Food Markets and Diets in the Democratic Republic of the Congo

Abstract

To inform policy makers concerned with food security, this paper relies on the 1–2–3 household survey (2004–2005) to provide a geographical overview of Congo’s food markets and dietary status. The results of this descriptive study point to inefficient domestic food markets, with Kinshasa being a case in point: it is deficient, poorly connected to its hinterland and highly dependent on imports. Food markets in the Kasaï provinces and the northeastern region are two minor exceptions, while the most competitive food producers are found in Équateur and North Kivu. Despite differential access, five diet types can be identified, with the most energy rich (cassava/palm oil) being consumed in Maniema, Orientale, Équateur, and rural Bas-Congo. In contrast, households in South Kivu and Kinshasa suffer from large calorie deficiencies, which is due to low purchasing power for the former and relatively higher food prices and more demanding social norms for the latter.

Keywords

Decentralization food markets diets 1–2–3 Survey Democratic Republic of the Congo

Introduction

Territorial decentralization is a popular instrument used by developing countries to comply with key donor requirements of good governance and effective policies (Bardhan, 2002). Under certain conditions, decentralization may indeed spur more grassroots political and social participation and commitment, more locally adapted policies, and by consequence, more effective development outcomes. Since 2006, the Democratic Republic of the Congo (DRC) is engaged in such a process of decentralization. Defined by the new constitution (RDC, 2006), this reform not only entails the transfer of a substantial number of functions from the central to lower tiers of public administration but also stipulates the reshuffling of provinces. By 2015, despite lacking important constitutional provisions, the DRC has officially moved from 11 to 26 provinces (Englebert, 2012; Marysse, 2005).

How provincial and central governments will fare in their efforts to generate and redistribute wealth, fight poverty, and ensure food security for all Congolese people greatly depends on the spatial attributes of available information. Indeed, the geography of the country with its tiny coastline in the west, its asymmetric location of the capital city, and its poorly accessible tropical forest in the center, calls for a geographically refined approach to guide future policies (Pourtier, 2008). This claim is further underscored by the country’s unequal provision of basic infrastructure and public services (Marivoet et al., 2018), its pronounced cultural and social heterogeneity (Marivoet, 2009), and its lingering conflict in various locations (especially in the country’s east), all which have shaped local communities in a particular way (Reyntjens, 2005; De Herdt, 2015; Kambale Mirembe, 2012; Titeca and De Herdt, 2011).

The main objective of this paper is, therefore, to meet, in part, the information needs of local, national, and international decision-makers in charge of defining socioeconomic and food policies for the DRC. As a matter of fact, little is known about the functioning of Congolese food markets and its direct implications on diets and calorie adequacy of its population. To illustrate the information deficit, the Food and Agriculture Organization of the United Nations since 2011 has no longer published country-level estimates for a common indicator like undernourishment—for reasons of data insufficiency (von Grebmer et al., 2015). Restricted by the sampling design of the 2004–2005 household and expenditure survey, this paper will therefore describe and discuss food markets and people’s diets across urban and rural areas in each of the 11 former provinces and 56 price zones. Although more recent surveys exist, issues related to data quality and accessibility currently prevent a more updated analysis of DRC’s food markets and diets.

The paper is structured as follows. The “Data and methodology” section introduces the household and expenditure survey data together with the descriptive tools to analyze and map economic food access and various diet types. In line with the methods described, the “Results and discussion” section presents and discusses the main results of this paper. Finally, the “Conclusion” section concludes.

Data and methodology

The paper makes use of the 2004–2005 household and expenditure survey, called Enquête 1–2–3 (henceforth: 1–2–3 Survey), conducted with more than 12,000 households by the National Institute of Statistics (INS, 2005).¹ The survey applies a multi-stage stratified sampling design (two stages for Kinshasa, and three stages for other urban and rural areas) with resulting estimates representative at sector level (statutory cities, provincial towns, and villages) within each of the 11 provinces. The numbering of the 1–2–3 Survey refers to the main thematic areas covered in three subsequent phases: (1) employment, (2) informal sector, and (3) consumption.

For the stated objective, this paper mainly relies on the third phase, which comprises diary and recall data on 12 consumption categories following the Classification of Individual Consumption by Purpose (COICOP).² Whereas, the diary data relate to an average period of 15 days, the recall period stretches to 6 or 12 months depending on the module. In total, around 1.3 million transactions have been recorded by all households in the sample, of which more than two-thirds (68%) are devoted to food coming from different sources, such as purchases, own production, and transfers received in-kind. Using metric conversion tables and derived standardized food prices, all food records could be assigned a metric quantity and monetary equivalent.³

To study the economic dimension of food access, general and specific price levels are estimated and geographically compared across 56 price zones, obtained by combining the three sectors of the country with 26 pools, the latter which were used to logistically organize the implementation of the survey. These price comparisons are then discussed using secondary data sources to gain insight into various food market dimensions, such as efficiency, the location of food flows, and import vulnerability.

Strictly speaking, the available survey data do not allow to qualify food markets as “efficient,” which requires knowledge about the full range of transfer costs to move goods from one location to another. In this analysis, however, food markets will be considered inefficient when prices of comparable food items differ more between two locations than what could be expected given their relative distance—pointing to unexploited opportunities for mutually beneficial trade at current transfer costs (exchange inefficiency), to unexploited opportunities to reduce transfer costs below the current level (operational inefficiency), or to both (Rashid and Minot, 2010).

To estimate the overall food price level, the GEKS (Gini–Eltetö–Köves–Szulc) aggregation rule is applied to convert the set of bilateral Fisher price indices into a transitive index (Deaton and Heston, 2010). Without the latter property, a price conversion from zones A–B followed by a conversion from zones B–C would not necessarily yield the same results as a direct conversion from zones A–C. Whereas, Equation (1) shows the generic formula to compute bilateral Fisher price indices with observed food prices p_i and expenditure shares s_i, Equation (2) presents the GEKS method that takes the geometric mean of all bilateral combinations using Kinshasa (j = 1) as the reference price zone.

