Asia regional synthesis: information for status and trends reporting on aquaculture^[100]

1 INTRODUCTION

The Bangkok Declaration and Strategy for Aquaculture Development Beyond 2000 (Aquaculture in the Third Millennium, 2001) emphasized the role of information in (a) the efficient management of the sector and (b) the collection and dissemination of accurate and verifiable information to improve the image of aquaculture.

The recommendation focused on improved information flows through (i) arrangements for sharing of data and information, (ii) strengthening national capacity for determining data needs and for collecting and managing data, (iii) providing mechanisms for better access to relevant and reliable information to stakeholders, and (iv) making effective use of new information technology.

The Conference identified five issues that needed to be addressed in order to achieve the above recommendations and suggested actions to address them. These are:

• poor understanding of the purpose of information and information activities,

• poor quality of data and information, which was elaborated as the result of irrelevance, unreliability and un-timeliness of the information;

• lack of internationally comparable methodologies;

• inadequate data analysis; and

• ineffective communication and presentation of information.

2 FRAMEWORK FOR THE REVIEW

A brief overview of the issues related to the use of statistical data and information is intended to provide a conceptual framework for this synthesis of six national reviews^[101]

Hierarchy of uses

A fundamental assumption is that the paramount purpose for data and information is to aid in decision-making. A decision has to be made in order to solve a problem. Therefore, the hierarchy of uses (and users) of statistical data, and the information that may be derived from the data, depends on the immediacy of the need for the decision and the number of those that would benefit from such decision. From a problem-solving standpoint, importance is based on the seriousness of the problem (seriousness being the function of severity or how large the damage a problem can cause if it occurs e.g. percent loss of production; prevalence or how widespread is its impact; and frequency of occurrence. In this light, the hierarchy of uses of national statistics (on aquaculture) would be as follows:

(i) for national and sector policy, planning and management - national and local governments;

(ii) for analysis of the sector, in other words, research - R and D sector, investors and entrepreneurs, industry, development agencies, farmers' cooperatives, aquaculture enterprises; and

(iii) for education, training, public information and advice to the sector - academia, government, training and extension providers, development agencies, mass media.

2.1 ATTRIBUTES OF STATISTICAL INFORMATION

Based on the Aquaculture Millennium discussions, there are two fundamental and one desirable attributes of statistical information:

2.1.1 Reliability

An intrinsic attribute of data and information, regardless of how they are to be applied, is reliability. A user will want to know to what extent the information can be trusted or "how close does it represent reality". It is the product of its adequacy and validity.

• Adequacy - This attribute is a function of the range of information or analysis that can be derived from the data presentation. A user will want to know to what extent, the data can be relied on to provide reliable interpretations and conclusions i.e. how far can it be extrapolated?

• Validity - This is a function of the representative-ness of the sample, generally a methodological and procedural issue. It primarily impacts on the extent to which the data or information can be applied for deriving conclusions about a specific population.

The importance of the above attributes, from the standpoint of government, may be summed up bearing in mind that planners and decision-makers need appropriate indicators of performance for the sector and indicators of future potential.

2.1.2 Relevance

This attribute is based solely on the purpose of the information, and would be determined by how applicable it is to solving a problem. Timeliness and the way the data is presented affect their relevance. Timeliness also impacts on validity, in the sense that information that loses currency has its applicability in decision-making eroded.

2.1.3 Complementarity

A third attribute, which is not basic but adds to usefulness, is the added value to statistical data. The usefulness of statistical information is enhanced by its being integrated with other types of information to provide a better picture and understanding of the status and trends happening in the sector.

3 DISCUSSION OF ISSUES

Before describing the constraints and weaknesses indicated or implied by the reviews, it would be useful to first establish those constraints and weaknesses that are fundamental to the problem of poor quality statistics and information; and those that merely add to the underlying constraints and weaknesses. This will help identify and prioritize those issues that need to be addressed and also determine how this can be done.

3.1 Fundamental constraints and weaknesses

3.1.1 Legal framework for aquaculture

The existence or absence of a separate legal framework for aquaculture and whether the law mandates responsibility for reporting to a distinct agency does not appear to affect procedures and quality of statistical outputs. However, having a strong legal requirement to register aquaculture operation significantly improve the information on farm number and location.

3.1.2 Administrative structures for aquaculture management

Some constraints particularly those related to reporting, collection and analysis of data, seem to be the consequence of the administrative structure for aquaculture (or fisheries) management. The structure, locus of responsibility and how far down the administrative system certain responsibilities are assigned, varies according to each country's overall administrative system (i.e. centralized, decentralized, state or provincial autonomy).

The degree of management responsibility assigned to the producer of the data and information does have an influence on accuracy of reports of farm data. A good example is Japan's system of requiring every FCA to prepare an annual report on the status and trends of aquaculture for management purposes. FCA's are originators of farm data reports as well as responsible for the management of the local aquaculture, and it is assumed that their inputs to the statistical system would be as accurate and complete as possible, in line with their own management requirement. This is an issue of the originator of the report having a high stake in the resulting output.

