14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022].

See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Journal of Agricultural Economics, Vol. 72, No. 3, 2021, 782–804
doi: 10.1111/1477-9552.12423

Demand for Imported versus Domestic

Fish in Nigeria
Lenis Saweda O. Liverpool-Tasie , Awa Sanou ,
Thomas Reardon and Ben Belton
(Original submitted May 2020, revision received November 2020, accepted Decem-
ber 2020.)

Abstract
Fish is among the most important animal-sourced foods in Africa and is crucial in
combatting malnutrition. Fish demand in Africa has far outpaced supply as the
import share rose from 16% in 1970 to 39% by 2017. Little is known about who is
consuming the imports: rural versus urban, rich versus poor. This is the first fish
consumption analysis in Africa distinguishing imported and domestic fish, and
within domestic fish, fresh versus traditional-processed. We analyse three rounds of
nationally representative data from Nigeria, disaggregating the richer South from
the poorer North, and urban and rural. Frozen (imported) fish accounted for 34%
of urban fish consumption in the North (23% for rural), compared with 67% in
urban areas in the South (54% for rural). The large difference in frozen fish con-
sumption between regions is due mainly to differences in income and refrigerator
ownership. For other fish forms (fresh, dried, smoked), regional differences are far
less pronounced. Income and price elasticities confirm that imported fish have
become deeply incorporated into fish consumption habits. From a policy perspec-
tive, this intensifies concerns about import bills as fish demand grows. However,
our elasticity results show that Nigerian consumers are keen to consume fresh fish
as incomes increase, and that demand for smoked and dried fish also remains
strong at high levels of income. Promoting aquaculture is a promising policy path
to reduce import dependence. Domestic capture fisheries remain a major source of

1
Saweda Liverpool-Tasie, Thomas Reardon and Ben Belton are all at Michigan State Univer-
sity, East Lansing, MI, USA. Email: [email protected] for correspondence. Ben Belton is also at
WorldFish, Penang, Malaysia. Awa Sanou is an applied economist in the thematic research
department at the MasterCard Foundation in Canada. We thank the anonymous reviewers and
the editor David Harvey for their constructive comments on an earlier draft. We acknowledge
and appreciate financial support for this work from the United States Agency for International
Development (USAID) under the Feed the Future initiative through the Nigeria Agricultural
Policy Project, Associate Cooperative Agreement Number AJD-620-LA-15-00001 and the Sus-
tainable Intensification Innovation Lab (SIIL), Cooperative Agreement AID-OAA-L-14-00006.
We also appreciate financial support from the US Department of Agriculture National Institute
of Food and Agriculture and Michigan AgBioResearch and the CGIAR Research Program on
Fish Agri-Food Systems (FISH).
Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
on behalf of Agricultural Economics Society.
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License,
which permits use, distribution and reproduction in any medium, provided the original work is properly cited and
is not used for commercial purposes.
 14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Demand for Imported versus Domestic Fish in Nigeria 783

fish, making it important to maintain productivity at sustainable levels through bet-

ter management.

Keywords: Africa; demand systems; fish; imports; Nigeria; EASI.

1. Introduction
Developing countries’ share of consumption of global fish output rose from 61% in
1990 to 78% in 2017, according to FAOSTAT. Despite the rising dominance of devel-
oping regions in world fish consumption, Bush et al. (2019) note that demand for fish
in developing countries is under-researched. However, the gap in the literature is dis-
tinctly different between Asia and Africa.
Asia has been the focus of the great majority of fish demand literature for develop-
ing countries. The rise of Asian fish demand and supply occurred earlier than Africa.
Asia also dominates developing country fish output: in 2014 Asia’s share of fish out-
put from developing countries was 84%, with Africa only 10% and Latin America
and the Caribbean, 4%. A wave of household survey studies in the past two decades
showed that the rapid increase in Asian fish consumption was driven by rising
incomes, falling fish prices, and shifting preferences associated with urbanisation, life-
style changes and employment: Huang and Bouis (2001) for Taiwan; Hovhannisyan
and Gould (2014) for China; Dey et al. (2011), Toufique and Belton (2014) and Toufi-
que et al. (2018) for Bangladesh.
Africa has received far less attention in the developing country fish demand litera-
ture. This may be because of a perception of African fish production being stagnant
and small, which was largely the case in the 1970s and 1980s when its share of world
fish production was about 5%. Yet FAOSTAT data in Table 1 show that from 1970
to 1990 African fish production rose 1.25 fold; from 1990 to 2010 it rose 1.5 fold; and
if the growth rate from 2010 to 2017 holds, production will rise 2.0 fold from 2010 to
2030. From 1970 to 2017, fish output in Africa rose 2.5 fold. But as African popula-
tion rose 3.5 fold, per capita consumption of fish should have fallen over time if it had
been only sourced from domestic production. Instead, consumption hovered around
10 kg per capita for those five decades. The gap between demand and supply was met
by rapidly rising imports, which rose from 16% to 39% of African fish consumption
over those decades. We discuss imports in more detail below.
Fish is crucial to nutrition in Africa (Chan et al., 2019). Ensuring adequate levels of
animal-sourced food consumption is considered to be one key to combatting malnu-
trition (Headey et al., 2018). Africa experiences high levels of food insecurity and mal-
nutrition (Akombi et al., 2017), and fish is among the most important animal-sourced
foods across most of the continent (Desiere et al., 2018). This is also the scenario in
Nigeria, where rates of malnutrition are high and fish is one of the main animal-
sourced foods (Kuku-Shittu et al., 2016; Ogundari, 2017). We analyse nationally rep-
resentative data for Nigeria – the most populous country with the largest economy in
Africa. We find that fish consumption accounted for about 35% of consumption
expenditure for animal proteins in 2015 and constituted about 10% of food consump-
tion expenditure by the average Nigerian – as much as any of the individual main sta-
ples (rice or maize or tubers or pulses). Changes in fish consumption therefore have
important implications for food and nutrition security in Nigeria (Bradley et al.,
2020).

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
on behalf of Agricultural Economics Society.
 784

Table 1

on behalf of Agricultural Economics Society.

Sub-Saharan African fish production (capture + farmed), imports, exports and apparent consumption per capita, 1970–2017

1970 1980 1990 2000 2010 2017

Production in millions of tons 3.1 2.9 3.9 4.9 5.8 7.7

Imports in millions of tons 0.4 1.2 2.0 1.6 3.8 3.3
Percentage (%) of imports in apparent consumption 16 32 38 33 48 39
Exports in millions of tons 1.0 0.4 0.6 1.6 1.6 2.5
Apparent consumption (production + imports – exports) in millions of tons 2.5 3.7 5.3 4.9 8.0 8.5
Population (millions) 267 350 465 600 776 928
Apparent consumption per capita (in kg) 9 11 11 8 10 9

Source: Authors’ calculations from FAOSTAT data.

Lenis Saweda O. Liverpool-Tasie et al.

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
 14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Demand for Imported versus Domestic Fish in Nigeria 785

Despite the importance of fish consumption in Africa, particularly for addressing

