«Malaba Mutukula Lwakhakha May 2007 FOREWORD The Informal Cross Border Trade (ICBT) Survey is an economic survey covering unrecorded trade ...»
The first stage involved determining the total number of customs stations to constitute the population size. The customs stations that are known to have informal trade transactions and are strategically situated at the frontier between Uganda and her neighbors comprised the entire population. This implies that inland ports were not considered for monitoring. Others customs stations excluded are those located in insecure places and do not experience any informal trade activities. Other considerations in defining the population were availability of supporting government institutions (like immigrations, revenue offices and police stations), easy accessibility and the volume of unrecorded trade involved.
The sampling frame therefore consisted of a list of all customs stations (twenty seven) in the population domain selected using the above criteria. This was made possible by prior assessment visits by the technical team. The fourteen customs posts were then purposively sampled for monitoring in order to generate estimates for unrecorded trade.
Overall, the customs posts selected for monitoring constitute approximately 90 percent of both formal and informal trade flows and are spread all over the five neighbouring countries.
The use of purposive sampling technique in selecting border posts for monitoring was found appropriate in this peculiar situation. Purposive sampling involves selecting sample units according to a purposive principle to achieve desired objective. Although this method is subject to bias and less efficient compared to stratified sampling, proper application of it with adequate knowledge of the population characteristics could provide a representative sample to generate sufficient information on the population parameters.
The technique was found useful in minimizing resource wastage on sampling units (border posts) that are insignificant or experiencing minimal informal activities, which would have otherwise been selected in the random process.
2.2 Selection of weeks for monitoring
Due to resource constraints, it was not possible to monitor ICBT activities for a full month.
The next stage therefore required specification of two weeks randomly selected from each month for monitoring. The method devised was to divide the month into four weeks, whereby two weeks in a month were monitored and the flows for the remaining two weeks estimated. Ideally, the selection of the weeks to be monitored was supposed to be random to avoid bias. However, due to financial constraints again, it was only possible to select the first week randomly and the subsequent week would follow automatically. This 5 was done to minimize the double costs involved in moving the survey team twice to the field if the weeks were not following each other. For instance, if the month of August was monitored say the second and third weeks, it is denoted as 8.2 and 8.3 respectively.
2.3 Direct Observation Technique The direct observation method of data collection was the most cost-effective way of gathering data under border conditions, which are far from ideal. Under this technique, enumerators were positioned strategically at border posts to record all merchandise entering or leaving the country by observation. All traded goods that were not recorded or officially cleared by the customs authorities were recorded at a point of crossing the customs stations in the counter books and later transferred in aggregated form into the “Summary Form A”.
Prior to fieldwork, the survey team comprising of Enumerators, Supervisors and Coordinators underwent a one-week training on the survey instruments. Field pre-tests on quantification and different packaging was done including developing different units of measures according to international standards. The enumerators were equipped with the metric system techniques of estimation and trained on how to obtain information from those involved tactfully. Besides, nearly all monitored stations had a weighing scale for accurate recording of information on all goods transacted whose weight needed verification.
2.4 Up-rating of Survey Results This was necessary in order to generate monthly estimates from two weeks data for each month monitored.
The up rating of survey results was based on the following assumptions:
(a) The supply and demand for industrial and other products from either side of the border were fairly constant throughout the months of border monitoring.
(b) The supply and demand of agricultural products fluctuate depending on the season, whether planting or harvesting season and the day of the week (e.g. market day). The average value of flows (imports/exports) for a day of the week, say Tuesday is multiplied by the number of times Tuesday occurs in a month. The procedure is repeated for all the days of the week and a sum of the values estimated to get the monthly estimates.
(c) Trade transaction through other crossing points in the neighborhood of the monitored border stations were estimated individually based on qualitative monthly reports that were compiled by the supervisors after every monitoring month. Therefore, the reported percentage of (a) and (b) yields the estimated trade that crossed via the neighbourhood of each monitored border station.
2.5 The Up-rating Model
Under assumption (a) above, for industrial and other products with constant trade flows, consider a given month having n days with a daily average value of industrial and other products of µi. The total value of inflows/outflows of industrial and other products in a
month are therefore mathematically presented as:
Ai= n µi --------------------------------------------------. (1)
Therefore, the aggregate value of inflows/outflows during the survey period is the sum of the estimates of the five months monitored.
Where i = month monitored and AT are total export/import flow for industrial and other product categories.
Equation (2) represents estimated total value of informal exports/ imports of the industrial and other products traded during the 5 months of border monitoring.
These are informal trade flows (exports and imports) of goods in industrial products and other products category that passed through the monitored borders during the full days of the months of monitoring.
To up rate informal trade flows of agricultural and other agricultural products during the five months of the survey, assumption (b) is taken into consideration. The monthly aggregate of agricultural trade flows can be expressed as the sum of product of the number of particular days in a month and the average imports/exports for the day of the week.
Let dj represent the number of particular days in a month, say four Mondays in March 2006 and ¥j the daily average value of agricultural exports/imports of a given day computed from the observed trade figures. Then, B = dj ¥j---------------------------------------------------------------------------- (3) Where B, stands for the monthly total value of trade for a given day, say Monday in a month of agricultural exports/imports (i.e. total of all Mondays).
