INTER-REGIONAL MAIZE MARKET INTEGRATION IN TANZANIA
David A. Nyange
Department of Agricultural Economics
And Agribusiness
Sokoine University of Agriculture
Introduction
-
Food markets in Tanzania, like in other developing
economies, are suspected to be segmented due to inadequate infrastructure.
-
Two trading markets are assumed integrated if
price change in one leads to identical price response in the other.
-
Associated with market integration is the degree
of price transmission, which may have an effect on the speed of traders’
response to move food to deficit areas, especially during emergencies such as
drought, floods or pestilence, before the affected people suffer from hunger.
-
Understanding the degree of market integration
is crucial to appropriate formulation of food security program and policies
such as emergency stock and trade.
-
Market based policies for poverty alleviation
and food security, could be more effective if markets are integrated.
-
If markets are integrated the effect of policy
intervention in one market would be transmitted to other markets. On the other
hand, if markets are not integrated (i.e. segmented) each market would need its
own policy or program which is costly.
-
For example, if there is maize shortage in the
country and markets are well integrated, importing food to the city port of Dar
es Salaam would be sufficient to alleviate food shortage in other markets since
price change due to increased supply in Dar es Salaam would be transmitted to
other markets and make food trade adjust accordingly.
-
Factors
affecting the degree of market integration differ with commodities and from
place to place but most common factors are transportation cost and availability
of market information.
-
Maize is a staple in Tanzania and is produced in
most parts of the country. However, only 6 regions among 20 are self-sufficient
and provide surplus for other regions. The maize surplus regions (in descending
order) are: Iringa, Mbeya, Rukwa, Ruvuma, Arusha and Singida.
-
Inter-regional maize trade is thus of great
importance due to the presence of few surplus and many deficit regions.
-
Average annual maize production is 2.5 million
tons and the 6 surplus regions contribute 50 to 62 percent of the country’s
production.
-
The Marketing Development Bureau estimates the
marketable surplus of maize to be 25 percent (50% for rice).
-
During years of maize shortage, the government
releases grain stock from its Strategic Grain Reserve (SGR). Besides SGR. While
private traders and relief agencies also import food.
-
Food
shortages are sporadic but usually occur in intervals of 4 or 5 years.
Objective
-
To analyze empirically the degree of integration
among geographically dispersed maize markets in Tanzania with a focus on the
impact of Dar es Salaam and Mwanza prices on other markets.
-
If
markets are integrated, emergency food relief program could focus on these two
major markets and be sufficient to alleviate food shortage elsewhere in the
country
Hypothesis
It can be hypothesized
that the cities of Dar es Salaam and Mwanza are focal points for maize price
formation in Tanzania because of the following common reasons;
-
The cities are the largest urban centers in the
country (Population: Dar es Salaam=3; milion, Mwanza=1million)
-
Both cities have ports (i.e. trade centers for
import and export of food grains)
-
Have highest per capita income which implies
effective demand
-
Have food grain processing plants
-
Have
large livestock feed processing factories
Methodology
a) Correlation
Analysis
-
Simple but cannot capture the time lag in price
transmission
b) Ravallion
Model
Assumes a radial
distribution of markets where one central (reference) market with price R, is related to n feeder (regional) markets.
and (1)
for i = 1,...,n (2)
Where there are n regional markets with prices P ; and R is the reference market price. 
is a vector for other
exogenous variables which might influence price formation in market i, such as seasonal changes and government
policy.
The dynamic form of
equation2 with l lags is:
(3)
Ravallion model
modified by Timmer is given by;
(4)
Where;
= the logarithm of the regional market for month t; 
= the logarithm of the reference market for month t; 
= a matrix of exogenous seasonal, regional or other variables
that might influence regional price formation independent of the reference
market, 
= estimated parameters, and
= random error term.
With one lag and without
x variables, Timmer suggested an index of market connectedness (IMC) defined as
the ratio of the regional market coefficient (
) to the reference market coefficient (
). That is:
Or 
.
Benchmark for
Evaluating Market integration
1) IMC
indicates the contribution of the regional market and the reference market past
prices on current regional prices. IMC with values of less than one is an
indication for short-run integration.
2) 
is a measure of the
degree to which the price change, in the reference-market is transmitted to the
regional market. This parameter measures long run market integration and its
value, which is expected to be equal or close to 1.