P_{F}^{A B} = {[(\sum_{i = 1}^{N} s_{i}^{A} \frac{p_{i}^{B}}{p_{i}^{A}}) / (\sum_{i = 1}^{N} s_{i}^{B} \frac{p_{i}^{A}}{p_{i}^{B}})]}^{1 / 2}

(1)

P_{F}^{A} = {[\prod_{j = 1}^{M} P_{F}^{1 j} P_{F}^{j A}]}^{1 / M}

(2)

for all food items i of N observed in price zone j = 1, A, B, . . . 56

While the combination of nominal income levels and GEKS Fisher price indices allows to estimate food purchasing power, it does not indicate what people finally decide to purchase and consume, which is driven by at least two other factors.⁴

First, real food budget levels depend not only on food prices but also on prices and preferences of nonfood items. If prices for the latter are relatively cheaper than prices for the former, households will allocate their total budget more in favor of nonfood goods. Although the 1–2–3 Survey data do not include standardized nonfood prices, it is generally assumed that nonfood goods are relatively cheaper in cities compared to rural areas. In addition to price, prevailing social norms or individual aspirations to make life more pleasant or less boring may also govern budget allocations between food and nonfood. In this respect, one can observe poor households being compelled or willing to spend considerable amounts on funerals, weddings, dowries, TV sets, or mobile phones (Banerjee and Duflo, 2011) or, which may sound more acceptable, on educating their children when the state is only partially financing this sector (Titeca and De Herdt, 2011). To capture these nonfood aspects, differences in average nonfood budget shares will be analyzed across geographical areas.

Second, individual and social preferences are equally at play within the food dimension, that is when allocating the available food budget over different food items. People not only choose their diet as a function of prevailing food prices (cf. above) but also depending on nutrition knowledge, convenience and taste, or because certain food items are socially preferred (Irz et al., 2016; Young, 2021). Indeed, if food preferences were irrelevant and the only requirement involved a minimal daily calorie intake (with 10% of calories coming from protein and 15% from fats), then a food budget of US$0.12 per day would be sufficient in the DRC to buy a diet consisting of only palm oil and peanuts, way too far from a realistic meal.

Accounting for spatial differences in prices and preferences, various diet types are identified and mapped based on the mean relative calorie contribution of individual food items observed across province and sector. These shares are obtained by converting individual metric food quantities into their calorie equivalents using food composition data (Dos-Santos and Damon, 1987). In addition, inspired by Arimond and Ruel (2004), household dietary diversity is estimated by counting the number of food groups consumed (out of seven) within a 3-day reference period. Despite their simplicity, dietary diversity scores (DDSs) consistently prove to be positively associated with diet quality and child nutritional status (Jones et al., 2013).

To account for differences in size and composition, calorie availability at the household level is expressed in adult male equivalents (AMEs). The selected AME reference is a male person between 30 and 60 years old with a moderate physical activity level, which on average requires 2900 kilocalories per day, whereby exact food energy requirements of all family members are expressed as a fraction of this reference (FAO/WHO/UNO as published in the work of Smith et al., 2006). A similar AME approach is followed within the monetary dimension of food access, but which also includes an exponent to account for economies of scale.⁵

Results and discussion

Economic food access

The adverse economic geography as illustrated in the introduction becomes directly tangible when comparing food price levels across the country. Based on price data derived from the 1–2–3 Survey, Figure 1 presents the GEKS Fisher food index for 56 price zones and should be read as follows: if the general cost of food equals 100 Congolese Francs (CF) in the city of Kinshasa, then the same cross-weighted basket will cost CF 62 in Matadi, and, respectively, CF 36 and CF 34 in the smaller towns and villages around the country’s main harbor.⁶ This example of food prices being 1.6 times more expensive in Kinshasa compared to Matadi is already telling, knowing that this intercity connection is one of the most economically integrated areas of the country.

Figure 1.

GEKS Fisher food price index by price zone, DRC (2004–2005).

More in general, one can easily conclude from Figure 1 that Congo’s food markets are far from efficient. Whereas food prices in Kinshasa are on average twice as high as those in Bas-Congo, South Kivu, and both Kasaï provinces, they are more than three times as high as food prices in the rest of the country—with Équateur clearly being the cheapest province of all. In the villages around Gbadolite, for example, food prices can be as low as 13 percent of Kinshasa’s prices. These observations of course stem from the uneven location of the country’s capital within its domestic market as well as from the adverse topographical nature of the country overall.

In addition to physical geography, many other constraints and transaction costs, such as taxation (legal and illegal), red tape, risk, or imperfect information explain the huge food price differentials between Kinshasa and its hinterland provinces (Goossens et al., 1994). At the same time, this price variation may also reflect relative food shortages and surpluses. The higher food prices observed in the capital city may, for instance, point to a structural deficit of local suppliers in Kinshasa’s periphery to feed a population estimated at 5.8 million people in 2005 (RDC, 2015).

As a result of such structural deficit combined with inefficient domestic markets, it should not surprise that food imports from abroad are quite substantial. This import dependence can be inferred from Table 1, which presents the ten most consumed food items in Kinshasa together with the corresponding prices observed in the capital and in each of the five cheapest pools of the country. Clearly, aside from maize and cassava (two of the basic components to prepare the national dish fufu na pondu), all food items consumed in Kinshasa are either imported or locally produced based on imported ingredients. An example of the latter is the baguettes, which are produced by one of the four industrial bakeries in town using imported wheat. These small breads are sold on the streets for less than CF 100 per piece (200 grams) and therefore have become a popular and cheap outdoor snack. Another source of calories is imported rice coming mainly from Asia (Tollens and Biloso, 2006) and consumed at CF 307 per kilogram. Given its intermediate price in between the two ingredients of fufu, rice with its 7.7% share nowadays takes up an important proportion of the total food outlays of the inhabitants in Kinshasa.

Table 1.

Prices of top 10 food items consumed in Kinshasa, DRC (2004–2005).

