M. B. GRAINGER
This article is based on a report by M. B. GRAINGER, who served with FAO, by arrangement with the United Nations Economic Commission for Africa, in order to initiate this study.
Problems connected with FAO's African timber trends study
THE ONE supreme advantage which wood enjoys over most other raw materials is its ease in renewal of supply - but unless future requirements can be foreseen far enough in advance this inherent advantage cannot be rationally exploited. In the case of Africa, relatively few countries know how much wood is being consumed at present - much less what will be needed in 30 to 40 years' time under the dynamic impact of economic and social development. This is why the concept of a timber trends study has become so important to African forestry, for it is the only means of providing the "chart and compass" which are essential if a sound sense of direction is to be maintained in all forest planning and policy decisions.
In 1961 a timber trends study was launched in Africa as a joint undertaking of FAO and the United Nations Economic Commission for Africa. It is designed to provide each country with the following information:
(a) an estimate of current wood consumption by major categories;
(b) a forecast of future requirements for the years 1975 and 2000;
(c) an assessment of the prospective supply/demand outlook.
In most African countries statistical data on wood consumption and forest resources are extremely meager, the forest estate is frequently involved in current administrative changes, and the pressure on land resources is increasing. The timber trends study will attempt to collate all the relevant facts and to lay down broad guides for policy formation at the national level. The emphasis is being placed on the individual African country as the basic unit, because it is only at national level that effective action can be taken in forestry matters.
In addition to following, wherever data are adequate, the obvious total apparent consumption approach (production plus imports, minus exports), the study is directed at obtaining details of consumption by end uses and by consumer categories. The approaches being used are:
1. a very limited sample survey to determine the pattern and extent of wood consumption by the
rural sector of the population in housing, furniture, cooking, agriculture, etc;2. the assembly of building statistics for urban centers, and the computation of wood consumption by using previously established wood-to-value or wood-to-floor area relationships;
3. a series of direct inquiries to large consumers of wood such as railways, mining companies, public
authorities, wood-working industries, etc., including the use of wood in packaging (at the manufacturing stage and/or at the point of final consumption).
The modus operandi is for the national forest authority to collect the basic information and forward it to FAO headquarters for collation, analysis and interpretation. All available forest resource data will be summarized, but no new survey work is being conducted as part of the study, because of the length of time required to produce substantial results.
In very recent years, six countries of Africa have undertaken national wood consumption surveys and forecasts. From 1959 to 1961 a team of three FAA officers directed such studies in Uganda, Kenya, and Tanganyika. One officer, provided by German bilateral aid and supervised by FAO, undertook a similar task in Ghana as an integral part of the African timber trends study. In the Sudan, another FAO officer is actively engaged in the same work. In all these countries, the local costs of transportation and of enumerating and clerical staff have been borne by the countries themselves. South Africa is nearing completion of its own national timber trends appraisal and will supply its findings for the continent study.
These studies have been distinctly of a pioneering nature, for they have been conducted where information has been far from complete and historical series were inadequate for trend analysis. Much has been learned concerning suitable methodology and many facts stand out boldly.
Perhaps the most striking point is that the relatively low consumption of individual rural households aggregates so large in the national total, in contrast to the much more obvious urban public and commercial use. In Uganda and Tanganyika, from one half to two thirds of the sawnwood is consumed by this household group. Even in Kenya, with its urban development as an industrial and commercial center, the rural household sector accounts for one third of sawnwood consumption. Simple joinery - doors, window frames, and shutters - and furniture are the most important end uses. The consumption of building poles in these households accounts for very large volumes of wood - from one half to four fifths of the country's industrial wood requirements, and in total their fuelwood use is enormous. Most of the poles and fuelwood used to come from unreserved bush forests.
Household consumption of sawnwood proved to be closely related to the general level of cash income in the community as well as to the local supply situation. A more than proportional part of increased income is spent on sawnwood, as opposed to the situation in more industrialized countries where expenditures in sawnwood do not tend to increase as rapidly as income or where per caput consumption may even decline as income rises. Sawnwood, of course, replaces rougher pole construction in most instances; this trend transfers importance from the unmanaged bush forest to the managed forest reserves.
In urban areas furniture and sawnwood roofing materials are most important. Even the typical urban structure of concrete utilizes in construction important quantities of poles and sawn material for formwork or shuttering.