3.1.3 Linkages between monitoring and planning and management

Linkages between monitoring and planning and management may or may not have an influence on the quality and relevance of information. It can be generalized that the more urgent the need for a decision is the more critical is the requirement for reliable and meaningful information. Decision makers that have to deal with more localized and immediate problems, put a premium on highly accurate and very relevant data and information. It follows that they would require that the data reported to them are accurate and that their fidelity was ensured throughout the processing, analysis and presentation process.

In China, farmers or aquaculture companies in some instances tend to "over-report and bureaucrats are reported to inflate the reports from farmers for reasons that they think would benefit their companies or themselves". This tendency as well as underreporting, in Thailand and the Philippines is related to the issue of misunderstanding of the importance of accurate statistics and lack of priority, which is a product of poor linkages between monitoring and planning and management.

It can be noted, that the independence or neutrality of the statistical agency from the technical agency (i.e. fisheries bureau or agency or department) tends to ensure that data from the source are faithfully reported during processing (in other words, not distorted deliberately).

3.1.4 Coordination among agencies concerned

Coordination is made necessary because monitoring and management frequently reside in different agencies or offices within the same agency. Lack of or weaknesses in coordination does not necessarily result in poorer statistics, but a stronger linkage does encourage the monitoring body to ensure cost-effectiveness, quality, timeliness and overall relevance of statistical outputs, and relevance of the output to the users. In this respect, the Philippines has constituted a Technical Working Group to put into users and producers together and address issues and concerns on how to strengthen statistics.

The situation in Vietnam where three agencies collect statistics (the Ministry of Fisheries on coastal aquaculture and marine fisheries; the Ministry of Agriculture and Rural Development on agriculture and inland fisheries including freshwater aquaculture; and the Office of General Statistic on all national statistics), might be a good case for studying the result of such a system on the various concerns about institutional arrangements and coordination, institutional priorities, resource allocation, impact on the communes and farmers who have to report more than once the same set of data, and to the field workers who have to ask farmers and communes probably the same or similar sounding set questions, and whether there are significant variations of the data collected by each agency. (The national review failed to deal with this issue so that this synthesis would not speculate on the effect of the Vietnamese system).

3.1.5 Absence of a crosschecking or validation mechanism

A common theme is "we know the data is inaccurate, but how inaccurate is it?". Lack of data validation mechanisms leaves the reliability of raw data open to doubt and undermines confidence in all subsequent analyses. This reduces its usefulness to policy makers and may lead to inappropriate allocation of resources. Even when improvement of the data could not be achieved, indicators of accuracy would allow policy makers to assign a relative weighting to the value of statistical information in making decisions. Most importantly, incorporation of validation mechanism may not necessarily incur higher cost but can effectively improve the quality of data.

3.1.6 Over-centralized processing

Over-centralized processing is another issue - the ability to undertake basic first step processing at the local level not only enables the local level to benefit form the preliminary results (for use in their management and planning) but also reduces the burden at the central level, speeding up the aggregation of the national figures. There may also be some value in the increased sense of ownership at the local level and improved understanding of the purpose of the information collection

3.2 Non-fundamental constraints and weaknesses

3.2.1 Standard procedures and methodologies

The presence of guidelines and the capability to implement standard procedures and methodologies are not a constraint. All the countries that reported on this aspect (except Vietnam, which did not deal with it) follow standard statistical procedures and methodologies and have the systems and the manpower to implement them properly. The weakness appears to be the inability of sampling procedures to cope with systems growing multiple species and where species are farmed using various different systems, which characterize much of the small-scale aquaculture in Asia. There are instances when the sampling frame does not include a new species (introduced Taiwan tilapia in Japan or P. vannamei in current Thai reports) or culture system (marine cage in China), or with rapid changes in culture management system (as in the Philippines) but these are not technical manpower capability and/or ignorance of methodology issues.

3.2.2 Data processing and analytical capability

This is a constraint only in terms of the speed by which data goes through the system. It is not due to lack of capable personnel. It is because of the bureaucratic protocols, or the hardware being used. Field offices generally are ill-equipped, which can hardly be regarded as a constraint because technology (hardware and programs) are now readily available. The bureaucratic procedures is another matter: the passing on of data from one level to the next, and for crosschecking and validation take time.

3.2.3 Basic packaging and presentation of the information

This is not a constraint either; the reviews indicated that statistical data and information are properly packaged and presented, again following standard formats. However, it appears that beyond the standard packaging and presentation, there is little attempt by the agency in charge of statistics to add value to the information through a more extensive or a deeper analysis, as for instance using non-statistical information to provide a more holistic picture. Intermediate users such as researchers will have to do it themselves. For instance, researchers doing economic analyses on a commodity such as shrimp would still have to conduct their own survey to collect data on costs and returns. However, systems such as Japan's are able to integrate production information with those on number of aquaculture establishments, number of aquaculture facilities, area under culture, which are compiled by prefectures, released as a part of an annual report of fisheries production and contains some more detailed analyses.

