Case of Nile Perch in Lake Victoria
                                           Petro Sauti Magai1


The study assessed and analyzed the sustainability of Tanzanian Nile Perch fishing industry
using Mwanza region on Lake Victoria as a case study. Focus was on showing how fish exports,
catch per unit effort and fish price impact on fish catch and sustainability. Data was collected
using interviews seeking to appraise the adequacy of the policies to effect sustainability of
fishing in the regions around the lake. Times series data was also collected from reports on
fishing activities in these regions and analyzed using E-Views 3.1. The study established that due
to open access nature of the fishery the fishing effort has increased tremendously which has led
to over fishing indicated by dwindling in the fish catch dominated by undersized fish.
Furthermore, the international fish demand especially by European Union countries and trade
with the neighbouring countries and the inadequacy of policies, legislation and enforcement
have facilitated the un-sustainability of fishing.

    A Paper to be presented on African International Business and Management Conference
                                      (AIBUMA 2010 Conference)

                         25th - 27th August, 2010; Utalii Hotel, Nairobi Kenya

                                              Nairobi, Kenya

                                                 May, 2010

 Mr. Petro Sauti Magai; University of Dar es Salaam Business School, P.O. Box 35046 Dar es Salaam, Tanzania. He
can be contacted through his mobile phone +255 713757989, E-mail, /


Fishing as one of the very important activities, playing a greater role of food security,
employment to the local inhabitants around Lake Victoria and foreign exchange earner to both
the exported of the Nile perch and the nation at large. The Nile perch and two other Tilapiine
fishes were introduced to Lake Victoria in East Africa in the 1950s and early 1960s (Prado, et al.,
1991) to convert the small sized abundant haplochromines into larger table fish and to fill an
ecological niche gap which had been left by the overfished native tilapiine species . In the
beginning the introduced species had minor impacts on the fisheries, but in the late 1970s,
increase in their biomass surpassed all other fish species. Currently, there are, in principle, three
commercial species; Nile perch, Dagaa and Tilapia, accounting for 60 percent, 20 percent and 10
percent of total landings respectively (Kyomuhendo, 2001). Since the introduction of the Nile
perch fish production in Lake Victoria has grown significantly (Witte et al, 1991 & Kitchell et al,
1997). Catches climbed from about 335 mt in 1975, to a peak of 380,776 mt in 1990 (Geheb et
al., 2008) raising the economy of the fishers from the lake (Kitchell et al, 1997) whose number
increased tremendously to approximately 150,000 (DTIS, 2005). In addition, the Nile perch
fishery led to the establishment of fish processing factories for export trade in Kenya and Uganda
during the 1980‟s, and Tanzania in the early 1990‟s (Reynolds, et al., 1992).

Following an increase in the demand for the Nile perch fish in Western Europe and some parts
of Asia in the recent years, the export of Nile perch fillets has generated a lot of foreign currency
(Reynolds, et al., 1992). For example in 2003, fish exports totaled US$154mill., making up 15
percent of the country‟s merchandise exports, and ranking the second item after gold. The export
demand has driven up the price of the Nile perch and this has in turn lead to an increase in capital
investment not only this but also in fishing catching equipment leading to commercial
exploitation of the fish. With such high demands for Nile perch, the value of the fishery has risen
considerably. Labour inflows into the fishery have increased along with growing demand. By
2004, there were 153,066 fishermen in Lake Victoria (LVFO). The fishery also generates indirect
employment for additional multitudes of fish processors, transporters, factory employees and
others. By 2006, the total value of Nile perch exports from the lake was estimated to be US$ 250
million (LVFO, 2007). Concurrently, theree has been a swift expansion fish collection fleet

which ensures that even the landing beaches in more remote areas not easily accessed by road are
reaching (Medard and Wilson, 1996).