Studies which have been completed show conclusively the extreme importance of the large and rapidly growing consumption of sawnwood in rural households and point to the need for more accurate appraisals of the size and growth of demand from this sector.
Many countries have as yet been unable, or have not seen fit, to embark on wood consumption studies. All but a few countries are, however, co-operating in the continent-wide study, although the contribution of some is severely limited by shortages of staff and many other priority tasks. A forest economist of the Economic Commission for Africa has already made two extensive liaison tours throughout the region. On the basis of information gathered during these visits, some problems attending the study are discussed in the following sections, as they may be of practical value to other countries confronted with problems of a similar nature. Emphasis is placed on the sample survey, as it is the more novel and most intricate phase of the study.
One of the first requirements in such a study is reliable population statistics - yet a number of African countries have never taken a full census of their populations. National estimates without any district or even regional breakdown may be all that is available. This is inadequate if a sample survey is to be undertaken, for the rural survey is based on a stratification of the country into homogeneous areas - and it is necessary to have a population estimate at least for each such area or stratum.
In one of the countries visited preliminary preparations were already being made for the first complete census of population, to be taken later in 1962. A prior step, the preparation of a complete electoral roll, was being done by house-to-house canvass, using the official list of taxpayers as a guide. It was found that the total number of electors varied from about 2.3 to 2.6 per taxpayer for different districts and communities; also from a series of supplementary studies it was tentatively estimated that electors (21 years and upward) constituted 66 percent of the total resident population as a general average. In this case it was thus possible to get a very good estimate of local populations by taking the electoral roll district by district, and then adding the ascertained percentage for persons under 21. By this means population estimates, sufficiently accurate for the planning of a sample survey, could be made available in advance of the census. The first phases of the analysis work on the census should be completed in time for accurate population statistics to be used in the final stages of the timber trends study. The census office has made all its working papers for the electoral roll freely available to the forest service - thereby greatly easing the problem of determining population figures for the individual communities contained in the sample survey.
Once the initial problem of estimating population has been overcome, the sample survey must be well designed - and prima facie this will frequently pose a major problem to the typical forest officer, who will usually be familiar with the broad principles of random selection but not with the finer details of sampling designs. It is therefore of very material advantage when a country has one or more trained statisticians available to advise and guide forestry personnel in the design of the sample survey. In one of the countries
visited the central statistical office readily agreed to plan the entire sample survey for the forest service. This is a desirable approach wherever it can be arranged, for it relieves the forest service of a technical phase with which it is not usually familiar, and leaves it with the enumeration and measurement of wood-use which are much more closely allied to its normal work.
Occasionally it will be found that a sample survey cannot be undertaken. For instance, in the case of two territories the population was reported to have become, for a variety of reasons, resistant to any further surveys, and it was therefore necessary to have a "resting period" of at least two years. The issue then was how to obtain the information that a sample survey would have yielded. Fortunately, some very good information was already extant in one of these territories where over the previous few years the forest service had been steadily carrying out its own series of local wood-consumption surveys to provide a sound basis for regional forest working plans. These surveys have already covered approximately 60 percent of the country in such a systematic fashion that the results could be expanded to provide a national figure for rural wood use within the major categories; sawnwood, building poles, and fuelwood.
In the second of these two territories no such system of local surveys had been initiated, and hence the picture of wood-use has to be systematically built up on the spot. The federal statistical service has agreed to collaborate with the local forest authority in obtaining the most realistic assessment possible by pooling all available information and making such limited enquiries as might be necessary to fill the gaps. The federal statistics office will co-ordinate the data for the two territories and will also plan and supervise a sample survey in a third territory.
Regardless of the basis of the national estimate of wood consumption, it will be advantageous to have a close degree of collaboration between the statistical service and the forest authority. The statisticians will frequently have useful information gathered in the course of household budget surveys or as part of a national census of production or distribution. Thus in two countries information on urban fuelwood consumption was found to be available from recent household budget surveys, and one central statistical office offered to include an additional item in a pending survey of household budgets so as to secure information on the annual use of wood for domestic furniture in an important urban area which would not otherwise be covered.
Something should be said on problems encountered in the design stage of the sample survey. The basic approach in the African timber trends study is to stratify the country into homogeneous areas before selecting the sample; and the two major variables being used for this purpose are "wood availability" and "level of income."