malnutrition, examination of fish demand has been limited. There are few survey-
based analyses of fish demand in Africa, though exceptions include: Abdulai and
Aubert (2004) for Tanzania; Tambi (2001) for Cameroon; and local area studies such
as Amao et al. (2006) for Lagos State in Nigeria. Zhou and Staatz (2016) used Living
Standards Measurement Study (LSMS) data from around 2012 to estimate income
elasticities for fish as a general category compared with other food categories for West
Africa. Desiere et al. (2018) also used LSMS and FAO data to assess current and
future meat and fish consumption in a group of countries in sub-Saharan Africa. Gen-
schick et al. (2018) analysed urban Zambian fish consumption patterns of the poor
strata.
Moreover, there has been little research globally on the determinants of the form in
which fish is purchased. ‘Traditional forms’ include dried/salted, smoked, and fresh,
all of which were common prior to the advent of refrigeration and freezing. The main
non-traditional product form is frozen fish, which is thawed after purchase for use at
home or in restaurants. Fish consumption analyses have often treated fish (and ‘sea-
food’) as a homogeneous group of products and few studies differentiate either species
or form. There are some exceptions: Toufique et al. (2018) distinguish fish originating
from capture or aquaculture. Dey et al. (2008) distinguish dried fish from other fish in
Asia. In Europe and the US, Trondsen et al. (2004) distinguish processed from fresh,
and Verbeke et al. (2007) distinguish traditional preservation styles versus fresh. In
the United States, Muhammad and Hanson (2009) distinguish fresh and frozen cat-
fish. In Africa, studies of demand for different fish forms are either of a locality, or of
one species, or limited product forms (Kumar et al., 2005; Jimoh et al., 2013; Dauda
et al., 2016).
In sum, the African literature has not had a systematic analysis of: (i) consumption
of domestically produced versus imported fish; (ii) consumption of different forms of
fish, such as frozen, fresh, dried and smoked; (iii) consumption of fish over spatial cat-
egories such as agroecological zones, and regions with different levels of development.
These gaps are important for the following reasons.
First, unlike Asia, food imports are among the top policy concerns in Africa (Afri-
can Development Bank, 2016) due to their viewed foreign exchange burden and their
competition with the domestic fish sector. In Africa, the share of imports in total
apparent consumption of fish more than doubled over the four decades 1970s–2000s,
to a high of 39% by 2017 (Table 1). This compares to the import share (derived from
FAOSTAT) in all food for 2017 of 13% (Liverpool-Tasie et al., 2020). Despite the
importance of fish imports, no survey-based analysis of the patterns and determinants
of imported versus domestic fish consumption has been done for Africa. As discussed
below, imports are mainly in the form of frozen fish, and the latter are nearly all
imported, so there is a correlation between a lack of analysis of demand for different
forms of fish and demand for imported fish.
Second, food demand analyses using nationally representative surveys in develop-
ing regions are often focused only on the national level. However, in African countries
such as Nigeria, there are particularly sharp inter-regional differences in development
levels as well as consumption habits. We posit that this holds for Northern versus
Southern Nigeria. Northern Nigeria is often in the international news because of the
Boko Haram insurgency, but it has also long had severe development constraints and
lagged growth because of its semi-arid agroecology and lower education compared
with the much richer South, which benefits from more oil revenue and has higher

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
on behalf of Agricultural Economics Society.
 14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
786 Lenis Saweda O. Liverpool-Tasie et al.

education. Nigeria thus presents a pertinent case concerning how different – or similar
– food consumption transformation is in the two regions, and whether the ‘imported
fish’ phenomenon is driven more by the larger middle class of the South or is occur-
ring in both regions. This is more broadly interesting than just Nigeria: the dichotomy
of poorer interior and more developed coastal regions is found over a large part of
Africa.
To address these three gaps, we analyse consumption patterns, food expenditure,
and price elasticities of imported versus domestic fish consumption using data from a
nationally representative panel survey, the Nigeria Living Standard Measurement
Study-Integrated Survey on Agriculture (LSMS-ISA). It has data on the same house-
holds for 5 years over 2010 to 2015. The LSMS data allow stratification by urban and
rural as well as by North and South Nigeria. We also explore the heterogeneity of
demand across rural and urban areas as proxies for employment, lifestyle and prefer-
ences, found to be so important in the Asian studies noted above. Following the same
households over multiple years is our fourth contribution to the literature. The great
majority of fish demand studies in developing countries (and, to our knowledge, all
those in Africa) use cross-section analysis. We are able to track changes over time,
something not yet done in Africa for fish consumption, to see whether changes in the
two regions diverge or converge.
However, the Nigeria LSMS data do not directly indicate whether the fish con-
sumed is from imports or domestic sources. While fish indicated as fresh are, with
near certainty, from domestic capture or aquaculture (although these two sources are
not indicated in the data), processed fish can be either domestic or imported. We have
assigned dried and smoked fish to the domestic category because the great majority of
dried and almost all smoked fish are domestically produced except for a small amount
received in informal cross-border trade, and for Norwegian dried cod (stock fish)
imports.1 By contrast, we assign all frozen fish to the imported category because Nige-
ria lacks a significant fish freezing industry using domestic fish as inputs, and nearly
all the frozen fish purchased in the country are imported.2
The paper proceeds as follows. In Section 2 we discuss the data used. Section 3 pre-
sents a description of fish consumption patterns across Nigeria. Section 4 presents the
econometric approach featuring an Exact Affine Stone Index (EASI) demand model
and section 5 presents the associated regression results. Section 6 concludes.

2. Data
We use data from three rounds of the Nigeria World Bank Living Standards Mea-
surement Study-Integrated Surveys on Agriculture (LSMS-ISA); 2010/11, 2011/12,

1
Given the unconditional dried fish consumption by households derived from the nationally
representative LSMS data and extrapolated to annual national levels (assuming 5–7 people per
household) we compare the expected total dried fish demand to the amount of dried fish
recorded to have been imported by COMTRADE and find this to be between 4% and 6%
depending on the assumed household size.
2
We also conducted a rapid reconnaissance with frozen fish sellers in one to three fish markets
(and/or cold rooms) in the capital cities of six Nigerian states that cover the six geopolitical
zones of Nigeria (one per geopolitical zone). We asked over 50 frozen fish sellers in each state
(minimum of five per market) the source of their frozen fish. We found that 100% of their fro-
zen fish was imported.

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
on behalf of Agricultural Economics Society.
 14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Demand for Imported versus Domestic Fish in Nigeria 787

and 2015/16 . It includes about 5,000 households in each survey year surveyed twice a
year during the agricultural season and post-harvest season. This generates a panel as
the same individuals were interviewed during each period of data collection. The sur-
vey was nationally representative, covering rural and urban areas in the two geo-
graphical regions that also capture agroecological variation; the North and the South,
as discussed above. The data cover household demography, assets, production and
food consumption from own production, purchases and gifts.
We computed household food consumption expenditure as the total value of con-
sumption (from purchases, own production and gifts) for 10 categories of food used
to categorise all the food items available in the data: cereals and tubers, pulses, dairy,
beef and other meats, poultry and eggs, dairy products, fresh fish, frozen fish, smoked
fish, dried fish, and other foods. The choice of the categories was informed by other
consumption studies in developing countries (Fashogbon and Oni, 2013; Dolislager,
2017). We disaggregated fish consumption into the four categories reported in the sur-
vey: fresh, frozen, dried and smoked. Price indices were computed for each of the food
categories as a weighted average of transaction-derived prices of items included in the
specific group. Nominal prices and values were converted into real values using the
consumer price index (CPI) at the national level for each survey year with 2010 being
the base year.

3. Fish Consumption Patterns

Table 2 shows overall fish consumption in the country and the two regions. Several
points stand out. First, fish consumption is widespread and increasing: 59% of Nigeri-
ans ate fish in 2010 versus 72% in 2015. The North stayed steady at about 50%, while
the South leapt from 71% to 90% in the 5 years.
Second, per capita consumption of fish products (unconstrained) in the South in
2015 was more than double that of the North. Annual fish consumption per capita is
13 kg. This is slightly higher than the average apparent consumption per capita for
Africa as a whole (Table 1), but about half the global average of 20.3 kg (FAO,
2018).
Table 3 compares the North and South over 5 years, and disaggregates fish con-
sumption into frozen fish (largely imported), and fresh, dried and smoked fish (pri-
marily domestic origin).3 There are several striking points. First, the share of people
consuming frozen fish is far higher in the South compared to the North. Only 14% of
Northern fish consumers ate frozen fish in 2015, versus 62% in the South. This gap
grew over the 5 years. Frozen fish might be more accessible in the South because the
ports are close by and more households own refrigerators. Only 10% of households in
the North own refrigerators, all of them in urban areas, as compared to 30% of
households in the South (20% rural and 40% urban) (Table 4, below).
Second, there is surprisingly little difference between the North and South in terms
of the share of people eating fresh fish (14% and 12%, respectively). The share of
fresh fish in total food consumption expenditure is even closer, around 1.5% in both
regions. This similarity may be because capture fisheries in rivers and lakes in the

3
National average annual consumption per capital (unconstrained) of fresh, frozen, smoked
and dried fish is 15.6 kg, 22.7 kg, 8.7 kg, and 5.1 kg, respectively.

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
on behalf of Agricultural Economics Society.
 14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
788 Lenis Saweda O. Liverpool-Tasie et al.

