| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a International Food Policy Research Institute, 2033 K Street, N.W., Washington, DC 20006-1002
b Dep. of Agricultural and Resource Economics, Univ. of California, Giannini Hall 207, Berkeley, CA 94720-3310
c Dep. of Agricultural and Resource Economics, Univ. of California, Davis, CA 95616-8512
d Centro Internacional de Mejoramiento de Maiz y Trigo (CIMMYT), Apartado 6-641, 06600 Mexico DF, Mexico
* Corresponding author (p.pardey{at}cgiar.org)
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIAL AND METHODSRESULTS3DISCUSSIONGLOSSARYREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTS3DISCUSSIONGLOSSARYREFERENCES |
|---|
These questions all have economic dimensions, and answering them with any precision is problematic. Moreover, quantifying the benefits of conservation and attributing them to different functions are difficult given the variables involved and the need to draw conclusions largely by inference (see, for example, Pardey et al. 1996). Many modern genebank facilities are so new that insufficient time has elapsed for breeders to establish usable time series of realized gains attributable to their establishment.
Beyond immediate agronomic values that are in principle estimable, conservation of crop genetic diversity also has value in terms of yet-unidentified future demand ("option value") and the sheer value of its very existence as opposed to extinction ("existence value"). Though methodologies do exist to assess the overall economic benefits from conserving seed, empirical results are bound to be very imprecise. The cost side, on the other hand, predominantly involves items that are at least estimable in principle from historical data relevant to existing genebank operations. If the total and marginal costs of the genebanking operations are judged to be less than any reasonable lower-bound estimate of the corresponding benefits, then it may not be necessary to confront the challenge of precisely estimating the latter, to establish the economic justification of the genebank operation, and to estimate the financial commitment required to underwrite the genebank operations in perpetuity.
The foregoing rationales motivate this study of the cost of estimating ex situ conservation. The example we consider is the genebank facility at CIMMYT, the International Maize and Wheat Improvement Center headquartered at El Batan, Mexico. The CIMMYT case offers an instructive opportunity for comparing the management of two different types of germplasm—maize and wheat—by two different crop programs housed under the same roof.
This study will look at the basic activities required to conserve an ex situ collection—including seed storage, seed health, germination testing, regeneration, and data-management—as well as other activities related to the addition of new accessions and the dissemination of conserved material to plant pathologists, entomologists, breeders, and other genebanks. We use methods that are designed to furnish upper bounds on the relevant cost concepts as conservative thresholds for the benefits needed to justify the genebanking operation as a whole, as well as conservation of additional accessions (samples of seed of a given variety). The growth in size of the CIMMYT wheat and maize collections is illustrated by the data in Table 1.
|
The CIMMYT maize holdings are based on a collection first assembled as part of the joint Rockefeller Foundation–Government of Mexico program initiated in 1943 to improve the productivity of basic food crops in Mexico. A Mexican seed bank was established in 1944 by the Office of Special Studies, and by 1947 its maize collection had grown to more than 2000 samples (mainly landraces). By 1985, a further 1500 samples had been added to the CIMMYT collection, which grew at a more rapid rate thereafter to its present size of more than 17000 accessions (Table 1).
The construction of a new genebank facility as well as renovation of the ancillary offices was completed in 1996. For the first time in CIMMYT's history, the maize and wheat collections were consolidated into a single modern facility for medium- and long-term storage. The main structure of the new genebank facility consists of a two-story, fortified-concrete bunker. The upper (ground) level of the storage rooms houses the active collection, held at just below freezing point (-3°C), and 25 to 30% relative humidity. This constitutes the "working" part of the bank, from which seed requests by CIMMYT and other scientists are filled. The lower (below-ground) level consists of the base collection stored at -18°C, primarily for long-term storage. The seeds are stored on movable shelves to optimize use of the available space.
The size of the seeds is an important source of distinction between the maize and wheat holdings, and is a distinction that has significant management and cost implications. A stored sample of wheat at CIMMYT is 250 g in the working collection (about 7000 seeds), and 100 g in the base holdings (around 3000 seeds). A working sample of maize is 3 kg (from 6000–10000 seeds) and a base-collection sample is about 1 to 1.5 kg (about 2000–5000 seeds). Wheat accessions are stored in aluminum-laminated bags about the size of a 1 lb bag of coffee, while maize accessions are stored in 1 gal plastic containers in the active collection and aluminum-laminated bags in the long-term collection. The new facility allocates 240 m3 of both medium- and long-term storage space to each crop program, sufficient to store 390 000 wheat accessions and 67000 maize accessions in the long-term collection. If present rates of growth in the size of the respective collections persist, it will take 53 yr to fill the space allocated to wheat and 50 yr to fill the space set aside for maize.