The former, which is based on the abundance of forest resources and the access to imported wood supplies, usually is not unduly difficult, since it requires merely that the country be divided under the headings "good," "fair" and "poor" wood availability. However, dividing a country according to the variations in per caput income is a much more difficult exercise. Even when the average per caput income is known for the country as a whole, regional or district estimates may not be available.
To cite a single example: in one of the countries visited there was no data whatever available on the regional breakdown of income levels - so it was a case of proceeding on the basis of informed opinion. Having stratified the country into three wood availability strata, the forest service sought the opinion of the Director of Agriculture on the subject of variations in living standards throughout the country - and the general pattern soon began to take shape. While discussions were proceeding, the Director of Co-operatives was brought into the meeting, and the considered views of these two officials - who were familiar with the regional economy of the country - resulted in boundaries being drawn on the map to delineate four levels of income: "high," , "medium," "low," and "very low." The major contributing factors were: the presence of diamond fields in the high income region; very poor soil fertility in the very low income region; predominant livestock farming in one region; and a particular pattern of agriculture and fishing in another.
This illustrates the value of discussions with agricultural officers and other people whose occupations make them familiar with the broad economic pattern of the country in which they are working. When first confronted with the problem of stratifying for level of income such officers (quite naturally) are liable to take the view that there is not sufficient evidence on which to base a useful appraisal. But, as the discussion proceeds, it will usually be found that the picture begins to take shape. Differences in soil fertility, farming patterns, industrial activity, climatic factors, access to markets, land tenures and other characteristics point to economic contrasts. Considering such indicators and consulting other authorities where possible, the local agricultural service should be able to draw up a stratification of the country which, even though it be somewhat crude, will be more suitable than dismissing the problem as too difficult, because a precise statistical basis is lacking. In the country just referred to, the four "income" strata were superimposed on the three "wood availability" strata, and the result was seven strata for sampling purposes. From each of these strata the forest service proposed to draw a random sample of two rural communities. In another ease it was found that the central statistical agency had somewhat rough data showing average income levels on a provincial basis for the entire country; these data showed the broad order of variation in per caput income and thus could be used as a basis for stratification.
Obtaining material for a suitable sampling frame can sometimes be a problem. In this respect any country which has recently carried out a census of population, or is currently preparing for such a census, will be in a favorable position. Several of the countries visited are taking a full-scale census of population in 1962, and the planning of the census will provide data which can be used for a sampling frame. Close co-operation with statistical and census officers is very important at this stage - for all the units from which the sample is to be selected need to have population figures, or at least reasonably good estimates, attached to them. The selection of the actual areas to be sampled may be a fairly complex technical task, and guidance from a statistical officer is advisable.
The public relations phase, which begins before technical field work, is most important. People everywhere are often suspicious of visits by parties of government employees; sometimes they associate requests for detailed information with taxation and other of the less popular activities of the state. Hence, particular care is necessary to allay suspicion and pave the way for the co-operation of the community. It is very important that the district officer (or his equivalent), and through him the community chief or headman, should be approached well in advance of any fieldwork in order that it may be explained what is proposed and how the information obtained will eventually benefit both the community and the country. To have the project understood and accepted by the persons who are interviewed is essential for good results. A single brief example in a slightly different setting illustrates both the objective and the technique. All around a district which was visited in a country preparing for its first census of population, a bold poster appeared which read:
|
HOW MANY SCHOOLS WILL YOUR |
The enumerators are key men in any sample survey and therefore must be selected with care. In the African timber trends study the practice is to recommend each country to have as many two-man teams as may be needed to deal effectively with the major variations in language, religion and tribe. The enumerator will achieve little if he is not accepted by the group he is dealing with. Often, enumerators can be found within the ranks of the local forestry organization. Wherever this is possible there is one major advantage: the enumerators will start with a certain basic understanding of wood and how it is used. In Liberia, the College of Forestry has assumed the major responsibility for the timber trends study, and it will draw its enumerators for the sample survey from among the student body which has representatives of the principal ethnic divisions of the country. In another country which is undertaking a sample survey the forest service and the collaborating central statistical office propose to employ the students of the forestry school as enumerators. The sample survey, of course, will provide a very useful practical exercise as an adjunct to more formal training. Enumerators with experience in agricultural or demographic census surveys are often available and are most suitable. School teachers with this experience are frequently available during vacation periods. Regardless of their background, a brief training of the enumerators in the use of the questionnaires is essential.