Food item		Food share (%)	Mean standardized food prices (CF per kilogram) for Kinshasa and the five cheapest pools
Food item		Food share (%)	Kinshasa	1^st	2^nd	3^rd	4^th	5^th
1	Cassava flour	9.4	439	36 Lisala	44 Lubum’shi	64 Kikwit	67 Dilolo	85 Kenge
2	Frozen fish (mpiodi)^a	8.7	629	468 Matadi	629 Kinshasa	636 Bandundu	657 Kananga	701 Mbanza-N.
3	Import rice	7.7	307	97 Goma^b	261 Matadi	286 Mbanza-N.	307 Kinshasa	318 Lubum’shi
4	Bread (“baguette”)	7.0	452	402 Mbanza-N.	410 Matadi	452 Kinshasa	474 Lubum’shi	552 Mbandaka
5	Maize flour	5.6	281	50 Gbadolite	65 Lisala	65 Manono	75 Kolwezi	88 Lubum’shi
6	Palm oil	4.2	255	57 Manono	62 Gbadolite	65 Isiro	70 Boende	75 Kikwit
7	Sugar (crystallized)	3.3	388	289 Lubum’shi	369 Bandundu	369 Kolwezi	388 Kinshasa	399 Matadi
8	Dry corn (husked)	3.1	143	46 Kolwezi	50 Kikwit	76 Lubum’shi	83 Bunia	89 Mwene-D.
9	Chicken (frozen)	2.9	1346	1018 Matadi	1346 Kinshasa	na	na	na
10	Cassava leaves	2.6	384	12 Mbandaka	13 Gbadolite	20 Lisala	24 Isiro	25 Kisangani
		54.5

Notes: Lubum’shi: Lubumbashi; Mbanza-N.: Mbanza-Ngungu; Mwene-D.: Mwene-Ditu; na: not available.

There were two other pools, Lubumbashi (village) and Kikwit (city), where mpiodi could be purchased at significantly lower prices. Given the geographical location of both pools and the fact that mpiodi is an imported mackerel captured in the territorial waters off the coast of Mauritania and Namibia, these price observations were labeled as unreliable and discarded from Table 1.

This low price for imported rice in the pool of Goma probably reflects the effect of humanitarian aid assistance.

Source: Based on the 2004–2005 1–2–3 Survey (INS, 2005).

Furthermore, as a source of protein, the inhabitants of the capital also rely on cheap imported mpiodi (8.7%), which is a sort of mackerel captured along the Atlantic coastline, and to a lesser extent chicken (2.9%), imported from Brazil and Europe. Both food items are generally of inferior quality, come in frozen form and are then stored in cool chambers across the country. In particular, the very low price of mpiodi (CF 629 per kilogram) guarantees at least a minimal protein intake for many households, but at the same time, it also complicates the prospects of artisanal fishers to enter the market in Kinshasa. Finally, whereas the country around independence was still producing enough sugar and palm oil for its own domestic needs, it is now importing large quantities from Brazil, Malaysia, and the European Union (Tollens and Biloso, 2006). Given the fact that all these imported food items are shipped through the country’s main harbor, price levels from Matadi to Kinshasa are often gradually increasing. As a result, the inhabitants of Mbanza-Ngungu (a town halfway between Matadi and Kinshasa) are able to benefit from their location with prices of mpiodi, imported rice, and bread each time falling within the price range of the five cheapest pools.

Based on road rehabilitation impact studies, Tollens and Biloso (2006) were able to track the origins of domestically produced food consumed in the capital city. For instance, Kinshasa is mainly supplied by its neighboring provinces with cassava from Kikwit, Mbandaka, and the Bas-Fleuve district and with maize from Kikwit and Lisala (Tollens and Biloso, 2006). Looking at corresponding price levels for these two food items, Table 1 again underscores the poor domestic market efficiency with cassava flour being seven times more expensive in Kinshasa than in Kikwit. The same ratio is somewhat lower for maize flour shipped from Lisala (281/65) and for husked dry corn from Kikwit (143/50), but it reaches an extreme level for cassava leaves from Mbandaka (384/12). Probably, the latter price difference is due to the highly perishable nature of leafy vegetables, which also explains the popularity of horticultural activities in the peri-urban areas around Kinshasa. Be it in terms of increased food security or improved livelihoods, the pronounced price differentials reflect the huge potential in reducing market inefficiencies both for the inhabitants of Kinshasa and for the many farmers throughout the country.

Although market efficiency is generally low throughout the country, the picture is somewhat more variable in two regions. First, given similar price levels and their more homogeneous ethnic and linguistic background, together with the relatively small intercity distances, markets appear more efficient in a region covering the provinces of Kasaï-Occidental and Kasaï-Oriental. Indeed, except for the diamond-rich pool of Tshikapa (where food prices are markedly higher), all cities in both Kasaï provinces seem to be well connected to their rural hinterlands which can be inferred from the smaller price differentials between sectors. Yet food prices in general are also relatively higher in both provinces compared to the rest of the country, which may point to a local production shortage. This observation is confirmed by the same study above and primarily relates to the production of maize, being the single most important food item consumed in Kasaï. As a result, maize is typically imported in large quantities from neighboring southern states, such as Angola and Zambia, and to a minor extent from Bandundu (Tollens and Biloso, 2006).

These observations are confirmed by Table 2, which displays prices of the ten most consumed food items in the urban sector of both Kasaï provinces and corresponding data for the entire urban sector. A remarkable difference with the Kinshasa diet concerns the consumption of maize flour: not less than a quarter of total food outlays in the Kasaï region is spent on this food item. As a result, prices for maize flour on average are 7.7% higher in Kasaï (CF 211) compared to the rest of the urban sector (CF 196). Despite the marked difference in demand, this price gap is rather small, which may indeed point to sizable and cheap imports of maize flour to suppress local prices in Kasaï. In contrast, price differences for this product between the cheapest and most expensive pool in Kasaï are considerably lower (3.1) than for the urban sector in general (7.2), which supports the observation of a more efficient local food market.

Table 2.

Prices of top 10 food items consumed in the urban sector of Kasaï-Occidental and Kasaï-Oriental, DRC (2004–2005).

