A. Carlisle and A.H. Teich
A. CARLISLE is Programme Manager, Petawawa Forest Experiment Station, Chalk River, Ontario, Canada.
A.H. TEICH is Research Scientist with the Harrow Research Station, Ontario, Canada.
In recent times both scientists and natural resource managers have had to contend with the paradox that, although society realizes that science and products from natural resources are the life blood of civilization, there is an increasing reluctance on the part of governments to invest heavily in research or natural resource management, especially when this investment only pays off in the long term. Whatever the reasons for this situation, it is a fact that must be faced, and this, together with inflation, has resulted in keen competition for dwindling research and development money in all sectors. This in turn has resulted in an increased demand for programme justification, so that sometimes one seems to spend more time justifying research than doing the research. In this context of limited resources, it is vital that programmes are designed for maximum efficiency. If they can be made to pay off in the short term all the better.
Tree improvement programmes are very much a part of this picture. They can no longer be carried out as a matter of faith that tree improvement is the right thing to do and will probably pay off in the long run. This has to be demonstrated quantitatively, either by practical demonstration or predictive procedures. Most tree improvement programmes are handicapped by being long term and expensive, and may be looked upon with disfavour unless it can be shown that the probability that benefits will substantially exceed costs is very high. As Porterfield (1977) points out, because tree improvement is such a long term, expensive business, economic evaluation of the work is vital.
The purpose of this report is to discuss in general terms the contribution that economic analysis can make to tree improvement programmes. Other papers in this session by Reilly and Nikles (1977), Porterfield (1977) and van der Meiden (1977) deal with individual problems in greater depth.
Economic analysis has been used in tree improvement programmes in four ways:
· Programme justification - Does tree improvement pay?· Programme optimization - What is the best tree improvement strategy in a particular situation?
· Selection of management alternatives - Should we invest in tree improvement, or silviculture, or mill process improvement, or all three?
· Identification of major cost sources in research and development programmes - Where does all the money go?
The problem with all these analyses, however, is that they are all in cash terms, and exclude many non-quantifiable benefits (e.g. crop security, product uniformity, aesthetics, etc.) that may be as important as the crop cash value, or more so. On the other hand costs can usually be computed precisely, so that we end up comparing minimal benefits with true costs. Fortunately most of the analyses carried out so far indicate strongly that tree improvement pays, in spite of this handicap, so there is no real problem in programme justification. However this imbalance is a potential source of error in programme optimization and selection of management alternatives, particularly the latter where the benefits of some alternatives are more easily quantified than others.
Tree improvement research and development can only be justified if they contribute appreciably to the success of operational tree planting and seeding programmes, or to improvement of trees used in urban situations for amenity purposes. A national relying wholly upon natural forests and natural regeneration for its forest products would not be able to justify a tree improvement programme. In contrast, such a programme is vital to a country relying wholly upon forest planting and seeding programmes to increase or maintain timber yield. Tree improvement has long been important to western Europe where natural forests have diminished considerably. However, even in a situation where planting and seeding programmes create a potential demand for improved tree seed, tree improvement - which aims at increasing crop security and yield, and producing the right kind of product - is only one of several management alternatives. It must be shown that it pays.
The majority of the reports in the literature of tree improvement economics are concerned with showing whether or not tree improvement programmes are profitable. Most of the authors use the method of discounted profits, comparing the expected profitability of the enterprises by comparing current values of future profits. For example, depending on the interest rate, a profit for which one has to wait ten years may only be worth half the amount now. The calculations use the compound interest equation. Also, most of the reports assume that it is necessary to allow for interest on tree improvement and tree establishment costs, adopting the attitude that one could just as well choose to invest in bonds as in tree improvement. This concept has met with some resistance from foresters who feel that this greatly inflates long term costs, and has resulted in a controversy, a red herring that has drawn attention away from the main fact that, even allowing for this cost weighting, tree improvement still seems to pay, provided the research and development results are ultimately used in forestry operations.
This report is not intended to be a comprehensive review of the literature dealing with tree improvement profitability, as this literature has been reviewed in other papers such as those by Carlisle and Teich (1971, 1975) and Dutrow (1974). However, in brief, although the potential of a tree to grow rapidly is controlled by several genes and is difficult to select and breed for, in most tree species we can expect at least 10% gain in growth, and in some 15% to 25% growth gains (e.g. Yeatman, 1974; Holst, 1971; Teich, 1973; Nikles. 1969). Trees can be selected for hardiness and, in some cases, resistance to insects or diseases such as Cronartium fusiforme or Scleroderris lagerbergii (Porterfield, 1964; Teich and Smerlis, 1969). Tree form and wood properties can also be selected and bred for.