3.2.4 Distribution

This is another factor that is not a constraint. The system is in place for publication in hard copy and, in almost all of the countries, on the web. Specific agencies and institutions are identified that are end users or as intermediate users, and routinely provided with the statistical information. The information is also available on demand.

3.2.5 Classification of production system and disaggregation of species

As to the broader issues that might impact on the classification of systems harvested, all the countries have their definitions of aquaculture as well as systems that are similar or consistent with those of FAO. Some countries however maintain different classifications of the aquaculture environment (i.e. brackishwater in the Philippines, various sub-classifications of mariculture in China) but this is only a constraint to having comparability of data across countries.

It is in fact necessary for countries that have significant culture areas in these environments and where a major species is cultured in different environments (i.e. milkfish) to have these classifications.

3.3 Classification of constraints

The reviews indicated three types of constraint and their effect on the reliability and/or relevance and the overall usefulness of aquaculture statistics and information. Many of these constraints and their consequences are not mutually exclusive (and are common to the issues of information in the fisheries sector), e.g. inadequate resource allocation may also be due to government priority or budgetary status of government or an agency priorities.

3.3.1 Constraints caused by inadequate funding^[102]

(i) insufficient field staff - not enough enumerators;

(ii) field staff capability - enumerators training, hiring of enumerators that are inadequately trained;

(iii) data collection and sampling procedures - probability vs. non probability: resorting to non-probability sampling or monitoring rather than proper sampling;

(iv) failure to validate or "triangulate" information - i.e. crosschecking with other indicators such as market data, export statistics, trade information, census data or other agencies' records;

(v) limited coverage preventing comparability; and

(vi) failure to disaggregate species - the result of lack of trained enumerators or lack of time or because sampling frame does not cover it.

3.3.2 Constraints that arise due to institutional arrangements and priorities

(i) lower priority placed on statistical and information gathering;

(ii) inadequate information due to improper sources (informants) of primary information (i.e. informant), as well as the inability to accurately estimate production from continuous and batch harvest systems;

(iii) lack of disaggregation of capture from culture as a policy;

(iv) inflation of production data;

(v) late release or publication of information;

(vi) "neutrality" of agency tasked with statistics and information;

(vii) timeliness and accuracy as affected by the importance placed on a species or commodity (i.e. shrimp vs. other species);

(viii) lack of a system to crosscheck field data sent to processing centres (a supervision as well as decentralization issue - data processors office have no direct link with collectors or authority to supervise the field data collection personnel);

(ix) lack of reports on production from new systems (i.e. marine cages) or new species (i.e. Penaeus vannamei or introduced tilapia);

(x) over-collection of more parameters than are actually needed for sectoral management adding burden and cost to the system (an important point here is that data needs for management may not be included in data collection for national statistics and equally, the data collected for statistical purposes may be of limited value in management planning); and

(xi) priority placed on high value export commodities over other domestically marketed or lower value species.

3.3.3 Lack of understanding or ownership of the information

These that are related to farmers' and government officer's perceptions or understanding as to the purpose of statistical data, and other factors:

(i) underreporting or over-reporting of production data (at source);
(ii) distortion of data (within the system);
(iii) non-reporting of the production of a species (i.e. P. vannamei).

From the above constraints (as well as from the summary of the reviews (Annexes 1 and 2), we can derive indications on the key factors that impinge on reliability and relevance. There are two factors related to institutional issues and sociological issues and neither are mutually exclusive.

4 DISCUSSION

4.1 Institutional issues

Lack of resources is an indication of the relatively lower priority placed on statistics and information. However, this may be the result of a more fundamental reason arising from the lack of appreciation or poor understanding of the importance of statistics and information. A second issue is that if the statistics are not actually useful at the local level, then the local level is unlikely to place much importance in their rigorous collection.

While inadequate resources by itself has been shown not to be an issue related to the intrinsic reliability of the statistics, it does create weaknesses that in turn affect the adequacy of the data for purposes such as forecasts and developing costs and return analyses (both of these functions rely on having up-to-date information). The net effect is that information users such as researchers and industry analysts and managers as well as policy makers will not trust the overall reliability of the statistical data. Another effect is that these data users may have to conduct their own surveys while using the statistical releases only as historical information or indicators (another effect is the creation of multiple data sets that may be contradictory!). Weaknesses in statistical systems are self-reinforcing - that is, if data is perceived to be unreliable, it will not be used or taken seriously and this will tend to result in a general sense of the lack of value of the information. In such a situation is inevitable that institutions conclude that valuable resources should not be wasted on an activity that has little value.