Enumerators must be taught, for instance, to pay serious attention to the manner of their approach when interviewing people. The phrasing of questions and the manner of putting them are very important. Considerable care has to be taken to ensure that questions are properly understood by the householder being interviewed. The enumerator will often be able to check whether the reply is adequate by putting a further question or completely rephrasing the original question.
AS in any household survey, enumerators must be on their guard to detect underreporting associated with taxation fears, as well as overreporting prompted by extreme willingness to satisfy the enumerator. These are aspects of the training of the enumerators that need to be given much emphasis.
Supervisors are necessary to direct the enumerators in the field and to check their work frequently. The practice in all countries visited so far has been for the forest authority to select one or more experienced forest officers for this task in order to ensure a high standard throughout the field work.
Some communities present quite a problem to the enumerator unless clear guidance is given in advance. For instance, in parts of West Africa acute problems may arise at the interview stage in the large compounds of the hollow-square type which are in fact a group of households. For survey purposes the ultimate unit is the household. The general approach is to record the number of people in the household, the quantity of wooden furniture acquired by that household during the previous 12 months, the quantity of wood used for agriculture or miscellaneous purposes, the daily or weekly rate of fuelwood consumption, and particulars of any work done on buildings during the preceding five years - whether it be entirely new buildings, additions, or merely repairs. In the case of these extended households the enumerator's problem is very real. If a household is defined as the entire family group housed in the one building, it will often be found that the head man of the compound will know what work has been done on the fabric of the building within the previous five years - but that he will not know how much furniture each of the individual households has acquired, nor will he know the amount of fuelwood consumed by each. Examples show that in this type of extended family the head has great difficulty in giving the total number of people living in his compound. In a trial run in West Africa nearly 20 minutes of discussion, accompanied by the ticking off of names with fingers, were required before a reply was produced and even then there was doubt.
To question each of the included subhouseholds individually is time-consuming. Experience suggests that the best procedure in this type of community is to select for interview, as the ultimate sampling unit, one of the individual family households included within the extended household organization. At this level it will be relatively easy to determine the number in the household, the amount of furniture acquired during the previous 12 months, the quantity of fuelwood used on an average per day or any number of days, and any wood used for general agricultural or other purposes by this particular family unit. A problem arises with respect to wood used on the building itself - for the several households share one common building. However, each of the subhouseholds has its own allotted section of the combined dwelling - one room, two rooms, or three rooms depending on the number in the subhousehold and the size of the compound building. These rooms can be identified and measured - hence any work done on that portion of the compound dwelling during the preceding five years can be readily distinguished and related to the household unit being interviewed. Some parts of such a compound tend to be common to all or several of the included households, for example, gates, washrooms and even kitchens. If any work has been carried out on these communal facilities during the five-year period, it can be apportioned by expressing the area occupied by any specific subhousehold as a percentage of the entire floor area of the living quarters in the building.
There are many problems associated with the measurement of wood use. One of the most obvious is that wood may be only partially visible when in use, therefore due allowance must be made for items and parts of items which are concealed from view. Poles used for hut walls and roof supports must include an average amount for the part set into the ground, and the same applies to fencing. Individual poles in particular uses such as wall uprights or as rafters will often be of the same length, but will vary in diameter. In these cases, the measurements may be recorded on the basis of the pole of modal cross-section.
Below a certain undefined diameter roundwood ceases to be wood in the accepted sense of the term and becomes mere woody material. From a practical standpoint the enumerators should know the critical dimension below which items in use need not be recorded. Small saplings and vines are frequently used for the lateral lacing of the wall and roof structure - also in the case of compound enclosure fences. Obviously many of these will be 90 small as to be justifiably excluded from any count of wood-use, but there are no clear cut criteria for all regions. As a practical guide, however, it would be logical in most cases to exclude any material below, say, 2.5 centimeters (1 inch) in diameter on the grounds that volumetrically such items are insignificant in relation to the remainder (and in any case such sizes can be grown virtually as an annual crop).
A special problem for many countries is the amount of re-used material - for example, old packing cases made into joinery and furniture items. It is not always possible for the household head or the enumerator to recognize this type of material, and the most suitable approach is usually to make supplementary estimates of the amount of wood that is re-used. Special studies on freight receipts by boat and rail have proved useful in these estimates.