|
MATERIAL AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTS3DISCUSSIONGLOSSARYREFERENCES |
|---|
Economies of scope are cost savings resulting from diversifying the genebank operation, wherein fixed or quasi-fixed inputs can be shared across different aspects of the operation (Bailey and Friedlaender, 1982; Baumol et al., 1988). Input sharing can also extend beyond the genebank. CIMMYT's genebank has ready access to field operation and maintenance crews, seed-health staff and facilities, and various other services (e.g., fund raising and management, publications, and computer support) conducted as part of the center's primary crop-improvement mission. Thus, consolidating the wheat and maize collection in a shared facility as part of a broader crop-research operation offers the prospects of significant cost savings compared with maintaining each crop collection in separate facilities.
We identified three classes of costs: those that were sensitive to the scale of the operation (treated as variable costs), those that were not scale sensitive (fixed or capital costs), and a group of costs that were neither fixed nor variable, but lumpy nonetheless (quasi-fixed costs). Some per-unit costs varied according to the size of the genebank facility; others varied according to the number of accessions stored in the genebank, which is related to but different from the size of the facility. Per-unit costs also varied according to the number of accessions processed (i.e., the throughput) for germination testing, regeneration, and seed-health testing activities. The amount of throughput is linked to the number of accessions held, and depends on the various seed-management protocols which are affected by the specifics of each crop and the history of the operation. Finally, some elements of the cost profiles are sensitive to the number of accessions disseminated in a given year: again these costs are related to, but not necessarily determined by, the size of the genebank holding.
Cost of Capital Inputs
A breakdown of the capital costs related to the genebank facility and the costs of the equipment used in CIMMYT's genebank operation is provided in Table 2, on a current, replacement-cost basis. Column 1 of Table 2 gives the presumed service life of each class of capital. Costs that were common to storing the maize and wheat collections were allocated equally to each crop.
|
While the purchase price of the capital items indicates the investments required to replicate the CIMMYT genebank facilities, they are not directly usable for one of our primary purposes, namely to provide a representative annual cost of the CIMMYT genebank operations. The annual "user cost" of capital items was obtained by calculating the annualized, present-value cost of capital based on information about the purchase price of each capital item, and assumptions about their respective service lives, and the real rate of interest. Annualized capital costs are shown in the two right-hand columns of Table 2, calculated with an interest rate of 4%. We assumed a "one-hoss-shay" depreciation profile; the capital good survives intact until the end of its life and then disappears all at once. The algebra for these cost calculations is spelled out in Pardey et al. (1999).
Storing Seeds
Maintaining the storage areas in the genebank at a precise, stable, low-temperature, low-moisture (i.e., low relative humidity) regime is a reasonably costly exercise. The variable costs of controlling the climate in the storage areas include the cost of electricity to run the compressors, dehumidifiers, and fans, the labor costs of maintaining this equipment, and the related costs of operating an emergency backup power plant. Allocating these types of costs to the germplasm facility is difficult as they represent only part of the overall costs involved in operating the institute's physical plant. To arrive at the estimates in Table 3, we directly costed the energy required to maintain the genebank at its specified climate regimes, and also estimated the costs of a routine schedule of maintenance on the climate-control equipment and the backup power-generation unit (Table 3).
|
Germination Testing and Regenerating Seeds
Since stored seeds gradually lose their viability due to aging, their germination rates must be checked periodically. For wheat, the monitoring and regeneration procedures followed by CIMMYT begin with a germination test when processing newly introduced seed to the genebank or after its regeneration. A large share of the costs in assessing viability consists of the costs of the labor used to carry out the tests, but additional costs (including the costs of establishing and running a suitable laboratory with germination chambers) must be factored in as well. The operational costs associated with germination testing are reported in Table 3, along with the respective annualized capital cost from Table 2.
Wheat accessions at CIMMYT were regenerated in a screenhouse (for wild or selected other species) or in fields at El Batan (CIMMYT's headquarters) or in Mexicali, a field site located in the state of Baja California Norte near the California border.2 Most of CIMMYT's maize accessions obtained from tropical maize-growing areas of low and intermediate elevations are regenerated at Tlaltizapan, while El Batan is used for germplasm obtained from the tropical highlands.