Some of the country reviews are indicating that measures are being taken to address this situation. The establishment or strengthening of consultative and coordinative mechanisms that involve statistical agencies and user agencies to address weaknesses in the system are in already place or are being planned. This is a strong indication of the importance for ensuring that data users to appreciate the usefulness of reliable statistics and equally, that the information provided is applicable to their needs.

Statistical information may often be of little value if taken by itself, "adding value" to statistical information is achieved by the integration of statistical and non-statistical information to increases the value of the information package. The country reviews indicate that this is an ongoing effort and there are a number information products already being disseminated that combine production, market, and other relevant information. In some cases efforts are being made to tailor the information to various levels of users i.e. national, state, local and farm.

Coordination is a non-issue in terms of the reliability or relevance of statistical data. It is a broader issue of institutional linkages and cooperation, which needs to be addressed at a wider level. However, it is important to point out that researchers, advisers and industry analysts' confidence in statistical reports would improve linkages with the Research and Development sector. There would be no or little need for them to spend more research resources by having to conduct their own field surveys.

The question of timeliness and relevance is critical among decision makers, who, because of political pressure, invariably have to make decisions with or without the best available information (sometimes without any at all).

4.2 Sociological issues

The practice of deliberate overreporting or under-reporting of production and inflation of figures by farmers and/or officials seems to have the greatest impact on the reliability of statistics. This can be due to a number of reasons.

A system of cross-checks and validation could mitigate the effect of the practice of inflating figures, while a statistical tool could be devised to compensate for underreporting. This would require much study but could be a worthwhile undertaking by an academic institution.

To find a way to directly address this problem, it should be noted that the basic issue is that the individual's priority system places a low rating on reporting accurate figures. Avoiding taxes and being promoted are certainly more tangible and perhaps brings more immediate rewards than having a reliable set of statistics.

This is a sociological issue and needs to be addressed as such. There is indication from the Japanese system that appreciation of the importance of up-to-date and accurate data by the final users of the information, i.e. the farmers themselves, can be developed. This is achieved by devolving the management of aquaculture and troubleshooting of farming and resource problems. In this situation, the data collectors are also the data users and this ensures that the information is of reasonable quality and accuracy.

The same generic approach of stimulating perception of greater benefit and proving that it is true, can apply to the entire range of users and stakeholders of statistical information. The national reviews and the Aquaculture Millennium Conference of 2000 described various opportunities for doing so. A generalized recommendation is that there is a need for institutional arrangements that enable the various stakeholders to be more closely linked and to cooperate better by being aware of each others' needs and understanding the value of quality information.

This review sees no need to discuss in much detail the manpower and institutional capability issues, as these have been the subject of frequent discussion. The need for training in various areas of expertise and the need for facilities, equipment and programs are also well known. It might be concluded that quick rotation of manpower (making it necessary to recruit and train replacements) is also an indication of institutional priorities.

The need for equipment and programs and for following prescribed statistical procedures and methodologies are not a critical issue, although they can be costly relative to available developing country resources. It can be summarized that, if resources are available, the governments and workers in statistical agencies are now sufficiently informed to know which ones to acquire and do not need much training to learn how to use the packages and implement prescribed survey, sampling and field data validation procedures.

The country reviews also described ongoing and planned activities to address specific problems and weaknesses in their national statistical and information systems. A number of recommendations to improve systems and procedures as well as coordination among providers and users of statistical information were also elaborated

5 LESSONS DRAWN FROM THE REVIEWS

5.1 Flexibility

The Philippines' offers a positive lesson in having various options according to the level of resources available. If budgetary resources preclude the conduct of probability sampling, alternative acceptable methodology will be employed for selected provinces, and when resources allow, additional data collection such as surveys on the cost and returns of various farming system will be conducted. Persistence to one methodology regardless of financial resources available would lead to a considerable delay of the process or deterioration of the quality and consequently reduce the usefulness of the data to users.

5.2 Neutrality and professionalism

Having a separate specialized agency for agricultural statistics assures neutrality, professionalism in the agency, and focus on the mandate. The fact that it is a specialized statistical agency means there is no other function that might compete with it for staff attention and agency priorities, as would happen in an agency with various other activities. This latter lesson was pointed by the experience in Thailand and India, where statistics is only one of many other responsibilities of the Department of Fisheries. Flexible employment of various data collection methodologies mentioned above can only be possible when specialized staff who are capable of dealing with ad hoc shift of methodology are fielded for data collection.

5.3 Coordination

Having a single national agency in charge of agricultural statistics (forestry, fisheries and crop and livestock) can be expected to provide an effective coordination of national level and local level activities, which could easily become a patchwork of uncoordinated efforts. Additional benefits include its being a one-stop shop for information, and an efficient mechanism for integration - and cross-validation - of data from the various sectors of the economy. The Bureau of Agricultural Statistics (BAS) of the Philippines, as well as China and Japan's systems exemplifies this arrangement. (The opposite could be that of Vietnam). Such a national agency eliminates unnecessary duplication of efforts. It would additionally reduce the tendency to collect more data and publish information than needed.