Table 2
Annual fish consumption in Nigeria by rural and urban locations and region, 2010–2015

Year Total number of households North South All

Households consuming fish (%)

2010 9,246 46 71 59
2012 9,284 48 90 72
2015 9,124 49 90 72
National (kg/per capita)
2010 9,246 6.4 17.0 12.1
2012 9,284 5.7 15.0 11.0
2015 9,124 6.3 18.7 13.2
Urban (kg/per capita)
2010 2,994 8.7 16.1 14.1
2012 2,853 7.3 15.2 13.4
2015 2,904 8.6 18.6 16.0
Rural (kg/per capita)
2010 6,252 5.7 18.1 10.8
2012 6,431 5.2 14.8 9.4
2015 6,220 5.5 18.8 11.2

Source: Authors’ estimations from the LSMS-ISA 2010, 2012 and 2015 data. These values are
unconstrained and in nominal terms to allow for comparison to other studies and international
statistics.

North balance more abundant aquaculture and marine fisheries in the South in pro-
viding access to fresh fish.
Third, while dried and smoked fish are preserved product forms, often thought to
be eaten in arid areas far from supplies of fresh fish, Southern consumers eat more of
both (each consumed by around 25% of households) than those in the North (con-
sumed by 17% and 11% of households, respectively).
In the North, the share of consumers eating fish was much higher in urban areas
than in rural: 61% versus 45% in 2015. In the South, the shares of people eating fish
were similar in urban and rural areas, at around 90%. Accordingly, the share of fish
in overall food consumption is lower in the North (3%) than the South (12%) in
2015, likely reflecting higher incomes in the South as well as access to a wide variety
of fish (including the more expensive fresh fish) and different food culture traditions
(see Table S1). Comparing urban to rural per region shows that there was an increase
in the share of households consuming fish in the South in both rural and urban areas
from 2010 to 2012, but little change in the South from 2012 to 2015, or in the North
in any year. Given that relative incomes of urban and rural areas did not change much
over the 5 years, this appears to imply that access to fish has grown in the South, lead-
ing to increasing equality of fish market access between rural and urban areas. Food
supply chains are critical for fish consumption; approximately 95% of fish consumed
in both regions is purchased.
Fish expenditure by fish type, as a share of total fish budget (by region, and rural
versus urban) is presented in Table S1. We find that in the North, frozen fish were
34% of all fish consumption in urban areas, versus 23% in rural areas where about
half of fish consumption is dried or smoked fish. This could be because of differences

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
on behalf of Agricultural Economics Society.
 Table 3
Share of households consuming fish, share of fish in the food budget, and fish consumption per capita (by fish type and region, 2010–2015)

Fresh fish Frozen fish Smoked fish Dried fish All fish

North South All North South All North South All North South All North South All

on behalf of Agricultural Economics Society.

Households consuming fish (%)
2010 14.3 11.4 12.7 12.9 49.9 33 10.8 18.2 14.8 15.9 19.5 17.9 46 71 59
2012 13.8 9.4 11.3 15.8 66.5 44.7 9.3 19.6 15.1 15.9 23.9 20.5 48 90 72
2015 13.9 12.3 13 13.5 61.5 40.2 11 24.6 18.6 17.2 26.4 22.3 49 90 72
Fish as share of food budget (%)
2010 1.6 1.7 1.6 1.1 6.3 3.9 0.5 0.9 0.7 0.5 0.7 0.6 3.7 9.7 6.9
2012 2.1 1.8 1.9 1.5 7.6 5.0 0.5 0.8 0.7 0.7 0.8 0.7 4.7 11.0 8.3
2015 1.4 1.6 1.5 1.2 8.0 5.0 0.4 1.3 0.9 0.4 0.5 0.5 3.4 11.5 7.9
Per capita fish consumption (kg/capita)
2010 12.4 10.1 11.2 7.9 39.6 25.0 6.6 7.5 7.1 8.2 9.8 9.1 6.4 17.0 12.1
2012 9.5 7.9 8.6 8.6 33.1 22.6 4.8 6.9 6.0 6.9 7.3 7.1 5.7 15.0 11.0

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
2015 16.7 14.7 15.6 7.8 34.6 22.7 4.7 11.9 8.7 5.0 5.2 5.1 6.3 18.7 13.2

Source: Authors’ estimations from the LSMS-ISA 2010, 2012 and 2015 data. Note that these values are unconstrained and households can consume
Demand for Imported versus Domestic Fish in Nigeria

more than one fish form.

789

14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
 14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
790 Lenis Saweda O. Liverpool-Tasie et al.

in income and shares of families with iceboxes or refrigerators. Rural areas are often
poorly served by electricity grids or generators. Even when an area has electricity it is
often shut off for significant periods, making it hard for consumers as well as retailers
and wholesalers to store frozen fish. By contrast, in the South, frozen fish dominates
in both urban areas (67% of fish consumption) and rural areas (54%).
In both the North and South, the share of dried fish is slightly higher in rural areas
than urban. This appears to be due in part to the association of dried fish with lower
incomes and transportability and ambient storability that favour consumption of it in
less accessible rural locations, particularly in the North. Fresh fish consumption is
similar in urban and rural areas (around 25% in the North, and 10% in the South).
Smoked fish also has similar shares in urban and rural areas in both regions (roughly
18%).
Table 4 shows real prices. Frozen fish tends to be more expensive than the tradi-
tional processed forms of fish in both North and South. However, the price of fresh
fish is higher than the price of frozen fish (up until 2015). In 2015, the price of frozen
fish increased significantly (by about 50% in the South and 30% in the North) com-
pared to the previous round. The price increase is largely due to the higher exchange
rate from the currency devaluation following the oil price collapse in 2013/2014.4
Interestingly, despite the large hike in imported frozen fish prices, consumption per
capita dipped only slightly in the North but increased in the South. This shows that
frozen fish is not viewed as a luxury to be dropped when its price rises, but is likely a
price-inelastic product, having penetrated basic consumption habits at least of certain
consumer strata, particularly in the South.
Contrary to expectations, frozen fish is more expensive in the South than the North.
This is surprising because the North is further away from ports through which frozen
fish is imported. The result might reflect heterogeneity in the dominant kinds of fish
consumed per region. A rapid reconnaissance of fish markets conducted to add infor-
mation on frozen fish indicated that the North tends to consume cheaper frozen fish
such as herrings, whereas relatively more expensive frozen fish such as croaker are
consumed more in the South.
Controlling for inflation, the average price of fish (irrespective of form) more than
doubled between 2010 and 2015, from ₦319 to ₦666 per kg (Table 4). The price of
dried fish increased by over 170% in the North and over 230% in the South. This
appears to be partly driven by the Boko Haram insurgency in North-Eastern Nigeria,
home to the Baga market, one of the largest fish trading centres supplying dried fish
to the entire country (Mukhtar and Gazali, 2016).
The price of fish rose most sharply from 2010 to 2012 (70%). This price rise was
linked to a drop in fish consumption per capita of 10%. Fish prices increased, but less
sharply (by 20%), from 2012 to 2015 (Table 5). During that period, fish consumption
per capita rose, likely due to fish remaining the cheapest animal sourced food. The
price of fresh fish declined by 20% between 2012 and 2015, perhaps in response to
rapidly growing aquaculture production. This change was accompanied by an 80%
increase in consumption of fresh fish.

4
Crude oil prices declined from about $80 a barrel in 2010 to about $40 a barrel in 2015 (US
Energy Information Administration, 2017).