Food item		Food share (%)	Prices Kasaï (urban, CF/kg)				Prices DRC (urban, CF/kg)				Diff. (%)
Food item		Food share (%)	Min.	Max.	Mean	Max./min.	Min.	Max.	Mean	Max./min.	Diff. (%)
1	Maize flour	25.6	109	340	211	3.1	77	555	196	7.2	7.7
2	Cassava flour	10.9	102	310	185	3.0	44	439	183	9.9	0.8
3	Palm oil	8.6	187	786	544	4.2	65	816	382	12.6	42.6
4	Fried sardines (Ndak)	5.7	808	1996	1310	2.5	363	2396	1032	6.6	27.0
5	Salted fish (Bitoyo)	3.8	738	2701	1513	3.7	345	5944	1438	17.2	5.2
6	Dry corn (husked)	3.5	103	313	176	3.1	103	313	153	3.1	14.7
7	Cassava leaves	3.2	57	248	113	4.3	29	443	114	15.3	–1.2
8	Frozen fish (Mpiodi)	3.0	657	1098	859	1.7	331	1992	833	6.0	3.1
9	Multicolored beans	2.5	272	800	466	2.9	175	800	405	4.6	15.0
10	Sugar (crystallized)	2.2	440	1246	697	2.8	289	1420	610	4.9	14.3
		69.0				3.1				8.7	12.9

Notes: Diff.: difference; DRC: Democratic Republic of the Congo; max.: maximum; min.: minimum. For each food item, the minimum and maximum price estimates, respectively, refer to the cheapest and most expensive average price level observed in all pools concerned.

Source: Based on the 2004–2005 1–2–3 Survey (INS, 2005).

For the other food items, the same general tendency occurs, that is prices on average are higher but more equal within the urban sector of Kasaï compared to price levels recorded in all Congolese cities combined. Indeed, the ten most important food items are on average almost 13% more expensive in urban Kasaï compared to the entire urban sector. Except for cassava flour and cassava leaves, where mean prices are similar, this price difference ranges from 3.1% for mpiodi to almost 42.6% for palm oil, each time with the cities in Kasaï recording the higher price. At the same time, average price inequalities for the ten food items level around factor 3 for cities in the Kasaï region and around almost factor 9 for the entire urban sector.

In sum, combining price levels and ranges, the Kasaï economy is at the same time internally more efficient than, and externally more isolated from, the rest of the country. The latter aspect of isolation is also reflected in the poor diversity of the Kasaï diet: whereas, the top 10 food items in Kinshasa represent 54.5% of the overall food budget (see Table 1), this share amounts to 69.0% for the urban sector in Kasaï.

A second region of relative market efficiency is the area shaped by the provinces of Maniema, North Kivu, and Orientale. Again, Figure 1 indicates that food prices are more equal across the region’s constituent pools. In addition, food produced in rural areas seems to find its way relatively easily to local city markets, as prices in the latter are only marginally higher than in the surrounding villages. Knowing that this part of the DRC has been suffering from violent conflict and insecurity for many years, this counterintuitive outcome may relate to the intimate blending of warfare and economic activities, which is constantly fueled by an abundance of mineral resources (Reyntjens, 2009; Vlassenroot and Raeymaekers, 2004). In contrast, apart from increased economic exchange within this area, the protracted instability has greatly reduced commercial traffic with other regions outside the conflict zone. This was the case during the second Congo War (1998–2003), which resulted in economic fragmentation and a disruption of nearly all commercial activities (Tollens, 2003).

Table 3 repeats the previous approach by displaying prices of the ten most consumed food items in the urban sector of North Kivu, Maniema, and Orientale. Apart from a greater variety in sources of protein, the diet of the northeastern region strongly resembles the one of Kinshasa. In terms of food prices, the urban sector of North Kivu, Maniema, and Orientale is on average 14.8% cheaper compared to all cities combined. Except for local rice, fresh fish, and cassava leaves, the northeastern region has a comparative advantage over the total urban sector for all top 10 food items, especially for salted fish, palm oil, and multicolored beans. The lower prices may point to a more self-sufficient food economy or to a better connection with production centers outside the region or even across the border with Rwanda or Uganda.

Table 3.

Prices of top 10 food items consumed in urban sector of North Kivu, Orientale, and Maniema, DRC (2004–2005).

Food item		Food share (%)	Prices region (urban, CF/kg)				Prices DRC (urban, CF/kg)				Diff. (%)
Food item		Food share (%)	Min.	Max.	Mean	Max./min.	Min.	Max.	Mean	Max./min.	Diff. (%)
1	Cassava flour	9.5	128	206	167	1.6	44	439	183	9.9	–8.9
2	Local rice	8.1	226	417	327	1.8	144	491	313	3.4	4.6
3	Palm oil	6.9	65	621	233	9.6	65	816	382	12.6	–38.9
4	Multicolored beans	6.1	175	513	294	2.9	175	800	405	4.6	–27.4
5	Beef with bone	4.1	922	1,030	962	1.1	838	2,417	1,228	2.9	–21.7
6	Dried/smoked fish	3.9	469	1,842	851	3.9	285	2,185	1,009	7.7	–15.7
7	Fresh fish	3.4	642	1,161	810	1.8	145	2,784	759	19.2	6.7
8	Salted fish (Bitoyo)	3.3	448	1,261	832	2.8	345	5,944	1,438	17.2	–42.1
9	Cassava leaves	3.1	34	443	131	13.1	29	443	114	15.3	15.0
10	Sugar (crystallized)	2.7	430	611	489	1.4	289	1,420	610	4.9	–19.9
		51.1				4.0				9.8	–14.8

Source: Based on the 2004–2005 1–2–3 Survey (INS, 2005).

Similar to cities in the Kasaï region, the urban markets of North Kivu, Maniema, and Orientale are relatively more efficient given the limited gap in maximum and minimum average prices observed in all constituent pools. Indeed, compared to price data for all Congolese cities, the average price range (4.0) is more than half the one observed for the entire urban sector (9.8). Moreover, the same range would have been much lower without the high price differentials observed for cassava leaves and palm oil—two items which are difficult to transport given their perishable or liquid nature.

In sum, the urban sector of North Kivu, Maniema, and Orientale resembles the Kasaï region with respect to its degree of internal food market efficiency but differs regarding self-sufficiency or connection with external food markets. The latter observation is again underscored by the greater food diversity characterizing the diet of urban dwellers in North Kivu, Maniema, and Orientale: the ten most important food items here represent only half of the total food budget.

The above observations about the relative efficiency of food markets in the DRC can be reassessed from Table 4, which provides an overview of price levels observed in the three most competitive provinces for each of the 20 most consumed food items in the DRC.

Table 4.

Lowest mean prices and origin of top 20 food items consumed, DRC (2004–2005).