5.4 Decentralization of responsibilities

Two positive lessons from the Japanese system are the strict neutrality of the agency, which avoids the tendency to issue figures more for the mother agency's performance image than accuracy; and the advantage of decentralization of responsibilities. Placing of responsibility of reporting on the fish-farmers cooperatives encourages fast and accurate farm data. As discussed elsewhere, the devolution of management responsibility to the FCAs of the area's aquaculture and natural resources encourages the need to have accurate data for management purposes. Japan's predominantly mariculture and commercial-scale aquaculture sector has sharpened its statistical methodologies for the sector. It is important to note that the FCA's are excellent mechanisms for dealing with a limited number of species in a close geographical area (e.g. scallop farming in a bay), it cannot provide lessons on how to deal with freshwater aquaculture, where farmers are dispersed and often culturing a wide variety of species and have little or nothing in common and few reasons to associate.

The Philippines' system decentralizes the processing and analysis of data which allows a better and faster crosschecking and validation of reports. In contrast, all the processing and analyses are done by headquarters in the case of Thailand's DOF. The slowness of this arrangement is now compounded by the fact that the field staff are no longer under the administrative supervision of headquarters, which has made crosschecking of data with field staff virtually impossible.

5.5 Appropriate methodologies for collection and recording of farm data

There are at least two situations that relate to this issue: One is the need to cope with the wide variety of farming systems especially in Asia, the other simply has to do with the ability of respondents to recall production and other information during an oral interview.

As far as the sampling frame, the identification of aquaculture systems and the methodologies are concerned, the systems of Thailand and China deal effectively with some of the unique features of freshwater aquaculture in developing countries. These include numerous small subsistence farms, integration with other commodities and multispecies culture. However, none of the systems seem to be able to cope effectively with sporadic harvests and harvests for home consumption. It has been repeatedly demonstrated that recall type interviews are completely unreliable for collecting information regarding in historic events (such as batch or occasional harvesting). The tendency (in Thailand and the Philippines) is thus an under-reporting of the yield, which can be significant with the predominance of small subsistence farms in both these countries.

The effect of under-estimation through recall interviews also compounds the result of deliberate under-reporting of farm production. India's schedules for farm records do account for multi-species but not for batch or occasional harvests. It does appear that this is a research opportunity to enable the development of an appropriate methodology.

A basis for improving the precision of capturing farm level information could be the monitoring forms of the BAS, Philippines and the farm record sheets of India. They could be used to develop a better instrument to obtain farm data than an oral interview relying on respondent's recall. There is significant work already done in the agriculture and livestock sub-sectors and a methodological study, akin to what has been done in cropping systems, to compare the precision and cost-effectiveness of various types of obtaining farm data would improve the instruments and the methodology.

6 COMMENTS ON THE FISHSTAT AQ

The comments and suggestions received on the FISHSTAT AQ relate to:

• the details requested of the countries being different from the information that the country's system presently requires. For instance, the Philippines says it is unable to answer number of units for ponds and tanks, enclosures and pens, cages, raceways and hatchery or nursery output, as these are not covered by their regular surveys, which is focused on production, price per kilogram of produced species and aquaculture area by type of aquaculture farm;

• the inability of ground personnel or the difficulty created by the system, in which there are various people and agencies involved in the survey, to disaggregate the species by systems as in the case of India;

• the difficulty of keeping with FAO's time requirement in order to be able to provide the classification that the Questionnaire requires as with Thailand; and

• the general weakness of the Questionnaire in its inability to capture data for more varied and meaningful economic analyses. This last comment, received from Japan, provides useful suggestions to address what the reviewer perceived as aspects in need of improvement in FishStat AQ - from the perspective of the country user and in light of Japan's experiences in developing its statistical system. (These appear as Annex 3.)

7 RECOMMENDATIONS

The following recommendations are derived from the above discussion and the summary of trends (Annex 1 and 2). No reference is made on FISHSTAT AQ as Annex 3 is felt to adequately serve the purpose.

• ensure strong coordination between statistical agency and user agency, if these are under separate government bodies;

• encourage regular consultations among the statistical agency and various users (intermediate and end-users) of the information;

• in line with devolution of management responsibilities for aquaculture and aquatic resources management, encourage the decentralization of data processing and the participation of local stakeholders in ensuring reliability of data and their processing for local management purposes;

• strengthen legislation to improve clarity regarding aquaculture operations. As part of this, encourage basic-level, decentralized farm registration;

• improve usefulness of statistical data through addition of important (and streamlining/elimination of unnecessary) parameters that allow cost-and-return and other economic analyses, value-addition such as integrating them with non-statistical information, and early release of reliable data;

• develop a model for developing countries (particularly sensitive to multi-species and small-scale farms). The reviews indicate distinctive strengths and capabilities as well as advantages of each national system, which the model would incorporate. This model could be used as a basis for incremental improvements to, if not wide ranging reforms, of national systems. The subsequent efforts should include a cost-benefit study of adopting the model;

• investigate the potential for using a range of approaches to capture specific information needs (e.g. a mixture of techniques);

• improve the precision of instruments and methodologies to capture farm level data;

• identify information that can be captured through other mechanisms such as censuses, trade and market information, proxy indicators (such as feed sales, export tonnage, tax/license revenue); and

• assess the economic impacts of the various causes of unreliability of statistical data and identify the weak links in the statistical and information development and dissemination system and ways to strengthen them.