The amount of inputs such as irrigation, agrochemicals (including fertilizers), and management time varies according to seasonal and other factors. Many of these costs are not explicitly itemized in CIMMYT's accounting system, so we first estimated the typical quantity of each of the inputs used for preparing the land and then planting and harvesting the seed, priced each item accordingly, and then derived the corresponding costs (Table 4). The benchmark, field-related cost of regenerating seed at El Batan is $1073 ha-1 and $1009 ha-1 at Tlaltizapan. Recently, both the wheat and maize programs have out-sourced some of these regeneration and evaluation activities to other, non-CIMMYT field sites, which may help reduce these costs in the future. In our survey year, INIFAP (the Mexican national agricultural R&D agency) charged CIMMYT $1217 ha-1 for such services at a location near Mexicali.
|
Processing Seed Accessions for Storage
If regeneration is performed, processing a newly acquired accession to the genebank is much like regenerating an existing accession, but involves certain additional treatments. Newly acquired seed is inspected thoroughly upon arrival for any known or suspected seed-health problems. Wheat and maize seeds are then deep frozen until processing to kill any insects. The first regeneration is performed on specially quarantined introduction plots designed to control for pest problems. The seed-health unit inspects the plants during this process as well as the resulting seed. After harvesting from the introduction plots, maize seeds are collected as bulk samples and added directly to the genebank—wheat seeds planted out at El Batan undergo a further round of regeneration at Mexicali to improve the quality of the seed in readiness for storage. In addition to the seed-health aspects, various characterization and data-entry activities are performed before an accession is finally added to the collection.
It typically takes much more time to clean and sort manually and inspect maize seeds than it does wheat seeds: each ear of maize must be sorted individually by hand to remove broken or diseased seed. Although wheat seeds are intrinsically easier to handle, they do require comparatively more attention for aspects of seed health, as discussed in more detail below. Both maize and wheat accessions require a similar amount of labor to record relevant data in field books, but the higher planting density for (and smaller growth habit of) wheat affords it some efficiencies (time savings) compared with maize. Before placing them in storage, all seeds are dried to reduce their moisture content after harvesting and cleaning. The costs of operating and maintaining the drying equipment are included in Table 4, along with the annualized capital costs from Table 2.
Each wheat and maize accession is stored at CIMMYT headquarters in two sets of containers—one goes to the active collection, the other for long-term storage in the base collection. Each wheat accession is stored in an aluminum bag both for the active and long-term collections at a cost of $0.11 per bag; each maize accession held in the active collection is sealed in a plastic bucket costing $2.80 each, while each accession stored in the base collection is placed in two aluminum bags costing $0.15 each (the bags used for maize are the same type, but bigger than the bags used for wheat). In addition, a sample of each accession (10 g of wheat seed and 1.5 kg of maize seed) is prepared for backup storage in the U.S. National Seed Storage Laboratory (NSSL) at Fort Collins, CO.
Seed Health
All newly acquired material is subject to seed-health checks before being included in the genebank. The health of all out-going seed must also be certified; the cost schedules reflect this. However, we took care not to double count health costs—in all but exceptional cases the checks done at the time of introducing or regenerating maize seed suffice for subsequent shipments made from the collection. Wheat seeds are checked when acquired and again at the time samples are packaged for shipment. At CIMMYT, most of the relevant seed-health activity and the associated costs are the responsibility of CIMMYT's seed-health unit. The capital costs incurred by these activities are identified in Table 2, and the labor and other operational costs for the genebank operation are included in Table 3 as parts of acquisition and dissemination costs—recognizing that only part of the seed-health operation relates to accessions coming into and being shipped from the genebank, and so only part of the overall seed health costs are included here.
Seed Dissemination
Distribution from the genebank takes various forms. Some material is used by genebank personnel for characterization or evaluation purposes. Other material is distributed in response to individual request from breeders, plant pathologists, and others at CIMMYT or elsewhere. Seed is also sent to other genebank facilities, often in the context of CIMMYT's joint collection and conservation work with developing-country national agricultural research systems (NARSs). The cost of responding to such a diverse set of seed requests includes determining which seeds are most suitable to fill the request, and then assembling, treating, and packing the samples to be sent, as well as the associated shipping costs.
Some aspects of these costs are sensitive to the number of shipments made (as distinct from the number of accessions shipped), as well as the size and destination of each shipment. Relatedly, each shipment to destinations outside Mexico is subject to phytosanitary controls, which are reasonably time-intensive and costly. Aside from the cost of the certificates themselves (payable to the Mexican government), the staff of the CIMMYT seed-health unit and the genebank must prepare the necessary documentation and arrange for the shipment itself. In addition, shipments of seed from CIMMYT must be accompanied by a Material Transfer Agreement that assigns use rights to the seed and this documentation must be developed, logged, and tracked.