Food item		Food share (%)	Top 1		Top 2		Top 3
Food item		Food share (%)	Province	Price	Province	Price	Province	Price
1	Cassava flour	11.6	Katanga	98	Équateur	127	Bandundu	136
2	Maize flour	7.1	Équateur	91	Katanga	109	Orientale	174
3	Palm oil	6.5	Équateur	116	Maniema	176	North Kivu	219
4	Multicolored beans	3.9	North Kivu	166	Katanga	221	South Kivu	271
5	Cassava leaves	3.2	Équateur	32	North Kivu	46	Bandundu	56
6	Frozen fish (Mpiodi)	2.9	Katanga	494	Bas-Congo	597	Kinshasa	629
7	Dry corn (husked)	2.7	Orientale	83	Katanga	89	Bandundu	96
8	Local rice	2.6	Katanga	168	Bandundu	236	Équateur	245
9	Cassava chips	2.5	Katanga	34	Kasaï-Ori.	63	North Kivu	63
10	Import rice	2.4	North Kivu	261	Bas-Congo	304	Kinshasa	307
11	Dried/smoked fish (other)	2.2	Équateur	565	Maniema	607	Kinshasa	630
12	Fresh fish (other)	2.1	Katanga	450	Équateur	512	Orientale	542
13	Salt	2.1	North Kivu	212	Bas-Congo	282	Équateur	333
14	Sugar (crystallized)	2.1	Kinshasa	388	North Kivu	452	Katanga	473
15	Fried sardines (Ndakala)	2.0	North Kivu	353	Maniema	770	Équateur	770
16	Salted fish (Bitoyo)	1.9	South Kivu	651	North Kivu	697	Orientale	984
17	Plantain bananas	1.8	Bas-Congo	46	North Kivu	68	Équateur	77
18	Meat (other)	1.7	Équateur	343	North Kivu	420	Kasaï-Ori.	431
19	Cassava tuber	1.4	Équateur	22	Bas-Congo	46	Orientale	47
20	Bread (“baguette”)	1.3	Bas-Congo	435	Kinshasa	452	Katanga	474

Notes: Kasaï-Ori.: Kasaï-Oriental. All prices are expressed in Congolese Francs per kilogram.

Source: Based on the 2004–2005 1–2–3 Survey (INS, 2005).

First, the isolated and deficient nature of the Kasaï economy is indicated by its limited occurrence in Table 4. Indeed, only for cassava chips and non-specified meat, respectively, is Kasaï-Oriental the second and third cheapest producer. Given the region’s preference for maize over cassava, the first observation is surprising, while the second is in line with the study by Tollens and Biloso (2006), which identifies the region east of Mbuji-Mayi as an important producer of beef. Second, Kinshasa’s particular connection to global food markets to compensate for its inefficient domestic food economy is confirmed by Table 4. The five products for which Kinshasa figures among the three cheapest “producers” mainly concern imported food items, such as mpiodi, imported rice, sugar, and bread. Third, the relatively low prices observed in the northeastern region of the country are driven mainly by North Kivu, being the cheapest province for multicolored beans, imported rice, salt, and fried sardines and among the three cheapest for not less than 11 food items. Conform Table 1 above, the cheap price of imported rice appears to be due to humanitarian food aid offered through Goma, the provincial capital of North Kivu. Orientale is also among the three cheapest provinces for five food items, including fish, cassava tubers, maize flour, and dry corn, for which the province can rely on favorable biophysical conditions or supplies from Lisala in the west or Bunia in the east (Tollens and Biloso, 2006). Maniema, with only three occurrences in Table 4, is much less competitive compared to the other two provinces within the northeastern market.

The most competitive overall producer is, however, not North Kivu but Équateur: for not less than 12 food items within Table 4, Équateur figures among the three cheapest provinces. As a matter of fact, to prepare the national dish, one should best live in Équateur given that cassava flour, maize flour, palm oil, and cassava leaves are sold at very cheap prices, while the same is true for dried and smoked fish, and meat. Remarkably enough, the lowest price of cassava flour is found in Katanga, where diets typically contain much more maize. More in general, with its ten occurrences in Table 4, Katanga could also be a competitive supplier for several food items, such as cassava flour, maize (flour and corn), and multicolored beans.

Based on this geographical profile and depending on local conditions, the most straightforward strategy to increase economic access to food involves the development of market linkages with Équateur and North Kivu. One solution to unlock the province of Équateur is to improve access to fluvial ports, dredge the Congo River, and increase its navigability to reduce the economic cost of transporting goods along this river to the main consumer centers in the east (Kisangani) and the west (Kinshasa). This recommendation is supported by Ulimwengu et al. (2009), who apart from identifying Équateur as one of the high-production areas, underscored the importance of river transport investments next to road rehabilitation. Other possibilities involve the increase in competition among the few shipping companies that currently control most of this inland maritime trade or to promote agro-processing in river ports to provide a solution for the highly perishable food items (Ulimwengu et al., 2009). For North Kivu, similar recommendations apply, however, these should be combined with increased efforts to end violence and insecurity while trying to maintain some of the established commercial linkages between economic actors in the region.

Diets and calorie intakes

Despite the high spatial variation in prices, five broad diet types that can be identified in the DRC. Whereas, Figure 2 indicates their location, Figure 3 gives an account of their average composition.

Figure 2.

Location of five major diet types, DRC (2004–2005).

Figure 3.

Composition of five major diet types, DRC (2004–2005).

The first diet can be labeled as “cassava based” given that almost half of all calories are obtained through this tuber (diet 1). The other components in this diet include maize and palm oil, recording shares, respectively, around 20% and 10%. This food basket is most typical for the province of Bandundu but can also be found in the rural sector of Kasaï-Occidental and Katanga. The latter observation slightly goes against general expectations as both Kasaï and Katanga diets are more maize oriented. However, the vital imports of maize from southern states as discussed above may be insufficient and primarily targeted at city centers.⁷ As a result, rural households (especially in the more isolated Kasaï-Occidental) may have shifted more to cassava for their daily calorie intake.

