This case study represents a considerable effort to adjust the system of national accounts in Zimbabwe to include the consumption of domestic fuelwood. The results suggest that using the specific methods, data and assumptions in this study, there is no significant difference between GDP and an adjusted NDP. This case study attempted to apply the Vincent- Hartwick approach of using marginal rather than average costs.
This case study illustrates the difficulties and limitations of estimating net investment and net domestic product in developing countries. First, there were serious data gaps. Improvements in data availability over previous natural resource accounting studies in Zimbabwe for forestry (for example Crowards, 1994) were not as great as initially hoped. Serious constraints were imposed by gaps in growth and yield data on indigenous forests despite major progress by the Forestry Commission in its mensuration programme. While better data are now available on forests cover types, the corresponding data on age class structures, growth and yield still need further developed. The computation of marginal cost, central to Vincent and Hartwick’s approach, was hampered by inadequate data. Better data are required which relate the opportunity cost of collecting domestic fuelwood to distance travelled. Logically, one would expect collection times to rise as deforestation increases. Adger (1993) in a preliminary study of natural resource accounting for Zimbabwe, suggested that average extraction costs for fuelwood were rising. He was unable to derive estimates of marginal extraction costs. These data could be gathered over time through ongoing national surveys under the auspices of the Forestry Commission’s District Offices if funds were made available either to the Commission or the CSO. In this study, the estimated elasticity of marginal cost was quite low. Better measures of value (especially if higher) would change the results dramatically.
The Forestry Commission appears to have many unpublished studies and several were useful in this current research. This raises the question of research in different agencies in Zimbabwe being easily accessible and how institutions can be strengthened to improve sharing of research results. Considerable time was spent in this study soliciting information from various agencies.
The Central Statistical Office does maintain a comprehensive set of national accounts. Experience by the authors suggests that Zimbabwe’s national statistics situation is better than many other countries in the region. However, the fact that current data on depreciation were not available forced the study team to make indirect estimates. While these estimates are felt to be reasonable, there is no way to test if they actually mirror capital depreciation. The CSO suffers the problem shared by many statistics agencies of a time lag between data collection and publication. Largely in response to criticism of long time lags in national statistics by local businesses, the CSO recently published national income data up to 1996. One anomaly however is the inclusion for the first time of estimated national income from the informal sector. This has caused some concern in business circles as to the accuracy of earlier data sets and the issue of comparability. A further problem, again not unique to Zimbabwe, is the level of aggregation of national income statistics. One of the original objectives of the study was to try and adjust in turn, the energy and forestry sector national income accounts. Unfortunately, the energy sector in the national accounts combines water and electricity sectors. This situation provided an insurmountable challenge when trying to isolate wood energy and estimate depreciation. Similarly, forestry had until 1996 been part of agriculture and fishing. Now, this sector is reconfigured as agriculture, fishing and hunting. Forestry is considered by the CSO to be part of agriculture and disaggregating it out proved difficult within the time frame of the study. Therefore, a broader macro-economic adjustment of national income was done.
This study illustrates methods of using a patchy database to estimate net investment and net domestic product. Further refinements could be made in the future. For example, the focus of this study was only on fuelwood. A wider array of forestry related costs and benefits could be evaluated in future studies. As an example, the benefits of forests for carbon sequestration and prevention of soil erosion and river siltation could be assessed in terms of net domestic product. On the cost side, the impact of cutting trees for other products not usually captured in GDP such as domestic building materials, woodcrafts, etc. could be added. Restricting the case study to fuelwood may limit the utility of the research. Although data on these other forest benefits and costs would be lacking in some areas, a long-term research programme could generate some useful information for policy makers.
Given the problems of data and the subsequent need to make a range of assumptions and estimates, why bother estimating NDP within a natural resource accounting programme? Using NDP instead of GDP in an economy highly dependent on natural resources such as Zimbabwe is vital. As shown by the results, the likely adjustments for fuelwood depletions average only around 0.16 percent of GDP (calculated as Hotelling Rent / GDP as a percent)9. This figure would however, appear to be a substantial underestimate given other uses of indigenous forests and issues not dealt with in this study. Still, this percentage adjustment to GDP is quite substantial and indicates some need for a forest policy review in Zimbabwe. The results do emphasise that using GDP as a measure of welfare can be misleading. The consequences could be severe by sending the wrong signals to policy makers. Development might be biased towards short-term goals, which could lead the country down an unsustainable path. For planning purposes, especially in central government agencies such as Planning Commissions and Ministries of Finance, using NDP rather than GDP provides a better measure of natural resource capital and the impact of drawing down this stock on the national economy.
In a perfect world, Zimbabwe would either have adjusted national accounts or a system of satellite accounts for planning and policy purposes. However, as this study has amply demonstrated, the world is not perfect. Questions must be asked about priorities for operational programmes either funded internally by the government, or subsidised by donors. Rather than agonising over whether to use marginal or average costs in calculating net domestic product, perhaps a point of departure in some developing countries is to focus scarce financial and technically-trained human resources on developing better data-bases.
A first stage would be to establish an inventory of change data on natural resources, initially in biophysical terms, and then gradually adding measures of value to the resources and changes over time. The data base programme could be supported by a slightly expanded data collection effort in the CSO as well as existing geographical information systems (GIS) and low-cost remote sensing. With a better database, adjustments to the system of national accounts for many natural resources would be much easier.
One objective of this study has been to identify data gaps in. These gaps could be turned around into targets for new programmes to strengthen database development and management in Zimbabwe and elsewhere. Data must have a purpose, otherwise there is no point is expending resources to collect it. In any data base programme, goals and objectives for identifying data sets, collecting, and using data must be clearly formulated at the outset. Data base development programmes could identify establishing data sets for national income account adjustments as one objective.
This study has demonstrated some of the advantages, disadvantages, and data limitations of trying to adjust GNP for natural resource utilisation. The methods of Vincent and Hartwick can be used, even where data gaps exist. However, to fully capture the benefits of this policy tool, better data are required. This must be a priority in future donor-funded development programmes. Further case studies should be carried out for other natural resource sectors as part of a broader programme of identifying data gaps and refining methodologies.