Table 5 summarizes the shipments made from the genebank since 1987, along with information on the number of shipments, as distinct from the number of accessions shipped. More wheat than maize accessions are shipped abroad and there are fewer wheat shipments per year. Thus the average number of accessions per shipment is 220 in the case of wheat and just 34 for maize (Table 5).
|
The backup collections for wheat are shipped and stored in "black-box" fashion. A 10 g (around 350 seeds) sample of each wheat accession is prepared, labeled, and packed in aluminum foil bags and then put into a cardboard box containing up to 400 accessions. The boxes are airfreighted to the backup facility where they are stored. The expense of preparing the samples and packing each black box are included in the costing calculations: freight costs from CIMMYT to NSSL for the 1996 shipment of black boxes totaled $342 for 35000 duplicates. Wheat duplicates are cumulated and shipped on a periodic basis to save shipping costs.
CIMMYT's maize holdings are duplicated and stored as an integral part of the NSSL collection, rather than in black-box fashion as for wheat. All new introductions and regenerated accessions are shipped to NSSL annually, and about 80% of the CIMMYT maize collection was backed up at NSSL by 1996. Between 1500 and 2000 accessions are shipped each year in cloth bags after the regenerated seed is dried. The NSSL repack and store the accessions in aluminum bags. CIMMYT identity numbers are entered into their data-management system, along with information on the amount of seed in storage and its germination status.
Data and Information Management
Fundamental to the genebank is the management of the information that describes each accession. However, operationally (and for costing purposes) it is difficult to separate data and information used in the effective management of genebanks from the data that are generated by and facilitate the breeding program at CIMMYT and elsewhere in the world. Some of the data serve multiple purposes. Standardizing accession identification numbers, common protocols for recording and reporting performance-evaluation data, and compatible software procedures for recording, storing, retrieving, and analyzing such data can yield significant benefits in the use of this information for both seed conservation and breeding purposes.
Routine operations of the genebank include entry of "passport" data (detailing the source and origin of the seed) at the time the accession is introduced to the collection, the processing of field book observations collected upon introduction and at all subsequent regenerations, and database maintenance to track timing and location of storage, seed viability history, and stock levels of each accession. Barcode labeling of each maize accession in the genebank is being introduced to streamline this process. The maize bank has also begun scanning a picture of each ear of corn, for eventual incorporation into its database.
|
RESULTS3 |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTS3DISCUSSIONGLOSSARYREFERENCES |
|---|
|
|
The annualized costs of the physical capital used to maintain CIMMYT's present wheat and maize holdings are quite similar, even though about seven times more wheat than maize accessions are currently stored at CIMMYT. Perhaps not surprisingly, the storage component accounts for the majority of these capital costs: over 65% for both crops. About 30% of the capital expenses are incurred in regenerating the seed. The spending each year for lumpy labor services (i.e., quasi-fixed labor) is slightly more for wheat ($121648) than for maize ($119261). In fact, it is the cost of senior scientific and technical staff that constitutes the largest share of the fixed costs for both crops (64% for wheat, 67% for maize).
While there is little difference between CIMMYT's wheat and maize operations in the annual cost of lumpy labor and physical capital inputs, Fig. 1 highlights the substantial differences in the structure of variable costs. The maize program spends considerably more each year than the wheat program on regenerating its holdings. Indeed, regenerating seed accounts for almost 50% of the variable costs for maize and only 23% for wheat. These differences are largely attributable to the substantially higher amount of labor required to regenerate maize while minimizing genetic drift in this heterogeneous, out-crossing plant.
Costs on the Margin
Given the genebank is operating well below capacity, the average variable and noncapital costs per accession detailed in Table 7 provide upper-bound estimates of the corresponding marginal costs.6 It is these marginal costs that are central to assessing the economics of changes to the genebank operations on the margin or over the short run. For example, what is the cost of storing an existing accession for one more year, or, equivalently, what is the benefit in terms of cost savings from eliminating a duplicate accession from the genebank? The answer depends, obviously, on the crop in question, and perhaps less obviously on the excess capacity of the genebank as well as the state of the sample, including its time in storage, time to last regeneration or germination test, and the like. If the sample is known to be viable, the marginal variable costs of holding over an accession of either crop for one more year is negligible—just $0.11 for each accession of wheat and $0.52 for an accession of maize. If we consider the average noncapital costs (which include quasi-fixed as well as variable costs), the corresponding figures are $0.19 for wheat and $0.93 for maize. However, if the viability of the seed needs to be checked and the sample regenerated because it failed the test, the cost of keeping it for another year jumps dramatically to $3.45 for each wheat accession and $109.63 for each sample of maize. Clearly there can be substantial cost savings from eliminating duplicate accessions. In fact it would be economical to spend upwards of $109 to ascertain if a maize accession was duplicated in the CIMMYT holdings—or, perhaps, for that matter held in collections at other sites, given the cost of shipping in seed, if needed, is comparatively low and falling (Table 7).
|
Table 7 also provides the costs of distributing accessions, which includes the cost of maintaining seeds in medium-term storage as well as the costs incurred in shipping seeds. If there is sufficient stock in the active collection, the cost of distributing an accession is $4.16 for wheat and $15.27 for maize. The larger size of the maize seed samples contributes to their higher cost of dissemination. If the seed stock is insufficient so that regeneration is required for distribution, the cost jumps to $7.42 for wheat and $123.98 for maize.