The difference between
these two indicators is that 
shows the percentage
of the price change in the reference market is transmitted to the regional
market price, whereas IMC indicates the relative percentages of the current
regional price that are originating from regional market and reference market
past prices
Data and Analysis
-
Monthly maize prices from 10 markets (from MDB)
-
The
price series run from September 1991 to April 1996
-
All prices are quoted in Tanzanian Shillings per
Kilogram (Tsh./Kg)
-
Ten markets were selected represent various agro-ecological
zones
-
From the 10 markets, 17 trading market pairs
were identified
-
To eliminate the effect of inflation, prices
were deflated using the consumer's price index (CPI)
-
Ravallion model (equation4) was estimated for
each of the 17 pairs of trading markets
Results
a) Mean Prices
Table 1: Tanzania:
Average Monthly Real Price of Maize in Selected
Trading Markets,
September 1991 to April 1996.
|
Market |
Mean Price Tsh/Kg |
Coeff. of Variation |
Months (n) |
|
Dar es Salaam |
0.963 |
0.173 |
56 |
|
Morogoro |
0.870 |
0.216 |
53 |
|
Iringa |
0.835 |
0.198 |
54 |
|
Mbeya |
0.701 |
0.154 |
53 |
|
Njombe |
0.583 |
0.259 |
53 |
|
Tanga |
0.862 |
0.316 |
52 |
|
Arusha |
0.894 |
0.314 |
52 |
|
Mwanza |
0.932 |
0.310 |
52 |
|
Tabora |
0.778 |
0.360 |
51 |
|
Kasulu |
0.656 |
0.353 |
46 |
Source: Calculated
from the Marketing Development Bureau price data
Figure
1: Maize price (in 1/100Tsh) per kg along Dar es Salaam-Mbeya market channel
Figure
4.3: Maize price (in 1/100Tsh) per kg along Mwanza-Mbeya market channel
-
The highest average price is observed in Dar es
Salaam market, followed by Mwanza.
-
Prices
declines the further the market is from Dar es Salaam or Mwanza
-
Such price patterns support the hypothesis that
Dar es Salaam and Mwanza are focal points of maize price formation in the Eastern
and Western zones of Tanzania, respectively.
b) Spatial Price
Differential and Bivariate Correlation
Table2:
Tanzania: Average monthly Real Price Differentials and Correlations Among Maize
Trading Markets, September 1991 to April 1996.
|
Market Channel |
Road Distance (Km) |
Mean Price Difference (Tsh/Kg) |
Coeff. of Variation |
Corre- lation |
n |
Railway (l=linked or near nl=not linked) |
Road (condition p=paved np=not paved) |
|
Morogoro-DaresSalaam |
193 |
0.102 |
1.304 |
0.717 |
53 |
l |
p |
|
Iringa-DaresSalaam |
503 |
0.135 |
1.015 |
0.643 |
51 |
nl |
p |
|
Mbeya-DaresSalaam |
893 |
0.270 |
0.493 |
0.575 |
53 |
l |
p |
|
Njombe-DaresSalaam |
722 |
0.380 |
0.342 |
0.683 |
53 |
l, near |
p |
|
Arusha-DaresSalaam |
649 |
0.074 |
2.365 |
0.821 |
52 |
l |
p |
|
Iringa-Morogoro |
310 |
0.045 |
3.356 |
0.643 |
51 |
nl |
p |
|
Mbeya-Morogoro |
700 |
0.183 |
0.678 |
0.749 |
50 |
nl |
p |
|
Njombe-Morogoro |
505 |
0.287 |
0.557 |
0.574 |
50 |
nl |
p |
|
Tabora-Mwanza |
363 |
0.162 |
1.111 |
0.802 |
47 |
l |
np |
|
Arusha-Mwanza |
826 |
0.024 |
11.292 |
0.528 |
49 |
nl |
np |
|
Mbeya-Mwanza |
1363 |
0.239 |
1.084 |
0.464 |
49 |
nl |
np |
|
Mbeya-Kasulu |
1329 |
-0.028 |
7.429 |
0.458 |
43 |
nl |
np |
|
Tabora-Kasulu |
360 |
-0.129 |
1.736 |
0.668 |
43 |
l, near |
np |
|
Mbeya-Tabora |
1242 |
0.096 |
2.520 |
0.515 |
48 |
nl |
np |
|
Arusha-Tanga |
435 |
-0.038 |
3.921 |
0.859 |
49 |
l |
p |
|
Tabora-DaresSalaam |
1078 |
0.183 |
1.143 |
0.630 |
51 |
l |
np |
|
Tabora-Morogoro |
885 |
0.087 |
2.611 |
0.560 |
48 |
l |
np |
Source: Calculate from
MDB data
-
For a pair of trading markets, the spatial price
differential is not expected to be zero, due to marketing costs of which the
most important is transport cost
-
The effect of distance on prices is evident as
spatial price differential increases with distance.
-
The negative sign in some trading markets may be
due to the failure of this method to capture time lag, which may exist during
price transmission.
-
Coefficient of variation for the spatial price differential
may shed some light on the degree of market integration as it indicates the
stability of margins. Integrated markets are expected to have stable margins
(spatial price differential) since prices move together as the local market is
influenced by the reference market price.
-
Arusha-Mwanza and Mbeya-Kasulu have unusually
large coefficients thus suggesting that they may not be integrated. These two
market pairs are the worst linked in terms of transportation infrastructure,
with neither paved roads nor railways.
-
Bivariate correlation seems to be inversely
related to distance