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
on behalf of Agricultural Economics Society.
 14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Demand for Imported versus Domestic Fish in Nigeria 791

Table 4
Descriptive statistics for key regression variables

2010 2012 2015

North South All North South All North South All

Mean prices (Naira/kg)

Price of other food 394 553 480 361 478 427 1215 1313 1270
Price of cereals and 230 392 318 221 323 279 1621 1138 1352
tubers
Price of pulses 210 251 232 342 440 398 435 466 452
Price of beef and 610 710 664 926 979 956 1289 1379 1339
other meats
Price of poultry 492 566 532 1070 843 940 852 788 816
and eggs
Price of dairy 414 949 704 659 1495 1136 1347 2142 1789
products
Price of fresh fish 360 578 478 1025 820 909 763 699 727
Price of frozen fish 324 452 393 561 614 591 723 901 822
Price of smoked 189 260 228 302 313 308 712 435 558
fish
Price of dried fish 156 194 176 383 364 372 436 655 558
Price of fish 257 371 319 568 528 545 659 673 666
Other demographic characteristics
Education (0/1) 0.6 0.7 0.7 0.6 0.7 0.7 0.6 0.7 0.7
Female head of 0.1 0.2 0.1 0.1 0.2 0.2 0.1 0.3 0.2
household (0/1)
Own refrigerator/ 0.1 0.3 0.2 0.1 0.3 0.2 0.1 0.3 0.2
freezer
Household adult 5.1 3.7 4.3 5.3 3.6 4.3 5.6 3.7 4.5
equivalent
Total expenditure 26 141 88 56 145 106 53 213 142
on fish (Naira/week)
Total expenditure 683 902 802 1,032 1,333 1,204 1,270 1,861 1,599
on food (Naira/
week)
Total expenditure 412 909 677 614 1514 1116 327 741 557
on non-food items
(Naira/week)
Total expenditure 1,085 1,761 1,451 1,637 2,760 2,277 1,597 2,602 2,156
on food and non-
food items (Naira/
week)
Number of 4,395 4,851 9,246 4,517 4,767 9,284 4,527 4,597 9,124
observations

Source: Authors’ estimations from the LSMS-ISA 2010, 2012 and 2015 data. The prices are real
prices created using CPI from the World Bank: 2010 is the base year, CPI for 2012 and 2015 are
124.382 and 158.943 respectively. Expenditure values are per capita. Due to significant variation
in prices and components of categories such as other foods and cereals and tubers, median
prices were also computed and show similar trends but less dramatic changes.

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
on behalf of Agricultural Economics Society.
 792

Table 5
Changes in fish consumption per capita and real prices, by fish type, time period, rural and urban locations and region

2010–2012 2012–2015 2010–2015

North South National North South National North South National

on behalf of Agricultural Economics Society.

Change in consumption per capita (%)
Fresh fish –23.7 –21.8 –23.2 76.7 85.2 81.5 34.8 44.9 39.5
Frozen fish 10.0 –16.4 –9.8 –9.4 4.6 0.6 –0.4 –12.5 –9.2
Smoked fish –26.4 –8.7 –15.4 –3.0 73.4 45.2 –28.6 58.4 22.9
Dried fish –16.1 –25.3 –21.4 –28.0 –29.5 –28.9 –39.6 –47.4 –44.1
All fish –11.7 –11.6 –9.4 10.8 24.2 19.6 –2.2 9.8 8.4
Change in real prices (%)
Fresh fish 184.72 41.87 90.17 −25.56 −14.76 −20.02 111.94 20.93 52.09
Frozen fish 73.15 35.84 50.38 28.88 46.74 39.09 123.15 99.34 109.16
Smoked fish 59.79 20.38 35.09 135.76 38.98 81.17 276.72 67.31 144.74
Dried fish 145.51 87.63 111.36 13.84 79.95 50.00 179.49 237.63 217.05
Lenis Saweda O. Liverpool-Tasie et al.

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
All fish 121.01 42.32 70.85 16.02 27.46 22.20 156.42 81.40 108.78

Source: Authors’ estimations from the LSMS-ISA 2010, 2012 and 2015 data. The prices are real prices created using CPI from the World Bank: 2010 is
the base year, CPI for 2012 and 2015 are 124.382 and 158.943 respectively.

14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
 14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Demand for Imported versus Domestic Fish in Nigeria 793

4. Econometric Analysis
4.1. Empirical approach
Our demand analysis focuses on four forms of fish (fresh, frozen, smoked and dried)
with other foods (beef, cereals and tubers, pulses, dairy, poultry, and eggs and ‘other
foods’). We assume, as is usual, that household food consumption is determined in a
two-stage budgeting process (Deaton and Muellbauer, 1980). In the first stage, house-
holds allocate their total household consumption expenditures to food versus non-
food items conditional on prices, income and household characteristics. In the second
stage, the household allocates food consumption to different food types including the
various forms of fish. We allow substitution between different fish products and other
protein sources or other food groups, and estimate a full food demand system.
Preliminary local polynomial regressions (Figure 1) of fish consumption shares and
total food consumption (in logs) for the North and the South5 show that the con-
sumption shares for most fish forms are non-linear – but also show that they are not
quadratic in total food consumption (as a proxy for income). Thus, neither the Linear
Approximate nor the Quadratic Almost Ideal Demand System model is appropriate
for the analysis of fish demand in Nigeria. We address this by using the more flexible
Exact Affine Stone Index (EASI) demand system of Lewbel and Pendakur (2009). The
preliminary analysis also shows clear regional differences in the Engel curves.

4.2. The EASI demand system

We apply the EASI demand system of Lewbel and Pendakur (2009) with a pooled
cross-section data.6 The EASI demand system does not impose any particular func-
tional form on the relationship between income and food consumption but allows for
arbitrarily complex Engel curves. In addition, it allows us to control for individual
preference heterogeneity across households and time-specific factors rather than leave
them as part of the error term as is done in other models (Lewbel and Pendakur,
2009).
We assume that: households have demographic and other characteristics that affect
food preferences, including fish products, in vector z; households have log nominal
total food consumption x; they face a vector of log prices p. Households then choose
a vector of consumption shares w, to maximize utility subject to the household budget
constraint.
Lewbel and Pendakur (2009) and Pendakur (2008) show how the Hicksian budget
shares associated with the households utility function expressed as a function of p and
z and utility level u can be expressed as a function of log real consumption with an
implicit Marshallian consumption shares function. We follow Lewbel and Pendakur
to define an implicit utility function y which only depends on observed variables. The
implicit utility function is used to derive the implicit Marshallian consumption (bud-
get) shares as follows:

5
The advantage of this approach is that the relationship is modelled as linear in the neighbour-
hood but may vary across values of the log of total fish consumption. The degree of polynomial
smoothing used here is 1, meaning that the graphs are a locally weighted least squares model.
6
For the estimation of the EASI the panel data are treated as a pooled cross-section. However,
for the data processing using a multivariate probit estimation approach before application to
the EASI, we fully exploit the panel structure of the data.

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
on behalf of Agricultural Economics Society.
 14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
794 Lenis Saweda O. Liverpool-Tasie et al.

North
Food budget share of fresh fish Food budget share of frozen fish

.4
.4
.3

.2
Frozen fish share
Fresh fish share
.2

0
.1

-.2
0
-.1

-.4
0 2 4 6 8 0 2 4 6 8
Log of food expenditure Log of food expenditure

95% CI lpoly smooth: Fresh fish 95% CI lpoly smooth: Frozen fish

Food budget share of smoked fish Food budget share of dried fish
.4

.3
.2

.2
Smoked fish share

Dried fish share

0

.1
-.2

0
-.4

-.1

0 2 4 6 8
Log of food expenditure 0 2 4 6 8
Log of food expenditure
95% CI lpoly smooth: Smoked fish
95% CI lpoly smooth: Dried fish

South
Food budget share of fresh fish Food budget share of frozen fish
.2

.8
.15

.6
Frozen fish share
Fresh fish share
.1

.4
.05

.2
0

0 2 4 6 8 0 2 4 6 8
Log of food expenditure Log of food expenditure
95% CI lpoly smooth: Fresh fish 95% CI lpoly smooth: Frozen fish

Food budget share of smoked fish Food budget share of dried fish
.3

.4
Smoked fish share

.3
.2

Dried fish share

.2
.1

.1
0

0 2 4 6 8
Log of food expenditure 0 2 4 6 8
Log of food expenditure
95% CI lpoly smooth: Smoked fish
95% CI lpoly smooth: Dried fish

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
on behalf of Agricultural Economics Society.
 14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Demand for Imported versus Domestic Fish in Nigeria 795

Figure 1. Engel curves.Note: Conditional on the household having consumed fish.Source:

Authors’ estimations from the LSMS-ISA 2010, 2012 and 2015 data. [Colour figure can be
viewed at wileyonlinelibrary.com]