We can also answer the following question from Table 7: does it cost CIMMYT less to keep an accession in medium-term storage for another year or, alternatively, discard the holding if the same accession can be acquired as needed? According to our estimates, it is clearly cheaper to hold on to an existing accession of wheat and maize for one more year ($0.10 for wheat and $0.77 for maize) than to introduce that same sample from elsewhere (estimated to be $1.96 for wheat and $4.84 for maize, including the costs of acquisition and long-term stroage), providing the existing accession needs no regeneration. If an existing accession needs regeneration because its sample size is too small, for example, then it is cheaper to introduce an accession that requires no regeneration for both crops.
This type of cost calculus, and its implied management responses, are even more complex if we allow for the interplay of time and costs. The dissemination data in Table 5 suggest that many genebank accessions sit untouched for many years. Indeed, it is the option value of these accessions rather than their more immediate use value that is the justification commonly cited for establishing and maintaining a genebank. However, that option value can only be realized if at some future date the sample is called upon for breeding or other research purposes. Rather than comparing the cost differentials of holding on to an existing accession versus introducing that same accession in the current year, a more subtle but perhaps even more relevant question is the following: if an accession will be first utilized n years from now, how long must that delay n be before it is economical to rely on introductions from elsewhere rather than to maintain an existing holding?7
The estimates in Table 7 indicate that regeneration costs are high, especially for maize. If an existing maize accession requires regeneration and the same accession is known to be stored elsewhere (but also requires regeneration upon introduction), it may be more economic to discard the accession from medium-term storage unless it is utilized within 2 to 5 yr. The cutoff period for wheat under the same situation is 8 to 14 yr, depending on the interest rate. Since the costs of introducing a wheat accession to the collection are large compared with the costs of storage, it is more economical to conserve existing accessions deemed useful in the near future. In general, if accessions are unlikely to be used within a decade or two, it is better to store those accessions in a single facility and distribute them to local genebanks when requested, assuming transportation costs and other quarantine barriers are not prohibitive.
Costs in the Very Long Run
Most of the figures above refer to the costs of conserving an accession for one more year, with the notion that decisions taken now can be revisited the following year. However, genebanks may well want, or be required, to guarantee safekeeping of samples in perpetuity; for example those accessions held in trust by the CGIAR centers by way of their commitments to the United Nations Food and Agriculture Organization (FAO). The cost of such a guarantee obviously depends on the state of future technology, input costs (including the rate of interest), storage capacity, and regeneration intervals.
Table 8 shows the average costs of conserving wheat and maize accessions in perpetuity, assuming costs are constant over time in real (inflation-adjusted) terms. We considered the present values of the costs of conserving an existing accession and a newly acquired accession with different regeneration regimes and different real rates of interest (2, 4, and 6% per annum, which were deemed to span the relevant range). Routine testing for the viability of seed samples was assumed to begin 10 yr after introducing a new accession to the collection, with retesting every 5 yr thereafter. Dissemination of an accession is assumed to be made every 10 yr.
|
These estimates are based on a 50-yr cycle of regeneration in perpetuity. In fact, the longevity of seed in the new CIMMYT genebank is uncertain: seeds in the base collection can be viable for 50 yr or even 100 yr, while those in a less advanced facility may remain viable for only 25 yr. If the conservation objective can be achieved with a regeneration cycle of 100 yr without initial regeneration, the present value of an in-perpetuity commitment to conserve seed is only $9.80 per accession for wheat and $43.24 for maize at a 4% rate of interest. Reducing the regeneration interval to 25 yr increases the costs to $11.69 for a sample of wheat and $106.30 for maize. These types of data can be used to assess the benefits from upgrading a genebank facility and thereby increasing the storage life of the seed. Lengthening the regeneration cycle from 25 to 100 yr reduces the present-value average cost of conserving an accession of wheat by $1.89 (i.e., 11.69–9.80), and by $63.06 (i.e., 106.30–43.24) for an accession of maize. The cost savings from such an increase in cycle duration would total $232 470 for the 123 000 accessions of wheat and $1 072 020 for the 17 000 maize accessions currently housed at the El Batan facility—a total exceeding the capital cost of the new CIMMYT genebank facility. Moreover, these savings in cost are a lower-bound estimate of the benefits from improved seed storage. They ignore the benefits derived from increasing the safety of the collection and lowering the rates of genetic drift that resulted from moving the collection to this new facility.