Annex 1

SUMMARY OF TRENDS FROM THE NATIONAL REVIEWS

1. THE SETTING

1.1 National practice used to identify aquaculture separately from fisheries?

The key aspects of the FAO definition of aquaculture as outlined in the FAO Technical Guidelines for Responsible Fisheries No. 5: Aquaculture Development are i) some kind of intervention in the rearing process and ii) individual or corporate ownership of the stock. All countries surveyed use definitions of aquaculture that are broadly comparable with the FAO definition in this regard.

1.2 Administrative structure responsible for aquaculture development, monitoring and management and whether different for marine and freshwater environments?

Administrative structures for fisheries reflect the prevailing government structure. However, there is a trend towards decentralised administration of fisheries by local government with some national level policies and programmes from the central government.

1.3 Separate legal framework for aquaculture or if it is included under fishery of agriculture law and whether the law stipulates reporting responsibilities

Aquaculture is addressed under the prevailing fisheries law; there is no separate legal framework in any of the countries surveyed. However, in some cases aquaculture is discretely recognised within the fisheries law (e.g. China, Philippines). At the time of writing (December 2003) Vietnam's fisheries law is under consideration by Parliament, it has not yet been implemented although there are several national regulations and policies (such as Sustainable Aquaculture for Poverty Alleviation) that pertain to aquaculture. In some cases aquaculture is also regulated in some cases by non-fisheries laws, such as the Environment Protection Act of India.

Reporting responsibilities are not necessarily set under fisheries law per se; for example reporting responsibilities are set under the Statistics Law in China, and under the Fisheries Cooperative Association Law in Japan. Details were not provided for India where fisheries are largely administered under state law.

1.4 Is there linkage and coordination between monitoring and planning and management?

All countries reported a linkage between monitoring, planning and management except for Thailand, which indicated that the linkage is not clearly implemented. Linkage appears to mainly occur at the level of provincial/state government (China, India, Japan, Viet Nam) with a feedback loop to the national government used to formulate nationals plans (e.g. India, Viet Nam). The feedback may take the form of national coordination meetings between national and provincial governments (India) or submission of reports by provincial government. However, the strength of these linkages both within and between levels of government is not clear.

1.5 Are reports on status and trends of aquaculture routinely prepared for management purposes?

All countries surveyed prepare regular reports at least annually, some semi-annually or quarterly (Philippines).

1.6 Main purpose of aquaculture production and the intended use?

The main purpose varies between countries and regions within countries. As a general trend the commercial sector is on the rise. Commercial aquaculture dominates in Thailand and Japan, however, subsistence aquaculture still dominates in China and is still found in the poorer areas of Thailand as well. It is still significant in Vietnam.

1.7 Main species produced and the culture methods and facilities used?

See individual country reports for details. 'Traditional' species still dominate the production of all countries but there is an increasing trend towards the adoption of new species and high value species such as shrimp, crab and marine finfish in all countries.

2 CURRENT STATUS OF NATIONAL AQUACULTURE DATA COLLECTION AND COMPILATION OF STATISTICS

2.1 Are aquaculture production statistics currently collected?

All countries surveyed collect aquaculture statistics except for India, which does not disaggregate them from fisheries statistics. However, India has plans to collect aquaculture data separately.

2.2 How often are the data collected and on what time basis? Provide a timeline indicating the approximate schedule from data collection to data availability

Data are collected at least annually by all countries, but some collect more frequently (for example, India and Philippines collect data quarterly). In some cases countries may collect certain data more frequently than others. For example Japan collects data annually except for seaweeds which are assessed quarterly.

2.3 Indicate parameters on which data are collected

The parameters are highly variable between countries but generally include some sort of assessment of biomass, price, area under culture and socio-economic details of farmers.

2.4 Who are the data clients/users? Are they involved or consulted in the planning process for collection of statistics and other information to meet their needs?

Government are the main users identified for planning purposes. The research community and large commercial sector players were also identified as important users. Little information was provided on feedback mechanisms. The Philippines has a Fisheries Technical Group that conducts consultations with regional offices to address concerns and strengthen data collection. India has an annual planning meeting of state and national fisheries agencies that could serve this function. Both India and China have recently planned or made improvements to their statistics collection systems.

