behavior of each individual to becomemore productive. Increased productivity can comefrom either larger-sized vessels or fishing existingvessels more effectively.

The maximum weight a fish stock can consis-tently produce from generation to generation un-der existing environmental conditions is called themaximum sustainable yield. In the Ricker (1975)model, between the carrying capacity and the max-imum sustainable yield more fishing effort actuallyincreases the surplus production that can be taken

CATCH DISTRIBUTION AND RESOURCE SCARCITY 275

TABLE 3.—Annual growth in number of vessels and their capacity (number offish) in a hypothetical fishery.

Vesselcapacity 1 2

100 2 3108 2117126136147159172186200

Total3 2 5(200) (516)

Number of vessels inyear

3832

13(1,358)

4

22832

35(3,667)

5

3522832

70(7,462)

6

603522832

130(14,064)

7

35603522832

165(18,694)

8

2035603522832

185(22,199)

9

102035603522832

195(24,969)

10

5102035603522832

200(27,467)

0 Values in parentheses are total capacities for all vessels.

from the stock. In the hypothetical fishery, thereis no natural variability. To move the fishery to-ward the maximum sustainable yield, growth inthe number fishing and in individual catching ca-pacity is desirable. During the development stageof the fishery, each new fisher benefits, and thecommunity overall is more productive as the stockproduces more. The benefits received by individ-uals brings more effort. When fishing effort exceedswhat is necessary to take the maximum sustain-able yield, the surplus production declines. Con-tinued increase in fishing will, in each successivegeneration, further reduce the fish stock.

By the seventh year in the hypothetical fishery,165 fishers have a capacity to catch 18,694 fishannually (Table 3). The next year, the capacityincreases to 22,199 fish. Assume that the maxi-mum sustainable yield is 17,067 fish at this effortlevel. Following the logic of maximum sustainableyield, this means that each vessel must fish at lessthan maximum capacity to maintain the fish stockat the maximum sustainable yield and help pre-vent decline in the stock size. Assume a manage-ment program capable of restricting catch to themaximum sustainable yield, and assume thatmanagement rules require every vessel to cut backequally. The assumption that every vessel mustbe treated equally is common to most fishery man-agement systems. Cutting the time for fishing af-fects every fishing unit equally. Seldom do man-agement rules differentiate between the sizes offishing units. A reduction in effort of 10 fish pervessel (165 fishermen x 10 fish = 18,694 fishingcapacity — 17,067 maximum sustainable yield)achieves this objective. This represents a 10% cutfor the new entrants, but only a 6% cut for thevessel with the largest capacity. These reductions

make no allowance for differences in economicefficiency in the operation of each vessel.

For the future, if the fishery is to avoid declinein stock abundance, no new vessels can enter thefishery. Additionally, the technical efficiency ofthose fishing cannot change. Fishermen are in-novators. Their innovations give them an advan-tage in a fishery constrained by new managementrules. Controlling the number fishing is not enoughto solve the problem: technological change, whereresources are limited, also increases effort.

Assume that the number of fishers and theirtechnical efficiency continues to increase, even atthe slower rate shown in Table 3. In the 10th year,the catch capacity of 27,467 fish must be reducedby 10,400 fish to sustain the fishery. This repre-sents a reduction of 52 fish per fisher. For the newentrants, this is a 52% reduction in their catchingability. For the two fishers who have fished all 10years, the reduction is 26%.

Table 4 shows the catch distribution in the 10thyear with a reduction of 52 fish per vessel. If catchdistributions were equal, each group of vesselswould land the same amount relative to their num-ber. The new entrants represent 2.5% of the fishers,but land only 1.4% of the catch. The 1% of vesselsthat have fished all 10 years land 1.7% of the catch.