Table 8 also provides the cost of distributing accessions from the genebank. At CIMMYT, seeds are maintained in medium-term storage to facilitate dissemination upon request. Since seed viability under medium-term storage is shorter than that under long-term storage, we assumed the regeneration cycle to be 25 yr, rather than the 50 yr for seeds in the long-term facility.8 Assuming that an accession is disseminated every 10 yr, Table 8 shows that the present value of the distribution costs is $27.12 for wheat and $274.21 for maize, with a 4% interest rate.
Total Costs in the Short and Long Runs
Table 9 illustrates the total costs of conserving seeds over different time horizons. Abstracting from the costs of characterizing, disseminating, or documenting seeds, the marginal variable costs of holding the existing collections of maize and wheat for one yr in long-term storage are modest indeed—only $41052 for wheat and $86875 for maize.9 These marginal costs cover the cost of maintaining and operating storage equipment: they take as given, and thereby exclude fixed costs in the form of physical capital and lumpy labor inputs (defined here, as the overhead of labor in the form of managerial staff). Including annualized fixed costs provides an estimate of the average costs of 1 yr of storage of all accessions—for wheat the costs are $149032, for maize they are $192182 (Table 9).
|
Setting aside the cost of capital, it takes $5468374 in total labor (including the labor of senior scientific staff) and operating costs to conserve the entire wheat and maize holdings for 40 yr, $6907034 if the seeds are saved in perpetuity. This figure includes much more than the labor and operational costs required to simply store the seeds in the genebank. It factors in the costs of checking the viability of the seeds, periodically regenerating the samples (here, the regeneration cycle was presumed to be 50 yr, although we based our calculations on the fact that a certain share of seed is regenerated each year, reflecting the varying time in storage and the respective regeneration histories of each accession), plus the data-management costs required to manage the collection.
Separate from these long-term conservation costs are the costs of maintaining seeds in medium-term storage and disseminating on request to breeders and others outside of CIMMYT (although a sizable share is taken from the genebank by the managers themselves for prebreeding, for characterization purposes, and for backup storage at an off-site facility.) If the genebank continued to distribute seed at the rate typical of the past few years, this distribution function alone would cost about $4178702 in present-value terms over a 40-yr time horizon, and $5278065 in perpetuity.
Bundling all these costs together (i.e., including the seed storage, regeneration, duplication, information management, and distribution activities), we estimate that the capital, labor, and operational costs combined would total $11296385 over the life of the genebank and $14268320 in perpetuity. This represents the amount of money that would need to be set aside (at a 4% real rate of interest) to underwrite genebank activities at their current levels over the longer run, a sizable but not an especially large sum of money.
Economies of Scale and Scope
In addition to the 600000 accessions held in the 11 genebanks maintained by the CGIAR, there are about 5.6 million accessions stored in 1308 genebanks worldwide (FAO, 1998). Is there any economic gain to be had from consolidating these holdings into fewer facilities? A sense of the size of the gains from consolidating the world's wheat and maize collections can be had from the CIMMYT data. By world standards, the CIMMYT holding is large, but not the largest. The Institute of Crop Germplasm in China has a total of about 300 000 accessions in long-term storage while the National Seed Storage Laboratory in the USA holds 268000 seed samples in its collection (FAO, 1998).
Returning to Table 6, the figures in parenthesis give the average annual costs of conserving the CIMMYT collection presuming the genebank were full to capacity—specifically, storing 390 000 wheat accessions (compared with 123000 in 1996) and 67000 maize accessions (17000 in 1996). Operating the genebank at full capacity and allowing for savings through size and scale economies would involve an estimated annual cost of storage (net of regeneration and other expenses) of $289810, compared with the annual storage costs of the facility at its current capacity of $82190 (Table 6). However, if the genebank at full capacity were operating with the cost structure at current capacity, the annual storage costs would be $656140. This constitutes an annual saving in costs of $366330 ($179400 for the wheat holding and $186930 for maize) compared with storing the same number of seeds, say, in two separate facilities at the average per accession costs currently experienced by the CIMMYT genebank. These savings come from significantly increasing the size of the holding without incurring a corresponding increase in the annual fixed and quasi-fixed costs of storage.