While the rural inhabitants in Kasaï (-Occidental) and Katanga may encounter some difficulties in pursuing their preferred diet, this is not the case for their urban counterparts. The diet observed in these cities (and the villages in Kasaï-Oriental) is clearly “maize based” with more than 40% of all calories provided by this cereal (diet 2). Furthermore, cassava and palm oil are the second and third most important food items, with shares of 20% and 10%, respectively. As such and clearly visible from Figure 3, this second diet is almost the perfect mirror of the previous, although the consumption of rice in diet 2 is twice the share of that in diet 1.

A third diet type can be observed in the urban sector of Bas-Congo and Kinshasa, and concerns the importance of bread and rice (diet 3). Both food items account for about 25% of all calories consumed—rice being responsible for two-thirds. This observation clearly aligns with the discussion above and relates to cheap imports of wheat and rice shipped through the harbor in Matadi. Apart from these two cereals, the third diet also has important (and rather equal) shares of maize, palm oil, and cassava.

Another diet type can be found in North and South Kivu, and is characterized by multicolored beans (diet 4). Although this food item does not dominate the food bowl, multicolored beans represent an important share in total calories consumed—a characteristic absent in all other diet types. This share levels at around 18% and is somewhat higher in the rural sector, where this crop is cultivated (Tollens and Biloso, 2006). The most important food crop in both provinces is, however, cassava with a share slightly higher than 30%. In contrast, maize and palm oil are much less important, each of them responsible for around 10% of all calories consumed.

The last diet comes in three versions, but all share the characteristic of being dependent on cassava with large intakes of palm oil as a complement. Both food items indeed make up not less than 50%–65% of all calories consumed. Except for the province of Bandundu and the urban centers in Bas-Congo and Kinshasa, this diet can be found all along the Congo River basin. In Orientale and the villages of Bas-Congo, this diet also contains substantial, and almost equal, shares of plantain bananas, fish, and meat (diet 5a). In the province of Maniema, the dominance of cassava and palm oil within the food bowl is further supplemented with rice, which is responsible for another 15% of all calories consumed (diet 5b). The latter observation stems from the production of local rice in Maniema. Although important rice production takes place in Équateur as well (Tollens, 2003), it is rather the consumption of maize that further characterizes and adds another important slice of calories to the province’s diet (diet 5c). Moreover, this diet also contains a good amount of fish and meat (around 10%).

Accounting for spatial variation in prices and diets, Table 5 illustrates how food purchasing power of households is converted into actual calorie intakes, both expressed per AME, while pointing to two intermediate factors affecting this conversion. Whereas the first relates to the coverage of nonfood needs, captured by the mean share of total budget spent on nonfood items, the second entails the allocation of the available food budget over different items, thus corresponding to a certain level of dietary diversity.

Table 5.

From food purchasing power to calorie intake by sector/province, DRC (2004–2005).

Province	Daily mean expenditures^a (GEKS, CF/AME)		Mean nonfood share (%)		Mean dietary diversity^b (score on 7)		Daily mean calorie intake^c (kcal/AME)
	Urban	Rural	Urban	Rural	Urban	Rural	Urban	Rural
Kinshasa	561	na	45	na	5.81	na	1818	na
Bas-Congo	817	711	38	31	5.47	5.35	2212	3331
Bandundu	752	575	34	29	5.38	4.30	2511	2585
Équateur	876	898	33	31	5.52	5.08	3027	3593
Orientale	746	882	36	26	4.90	4.50	2156	3825
North Kivu	1359	782	45	28	4.89	3.98	2708	3408
Maniema	1202	992	29	30	4.83	4.45	3805	3472
South Kivu	583	429	40	21	5.27	4.41	1722	1427
Katanga	969	827	38	26	5.40	4.81	3256	3939
Kasaï-Oriental	912	679	36	30	5.32	4.61	3077	2732
Kasaï-Occidental	768	601	35	30	4.97	4.46	2690	2847
Total	784	737	40	28	5.44	4.60	2454	3186

Notes: GEKS: Gini–Eltetö–Köves–Szulc; na: not available.

These daily expenditure levels are expressed per adult equivalent and controlled for regional price differences using the GEKS Fisher food index as displayed in Figure 1. As a result, they reflect each region’s purchasing power for food.

Inspired by Arimond and Ruel (2004), dietary diversity has been obtained by applying a 7-point score card to each household’s food basket during a 3-day reference period.

For Kinshasa and Kikwit, the calorie intakes expressed per person are well in line with observations from other small budget surveys executed by PNUD-SOCOMEG in 2000 and FAO in 2002, while this correspondence is less strong for Lubumbashi and Kindu (Tollens, 2003).

Source: Based on the 2004–2005 1–2–3 Survey (INS, 2005).

Let us first compare Kinshasa and the rural sector of Bandundu. Whereas, mean purchasing power for food is very similar in both geographical areas (around CF 570 per day/AME), calorie intakes are much lower in Kinshasa (1818 kilocalories per day/AME) compared to rural Bandundu (2585 kilocalories per day/AME). The reasons for such a difference in calorie intake can be observed from the two intermediate factors: not only do Kinshasa dwellers spend a much larger fraction of their income on nonfood items (45%), but their diet is also much more diverse (5.81) than the one observed in rural Bandundu. The higher nonfood share may relate to the relatively lower prices for nonfood goods in urban settings, but may also indicate that social norms regarding food and nonfood are more demanding in the capital city. By contrast, the range of food and nonfood goods available to rural households in Bandundu is probably much more limited, which explains their poorly diversified diet (4.30) as well as their low nonfood share (29%). Simply stated, households in rural Bandundu spend the lion’s share of their budget on a few food items only; most probably cassava, maize, and palm oil—all three very rich in calories.

The latter comparison between Kinshasa and rural Bandundu is a good illustration of the general difference between urban and rural areas. Sharing roughly the same mean food purchasing power (CF 784 vs. CF 737),⁸ both the average share spent on nonfood items and mean dietary diversity are markedly higher in cities compared to villages, while calorie intakes are 30% lower in urban (2454 kilocalories) compared to rural (3186 kilocalories) areas.