YEAR-OLD POPLAR CUTTINGS IN IRAQ making tree breeding pay
All of these contribute to crop value and security or to efficiency of harvesting and manufacture of forest products. In the literature the evidence that these benefits considerably exceed the costs in cash terms is overwhelming. Davis (1967) found that for pines in the south eastern USA, even small increases in yield of 2.5% to 4.0% more than paid for the cost of improved seed, and other papers support this view (Perry and Wang, 1968; Bergman, 1968). We expect far greater gains from tree improvement programmes. Various rates of return on investment in tree improvement programmes have been computed, ranging from 13.2% for poplar improvement programmes in the Netherlands and 12% to 20% in some of the USA southern pine programmes (e.g. Anon, 1975; Swofford, 1968; Porterfield, 1977) to the more modest 4.0% to 6.8% for red pine (Pinus resinosa) and jack pine (P. banksiana) in the USA (Lundgren and King, 1965) and 6.7% for white spruce (Picea glauca) in Canada (Carlisle and Teich, 1971). Davis (1969) showed that even a small (5%) increase in wood quality due to tree selection would raise mill profits by 15% to 41% by increasing yield of product per unit of wood processed and reducing mill processing time.
The report by van der Meiden (1977) describes the spectacular profitability of poplar (Popolus) breeding research in the Netherlands; the programme resulted in an increase in poplar wood output of 45% and an increase in cash output of over 50%. Also Reilly and Nikles (1977) clearly demonstrate the profitability of Pinus caribaea improvement in Australia.
None of the studies carried out so far indicate that tree improvement does not pay; it is usually just a question of which tree improvement strategy pays best. Again it must be emphasized that this favourable picture exists even though the benefits included in the calculations are minimal, excluding non-quantifiable benefits such as increasing crop security.
In recent years the research on the economics of tree improvement programmes has shifted its emphasis from programme justification to programme optimization. The pioneers in this field are Van Buijtenen and Saitta (1972) and Porterfield (1974, 1975), all concerned with the southern pine programme of the United States. They used linear programming to "optimize the allocation of limited resources" available for tree improvement (Van Buijtenen and Saitta, 1972). The models assume a linear relationship between variables, absence of interactions between variables and constancy of the coefficients used. Van Buijtenen and Saitta (1972) used such a model to examine the consequences of roguing in seed orchards, and found that in their case (with southern pines) the maximum amount of roguing that could be done profitably was about 50%. They found that the use of a linear programming model forces the researcher to think systematically. More recently, Porterfield (1974) applied a multiple trait analysis to the southern USA loblolly pine (Pinus taeda) programme. A great deal is known about the genetics of this species and the economics of the operational programmes. A set of tree improvement goals were specified by the forest managers, and the mathematical model was used to find the best combination of tree traits to be selected and the optimum selection intensity for each trait. The traits used were bole volume, stem straightness, crown shape and size, wood specific gravity and resistance to fusiform rust (Cronartium fusiforme). The managers ranked these traits in order of importance, and decided upon a minimum desired percent gain. Porterfield used his model to optimize tree trait combinations and selection intensities so that the genetic improvement came close to programme goals. The economic return from the programme was also computed; using progeny testing and roguing the internal rates of return ranged from 10% to 14%, and without progeny testing, 8% to 13%. It was conservatively estimated that volume gains of 20% or more were possible by manipulating roguing intensity and wild stand selection expenditure. Reilly and Nikles (1977) report that in a Pinus caribaea var. hondurensis seed orchard in Australia, an internal rate of return of at least 14% was obtained without roguing, and a further 5% by effectively roguing.
Porterfield (1977) points out the a major part of on-going tree improvement programmes should be the examination of the effects of parameters upon profits, and consideration of the merit of selecting for each trait and the contribution of practices such as roguing.
Techniques of programme optimization are of great value in tree improvement programme planning and fund allocation. However, in order to use them efficiently, a great deal of basic genetic and economic information is needed, and this is sadly lacking in some tree improvement programmes.
In order to optimize returns from tree improvement research and development programmes it is necessary that the trees be capable of responding to selection and that the results from the programmes be extensively used. A small gain in growth or hardiness is of far greater value if the tree species concerned is planted or seeded over large areas than a much greater gain in a species planted on a limited range of sites. Before we can consider optimization of a tree improvement programme for a particular species, we must be sure that it is the right species, and that our eggs are in the right basket. Planning tree improvement programmes, therefore, requires a knowledge of species variability and an accurate prediction of the extent of future operational use, information that is not always readily available.