Given that methodologies are generally set centrally but collected/administered by provincial authorities, the lack of dialogue between agencies or levels of government could be an issue. No mechanisms were identified for consulting non-government users of statistics.

2.5 What institution(s) is responsible for statistical data collection?

Collection is generally carried out either by fisheries agencies directly through provincial or local level staff (India, Thailand, China), or by a dedicated government statistics agency (Japan, Philippines). National-level staff are seldom directly involved in data collection. In the case of Japan, the Ministry of Agriculture has established an autonomous statistics collection body specifically to separate the function from the fisheries agencies to avoid any potential bias or distortion in reporting by government officials, as reported to occur in China.

2.6 Are there in place different methods of estimation for different production systems? e.g. intensive cage culture, semi-intensive culture in ponds etc?

Generally the methodologies are standardised within production systems, but are variable between production systems.

2.7 Provide definitions for classification terms used for data collection

There is general consensus over the definition of freshwater and marine water, but there is variation between countries in how the intertidal zone is treated. Some countries break it up into a number of zones, for example China divides mariculture into i) shallow sea culture; ii) seaport or bay culture; iii) tidal field culture. This should not be a serious barrier to aggregating data for international comparisons but a consistent approach is needed.

2.8 Describe methodology for data collection for each production typology if applicable

See papers for details on individual country treatments of different production systems.

3 DATA QUALITY, PROCESSING AND ANALYSIS

3.1 What are the key problems in collecting high quality statistical data on aquaculture?

The fundamental problem is the collection of high quality primary data from the field. This underpins the reliability of all subsequent analyses. This is the key to the success of the Japanese system, which can obtain very good market-based estimates due to the national system of Fisheries Cooperative Associations (FCAs) and their role in administration/marketing the produce of their members.

Other countries face a far more difficult situation due to the dispersed nature of farms and numerous/highly complex marketing and distribution systems, and inaccessibility of some regions. Other common issues were a lack of human resources and funding with which to collect field data, which may restrict the frequency of data collection or lead to downsizing of sampling (Philippines). In Thailand, data collectors have been moved out to the provinces where they are often given other assignments and reducing the capacity of the head office to supervise their activities, giving concerns about reliability of data.

Deliberate misreporting is a common issue. For farmers this is mostly related to fear of taxation or regulation (which would lead to under-reporting) but in the case of China farmers may misreport in order to promote their products. There have also been instances of Chinese officials exaggerating reports in order to appear successful and obtain promotions. Japan has addressed this issue by establishing an autonomous Department of Statistics under the Ministry of Agriculture, so that fisheries staff are not involved.

A need to establish mechanisms to validate field data, to increase confidence in its reliability was expressed in the papers on China, Thailand and India.

3.2 Assessment of current quality of statistical data?

Data comparability is generally good as all countries reported that methodologies have been standardised within country. Analyses are similarly not a constraint and adequate expertise is available. However, the completeness and quality of raw data is suspect in most countries (except Japan) due to constraints on the collection and validation of field data, as described above.

3.3 Processing, storage, compilation and distribution of statistical data - is it distributed to the users identified in the previous section, how and in what form?

Data is distributed in all countries in the form of printed publications (at least an annual report or yearbook) and distributed to government agencies and researchers at national and provincial level. There is increasing interest in use of electronic media to disseminate data (such as CDs). Thailand, Philippines, China and Japan all make data available on the web, although permission is currently required to access it in China.

Japan also makes use of mass media - television, radio, newspapers - a short report is produced as results become available in April each year.

3.4 Analysis of statistical data: How is this done and by whom? Is information from other sectors/institutions outside aquaculture used to provide a more holistic status and trends reporting? How are the major issues identified and development potential/prospects estimated?

Analysis is carried out by both fisheries agencies (India, Thailand, China) and dedicated statistics agencies under the Agriculture portfolio (Philippines, Japan), depending on country. All countries except Thailand follow a decentralised approach to collating and analysing data, with analyses taking place at the provincial or state level. Higher levels of government collate the information provided by provincial authorities to generate the national view.

The major issues in Japan are identified and solved at the prefecture level (in that case through the joint efforts of the prefecture government and the Fisheries Cooperative). It is reasonable to expect that provincial governments in other countries, where they have the main administrative responsibility for aquaculture, will also perform this role.

There is no clear trend in use of external information to aid analyses, since most authors did not respond on this issue. Japan indicated that no external information is used in analyses but China indicated that some data such as customs data is used.

3.5 Presentation/packaging: Are statistical data analysed and packaged to provide information useful for management purposes, thus promoting their use by managers and policy makers?

All countries publish at least an annual statistics report, some more frequently (Philippines is quarterly, China semi-annually), and an analyses of the statistics is generally included.

3.6 Are there any metadata available - methodological notes, other sources, etc.?

India publishes manuals on data collection and catalogues of commercially important species from time to time. Japan and China also publish some methodological notes from their statistics agencies. The other countries did not respond.