The commercialization of the fishery also has disrupted the socio-economics of the local
communities bordering the lake whose livelihood was very much depending on fishing.
Commercial fishing operations, have displaced many local small scale artisanal fishers from their
traditional fishing related occupations and brought them into the cash economy or turned them
into economic refugees. Currently, there is diminishing access to the fishery resources for small
scale fishers due to heavy investment of outboard motors and other fishing equipment by
commercial fishers. Artisanal fishers turn on to catching species other than Nile perch or choose
to become crewmembers for successful fisher who often are financed or hired by the fish
processing factories. Others artisan fishers become involved in illegal beach seining or in
exploiting Nile perch by long lines.

However, the long-term outlook of the fishery is less clear, as overfishing has reduced Lates
niloticus stocks, depicted as declining catch rates and average size of the landed fish (Lokina,
2004), because fishers are using illegal fishing gears that catch predominantly immature fish.
Illegal fishing practices include the use of beach seines and gillnet of under five inch mesh sizes
(Kyomuhendo, 2002). For that reason, fisheries resources should be managed in a sustainable
manner so that they continue to support the livelihoods of the people and contribute to the
nation's economy (Magai, 2005).

The future of the Nile perch export trade is by no means clear as it threatened by many factors
including over-fishing environmental issues and problems of quality and market access to the
European. A crash in the fishery is frequently predicted, and would undoubtedly have severe
consequences for the estimated 1 million people who depend on Nile perch for their livelihood.

Therefore, the aim of this study was to investigate on the factors (variables) that may be
contributing to the un-sustainability of the Nile perch fishery on the Tanzanian part of Lake
Victoria. The model being developed was trying to establish which factors among Amount of
fish catch (AFCT), fish price (FPRT), fishing industry growth rates (FIGR) and catch per unit
effort (CPUE were more important in predicting the MSY of the Nile perch fishery .

Methodology of the study
This study used the time series secondary data for the variables appearing in equation 2 for the
period of thirty years from 1977 to 2007, and the study was limited to the Nile perch fish catch.
Data on amount of fish catch (AFCT) from 1985 to 1995 were obtained from Kyomuhendo
(2002), and from 1996 to 2007 were obtained from Economic Survey of 2008. Data on catch per
unit effort (CPUE) from 1985 to 2000 were obtained from Kyomuhendo (2002), and from 2001
to 2007 were obtained through extrapolation. Data on fish export (FEXP) from 1997 to 2004
were obtained from Economic Survey (2005) and data from 1996 back to 1985 were obtained
through extrapolation. Data on fishing industry growth rates (FIGR) from 1988 to 2005 were
obtained directly from Economic Survey (2005), while the data from 1986 back to 1980 were
compiled from other literature. Data on fish price (FPRT) also from 1997 to 2007 were obtained
from Economic Survey (2008) and other information was compiled from the Department of
Fisheries, Ministry of Natural Resources and Tourism.

Several variables including as the Amount of fish catch (AFCT), fish price (FPRT), fishing
industry growth rates (FIGR) and catch per unit effort (CPUE) were used to fit into a model
which intends to predict the status of the Nile perch under different conditions of the variables,
The E - View 3.1 programme was employed for econometric analysis.

Conceptual Framework of the Model

The theoretical framework for this study was based on the “Maximum Sustainable Yield” (MSY)
The model dwell on management goal which aims at producing maximum sustainable returns of
fish catch in a long term without harming the fish stock‟s capacity to replenish itself. Always the
amount of fish catch depends on the effort. Therefore, in order to have a sustainable fishery a
goal will always be to have an appropriate effort that will produce the MSY. This can only be
achieved through some legislative measures such as having some control on the type, size and
amount of fishing gear to limit access to the fishery through licensing, imposing limits such as
fishing quotas and imposing the minimum size of the fish to be caught and landed. All of these
measures are incorporated in the study variables.

The amount of fish catch under maximum sustainable yield of the fish stock is defined to be a
function of the following variables:
  CPUE = Y ⁄ E………………………………………….……………………...…..……. …...(1)

         Whereby;      Y       =       Yield or fish catch,

                       E       =       Effort invested to catch fish.