Our calculations imply some economies to centralizing storage of all cultivars of a crop and avoiding excessive duplication of storage facilities. Given the relatively modest cost of black-box or other forms of safety duplication, conservation economics and security imperatives can be jointly satisfied with one central genebank and duplicates held in other parts of the world. One possible scenario is to set up one central genebank for long-term conservation and various local genebanks for active collections. However, the best scenario depends on transport and communication costs, on the relative conservation costs of active collections and long-term collections, as well as on the different effects of the environment on different crops, issues for further study. At least one duplicate set should be at a location in which the prospects of political embargoes, military actions, or terrorism that could disrupt international access are extremely remote.
|
DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTS3DISCUSSIONGLOSSARYREFERENCES |
|---|
We estimate the marginal variable cost of conserving an accession of wheat in long-term storage for one more year is just $0.11, and for maize is $0.52. This cost includes the costs of the electricity and labor for operating the storage plant and equipment, but not the corresponding capital costs, nor the costs of replenishing, germination testing, or regenerating a holding. The average cost of holding an accession for an additional year (i.e., including all the variable costs and the quasi-fixed cost) is $0.19 per accession for wheat and $0.93 for maize if the seed samples are viable and need no regeneration. If regeneration is required the average cost per accession rises substantially to $3.45 for wheat and $109.63 for maize.
Advances in technology have eliminated much of the location-specificity of ex situ genebanks. Complete climate control means independence from local weather when storing the seed, and advances in communications mean the bank should in principle be accessible worldwide if it is served by modern telecommunications and express mail facilities. Regeneration requirements may place a premium on location, but decisions as to where to physically store the seed and where to regenerate it are separable. Thus, if questions of security, undue delays because of phytosanitary controls, and political interference in access can be satisfactorily resolved (and presuming the benefits in knowledge spillovers among breeders and genebank managers working in close proximity are insufficient to affect the story), the argument for consolidating crop accessions held for long-term conservation purposes in genebanks worldwide in just one or a few sites can be justified on economic grounds. By way of example, we estimate that the costs of storing seed in CIMMYT's long-term facility operating at full capacity (i.e., 390000 wheat accessions and 67000 accessions of maize) would annually save $366330 through size and scale economies alone, compared with the annual costs of storing the same amount of seed, say, in two facilities with cost structures equivalent to the CIMMYT facility operating at its current capacity (123000 wheat and 17000 maize accessions). In present value terms this would generate a cost saving in the order of $7.54 million over the 40-yr life of the genebank (assuming a 4% rate of interest). The physical security problem seems solvable by present black-box or other off-site arrangements for storing duplicates. In any event, failure to consolidate holdings (at least among the CGIAR genebanks) carries with it a sizable price tag in terms of foregone cost savings.
Genebanks are generally seen as a means of conserving seeds for the long run. Taking this idea at face value, we estimated the costs of conservation over various time horizons stretching from 1 yr to forever. Keeping CIMMYT's present collection intact for the life of the genebank would cost a total of $7.02 million in present-value terms under baseline assumptions about the rate of interest and various regeneration and other conservation protocols—$3.06 million for the wheat collection and $3.96 million for the maize collection. This includes the costs of the genebank facility, the periodic replacement of equipment as it wears out, and the annual costs of the labor and materials required to store, germinate test, and regenerate seed as necessary. Holding the collection in perpetuity (with periodic capital replacement) would increase the total cost to $8.86 million.
Common wisdom would suggest that conserving seeds is a capital-intensive affair. However, when viewed long-term, our figures suggest otherwise. While CIMMYT's facility represents a sizeable investment in buildings, plant, and equipment totaling about $1.57 million, the annualized cost of capital represent 22% of the annual costs of the operation compared with 65% going to labor expenses. Other operational costs make up the remaining 13%. In fact it is the lumpy labor costs of senior scientific and technical staff (at 41% of the annual costs) that are the largest fixed or quasi-fixed component of the genebank costs.
Taking a long-run perspective, it would require an endowment of $3.25 million earning a 4% rate of interest to underwrite the capital costs for the CIMMYT genebank into perpetuity (including the periodic replacement of the genebank buildings and related equipment). To underwrite all the conservation and dissemination costs of the CIMMYT genebank at its present scale of activity into perpetuity would involve an endowment totaling $14.26 million ($8.86 million for conservation and $5.28 million for distribution). Given the importance of the conservation of germplasm into the new millennium, such an endowment appears to be a bargain as an investment on behalf of coming generations.
|
GLOSSARY |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTS3DISCUSSIONGLOSSARYREFERENCES |
|---|
Costs
Capital costs
Durable costs, broken down into the subcategories physical capital such as refrigerators, freezers, drier units, vehicles, and computers; and human capital meaning here skilled permanent labor such as genebank managers.
Fixed costs
Costs that do not vary in the short term with variations in the scale of operations. For the purposes of this study, fixed costs are primarily physical capital costs.
Quasi-fixed costs
Costs that are neither fixed nor variable but are lumpy nonetheless (See Lumpy costs)—for the purposes of this study primarily human capital.
Lumpy costs
Costs that involve purchases of whole units that are not readily divisible.
Variable costs
Costs that can vary in the short term with variations in the scale of operation such as the cost of temporary labor, electricity, telecommunications, and shipping.