Regarding nonfood consumption, urban dwellers on average spend 40% of their income on nonfood goods compared to 28% for their rural counterparts. Except for Maniema and with differences in degree, the same observation applies to most provinces, which may point to nonfood goods being in general relatively less expensive in cities. In contrast, the higher nonfood shares may also indicate that social nonfood norms are more demanding in urban areas. Accordingly, Kinshasa and the cities in both Kivu provinces may be among those locations exercising the most social pressure on their inhabitants. At the other extreme, rural South Kivu, Orientale, and Katanga may be among the least demanding, with nonfood budget shares ranging between 21% and 26%. However, lacking standardized data on nonfood prices, it is unclear to what extent the differences in mean budget shares are due to variation in price levels or social norms.

Looking at dietary diversity is another way to gain more insight. Assuming that social norms uniformly apply to both food and nonfood items, variation in dietary diversity may reveal differences in social pressure irrespective of nonfood prices. Indeed, why else would the mean diet in Kinshasa be the most diversified of all, while its calorie content figures among the lowest of all? If social norms were completely irrelevant, the Kinshasa diet would probably contain more palm oil and peanuts, both being cheap and calorie-rich food items. Under the same assumption, the high nonfood share observed in urban North Kivu can be considered as a manifestation of relatively lower nonfood prices. Indeed, with a mean dietary diversity being the second lowest of all urban areas, social expectations regarding what people eat might be lower in the cities of North Kivu.⁹ Overall, with a score of 5.44 compared to 4.60 on a scale from 1 to 7, urban households on average have a more diversified food basket than their rural counterparts.

The profile observed at sector level, however, masks quite some spatial heterogeneity between provinces, each of which is characterized by a unique set of location-specific factors affecting households’ level of income, food purchasing power, dietary diversity and calorie adequacy. For instance, the inverse relation between dietary diversity and calorie intake at sector level does not necessarily hold for individual provinces. While it still very much applies to urban households in Maniema and Kinshasa, respectively, displaying the highest (second lowest) calorie intake (3805 and 1818 kilocalories) and the lowest (highest) dietary diversity (4.83 and 5.81), rural households in Katanga have the highest energy intake (3939 kilocalories) and also the third most diversified diet (4.81), and those in Bandundu are ranked second lowest with respect to both indicators (2585 kilocalories and 4.30). Remarkably, the inhabitants of South Kivu also have a distinctively lower calorie intake (1722 and 1427 kilocalories in the urban and rural sectors, respectively), combined with a relatively moderate dietary diversity.

The variable association between food energy levels and dietary diversity is not exceptional. In a study covering 12 Sub-Saharan African countries, Smith et al. (2006) suggested that both food insecurity indicators have quite different distributions, with possible discrepancies stemming from variations in socio-economic determinants, agricultural conditions, local market functionings, and cultural traditions. The findings of this paper very much corroborate these observations in that geographical location matters a lot in how economic livelihoods are shaped and co-determine food insecurity in the DRC. Among the cultural factors, this paper also emphasizes the importance of nonfood preferences and social norms influencing food consumption decisions of households.

Conclusion

This paper examined the geographical variation in access to food and resulting diets across the DRC based on the 1–2–3 Survey data (2004–2005). The importance of such a spatial perspective stems not only from the fragmented economic geography of the country, but also from the ongoing process of decentralization. This administrative reform is intended to improve the information and knowledge base regarding the exact location of food production, trade, and consumption while fostering local solutions for encountered problems. Meanwhile, to guide future policy design and help identify areas of interventions, the findings of this paper allow for a “thick” description and mapping of DRC’s food markets and corresponding food insecurity levels.

Following an extensive spatial analysis of food prices, this paper pointed to the very inefficient nature of DRC food markets. The situation of Kinshasa is a case in point: with overall food prices being twice to four times higher in the capital city compared to its hinterland, Kinshasa is highly dependent on, and extremely vulnerable to, food imports from abroad. Despite the poor functioning of food markets in most other locations, two exceptions stand out. First, the culturally more homogeneous provinces of Kasaï record more equal, but relatively high, food prices, which points to some degree of regional market efficiency combined with external isolation from the rest of the country. Second, the northeastern part of the country, apart from violence and hardship, also experiences more equal and relatively lower food prices. Furthermore, since Équateur and North Kivu are among the most competitive food producers, their further development and integration within the national food economy in general and with food deficient regions in particular should be among the more promising strategies to help address food insecurity in the country.

Despite the variation in economic access to food, five diet types could be distinguished, with the one based on cassava combined with large complements of palm oil being the most energy rich. As a result, households that rely on this diet on average have higher calorie intakes and mainly live in the provinces of Maniema, Orientale, Équateur, and rural Bas-Congo. Next to diet preferences, differences in income level and aspects affecting nonfood budget decisions also determine calorie intake of households. As such, those who live in Katanga and North Kivu have enough to eat as well, thanks to their higher income levels, while urban households in Bas-Congo and Orientale generally eat too little partly as a result of their relatively higher nonfood expenditures. The lowest calorie levels, however, can be found in South Kivu and Kinshasa, the former because of low purchasing power of its inhabitants and the latter due to higher food prices combined with more demanding social norms governing the capital city.

Although some general profiles could be discerned, the findings of this paper very much point to the importance of geographical location in shaping economic livelihoods and determining food security levels across different sectors and provinces in the DRC. In addition, location not only matters with respect to factors closely linked to the food dimension, such as biophysical attributes and food market efficiencies which, respectively, affect agricultural output and local food supplies, also nonfood preferences and social norms may have an important (indirect) effect. As a result, public and private policies aimed at improving food security should adopt a comprehensive and location-specific lens—a view recently underscored again by the 2021 UN food system summit.¹⁰

Footnotes

Acknowledgements

The author is grateful to John Ulimwengu and Ousmane Badiane for their insightful and constructive comments on earlier versions of this paper. The paper also benefited from inputs and remarks made by several participants at the International Conference on Nutrition and Food Production in the Congo Basin (Royal Academy for Overseas Sciences, Brussels, September 30–October 1, 2013). Despite these acknowledgments, any errors remain the author’s sole responsibility.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Wim Marivoet

Notes

Author biography

Wim Marivoet (w.marivoet@cgiar.org) is a Research Fellow in the Africa Division of the International Food Policy Research Institute (Dakar-Senegal).

References

Arimond

Ruel

(2004) Dietary Diversity, Dietary Quality, and Child Nutritional Status: Evidence from Eleven Demographic and Health Surveys. Washington, DC: Academy for Educational Development, Food and Nutrition Technical Assistance Project.