Linear programming has a broader application than just the examination of alternative tree improvement strategies to optimize programmes; it can also be used to evaluate the effects of genetic selection and silviculture on wood and linerboard quality compared with the effects of technological control of linerboard quality during manufacture. It helps the manager to understand where to invest his money. An interesting study was carried out by T.L. Hart and A.E. Ferrie (1972) of MacMillan Bloedel Ltd.'s Operations Research Department in connection with the southern pine kraft linerboard programme. The aims were to "evaluate the extent to which quality of woody raw material can be modified by genetic and silvicultural means, in light of the investment required and the return expected", and to "compare the results obtained by this method of modification with the investment return picture of technological methods which will produce equivalent end-products". The example used was a hypothetical southern pine kraft linerboard mill producing 1000 tons (907 metric tonnes) of 42 pound (19 kg) linerboard per day, 345 days per year. The linerboard machine was fully integrated with a large scale pulp mill using batch digesters, and had an efficiency of 90%. Using the linear programming model to examine genetic, silvicultural and mill technology options, seven test cases were solved, using interest rates of 4% to 8%. The optimum solution for the forestry alternatives (genetic improvement, site preparation, fertilization) included, without exception, genetic improvement of such characteristics as volume production, specific gravity and disease resistance. On one soil type, a combination of genetic improvement and forest fertilization was the best option. Mill processing control of linerboard quality was emphasized more at the highest (8%) interest rate. The indicated return on the genetic improvement investment was 17%. It was found that in building the model and interpreting the results, many thought provoking questions were raised. The authors concluded that meaningful linear programming models can be constructed to examine forest management options and their effect on wood quality and volume and the interactions with mill processing. It is clear that models such as these are of great value in the decision-making process.
Again, this applies to a situation where there already exists a great deal of genetic information, economic data and processing techniques which can be incorporated into the model. For many species and situations this information is not available.
Even when subjected to this type of objective, rigorous scrutiny, tree improvement is shown not only to be economic, but to be the best of a number of options. By this time, surely, geneticists should be able to reduce their efforts of demonstrating to the policymakers that tree improvement pays, and move into the more rewarding field of programme optimization.
Economic techniques, ranging from simple record keeping and accounting to complex mathematical models, can be used to identify major sources of cost. Tree improvement is a costly business and very often economic analyses tend to study costs and benefits from the seed orchard stage to the harvesting, often forgetting that before the seed orchard stage there may have been a large investment in research. However, in Canada an approximate calculation (Carlisle and Teich, 1971) indicated that, in the context of a 100000 acre/year (40469 ha/yr) white spruce planting programme, an expenditure of $1.5 million over a 15 year period on research, and $23000 per year on seed production and collection, generates a potential economic benefit of about $832000 per year on moderately fertile sites. So tree improvement still appears to be profitable when one includes basic research costs. Nevertheless it is vital that research costs are kept as low as possible without affecting efficiency. One of our studies (Carlisle and Teich, 1975), using unpublished data compiled by Paul Viidik of the Canadian Forestry Service, described the sources of costs in provenance trials and arboreta at Petawawa Experiment Station in Canada over a 50 - year period. The total cost of the trials, in terms of present day wages, was about $2500 per acre or $6300 per hectare. By far the largest source of cost was in site clearance, at 325 man hours per acre (805 man hours per hectare) or, at present wage rates, about $1950 per acre ($4816 per hectare).
Governments become increasingly reluctant to invest heavily in the kind of research that pays off only in the long term. This means that scientists can find themselves spending more time justifying research than doing it
Utilizing cut-over land for tree improvement experiments and seed orchards following harvesting operations, and thereby avoiding land clearance costs, is obviously much more economical.
Using a mathematical model, it has been shown that, allowing for interest and inflation, accumulated establishment costs far exceed management cost in a plantation operation at the time of harvest, indicating that establishment costs should be one of the prime targets for cost reduction (Carlisle and Teich, 1971). Tree improvement can help reduce these establishment costs by producing fast growing, hardy trees and minimizing tree losses and the need for costly tree replacement operations.
Estimating true costs is often hindered by the lack of adequate cost records. It would be of great value if detailed costs could be kept of all aspects of tree improvement.
In tree breeding programmes the selection of parent trees can involve considerable expenditure. Van Buijtenen and Saitta (1972) reported that 11% of total costs were attributable to this source, and Reilly and Nikles (ibid) found that for Pinus caribaea in Australia it was as high as 30% of the costs (when roguing costs were included) in a programme of selection and use of superior phenotypes for cloning and seed orchard establishment.
It is clear that economic techniques can play a useful role in tree improvement in programme justification, programme optimization, selection of management alternatives and major cost source identification. The most promising current trend is the increasing use of linear programming models to optimize tree improvement research and development; thereby increasing programme efficiency. However, these techniques require a great deal of genetic and economic input, and information in these areas is all too scarce in some programmes. The identification of these knowledge gaps and subsequent provision of the information should be of high priority in the future, so that limited research and development resources can be used most effectively.
References are at the back of the magazine following the last article.