  AFCTmsy = f (FPRT, FIGR, CPUE, FEXP, є)………………………………...……………..(2)

  Where; AFCTmsy = Amount of fish catch under maximum sustainable yield,

          FPRT= Fish price in Tanzanian shillings,

          FIGR = Fishing industry growth rates,

          CPUE = Catch per unit effort,

          FEXP = Fish export

          Є = Dummy variable to capture other factors that may affect the sustainability of fish

  The linear relationship can be postulated to be as follows;

  AFCT msy = αo + α1FPRT + α2FIGR + α3UPUE + α4FEXP + α4є………………..………….. (3)

Data analysis

The data was tested for the normality distribution using Jarque-Berra test statistic, and it was
established that all variables were approximately normally distributed, and the correlation matrix
analysis showed a positive correlation for almost all of the variables except CPUE Vs FEXP. A
low positive correlation was revealed when CPUE was correlated against FIGR and FPRT. This
supports the situation that, the increase in any independent variables leads to increase in amount
of fish catches. Even though other variables are highly correlated, the phenomenon of
multicollinearity is inherited in most macro-economic variables and thus dropping a variable
depends on the seriousness of the correlation among variables during the estimating of the

The Augmented Dickey Fuller test for presence of unit root was performed on the variables in
levels and results are presented in Table 1 below. The statistical test results indicate that all
variables except FEXP becomes stationary at their first difference implying that the first
difference of the variables is ideal in the regression analysis. The results also show that all
variables except FEXP have unit roots, this indicates that all variables except for FEXP accept
the null hypothesis of unit root (Table 1)

Table 1: Unit Root Test Results: At levels
 Variables                       ADF Test Statistics              Order of integration

 AFCT                            1.498559                         I(1)

 FEXP                            -3.2643386                       I(0)*

 FPRT                            -1.014682                        I(1)

 FIGR                            -0.725954                        I(1)

 CPUE                            -2.560642                        I(1)

Note: (i) McKinnon (1980) critical values are used for rejection of the null of the Unit Root
(ii) I (0) = the variable is stationary; I (1) = a variable is integrated of order one
(iii) Critical values for ADT: * 1% = -3.6959; ** 5% = -2.9750; ***10% = -2.6265

Following the Engel – Granger theorem, Equation 3 can be reparameterized into an error
correction formulation. Therefore, to arrive at a more parsimonious equation, variable with lower
t – values in an over parameterized model were dropped out. This process is undertaken in order
to maximize the goodness of fit proxied by the R2 the model with minimum number of variables.
The general overparameterised model was estimated with maximum two lags. An error
correction term was introduced in the model. Hence equation 3 is re-specified to include error
correction term (ECM t-1).

ΔAFCTt = αo + α1ΔFPRT t-1 + α2ΔFPRT t-2 + α3ΔFIGR t-1 + α4ΔFIGR t-2 + α5ΔCPUEt-1

              + α6ΔCPUE t-2 + α7ΔFEXP t-1 + α8ΔFEXP t-2 + α9ΔECM t-1 +єt……………(4)

The above model with its associated lags was estimated using Ordinary Least Square (OLS) for
time series data covering 1977 to 2007. Thus, an error correction term lagging one period (ECM
t-1)   is included as one of the independent variables in the general over parameterized error
correction model of maximum sustainable yield equation. This term capture the long run
relationship by attempting to correct deviations from the long run equilibrium path. Its
coefficient can be interpreted as the speed of adjustment or the amount of disequilibrium
transmitted each period to their amount of fish catch. Table 2 represents the results of the
estimation of the over parameterized model.

Table 2: Results of general model ΔAFCT (-t)
Explanatory            Estimated
Variables              Coefficients
                                         Std. Error          t - Statistic      Prob.