R L
w ¼ ∑ br yr þ Cz þ Dzy þ ∑ zl Al p þ Bpy þ ɛ (1)
r¼0 l¼0

where p and z are the vector of J prices and L demographic variables and ϵ is a vec-
tor of error terms which include unobservable preference heterogeneity. y is a measure
of real total food consumption and is specified as the equal affine transform of the
Stone index-deflated log nominal consumption levels. That transform, by Lewbel and
Pendakur (2009), is:
lnx  p0 w þ ∑Ll zl p0 Al p=2
y¼ (2)
1  p0 Bp=2
The budget share expressed in equation (1) has all the desirable properties of tradi-
tional demand models with some added advantages. Similar to the Almost Ideal
Demand System (AIDS) model, these implicit Marshallian budget shares are linear in
parameters (thus easy to estimate) and have additive error terms including unobserved
preferences due to taste and time. However, while the AIDS budget shares are linear
in p, z, and y, the EASI budget shares are linear in p but are polynomials of any order
in zand y. Thus, as noted above, the EASI Engel curves can take any shape through
the addition of polynomials of any order in real consumption.7 The budget shares can
also include interaction terms such as py, zy, and pz’.
Subsequently, income and price elasticities of different fish forms across regions in
Nigeria can be estimated using the semi-elasticities in equations (3) and (4) below (i.e.,
the derivatives of the budget shares with respect to total consumption expenditures
(as a proxy for income) and prices) following Lewbel and Pendakur (2009).
ry0 wðp, y,z, ∈ Þ ¼ cl þ dl y þ Al p (3)

L
rp0 wðp,y, z, ∈ Þ ¼ ∑ zl Al þ By (4)
l¼0

These semi-elasticities in equations (3) and (4) are easier to present algebraically
and they can be converted into the relevant elasticities by dividing these expressions
by the budget share.

4.3. Resolving the zero consumption and endogeneity problems

Estimating demand systems for subgroups of food often faces the ‘zero consumption’
problem. Three main reasons for this problem are discussed in the literature. First,
households may never consume that product. Second, a limited survey period may

7
In our analysis R is equal to 3 and the budget shares are a third order polynomial in the log-
budget.

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
on behalf of Agricultural Economics Society.
 14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
796 Lenis Saweda O. Liverpool-Tasie et al.

not find the household consuming a product that it might consume in a period outside
the survey recall period. Third, a household might not report consuming the product
because it feels it would reveal its making a bad decision (e.g. because the product was
too expensive or they could not really afford it but still bought it) (Meyerhoefer et al.,
2005; Tafere et al., 2011). In the LSMS-ISA data, there is zero consumption for one
or more fish forms among 41%, 30% and 29% of the households in 2010, 2012 and
2015, respectively.
To address this, we employ a two-step procedure to estimate a system of equations
with limited dependent variables to obtain a synthetic dataset with imputed consump-
tion for households with zero consumption (e.g. Magrini et al., 2017; Tefera et al.,
2018). In the first step, we estimate the determinants of consuming different forms of
fish (and other food groups) with a Correlated Random Effect (CRE) multivariate
probit model, which accounts for correlation among the food groups (Wooldridge,
2010).8 The explanatory variables (zis) used in the estimation include a vector of log of
total household consumption expenditure on food, log of prices of the 10 food
groups, and demographic variables (education, gender, asset index, living in an urban
area, living in the north, household adult equivalent, round of data collection and the
mean of all the time-varying household characteristics. In the second step, we calcu-
late the cumulative distribution (φ(.)) and normal probability density functions (φ(.))
for each food group. This is then used to generate new consumption shares for all
food groups w∗it as:
   
w∗it ¼ Φi bθ0is zis wit þ δi ϕi bθ0is zis (5)
where wit is the budget share of food group i at time t and the estimated partameter
δi is the covariance between the first and second stage error terms. As mentioned
above, z0is refers to the explanatory variables explaining purchasing behaviour and θis
are the associated parameters for the i food groups from the multivariate probit
regressions.9 One challenge with this transformation (equation 5) is that the new con-
sumption shares w∗it no longer satisfy the additivity condition as required by demand
theory. We address this issue by reweighting the transformed shares (Steele and
Weatherspoon, 2016) to obtain w∗∗ it . This approach has two advantages. First, we do
not have to choose arbitrarily any of the fish groups as the residual category with no
specific demand. Second, it avoids obtaining negative consumption shares for the
good since it is possible that the sum of the other goods is greater than one when one
imposes the following condition:
k
w∗∗ ∗ ∗
it ¼ wit = ∑ wit (6)
i¼1

Not accounting for the endogeneity of the allocation of consumption across fish
forms with respect to the demand for fish as a product category relative to demand
for other food products, can lead to biased and inconsistent demand parameter

8
The correlated random effects (CRE) estimator allows for correlation between the time invari-
ant unobserved household omitted variable and included explanatory variables. One class of
CRE models allows for modelling the distribution of the unobserved household characteristic
conditional on the means of time-varying exogenous variables (Mundlak, 1978; Chamberlain,
1980).
9
Standard errors are clustered at the household level.

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
on behalf of Agricultural Economics Society.
 14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Demand for Imported versus Domestic Fish in Nigeria 797

estimates. To address this, we follow Lewbel and Pendakur (2009) and use an instru-
mental variables approach. Our instruments are logged prices and logged assets and
powers of both to the third order.10 P represents prices for each of the food groups.11

4.4. Accounting for time and regional effects on fish demand

As noted above, culture and income vary significantly between the two regions. The
importance of this variation on food choices is well acknowledged (e.g. Ma, 2015).
The regions differ in other characteristics as well. The major ports through which
imported fish enter the country are in the South. Also, while there are many lakes and
rivers dispersed across the country, the South is closer to the Atlantic Ocean, which is
a major source of fish from capture fisheries. The South has also experienced a rapid
growth in fish farming over the past decade. All these are likely to influence fish
demand. We account for regional differences by estimating the EASI demand model
separately for the North and South. Thus, we derive consumption and price elastici-
ties by region. We do the same for rural and urban areas.
Furthermore, we use a Linear EASI demand model that controls for time-specific
factors (such as season) that influence food choices and preferences for particular fish
forms. Thus, we estimate equation (7) for North and South Nigeria distinguishing
between rural and urban areas.
w∗∗j R j r L j L
ir ¼ αio þ ∑r¼1 bir yit þ ∑l¼1 gil zlit þ ∑k¼1 ajki lnpt
k
(7)
þ∑Lk¼1 ∑Ll¼1 sjki zlit lnpkt þ ∑Lk¼1 bjk lnpkt yit þ ∑Ll¼1 htj zlit yit þ ɛ ijt
To account for the possible heteroscedasticity of error terms and the simultaneous
determination of budget shares and total consumption, the estimation of the EASI
demand system in the software R uses an iterative linear three-stage least squares
(3SLS) estimator as in Hoareau et al. (2012) that is similar to Blundell and Robin
(1999).

5. Regression Results
The estimated expenditure elasticities from the EASI for different food items and fish
forms are reported in Table 6. Several points stand out.
First, as incomes rise, Nigerians consume more of all forms of fish: frozen, fresh,
smoked, or dried; imported or domestic. However, in Southern Nigeria imported fro-
zen fish has the lowest expenditure elasticity (compared to all other fish forms). This
shows it is likely a necessity in the South. By contrast, in the North, frozen fish has an
expenditure elasticity above 1, and hence is a ‘luxury’. This shows how deeply imports
have penetrated the basic fish consumption habits of the South, but in the North are
limited to the middle class.
Second, fresh fish have the highest consumption elasticities (among all fish forms)
and remain a luxury. A 1% increase in income is associated with an increase of 1.1%

10
The analysis was a three-stage least squares estimation and the budget shares are a third order
polynomial in the log budget with R = 3.
11
We recognise the possibility of price endogeneity due to the simultaneity between supply and
demand. We attempt to reduce concerns about this with our location controls (rural and urban)
and other household demographic characteristics.