Therefore, the monthly informal agricultural exports/imports aggregates for all days in a month are estimated as;
Where j represents day of the week, i.e. Monday, Tuesday… Sunday. Note that, the maximum number of times a day of the week appears in a month is 5 times.
Adding the monthly totals for 5 months we get the aggregate informal (unrecorded) agricultural flows as;
7 5 5
Where k, stands for the months monitored which were five in our case (March, May, June, September and November).
Equation (5) represents the estimated total value of informal exports/ imports of the agricultural products traded during the five months of border monitoring.
What remains is to estimate total informal traded goods that passed through the routes in the vicinity of the monitored border stations that were not captured by the enumerators.
From assumption (c) above, the percentages provided for each border post was multiplied by equation (2) and (5) to yield trade estimates through the neighborhood. For instance, informal trade through Busia neighbourhood alone was estimated at 25 %.
Other stations had percentages ranging from 5-25% as trade through nearby routes or “panya’s”. For Busia alone, this is expressed as, 5 5 5
Equation (6) represents informal trade flows (exports and imports) of goods in all categories that passed through the routes within the vicinity of the monitored sites but not captured by the fieldworkers.
A summation of the results from the three equations (2), (5) and (6) gives the up rated estimates of informal cross border trade figures. Hence,
Equation (7) shows the trade estimates from unrecorded/informal transactions with Uganda’s neighbours during the 5 months of monitoring.
2.6 Estimation of missing data for un-monitored months In order to show the magnitude of trade flows for un-monitored months, estimation is necessary to fill the existing data gaps. Filling the gaps would improve the analytical usefulness of trade data so as to allow easy integration of the figures into BOP and National Accounts Statistics framework. The practice of estimating missing trade data is in consonant with internationally accepted standards by international organizations such as UN, UNECA, World Bank, and IMF. The estimation methods stipulated by these organizations are documented in the book entitled, “Manual on Methods of Estimation of Missing International Trade Data in Africa (UNECA 1995).” Two techniques were found useful in attempting to estimate monthly flows that were missed out due to logistical constraints. These are linear interpolation and extrapolation models.
This method estimates intermediate terms of a sequence of which particular terms are known. Consider the line defined by the two points (X0, Y0) and (X1, Y1), and a third point
to be determined (X, Y) lies on this line only if the following relation holds:
(Y1-Y0)/(X1-X0) =(Y-Y0)/(X-X0), ……………………………………………. (8) Suppose that the value of X is known, but not that of Y, Solving for Y from 8 above Y = (Y1-Y0) (X-X0)/ (X1-X0) + Y0 …………………………………………. (9) Re-arranging (9) becomes Y = ((X-X0)/ (X1-X0)) Y1 + (1.0-((X-X0)/ (X1-X0))) Y0 …………………… (10) Equation (10) can be rewritten as;
Y= α Y1 + (1.0- α) Y0 ……………………………………………….. (11) Where α =(X- X0)/(X1- X0) ………………………………………………. (12) Equation (12) is the interpolation factor, while (11) is the linear interpolation model. The Linear interpolation model was applied to determine the values of informal trade transactions for the month of April, August, and October 2006.
2.6.2 Extrapolation Method The linear projection model is based on the assumption that there are no sudden or dramatic changes occurring on conditions affecting growth during the period under review.
The mathematical formula is thus, Yt+n = Yt +bn ………………………………………………………………(13) Where Yt+n is the value being projected, n units from time t Yt is the recent time interval of the historical data and the starting point of projection b is the average amount of growth or decline per unit of time.
n is the number of units of time(e.g. months, weeks, years etc) To use model (13) above, b is estimated using the formula below.
∑ (Yt-Yt-1)/m……………………………………………………………(14) b= i =1 Where m is the historical interval over which the average growth is calculated
2.7 Data Limitation
The ICBT survey had some of the following limitations:
(i) The survey did not cover all the border stations in the country. Some of the border stations that were not covered may be experiencing significant amount of informal cross border trade flows;
(ii) Trade occurring at night and beyond stipulated time of monitoring 7.00 a.m to
6.p.m was not covered;
(iii) The porous nature of the Ugandan borders and failure to monitor all cross border sites led to underestimation of the volumes of informal trade flows;
(iv) The survey took utmost two weeks of border monitoring instead of four weeks in a month, thus failing to capture the transactions for the entire months in which border monitoring took place;
(v) The Direct Observation technique did not accurately estimate the quantities of some traded items especially where assorted goods were involved in one package. Other estimation problems arose as a result of items being transported in packages not transparent, and those in bulk like sugar canes, fruits etc.
(vi) Assignment of values, quantities and units of measure to some commodities was not accurately done because of the nature of the goods traded at respective border stations.
Limitations (i) to (iv) rendered the survey procedures unable to cover 100 percent of the trade in all the monitored sites.
The following measures were taken to address limitations (iv), (v) and to a significant
extent (vi) stated above:
a) Data verification, which included validation, consistency checks and coding in line with international merchandise trade statistics recommendations.
b) Data up-rating Limitation (ii) has not been addressed because of failure to find conclusive technical solution at the moment. Limitation (i) will be addressed after visits mounted to border posts that were reported to be experiencing informal trade flows like the ones along DRC and Sudan border. Expansion of border posts to be monitored will partially tackle limitation (i).
3.0 BORDER PROFILES