In most of the area in and adjacent to the Sahara desert the population is largely nomadic. Here, the uncertain tenure of village and camp sites gives rise to problems both in planning and in carrying out a household survey. These problems exist in most of Libya, Somalia and Spanish Sahara, and much of Chad, Ethiopia, Mali, Mauritania, Niger, and the United Arab Republic. Nomadism tends to reduce consumption of wood for building purposes, for the fabric and furnishing of a nomad's dwelling have, of necessity, to be very simple and easily transportable. In Somalia, for example, the habitations of the bulk of the population scarcely employ any real wood at all. The framework of the dwellings is constructed from acacia roots bound and molded into hoops which are assembled with a few light braces and supports of roundwood poles, and the entire framework is then covered with woven grass matting. These structures are very easily dismantled for transport by camel, and while breakages do sometimes occur in transit, such losses tend to be more than balanced by the fact that periodic removal ensures relative freedom from termite attack.
It was not possible to institute a sample survey in this country. Although sawnwood was found to be covered quite well by trade statistics, fuelwood and roundwood had to be assessed indirectly by reference to what an average family unit would consume. In addition, there is some industrial usage and also a sizable trade in charcoal and sufficient information was available as a guide to the quantity of wood involved.
The heart of the African timber trends study is the sample survey of rural wood consumption which has just been discussed - and it is the most demanding phase from the standpoint of the time and effort needed to carry it out. The remaining phases are based on a series of centralized enquiries aimed at giving full national coverage. The success of these direct enquiries will depend very largely on three factors:
1. the completeness of the basic list of large consumers which must be compiled in advance;
2. the completeness of the records kept by these consumers and their suitability as a guide to wood used;
3. the zeal with which these enquiries are conducted.
Among other things, a careful watch has to be maintained for any duplication - for duplication can indeed occur. However, the danger that consumption will be omitted altogether is probably more serious than that there will be undetected duplication.
In some cases it may be difficult to obtain reliable data on the quantity of wood used by some of the large users. Financial and statistical records may not be sufficiently itemized; thus expenditures on construction in wood may also include labor costs, cement, and hardware, or a global cost figure may be for an entire job, done on contract, with no details given; or the data in the official financial records may group several projects under one expenditure figure. Detailed copies of individual orders placed and materials delivered will of course provide the information required in such cases - if they are available and someone has the time to search meticulously through hundreds of invoices for references to wood. Frequently, it is not possible for either the forest service or the consumer himself to allocate a man to such a time-consuming task and sometimes it will happen that complete papers for a recent year have not been retained. Perhaps the simplest solution where a large wood-user cannot quote his annual usage by reference to readily accessible data is to ask him to analyse his orders or invoices for a single typical month and then to multiply by twelve. A consumer who buys or uses wood at a fairly even rate over the entire year can probably estimate his daily or weekly usage quite readily. The complex activities of government pose a different and more involved problem. Here a central stores purchasing officer or unit can supply much useful information.
Reasonably good data may frequently be obtained from a percentage of the consumers who receive questionnaires while a significant proportion either cannot or will not supply information. Recording partial data without indicating what proportion of consumers are involved is inadequate. The supervisor should either indicate clearly what percentage of returns are accounted for or estimate the consumption of the defaulters and add it to the information received.
These remarks illustrate just a few of the problems and the possibilities.
The ultimate success of the African timber trends study will largely depend on the care and determination with which each participating country presses its enquiry into the current pattern of wood consumption. The situation in no two countries will be precisely the same, but many of the problems will be similar. Much has to be left to the initiative and resourcefulness of the forest officer supervising the study. Different problems call for different techniques, and if one method fails the supervisor must be prepared to try another approach rather than drop the particular line of enquiry altogether. As a guiding principle, each country should aim to have no significant gaps in its final wood consumption statistics. However, this does not mean that any information is better than no information. On the contrary, any figure which grossly misstates the facts is misleading and is worse than no information. Informed estimates are, however, to be preferred to a blank return.
Whatever the difficulties it encounters in participating in the study, each country can look forward to receiving more in return than it is required to give. The present and prospective wood economy of Africa will be pictured in proper perspective, and each country will be able to see where it fits into this picture and what action is necessary to take care of its future national requirements. The over-all aim of the study is to make a contribution to economic progress throughout Africa by enabling the establishment of national forest production goals and by revealing the opportunities for the development of forest industries.