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
on behalf of Agricultural Economics Society.
 14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
798 Lenis Saweda O. Liverpool-Tasie et al.

and 1.2% in expenditure on fresh fish in the North and the South respectively. This is
a strong indicator of the potential for domestic aquaculture to grow further as
incomes increase in Nigeria.
Third, the expenditure elasticities for smoked and dried fish indicate they are luxu-
ries in the South (Table 6) on average. This is surprising given that these are the tradi-
tional forms in which fish are consumed. By contrast, the lowest elasticity in the
North (among fish forms) is for the traditional largely domestic dry fish, indicating it
as a necessity rather than luxury. A closer look shows that rural consumers drive the
inelasticity of dried fish in the North, as their consumption elasticity is lower than
urban consumers for dried fish. Higher responsiveness among urban consumers in the
North might reflect some sort of quality trade-off between types of dried fish. Our
data cannot disaggregate between different types of dried fish such as stock fish, which
is much more expensive than other traditional dried fish. Similar explanations could
also explain the higher expenditure elasticities for dried fish in urban areas (compared
to rural areas) in the South.
Table 7 shows the own- and cross-price elasticities of different foods and fish forms,
by region, and by urban versus rural. As predicted by demand theory, compensated
own-price elasticities are negative for all the food groups and fish forms. The own-
price elasticities of all food groups (except dairy in the North and cereals, poultry and
dairy in the South) are inelastic.
When own price increases, households in the North tend to reduce the quantities of
fresh fish purchased the most while households in the South tend to reduce the quanti-
ties of frozen fish consumed the most, for both urban and rural areas in both regions.
For the North, this finding is consistent with our earlier finding that fresh fish is a lux-
ury with high consumption elasticities. Overall, southerners reduce the quantity of
smoked fish the least while Northerners reduce that of frozen fish the least because of
changes in own prices.
Finally, we compare the cross-price elasticities to see the substitutability and com-
plementarity among fish types as well as food groups. Though statistically significant,

Table 6
Expenditure elasticities by fish type, rural and urban locations and region

North South

Rural Urban All Rural Urban All

Fresh fish (domestic) 1.10*** 1.09*** 1.09*** 1.17*** 1.19*** 1.18***

Frozen fish (imported) 1.07*** 1.05*** 1.05*** 0.91*** 0.91*** 0.91***
Smoked fish (domestic) 1.07*** 1.09*** 1.08*** 1.04*** 1.09*** 1.06***
Dried fish (domestic) 0.95*** 0.97*** 0.97*** 0.99*** 1.05*** 1.02***
Cereals and tubers 0.97*** 1.10*** 0.99*** 0.84*** 0.83*** 0.83***
Pulses 0.81*** 0.75*** 0.81*** 0.90*** 0.84*** 0.88***
Beef and other meats 0.99*** 0.91*** 0.97*** 1.05*** 0.99*** 1.03***
Poultry and eggs 1.42*** 1.25*** 1.35*** 1.37*** 1.24*** 1.30***
Dairy products 1.19*** 1.08*** 1.15*** 1.15*** 1.08*** 1.12***
Other food 0.74*** 0.81*** 0.75*** 0.77*** 0.89*** 0.81***

Source: Authors’ calculation from the EASI model estimation. ***P < 0.01, ** P < 0.05, *
P < 0.1.

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
on behalf of Agricultural Economics Society.
 Table 7
Compensated own- and cross-price elasticities by rural and urban locations and region

Beef
Cereals and Poultry Dairy
and tubers Pulses other meats and eggs products Fresh fish Frozen fish Smoked fish Dried fish Other food

North – Rural
Fresh fish(domestic) 0.02*** 0.03*** –0.05*** –0.05*** –0.02*** –0.91*** –0.17*** 0.03*** 0.01*** 0.09***
Frozen fish (imported) –0.003** –0.06*** 0.02*** 0.007** 0.02*** –0.16*** –0.65*** –0.11*** 0.05*** –0.06***
Smoked fish (domestic) 0.009*** –0.04*** –0.05*** –0.08*** –0.002*** 0.02*** –0.11*** –0.73*** –0.002 0.05***
Dried fish (domestic) –0.03*** –0.003 0.002 –0.12*** –0.002** 0.002 0.05*** –0.01*** –0.84*** –0.05***

on behalf of Agricultural Economics Society.

Cereals and tubers –0.98*** –0.005 0.01*** –0.03*** 0.02*** 0.03*** –0.002 0.02*** –0.05*** –0.04***
Pulses –0.01*** –0.98*** 0.06*** 0.13*** –0.04*** 0.04*** –0.14*** –0.10*** 0.001 –0.006
Beef and other meats 0.01*** 0.06*** –0.92*** –0.10*** –0.05*** –0.10*** 0.05*** –0.10*** 0.02*** –0.04**
Poultry and eggs –0.01*** 0.06*** –0.04*** –0.90*** 0.01*** –0.04*** 0.01*** –0.07*** –0.08*** 0.07***
Dairy products 0.02*** –0.01*** –0.03*** 0.02*** –1.03*** –0.02*** 0.05*** 0.01*** 0.02*** 0.02***
Other food –0.03* –0.002 –0.03 0.09*** 0.01 0.09*** –0.07*** 0.06*** –0.04*** –1.01***
North – Urban
Fresh fish (domestic) 0.03*** –0.01* –0.03*** 0.01*** –0.02*** –0.93*** –0.17*** 0.03*** 0.02*** 0.11***
Frozen fish (imported) –0.02*** –0.01 –0.004 –0.01*** 0.03*** –0.21*** –0.72*** –0.10*** 0.03*** –0.10***
Smoked fish (domestic) 0.01*** –0.04*** –0.04*** –0.09*** –0.001 0.02*** –0.06*** –0.68*** 0.02*** 0.05***
Dried fish (domestic) –0.02*** 0.01* –0.03*** –0.08*** 0.01*** 0.01** 0.02*** 0.02*** –0.84*** –0.02***
Cereals and tubers –0.99*** –0.04*** 0.006 –0.002 0.01*** 0.06*** –0.01*** 0.03*** –0.02*** –0.02
Pulses –0.07*** –0.95*** 0.07*** 0.04*** 0.03*** –0.03*** –0.02** –0.11*** 0.01 –0.05

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
Beef and other meats –0.01** 0.08*** –0.83*** –0.03*** –0.07*** –0.07*** –0.01 –0.12*** –0.06*** –0.03
Poultry and eggs –0.01** 0.04*** –0.01* –0.93*** –0.01*** 0.02*** –0.01*** –0.16*** –0.10*** 0.09***
Demand for Imported versus Domestic Fish in Nigeria

Dairy products 0.01* 0.04*** –0.05*** –0.01*** –1.03*** –0.03*** 0.04*** 0.00 0.02*** 0.02*
Other food –0.03 –0.03 –0.02*** 0.07*** 0.01 0.12*** –0.06*** 0.07*** –0.03* –1.03**
South – Rural
Fresh fish (domestic) 0.04*** –0.01*** –0.09*** –0.07*** –0.04*** –0.84*** –0.02*** 0.01*** 0.01*** 0.09***
Frozen fish (imported) 0.01 –0.06*** 0.05*** 0.02*** –0.07*** –0.06*** –0.88*** 0.002 –0.01*** –0.02**
Smoked fish (domestic) 0.04*** –0.08** –0.06*** –0.10*** –0.004*** 0.02*** 0.01*** –0.86*** 0.02*** 0.06***
799

14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
 Table -0007
800

(Continued)

Beef
Cereals and Poultry Dairy
and tubers Pulses other meats and eggs products Fresh fish Frozen fish Smoked fish Dried fish Other food

Dried fish (domestic) –0.02*** 0.01*** –0.03*** –0.10*** –0.05*** 0.01*** –0.01*** 0.01*** –0.87*** 0.01*
Cereals and tubers –1.06*** –0.02*** 0.02*** –0.002 0.02*** 0.04*** 0.002 0.02*** –0.01*** 0.001
Pulses –0.03*** –0.95*** 0.07*** 0.03*** 0.02*** –0.03*** –0.06*** –0.12*** 0.03*** –0.01
Beef and other meats 0.04*** 0.05*** –0.97*** –0.02*** 0.01** –0.20*** 0.06*** –0.08*** –0.03*** 0.003

on behalf of Agricultural Economics Society.