C                      0.124139          0.059017            2.103451           0.0617

LAG1DLNCPUE            0.796954          0.210604            3.784130           0.0036

LAG2DLNCPUE            -0.216880         0.275082            -0.788421          0.4487

LAG1DLNFEXP            -0.207747         0.101727            -2.042202          0.0684

LAG2DLNFEXP            0.118934          0.142690            0.833514           0.4240

LAG1DLNFIGR            0.480869          0.167773            2.866180           0.0168

LAG2DLNFIGR            0.235432          0.220937            1.065607           0.3117

LAG1DLNFPRT            0.051862          0.063718            0.813932           0.4346

LAG2DLNFPRT            -0.405073         0.186997            -2.166202          0.0555

LAG1ECM                -0.509862         4.37E-09            -0.365423          0.0056

Note: D and LN in the beginning of the variables represent Difference and Natural Logarithm
            respectively, Adjusted R2 = 0.796, Prob. (F-Statistic) = 0.00087, DW = 2.258 and n = 30

As shown in the general model (Table 2), most of the insignificant variables whose T-Statistics is
less than 2, (for example DLNCPUE          t -2   with T-Statistic of - 0788421) have been eliminated
without losing valuable details and will not appear in the preferred model. The only significant

variables (whose T-Statistic is greater than or equal to 2) will be appearing in the preferred
model as shown in Table 3 below.

Table 3: Estimated OLS Results of the preferred model
Explanatory              Estimated          Std. Error         t - Statistic   Prob.
Variables                Coefficients

C                        0.160457           0.053671           2.989641        0.0092

LAG1DLNCPUE              0.922001           0.213707           4.314315        0.0006

LAG1DLNFIGR              0.427204           0.156778           2.724894        0.0157

LAG2DLNFPRT              -0.290755          0.103637           -2.805518       0.0133

LAG1ECM                  -0.200563          0.058585           -3.423401       0.0055

Note: D and LN in the beginning of the variables denote Difference and Natural Logarithm

respectively, Adjusted R2 = 0.754, Prob. (F-Statistic) = 0.00032, DW = 2.741 and n = 30.

Discussion of Results and Conclusion
The comparison between the general and the preferred model shows that, there is an elimination
of the insignificant variables without losing valuable details. The entire information criterion (see
Tables 2 and 3) shows improvement of results of the preferred model over the general model.
Furthermore, the results show that the standard error of the model was reduced from 0.00087
(Table 2) to 0.00032 in the preferred model (Table 3). Following the dropping out of non
significant variables in the general model, the preferred model passes through misspecification
and serial correlation tests, and report a significant F-test statistics which imply that there is an
improvement in the overall significance of the model.

Generally, all variables were significant, thus Catch per unit effort lagged once (DLNCPUE       t -1),

the Fishing industry growth rates (DLNFIGR t -1), and the Fishing price lagged twice (DLNFPRT
t -2),   all of these variables were obtained at first difference.

The obtained R2 `in the preferred model implies that the explanatory variables included in the
model account for more than 75.4 percent of all variations in the amount of fish catch

(DLNAFCT). Model specification was significant at first difference with F- statistic Probability
of 0.00032. The model has also proved to be not having a serious serial correlation since the DW
is given by 2.741.

Most of the variables in the model were considered in the determination of the amount of fish
catch in the lake as shown in Table 3. The coefficient of lagged once fish catch per unit effort
(DLNCPUE t -1) was positively envisaged and was statistically significant at first difference. This
result cement on the argument that, as the catch per unit effort increase the more fish are caught.
This means that, one unit increase in lagged once catch per unit effort on the lake (DLNCPUE          t-

1)   will increase the amount of fish catch by 92 percent.

In addition, the coefficient of the lagged once on fishing industry growth rates (DLNFIGR           t -1)

was positive as earlier anticipated, and was statistically significant at first difference. Likewise,
the industry growth rate also will increase the amount of fish catch by about 42 percent for a
single unit of effort.

The negative impact of the lagged twice of the fishing price (DLNFPRT           t-2)   on fish catch was
statistically significant at first difference as expected. This means that a rise in one unit of lagged
twice fish price leads to a decline of approximately 29 percent of the amount of catch on the
lake. This is true with respect to economic theories that, once the goods /services are more in the
market, their prices tend to decrease until the equilibrium is reached.