Inequality in catch distributions increases froma number of factors. One is the assumption thatthose who have fished longer have the advantageof an 8%/year growth in catch rate. The effect ofthis on the Gini coefficient is relatively minor. TheGini coefficient for the vessels and their capacitiesin Table 3 in years 6, 8, and 10 is 0.05, 0.06, and0.07, respectively. These are increases in relativeinequality of about 20% per year. The magnitudeis not large because in this example the range over

276 SMITH

TABLE 4.—Catch distributions in the 10th year of the hypothetical fishery shown in Table 3 showing number ofvessels, their catch capacity (number of fish), and the reductions required to maintain the maximum sustainableyield (MSY).

Number of yearsin the fishery

123456789

10All

Number ofvessels

5102035603522832

200

Vessel capacity100108117126136147159172186200

27,467

Catch limit tomaintain MSY*

485665748495

107120134148

17,067

Percent reductionin capacity tomaintain MSY

5248444138353330282638

Percent allowablecatch per vessel

0.280.330.380.430.490.560.630.700.780.870.50

a Based on a reduction of 52 fish per vessel.

which catches vary is not large. In most fisheriesthe range between the maximum and minimumcatches is much greater than the twofold increasehere. For the Oregon bottom-trawl fishery, therange in annual catch is from less than 1 tonne tonearly 1,000 tonnes.

Of much greater relative impact is the rule re-quiring equal sharing of the cutbacks. The Ginicoefficient for the vessel capacity to maintain themaximum sustainable yield in year 10 increases71%, from 0.07 to 0.12. All the vessels are madeless efficient. The less experienced and less inno-vative fishermen are hurt most. This is becausethose with larger catches give up a smaller pro-portion of their total catch. For the cutbacks tohave equal effect, each vessel needs to be cut backby the same percentage, not the same absoluteamount.

Relaxing the assumption of price inelasticity withrespect to supply does not change the results. Price,if it does not radically change incentives, will changethe magnitudes of the income curve but not itsshape. Further, changing management to allowmarket forces to work rather than cutting backeach vessel equally removes constraints on thetechnologically more effective vessels. They willcatch more and reduce the overall stock, and inthe process they will outcompete the other vesselsand increase the degree of inequality in catch dis-tribution. Many expert fishermen say they wouldrather fish when stocks are limited because theywill be relatively more successful than when stocksare abundant (Wix 1967).

When fishing effort is beyond what is necessaryto maintain the maximum sustainable yield, thegeneral welfare of fishermen does not increase.Rules restricting effort make the fishery as a whole

less efficient. To sustain the fishery, no new vesselscan enter and individual innovativeness has to beconstrained, or if innovativeness continues to beencouraged, vessels have to leave the fishery tokeep from fishing the stock beyond what it cansustain. The community of fishermen no longerbenefits from technological innovation that makesthem more effective. New technology merely altersthe competitiveness among fishermen.

In the Ricker (1975) model, once fishing effortreaches the maximum sustainable yield, any vesselmaking additional catch does so from redistribu-tion of catch among those fishing. This is true atall levels of effort for a Beverton-Holt (1957) mod-el. When catch can only be increased at the expenseof others, for every more successful fisher, therehas to be one or more who are less successful. Onefishermen cannot be made better off without some-one else being made worse off (von Neumann andMorgenstern 1953).

Implications for ManagementThese data suggest that, when a resource is lim-

iting, management is a no-win situation with re-spect to efficiency and equality. Restrictions toprotect the resource cause greater inefficiency andinequality. Estimating the capacity of the resourceto support fishermen, well ahead of the build-upof new fishermen and improved technical efficien-cy, is the one way to avoid the twin problems ofinefficiency and inequality.

All four cases and a simple model have beenused to question whether resource scarcity and theassociated management adversely affect efficientproducers more than those who are less efficient.The four cases given here show that limited re-

CATCH DISTRIBUTION AND RESOURCE SCARCITY 277

sources lead to greater inequality, as seen in Lorenzcurves and as measured by the Gini coefficient.When population and technological growth occurand resources are scarce, greater relative inequalityis the result. Data controlling the actual efficiencyeffects on each individual are only available in thesimple model, but all the cases suggest that themost efficient producers are least affected by re-source scarcity and management. The fishery over-all is undeniably less efficient, but changes in eq-uity affect small producers more.