Marginal cost
The change in total costs resulting from a marginal change in output.
Present value
The value of the sum of a stream of time-discounted future costs.
Depreciation
The reduction in the value of a capital asset through wear and tear over time.
Ex situ
Away from place of origin (as compared with in situ: in place of origin).
Genebank
The repository of seeds (hence germplasm) held within a cold storage facility. Also refers to the facility itself. (See Germplasm.)
Germplasm
The material that controls heredity—genes plus any other materials that affect inheritance.
Human capital
(See Capital costs.)
In perpetuity
The condition of being perpetual or unending. In this case, for example, it refers to an unending stream of seed conservation costs (including the costs involved in the periodic replacement of the genebank facility itself and related capital items) that would be incurred if the seeds were conserved forever. See Pardey et al. (1999) for more details of this concept in this context.
Landraces
Varieties developed through farmer selection.
Marginal benefits
The change in total benefits resulting from a marginal change in output.
Regeneration
The process of producing new seeds to replace or add to samples of old seed.
Scale economies
Reductions in the unit costs of conservation as the number of conserved accessions increases.
Seed-management protocols
The operational detail whereby CIMMYT stores, identifies, retrieves, and, especially, maintains the viability of its conserved wheat and maize samples.
Throughput
The number of accessions processed by the genebank through germination testing, regeneration, and seed-health activities.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTS3DISCUSSIONGLOSSARYREFERENCES |
|---|
2 To deal with Karnal bunt problems when transferring materials from the old to the new genebank facility, the regeneration of wheat during the primary survey year involved a two-step procedure rather than the normal one-step operation. Materials were first grown at El Batan to screen for disease-free seeds, and the clean seed was then shipped to and planted at Mexicali to produce seed of sufficient quality for storage. A total of 11000 accessions of wheat were regenerated in this two-step procedure and so we divided our regeneration cost totals by 22000 to derive the per accession costs reported here.
3 Earlier efforts in cost conservation activities are reported by Jarret and Florkowski (1990) and Epperson et al. (1997). Differences in the crops and conservation activities costed along with substantive differences in the estimation details preclude any ready comparisons. Specifically, our estimates are the first, to our knowledge, to place the costing exercise explicitly in a dynamic context.
4 We took the average number of accessions processed or stored over the past few years as our representative accession totals. For rescaling the new introductions, we used 5000 wheat accessions (and 1000 maize accessions), germination testing was 1500 wheat (4000 maize), regeneration was 6000 wheat (500 maize), dissemination was 13500 wheat (3800 maize), and duplication was 11600 wheat (1500 maize).
5 We opted not to round off the estimates presented in the text to facilitate cross-referencing with the tables—this should not be construed as implying any false precision.
6 For this discussion, we grouped the genebank functions into those concerned with conservation and those with distribution activities. Conservation services include maintaining seeds in long-term storage, monitoring viability and regenerating seed samples on a regular basis, and duplicating samples in an off-site facility. Distribution services include maintaining seeds in medium-term storage with more frequent regeneration, characterizing agronomic traits of samples, and disseminating materials upon request.
7 This break-even year, n, is found by solving
 |
8 We also assumed that there was no excess demand for certain accessions, which would require a more rapid rate of regeneration to produce sufficient stock.
9 These estimates and the ones to follow were developed assuming part of the holding is regenerated each year and part is tested for viability. For the multiyear, present-value figures we assumed a 50-yr regeneration cycle and a 4% per yr rate of interest.
Received for publication December 14, 1999.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTS3DISCUSSIONGLOSSARYREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. E. Dulloo, A. W. Ebert, S. Dussert, E. Gotor, C. Astorga, N. Vasquez, J. J. Rakotomalala, A. Rabemiafara, M. Eira, B. Bellachew, et al. Cost Efficiency of Cryopreservation as a Long-Term Conservation Method for Coffee Genetic Resources Crop Sci., October 22, 2009; 49(6): 2123 - 2138. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. D. Rubenstein, M. Smale, and M. P. Widrlechner Demand for Genetic Resources and the U.S. National Plant Germplasm System Crop Sci., March 27, 2006; 46(3): 1021 - 1031. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Dreisigacker, P. Zhang, M. L. Warburton, B. Skovmand, D. Hoisington, and A. E. Melchinger Genetic Diversity among and within CIMMYT Wheat Landrace Accessions Investigated with SSRs and Implications for Plant Genetic Resources Management Crop Sci., February 23, 2005; 45(2): 653 - 661. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Zohrabian, G. Traxler, S. Caudill, and M. Smale Valuing Pre-Commercial Genetic Resources: A Maximum Entropy Approach Am. J. Agr. Econ., May 1, 2003; 85(2): 429 - 436. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