The significance at 1 percent level of the lagged error correction model (ECM_1) included in the
model suggests that, there is correctly signed (negative). This implies that it would take 0.2 of
the year for the fish catch to come to equilibrium if an econometric shock of fish catch occurred
in the exogenous variables on the right hand side. This is an average low speed of adjustment
which implies that all errors/deviations are not corrected within one year and most of the time the
fish catch is not operating in a sustainable manner.

This study revealed that the catch per unit effort and fish export affect the amount of Nile perch
catch which in turn disrupt the sustainability of fishing industry, while only the fish price was
negatively correlated to amount of fish catch. All variables were found to be stationary in first
difference. The co-integration analysis indicated that there is a long run relationship between
different variables used in the study and they tend to agree with hypothesis of the study.

The study also revealed that the prices of the Nile perch continue to rise as harvests shrink,
making the perch a less affordable meal among low-income segment of the population. The
existence of a long run relationship between catch per unit effort and amount of fish caught may
intimidate the sustainability of fish resources, calling for restriction/limited access as an
appropriate policy for controlling open access which will eventually reduce the fishing effort.

The current situation on Lake Victoria fisheries reveals that the linkages between international
trade, fishery sustainability, and socio-economic development remain poorly understood and
inadequately addressed. For example, Lokina, (2004) in his study of Technical Efficiency and
Skipper Skill in artisanal Lake Victoria fisheries, revealed the lack of enforcement measure of a
minimum mesh size regulation, which was greatly violated by the fishers on Lake Victoria and
that lack of alternative employment opportunities had lead to a substantial level of over-
capitalization of the fishery..

Ogutu (2000) predicted a danger of the collapse of this fishery because of unregulated fishing,
neglecting advice from scientists including disbandling of the East African Freshwater Fisheries
Research Organization (EAFFRO). It should be known that management regimes often fail to
adequately control fishing effort or encourage the internalisation of costs, tariff liberalisation
may fuel unsustainable “scale effects”-increasing fish harvesting, exacerbating overexploitation
of fish stocks, increasing investment in harvesting capacity, and driving practices which are
ecologically detrimental.

Kyomuhendo (2004) also revealed that a severe over fishing problem exists and that the fishery
has never been managed for economic efficiency, all economic rent from this fishery has been
and continues to be dissipated. Pomeroy (1994) and Squires, et al., (2003), hold that strategies
for cooperation and community management can help to control fishing activities and promote
sustainable fishing practices, while Christy (1999) suggest that limited entry could be beneficial
for developing fisheries.

The absence of state support to the fishing sector such as subsidies for inputs and credits has
further marginalized the small and poor fishers in terms of access to sustainable resources. The
downward pressure on prices in countries where fishers are not subsidized like Tanzania, it costs
more to produce than it is possible to sell it for. This has been among the major causes of food

insecurity in the area and reducing household incomes. If subsidies were provided to the sector it
would have helped to create an enabling environment where fishers can fish sustainably.

In many of the poorest countries such as Tanzania often face serious obstacles for expanding
their participation in international trade and diversifying production and exports toward higher
value-added processed products. However, stringent standards, rules of origin and domestic
supply-side capacity constraints become the barriers which hamper the development of the
fishing sector. This state of affairs brings the need for fishing sector to be sustainably managed
towards economic efficiency in maximizing the net present value of fish catches over an
indefinite time period and the limitation of use and capital in the fishery. The open access nature
of capture fishery resources makes management both technically and politically difficult.

A first step towards sustainability would be coordination between the three countries involved. A
second would be to address property rights problems, which implies either some form of co-
management structure or the development of quota system (Eggert, 2001). However, as fishers
are very far from any quota system, a pragmatic approach could be a limited access system,
which in fact limits access (Christy, 1999).