Poultry and eggs 0.01 0.02*** –0.01** –0.99*** 0.01*** –0.09*** 0.02*** –0.08*** –0.07*** 0.10***
Dairy products 0.05*** 0.02*** 0.01*** 0.02*** –1.03*** –0.07*** –0.04*** 0.002 –0.04*** 0.06***
Other food –0.02 –0.04*** –0.03** 0.13*** 0.03*** 0.18*** –0.04*** 0.06*** –0.10* –1.11***
South – Urban
Fresh fish (domestic) 0.09*** –0.01** –0.09*** –0.04*** –0.04*** –0.87*** –0.03*** –0.02*** 0.01*** 0.13***
Frozen fish (imported) –0.01 –0.01 0.09*** 0.02*** –0.03*** –0.10*** –0.95*** –0.01** –0.08*** –0.01
Smoked fish (domestic) 0.03*** –0.08*** –0.07*** –0.05*** 0.004** –0.03*** 0.003 –0.82*** 0.02*** 0.06***
Dried fish (domestic) 0.01*** –0.004 –0.01*** –0.06*** –0.02*** 0.01*** –0.04*** 0.02*** –0.85*** 0.02**
Cereals and tubers –1.01*** –0.04*** –0.001 0.002 0.02*** 0.10*** –0.01*** 0.01*** –0.002 –0.011
Pulses –0.06*** –0.94*** 0.03*** 0.04*** 0.03*** –0.03*** 0.002 –0.13*** –0.01 –0.02
Beef and other meats 0.01* 0.03*** –0.93*** –0.01 0.003 –0.23*** 0.09*** –0.11*** –0.03*** 0.001
Poultry and eggs 0.02*** 0.04*** –0.002 –1.00*** –0.01* –0.09*** 0.03*** –0.07*** –0.08*** 0.08***
Dairy products 0.05*** 0.03*** 0.01** –0.01* –1.02*** –0.09*** –0.02*** 0.01** –0.03*** 0.03***
Lenis Saweda O. Liverpool-Tasie et al.

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
Dried fish 0.01*** –0.004 –0.01*** –0.06*** –0.02*** 0.01*** –0.04*** 0.02*** –0.85*** 0.02**
Other food –0.02 –0.03 –0.01 0.06*** 0.01 0.21*** –0.01 0.06*** 0.004 –1.15***
***
P < 0.01,
**
P < 0.05,
*
P < 0.1.

14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
 14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
the cross-price elasticities of different fish forms are extremely small. This implies they
occupy specific niches in local cuisines. Several fish forms are complements to poultry
products. In the South, frozen fish is the exception as it is substitute to poultry prod-
ucts. Households appear to consider the different fish forms and poultry products as
distinct food items, consistent with their use in different dishes in a given multi-dish
meal for a family, or over different meals in the day, or even in joint use in various tra-
ditional dishes. Surprisingly, imported frozen fish is the only substitute for beef and
other meats. Thus, as beef and other meat prices increase, more imported frozen fish
is consumed.

6. Conclusions
Our analysis of nationally representative food expenditure data in Nigeria yield sev-
eral key findings. First, fish is among the most important sources of animal protein in
Nigeria. It accounts for 10% of the total food budget and 35% of the budget allocated
to animal source foods of the average Nigerian, rising to 45% in the South.
Second, fish is the cheapest animal protein consumed, with a price significantly
lower than that of poultry and eggs as well as other meats, and less than half that of
dairy products. This status underlines the importance of fish in Nigeria for food and
nutrition security.
Third, there are substantial differences in fish consumption between the poorer
North and the richer South. In the South, 90% of households consume fish (account-
ing for 11% of total food consumption expenditures), versus 50% in the North (with
3% of food outlay). Conditional on consuming some fish, the per capita fish con-
sumption is about 1.2 times higher in the South than the North but not too different
between urban and rural areas in each region.
Fourth, frozen (imported) fish makes up 30% of total fish consumption in Nigeria.
This national figure masks a large difference between the more developed South (clo-
ser to ports for imports, with more refrigeration and higher incomes), with 40% of
fish consumption as frozen, compared with 13% in the North.
Fifth, the share of imported frozen fish in rural fish consumption is 20%, versus
35% in urban areas. Urbanization is associated with more consumption of imported
frozen fish. Rural fish consumption is much more skewed toward traditional forms
(dried, smoked) than frozen/imported, because of differences in access to and costs of
the different product types and refrigeration facilities. Urban consumers appear more
likely to shift to frozen and fresh fish (partly from rapidly growing aquaculture) and
pay more for it. Yet despite these general differences, there is still a non-trivial share
of frozen/imported fish consumed in rural areas, at levels similar to smoked and fresh
fish. However, smoked and dried fish together account for half of total fish consump-
tion in Nigeria, underlining the continuing importance of these product forms for
food and nutrition security.
Sixth, our EASI elasticity estimates show that while frozen imported fish is largely
a necessity in the South (but still a luxury in the North), domestically produced fish
(particularly fresh fish) remains a luxury with much higher elasticities. These results
indicate that if incomes increase in Nigeria, spending on most forms of domestically
produced fish will increase more than proportionately.
Together, these findings suggest that fish plays an important role in food and nutri-
tion security in Nigeria. This can be further supported with investment and interven-
tions to increase supplies of fish and reduce the cost of fish to the consumer. This is of
 14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
802 Lenis Saweda O. Liverpool-Tasie et al.

particular concern in the North where food security is low and still only about 50% of
households consume fish. The higher cost of imported products since 2015 and the
2019 devaluation of the naira have created a greater opportunity for domestic fish
production to compete with imported fish.12
The highly differentiated nature of demand for fish by product type and geographi-
cal region revealed here suggests that multiple policy responses may be required.
These could include: (i) Ensuring that trade restrictions are not imposed on imported
frozen fish, which are shown to make up a significant part of the food basket even in
rural areas of the North (where ~95% is purchased); (ii) Supporting the expansion
and increasing the productivity and efficiency of the domestic aquaculture sector to
increase supplies of fresh fish and produce raw material for fish smokers and driers;
(iii) Instituting governance arrangements and regulations that maintain the long-term
productivity of inland and marine capture fisheries at sustainable levels, to ensure
continued provision of fresh, dried and smoked fish from these sources.

Supporting Information
Additional supporting information may be found online in the Supporting Informa-
tion section at the end of the article.
Table S1

References
Abdulai, A. and Aubert, D. ‘A cross-section analysis of household demand for food and nutri-
ents in Tanzania’, Agricultural Economics, Vol. 31(1), (2004) p#. 67–79.
African Development Bank Feed Africa: Strategy for Agricultural Transformation in Africa
2016–2025. Abidjan: African Development Bank, 2016. Available at: https://www.afdb.org/
fileadmin/uploads/afdb/Documents/Generic-Documents/Feed_Africa-_Strategy_for_Agric
ultural_Transformation_in_Africa_2016-2025.pdf
Akombi, B. J., Agho, K. E., Merom, D., Renzaho, A. M. and Hall, J. J. ‘Child malnutrition in
sub-Saharan Africa: A meta-analysis of demographic and health surveys (2006–2016)’, PLoS
One, Vol. 12(5), (2017) pp. e0177338.
Amao, J. O., Oluwatayo, I. B. and Osuntope, F. K. ‘Economics of fish demands in Lagos State,
Nigeria’, Journal of Human Ecology, Vol. 19(1), (2006) pp. 25–30.
Blundell, R. and Robin, J. M. ‘Estimation in large and disaggregated demand systems: An esti-
mator for conditionally linear systems’, Journal of Applied Econometrics, Vol. 14(3), (1999)
pp. 209–232.
Bradley, B., Byrd, K., Atkins, M. A., Isa, S. I., Akintola, S. L., Fakoya, K. A., Ene-Obong, H.
and S.H., Fish in food systems in Nigeria: A review (Penang: WorldFish, 2020). https://digita
larchive.worldfishcenter.org/handle/20.500.12348/4210
Bush, S. R., Belton, B., Little, D. C. and Islam, M. S. ‘Emerging trends in aquaculture value
chain research’, Aquaculture, Vol. 498, (2019) pp. 428–434.
Chamberlain, G. ‘Analysis of covariance data with qualitative’, The Review of Economic Stud-
ies, Vol. 47(1), (1980) pp. 225–238.
Chan, C. Y., Tran, N., Pethiyagoda, S., Crissman, C. C., Sulser, T. B. and Phillips, M. J. ‘Pro-
spects and challenges of fish for food security in Africa’, Global food security, Vol. 20, (2019)
pp 17–25.