These data were not gathered to test this hy-pothesis. Following individuals and controlling forvessel size, captain and crew experience, and en-vironmental and market conditions would makea better test. Many people have catch distributiondata, and other analyses may provide additionalinsights.

If a majority in society is improving its absoluteposition, perhaps greater relative inequality is ac-ceptable. However, when cutbacks are required,rules specifying that all users of a resource be treat-ed equally cause more extreme inequality, partic-ularly among small producers, as shown in thesimplified model and the Rampur case. The re-quirement that rules treat each resource user equallywhen cutbacks are required discriminates moreagainst the newer entrants and smaller-scale users.

Can the inequality consequences of populationand technological growth when resource limits areexceeded be avoided? In a free-enterprise system,as long as there are opportunities elsewhere, theless successful can shift to other activities, but oth-er opportunities are often limited, particularly inthe vicinity of fishing communities. The effects ofresource limitations depend on what is happeningin other economic sectors. Economic perspectivesassume open systems in which unsuccessful fish-ermen shift to other occupations. Often, however,fishermen are place- or experience-bound. In theabsence of full employment, those restricted fromfishing have limited alternatives, which contrib-utes to greater inequality.

In fishery management a basic paradox exists.Economic growth incentives stimulate more effort,and technical efficiency increases catches. Greatereffort and efficiency make fishermen better off.These contribute some to inequality. When theresource is limiting, the incentives that promotegrowth in effort and technical efficiency becomecounterproductive. Catch becomes more unequaland competition among users increases. In thesimple model of a fishery, the rule to restrict catchby the same amount for everyone increases in-

equality much more than growth in numbers andtechnological efficiency.

Fishery management rules are usually appliedequally, which contributes to inequality becauseof the greater capacity and technical efficiency thatlarger producers have. Small producers are re-garded as inefficient and are thought to benefitfrom management rules that penalize efficient pro-ducers. The evidence presented here does not denyan inefficiency effect, but indicates that small pro-ducers are hurt more.

Despite assertions that, on balance, democracyand capitalism adequately weigh issues of equalityand efficiency (Okun 1975), when there is resourcescarcity, equality and efficiency both decline. Inany situation where the number of resource usersare increasing and the resource is decreasing, thenumber seeking the resource needs to be reduced.Each decision designed to reduce the populationof those using the resource becomes more difficult.The continued progress toward better technicalefficiency makes the situation worse. A commonsolution is to buy out the inefficient, small pro-ducers. From an equity perspective, buying out afew large producers would result in greater equalityin catch distribution than removing many smallones. Hilborn (1985) makes such a suggestion forthe British Columbia salmon purse-seine fishery.An argument against this is that it tends to restricteconomic efficiency.

Fishery management becomes critical when aresource cannot sustain the fishing effort, inequal-ity is increasing, and success in the fishery onlycomes at the expense of others. Unless opportu-nities exist for people outside the fishery, fisherymanagement decisions to restrict effort only ex-acerbate the growing inequality and conflict.Knowing this, an alternative is to plan ahead, wellbefore the resource becomes limiting, and to factorin estimates of growth in technical efficiency. Along-term plan should be developed that estimatesthe number of people that the resource can sup-port. The influx of new users should be controlledso it does not exceed the productive capacity ofthe resource. The argument against planning aheadin this way is that it limits freedom of enterprise.When resources are scarce, equality, efficiency,freedom of enterprise, and resource protection be-come competing goals: one can be optimized onlyat the expense of the others.

AcknowledgmentsPortions of this manuscript were presented at

the American Fisheries Society 1987 Annual

278 SMITH

Meeting in Greenville, North Carolina. Supportfor this research came, in part, from the citizensof Oregon and the Oregon State University SeaGrant College Program with funds from the Na-tional Oceanic and Atmospheric Administration,Office of Sea Grant, Department of Commerce,under grant NA85AA-D-SG095, Project R/ES-8.I appreciate the help of Susan S. Hanna in develop-ing the data on the Oregon bottom-trawl fishery.