Banerjee

Duflo

(2011) Poor Economics, a Radical Rethinking of the Way to Fight Global Poverty. New York: Public Affairs.

Bardhan

(2002) Decentralization of governance and development. Journal of Economic Perspectives 6(4): 185–205.

De Herdt

(ed.) (2015) Quelle reconstruction? Acteurs Et Enjeux De La Reconstruction Post-conflit En RDC. Report for Belgian Science Policy. Available at: www.belspo.be/belspo/fedra/ta/ta16rapp_fr.pdf (accessed 8 April 2016).

Deaton

Heston

(2010) Understanding PPPs and PPP-based national accounts. American Economic Journal: Macroeconomics 2(4): 1–35.

Dos-Santos

Damon

(1987) Manuel de nutrition Africaine. Paris; Douala, Cameroon: ACCT/IPD/Editions Karthala.

Drèze

Srinivasan

(1997) Widowhood and poverty in rural India: some inferences from household survey data. Journal of Development Economics 54: 217–234.

Englebert

(2012) Incertitude, Autonomie et Parasitisme: Les Entités Territoriales Décentralisées et l’Etat en République Démocratique du Congo. Politique Africaine 125: 169–188.

Goossens

Minten

Tollens

(1994) Nourrir Kinshasa, L’approvisionnement Local d’une Métropole Africaine. Paris: L’Harmattan.

10.

Institut National de la Statistique (INS) (2005) 1-2-3 Survey (2004–2005). Kinshasa, Democratic Republic of the Congo: Institut National de la Statistique.

11.

Irz

Leroy

Réquillart

, et al. (2016) Beyond wishful thinking: integrating consumer preferences in the assessment of dietary recommendations. PLoS ONE 11(6): e0158453.

12.

Jones

Ngure

Pelto

, et al. (2013) What are we assessing when we measure food security? A compendium and review of current metrics. Advances in Nutrition 4: 481–505.

13.

Kambale Mirembe

(2012) Governance of road infrastructure in Eastern Democratic Republic of Congo. IOB Working Paper 2012.10. Antwerp: Institute of Development Policy and Management.

14.

Marivoet

(2009) Decentralizing the Challenges of Poverty Reduction in the DRC. In: Marysse

Reyntjens

Vandeginste

(eds) L’Afrique des Grands Lacs. Annuaire 2008–2009. Paris: L’Harmattan, pp.245–268.

15.

Marivoet

De Herdt

Ulimwengu

(2019) Reviewing Drc’s Poverty Estimates, 2005-2012: Unprecedented GDP Growth without Trickle Down. London: Secure Livelihoods Research Consortium.

16.

Marivoet

Ulimwengu

El Vilaly

MAS

(2018) Understanding the Democratic Republic of the Congo’s Agricultural Paradox: Based on the Eatlas Data Platform. Addis Ababa, Ethiopia: ReSAKSS.

17.

Marysse

(2005) Decentralization issues in post-conflict Democratic Republic of the Congo (DRC). In: Marysse

Reyntjens

(eds) L’Afrique des Grands Lacs. Annuaire 2004–2005. Paris: L’Harmattan, pp.187–208.

18.

Pourtier

(2008) Reconstruire le territoire pour reconstruire l’État: la RDC à la croisée des chemins. Afrique Contemporaine 3(227): 23–52.

19.

Rashid

Minot

(2010) Are Staple Food Markets in Africa Efficient? Spatial Price Analyses and beyond. Food security collaborative working paper 58562. Ann Arbor, MI: Michigan State University, Department of Agricultural, Food, and Resource Economics.

20.

RDC (2006) Constitution De La République Démocratique Du Congo. Kinshasa, Democratic Republic of the Congo: Ministère du Plan et Révolution de la Modernité, Institut National de la Statistique.

21.

RDC (2015) Annuaire Statistique 2014. Kinshasa, Democratic Republic of the Congo: Ministère du Plan et Révolution de la Modernité, Institut National de la Statistique.

22.

Reyntjens

(2005) The privatisation and criminalisation of public space in the geopolitics of the great lakes region. Journal of Modern African Studies 43(4): 587–607.

23.

Reyntjens

(2009) The Great African War, Congo and Regional Geopolitics, 1996–2006. Cambridge: Cambridge University Press.

24.

Smith

Alderman

Aduayom

(2006) Food insecurity in Sub-Saharan Africa, new estimates from household expenditure surveys. Research Report 146. Washington, DC: International Food Policy Research Institute.

25.

Titeca

De Herdt

(2011) Real governance beyond the ‘failed state’: negotiating education in the Democratic Republic of the Congo. African Affairs 110(439): 213–231.

26.

Tollens

(2003) L’état actuel de la sécurité alimentaire en R.D. Congo: Diagnostic et Perspectives. Working Paper 2003/77. Leuven: Faculté des Sciences Agronomiques et de la Biologie Appliquée.

27.

Tollens

Biloso

(2006) République Démocratique du Congo: Profil des marchés pour les évaluations d’urgence de la sécurité alimentaire. Rome: Programme Alimentaire Mondial, Service de l’Évaluation des besoins d’urgence.

28.

Ulimwengu

Funes

Headey

, et al. (2009) Paving the way for development? The impact of transport infrastructure on agricultural production and poverty reduction in the Democratic Republic of Congo. IFPRI Discussion Paper 00944. Washington, DC: International Food Policy Research Institute.

29.

Vlassenroot

Raeymaekers

(eds) (2004) Conflict and Social Transformation in Eastern DR Congo. Ghent: Academia Press.

30.

von Grebmer

Bernstein

de Waal

, et al. (2015) 2015 Global Hunger Index: Armed Conflict and the Challenge of Hunger. Bonn; Washington, DC; Dublin: Welthungerhilfe; International Food Policy Research Institute; Concern Worldwide.

31.

Young

(2021) Measuring palatability as a linear combination of nutrient levels in food items. Food Policy 104: 102146.

Food Markets and Diets in the Democratic Republic of the Congo—A Geographical Overview (2004–2005)

Abstract

Keywords

Introduction

Data and methodology

Results and discussion

Economic food access

Diets and calorie intakes

Conclusion

Footnotes

Acknowledgements

Funding

ORCID iD

Notes

Author biography

References