12
Border closures over the 2019/2020 period are also said to have supported growth of domestic
production.

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
on behalf of Agricultural Economics Society.
 14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Demand for Imported versus Domestic Fish in Nigeria 803

Dauda, A. B., Ojoko, E. A. and Fawole, B. E. ‘Economic analysis of frozen fish demand in
Katsina Metropolis, Katsina State, Nigeria’, Journal of Fisheries and Aquatic Science,
Vol. 11(1), (2016) pp. 93–99.
Deaton, A. and Muellbauer, J. ‘An Almost Ideal Demand System’, American Economic Review,
Vol. 70(3), (1980) p#. 312–326.
Desiere, S., Hung, Y., Verbeke, W. and D’Haese, M. ‘Assessing current and future meat and
fish consumption in Sub-Sahara Africa: Learnings from FAO Food Balance Sheets and
LSMS household survey data’, Global Food Security, Vol. 16, (2018) pp. 116–126.
Dey, M. M., AlAM, M. F. and Paraguas, F. J. ‘A multistage budgeting approach to the
analysis of demand for fish: An application to inland areas of Bangladesh’, Marine Resource
Economics, Vol. 26(1), (2011) pp. 35–58.
Dolislager, M. J. Food Consumption Patterns in Light of Rising Incomes, Urbanization and Food
Retail Modernization: Evidence from Eastern and Southern Africa (Michigan State University,
2017).
FAO The State of World Fisheries and Aquaculture 2018 - Meeting the Sustainable Development
Goals (Rome: Food and Agriculture Organization of the United Nations, 2018).
Fashogbon, A. E. and Oni, O. A. ‘Heterogeneity in rural household food demand and its deter-
minants in Ondo State, Nigeria: An application of quadratic almost ideal demand system’,
Journal of Agricultural Science, Vol. 5(2), (2013) pp. 169–177.
Genschick, S., Marinda, P., Tembo, G., Kaminski, A. M. and Thilsted, S. H. ‘Fish consump-
tion in urban Lusaka: The need for aquaculture to improve targeting of the poor’, Aquacul-
ture, Vol. 492, (2018) pp. 280–289.
Headey, D., Hirvonen, K. and Hoddinott, J. ‘Animal sourced foods and child stunting’, Ameri-
can Journal of Agricultural Economics, Vol. 100(5), (2018) pp. 1302–1319.
Hoareau, S., Lacroix, G., Hoareau, M. and Tiberti, L. Exact Affine Stone Index Demand System
in R: The easi Package. Technical Report (2012).
Hovhannisyan, V. and Gould, B. W. ‘Structural change in urban Chinese food preferences’,
Agricultural economics, Vol. 45(2), (2014) pp. 159–166.
Huang, J. and Bouis, H. ‘Structural changes in the demand for food in Asia: empirical evidence
from Taiwan’, Agricultural Economics, Vol. 26(1), (2001) pp. 57–69.
Jimoh, W. A., Popoola, M. A., Ibrahim, G. A., Ayeloja, A. A., Ayanwale, A. O. and Akinosho,
G. A. ‘Evaluation of consumer’s preferences for fresh and frozen fish products in Ibadan
Metropolis Nigeria’, Journal of Agricultural Economics, Vol. 9(1), (2013) pp. 13–20.
Kuku-Shittu, O., Onabanjo, O., Fadare, O. and Oyeyemi, M.Child malnutrition in Nigeria:
evidence from Kwara State, Working paper, Vol. 33 (International Food Policy Research
Institute, 2016).
Kumar, P., Dey, M. M. and Paraguas, F. J. ‘‘Demand for fish by species in India: Three-stage
budgeting framework’, Agricultural Economics Research Review’, Vol. 18, (2005) pp.
167–156.
Lewbel, A. and Pendakur, K. ‘Tricks with Hicks: The EASI demand system’, American
Economic Review, Vol. 99(3), (2009) pp. 827–863.
Liverpool-Tasie L. S. O., Reardon T., Belton B. “Essential non-essentials”: COVID-19 policy
missteps in Nigeria rooted in persistent myths about African food supply chains. Applied
Economic Perspectives and Policy, (2020) http://dx.doi.org/10.1002/aepp.13139.
Magrini, E., Balié, J. and Morales-Opazo, C. ‘Cereal price shocks and volatility in sub-Saharan
Africa: what really matters for farmers’ welfare?’, Agricultural Economics, Vol. 48(6), (2017)
pp. 719–729.
Meyerhoefer, C. D., Ranney, C. K. and Sahn, D. E. ‘Consistent estimation of censored demand
systems using panel data’, American Journal of Agricultural Economics, Vol. 87(3), (2005) pp.
660–672.
Muhammad, A. and Hanson, T. R. ‘The importance of product cut and form when estimating
fish demand: The case of US catfish’, Agribusiness: An International Journal, Vol. 25(4),
(2009) pp. 480–499.

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
on behalf of Agricultural Economics Society.
 14779552, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1477-9552.12423 by Benin Hinari NPL, Wiley Online Library on [11/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
804 Lenis Saweda O. Liverpool-Tasie et al.

Mukhtar, Y. and Gazali, W. A. The dynamics of Fish Trade in North-east Nigeria: A Case Study
of Doron Baga (Universitätsbibliothek Johann Christian Senckenberg, 2016).
Mundlak, Y. ‘On the pooling of time series and cross section data’, Econometrica, Vol. 46(1),
(1978) pp. 69–85.
Ogundari, K. ‘Categorizing households into different food security states in Nigeria: the socio-
economic and demographic determinants’, Agricultural and Food Economics, Vol. 5(1),
(2017) pp. 8.
Steele, M. and Weatherspoon, D. Demand for Varied Fruit and Vegetable Colors (No. 333)
(2016) pp. 1-20. https://ageconsearch.umn.edu/record/235912
Tafere, K., Taffesse, A.S., Tamiru, S., Tefera, N. and Paulos, Z. 2011. Food demand elasticities
in Ethiopia: Estimates using household income consumption expenditure (HICE) survey data,
ESSP working papers 11, International Food Policy Research Institute (IFPRI). Available at:
https://www.africaportal.org/publications/food-demand-elasticities-in-ethiopia-estimates-
using-household-income-consumptionexpenditure-hice-survey-data.
Tambi, N. E. ‘Analysis of household attitudes toward the purchase of livestock products and
fish in Cameroon’, Agricultural Economics, Vol. 26(2), (2001) pp. 135–147.
Tefera, N. U., Desta, M. D., Rashid, S. and Kayitakire, F. ‘A Dynamic analysis of food
demand patterns and habit effects: panel evidence from rural households in Ethiopia’, Journal
of Agricultural Economics Open Access, Vol. 1(1), (2018) pp. 1–13.
Toufique, K. A. and Belton, B. ‘Is aquaculture pro-poor? Empirical evidence of impacts on fish
consumption in Bangladesh’, World Development, Vol. 64, (2014) pp. 609–620.
Toufique, K. A., Farook, S. N. and Belton, B. ‘Managing fisheries for food security: Implica-
tions from demand analysis’, Marine Resource Economics, Vol. 33(1), (2018) pp. 61–85.
Trondsen, T., Braaten, T., Lund, E. and Eggen, A. E. ‘Health and seafood consumption pat-
terns among women aged 45–69 years. A Norwegian seafood consumption study’, Food
Quality and Preference, Vol. 15(2), (2004) pp. 117–128.
Verbeke, W., Vermeir, I. and Brunsø, K. ‘Consumer evaluation of fish quality as basis for fish
market segmentation’, Food Quality and Preference, Vol. 18(4), (2007) pp. 651–661.
Wooldridge, J. M. Econometric Analysis of Cross Section and Panel Data (Cambridge, MA:
MIT Press, 2010).
Zhou, Y. and Staatz, J. ‘Projected demand and supply for various foods in West Africa: Impli-
cations for investments and food policy’, Food Policy, Vol. 61, (2016) pp. 198–212.

Ó 2021 The Authors. Journal of Agricultural Economics published by John Wiley & Sons Ltd
on behalf of Agricultural Economics Society.