ReferencesAitchison,J.,andJ. A. C. Brown. 1957. Thelognormal

distribution. Cambridge University Press, Cam-bridge, UK.

Alaska Commercial Fisheries Entry Commission. 1982.Alaska's fishing fleets: compilation of data on resi-dence of gear operators, vessel characteristics andfishery diversification patterns for some major Alas-kan fishing fleets, 1969-80. Juneau.

Atkinson, A. B. 1973. Wealth, income, and inequality.Oxford University Press, New York.

Atkinson, A. B. 1975. The economics of inequality.Clarendon Press, Oxford, UK.

Beverton, R. J. H., and S. H. Holt. 1957. On the dy-namics of exploited fish populations. Fishery In-vestigations, Series II, Marine Fisheries, Great Brit-ain Ministry Agriculture, Fisheries and Food 19.

Bronfenbrenner, M. 1971. Income distribution theory.Macmillan, London.

Carter, C. N. 1981. Multi-fishery activity in Oregoncommercial fishing fleets: an economic analysis ofshort-run decision-making behavior. Doctoral dis-sertation. Oregon State University, Corvallis.

Champernowne, D. G. 1973. The distribution of in-come between persons. Cambridge University Press,London.

Cowell, F. A. 1977. Measuring inequality: techniquesfor the social sciences. Wiley, New York.

Dewees, C. M., and G. R. Hawkes. 1988. Technicalinnovation in the Pacific coast trawl fishery: the ef-fects of fishermen's characteristics and perceptionson adoption behavior. Human Organization 43:224-234.

Hibbs, D. A., Jr., and C. Dennis. 1988. Income dis-tribution in the United Stales. American PoliticalScience Review 88:467-490.

Hilborn, R. 1985. Fleet dynamics and individual vari-ation: why some people catch more fish than others.Canadian Journal of Fisheries and Aquatic Sciences42:2-13.

Huppert, D., and C. S. Korson. 1987. Profile of thegroundfish fishery. Pages 7-12 in D. D. Huppert,editor. Limited access alternatives for the Pacificgroundfish fishery. NOAA (National Oceanic andAtmospheric Administration) Technical ReportNMFS (National Marine Fisheries Service) 52.

Karpoff, J. M. 1987. Suboptimal controls in commonresource management: the case of the fishery. Jour-nal of Political Economy 95:179-194.

Knapp, G. 1984. Allocative effects of regulations in theAlaska king crab fishery. University of Alaska, In-stitute of Social and Economic Research, Anchor-age.

Lewis, O. 1958. Village in northern India: studies in aDelhi village. University of Illinois Press, Urbana.

Lewis, R. C. 1973. Preliminary economic analysis ofmanagement alternatives for limiting harvest effortby Oregon commercial fishermen. Oregon Depart-ment of Fish and Wildlife, Portland.

Okun, M. A. 1975. Equality and efficiency: the bigtradeoff. The Brookings Institution, Washington,D.C.

Pelto, P. J. 1973. The snowmobile revolution: tech-nology and social change in the Arctic. Cummings,Menlo Park, California.

Ricker, W. E. 1975. Computation and interpretationof biological statistics offish populations. FisheriesResearch Board of Canada Bulletin 191.

Smith, C. L. 1979. Salmon fishers of the Columbia.Oregon State University Press, Corvallis.

Stander, J. M., C. L. Smith, A. V. Tyler, D. R. Bernard,R. S. Johnston, and F. J. Smith. 1982. Trawl user'smanual. National Marine Fisheries Service, North-west and Alaska Fisheries Center, Seattle.

Stephenson, G. O. 1980. "Pushing the highline": thediffusion of innovations in the Oregon otter trawlfishery. Master's thesis. Oregon State University,Corvallis.

von Neumann, J., and O. Morgenstern. 1953. Theoryof games and economic behavior, 3rd edition. Wi-ley, New York.

Wix, J. R. 1967. Some economic considerations inDungeness crab marketing. Master's thesis. OregonState University, Corvallis.