Christy, F., (1999) „Common Property Rights; An Alternative to ITQs’, FAO/Western Australia
Conference on the Use of Property Rights in Fisheries Management, 15-17 November.
DTIS Tanzania Diagnostic Trade Integrated Studies report, November 2005

Eggert, H., (2001) ‘Essays on Fisheries Economics’. Unpublished Ph.D Thesis, Department of
Economics, University of Gothenburg. Germany
Geheb, K., Kalloch, S., Medard, M., Nyapendi, A.-T., Lwenya, C. and Kyangwa, M. 2008. Nile
Perch and the Hungry of Lake Victoria. Food Policy, 33 (1): 85-98. See also catch statistics
provided by the Victoria Fisheries Organisation

Harris, C. K., Wiley, D. S. and Wilson, D. C. 1995. Socio-economic impacts of introduced
species in Lake Victoria fisheries. In Pitcher, T. J. and Hart, P. J. B. (Eds.) The impact of species
change in African lakes. London, Chapman and Hall: 215-242.

Kyomuhendo, K., (2002). A bioeconomic model for Uganda’s Lake Victoria Nile perch Fishery.
Norwegian College of Fishery Science.

Kitchell, J.F. Shindler, D.E. Ogutu-Ohwayo, R. and Reinthal, P.N (1997), The Nile Perch in
Lake Victoria: Interaction between predation and fisheries. Ecological Applications

Lokina, R.B., (2004), „Fishermen Compliance Behaviour to different conservation measures in
Lake Victoria fisheries‟. Paper presented on at the third Nordic Conference In Development
Economics, 17 – 18 June 2004.

LVFO (Lake Victoria Fisheries Organisation), 2005. Regional report on Lake Victoria Frame
surveys for 2000, 2002 and 2004. Jinja, Uganda, LVFO and the Lake Victoria Environmental
Management Project.

LVFO (Lake Victoria Fisheries Organisation), 2007a. State of fish stocks (See; Visited January, 2009.

Magai, P.S. (2005). Assessment of local communit activities on managing water resources and
its environment. Case of fishing along Lake Victoria in Ukerewe Island. Unpublished
dissertation thesis: University College of Lands and Architectural Studies (UCLAS). Dar-es-
Salaam, Tanzania.

Mbithi Mwika, S. (2006), Fisheries Access Agreement: Trade and Development Issues, ICTSD
Natural resources, International Trade and Sustainable development series Issue Paper No. 2,
International Centre for Trade and Sustainable Development, Geneva, Switzerland.

Ogutu-ohwayo, R. (2000). The biology, ecology and impact of Nile perch, Lates niloticus in Lake
Victoria, Kyoga and Nabigado and the future of the fisheries. Fisheries Research Institute, 343, Jinja, Uganda.
Prado, J., Beare, R.J., Siwo Mbuga, J. and Oluka, L.E. 1991. A catalogue of fishing methods and
gear used in Lake Victoria. UNDP/FAO Regional Project for Inland Fisheries Development
(IFIP), FAO RAF/87/099-TD/19/91 (En). Rome, Food and Agricultural Organisation.

Reynolds, J.E, D. Greboval and P. Mannini (1992) „Thirty Years on; Observation on the
Development ofthe Nile perch Fishery in Lake Victoria’, UNDP/FAO Regional Project for
Inland Fisheries Planning (IFIP).
Squires, D., Grafton, Q., Alam, F., Omar, I.H., (2003), „Technical Efficiency in the Malaysian
Gill net Artisanal Fishery’ Environmental and Development Economics 8, 481-504.

The United Republic of Tanzania (2008). “Economic Survey”. The Ministry of Finance and
Economic Affairs, Dar es Salaam – Tanzania.

The United Republic of Tanzania (2005). “Economic Survey”. The Ministry of Finance and
Economic Affairs, Dar es Salaam – Tanzania.

Wilson, D.C. and Medard, M. (1999) „The Implications for Fisheries Management of the
changing Situation in Lake Victoria Fishing Communities: Preliminary Findings’ Journal of
African Rural and Urban Studies.

Witte et al., 1992; Barel, C.D.N., Ligtvoet, W., Goldschmidt, T., Witte, F. and Goudswaard, P.C.
1991. The haplochromine cichlids in Lake Victoria: an assessment of biological and fisheries
interests. In Keenleyside, M.H.A. (Ed.) Cichlid Fishes: Behaviour, Ecology and Evolution.
London, Chapman and Hall: 258-279.


To top