4 NON-STATISTICAL INFORMATION

4.1 Is non-statistical information used to supplement the statistical data for status and trends reporting?

Non-statistical information is not used in Thailand or the Philippines, although Thailand recognised this as an issue that needs to be addressed. Others do, such as Japan (research relevant to planning), China (fisheries yearbook) and India (gear, environment, food security, livelihoods, sustainability, consumer preference and Japan).

4.2 Describe the main national non-statistical databases/information systems which are used or could be used in status and trends reporting

Sources cited include printed media - four magazines in Japan, the Census of Fishers and Farmers conducted and National Disease Reporting System by MOA in India, a Database of Fisheries Abstracts (Chinese journals) and websites.

4.3 Are there any key problems in collecting this type of non-statistical information?

Two issues were identified in the Chinese situation: i) most information is only available in hard copy and ii) some information is not available for public use. It was felt that data could be obtained from government agencies and private sector in India and Thailand.

5 DATA NEEDS AND OPPORTUNITIES FOR IMPROVING CURRENT INFORMATION ON STATUS AND TRENDS

5.1 What are the perceived national priority needs in terms of aquaculture information and why are they needed? Is the information currently collected or available from all sources meeting these needs?

Basic production data and trends were felt to be required for government planning purposes. In China, Japan and Thailand - to help anticipate and direct further development - along with farmer status (income etc). Market intelligence was also highly rated as a need for these countries and also in India, noting that accurate and timely market information is difficult to collect.

A variety of information tools - to support planning, decision making and industry development were seen as priorities in the India paper - perhaps better described as an information system. Needs identified included resource data, comprehensive GIS data sets, seed and feed data, disease management, value addition, post harvest technology and forward socio-economic projections of the aquaculture sector.

It was generally felt that information was meeting current needs at a basic level, although there was some dissatisfaction with timeliness and accuracy.

5.2 Is available information need-driven and user oriented, and is it accessible and used? What are the fundamental issues and constraints related to effective information utilization?

Information was generally regarded as needs driven and user oriented but there was dissatisfaction with the accuracy and/or timeliness in most countries (except Japan). There is insufficient detail in some information such as insufficient disaggregation by species, production systems or geographical areas (China). It was also felt that while macro-level information was useful to managers it is not useful to entrepreneurs operating at smaller levels, and that perhaps repackaging information to suit the needs of different user groups - or to make it more user friendly - would be beneficial (India).

5.3 Assess the strengths and weaknesses of the present methodology and processes for the collection, processing, analysis and dissemination of statistical data

Positive aspects included the broad coverage of statistical systems, standardised and consistent approaches, high comparability of data and the capability of resolving analyses at both the provincial and national levels. The processing and analysis of data were not seen as problems.

A number of common weaknesses were identified, most of which revolved around the capacity to collect high quality raw data from the field. Budget and human resource issues were universally identified as a constraint and the Philippines indicated that this impacted on both the frequency of sampling and on sample size. Misreporting by farmers (due to fear of taxation, etc.,) is a common issue. A clear need was expressed for mechanisms to validate field data to improve confidence in its accuracy (China, India, Thailand).

The appropriateness of extrapolating across non-homogeneous environments and production systems was identified as a threat to assessments (India).

5.4 Assess the adequacy of existing non-statistical information sources

Non-statistical sources were felt to be good in Thailand and Japan, with mass media providing coverage, and inadequate in China with a limited number of print publications providing such information.

5.5 What are the constraints and opportunities in improving quality aspects of information on aquaculture (statistical and non-statistical)?

Constraints are as discussed above - inadequate budget, lack of trained staff to collect field data, lack of coordination with other information sources, and deliberate misreporting by farmers and in some cases by officials. All of these constraints have one common impact: They reduce the accuracy and quality of the raw data.

The lack of mechanisms to validate field data was seen as a serious issue with one comment that if the quality of the field data could be improved the statistics would be much more useful (Thailand).

It was suggested that meetings of personnel to familiarise them with methodologies and reconcile/validate information would be useful (India). The increasing demand for information - including from government - was identified as an opportunity (China).

6 PLANNED IMPROVEMENTS OF INFORMATION ON STATUS AND TRENDS

6.1 Describe specific plans and actions if any to improve current information

The Philippines has established a Fisheries Technical Working Group to meet and conduct workshops with Regional Offices to address issues and concerns on strengthening fisheries statistics. Similarly, India has also made efforts to improve dialogue between (and within) data collectors and processors.

The Chinese MOA will launch a plan to amend the fisheries data collection and processing system in 2004. Changes will include: Greater disaggregation of species groups; removal of data from state owned enterprises (to be reported separately on a 5 year basis); reclassification of mariculture into marine waters, tidal flats and land-based culture; extension of freshwater classification system to include fence, indoor and cage culture; value of finfish, shrimp, crabs and molluscs will be included; 1990 price system will be abandoned; system will report on fingerlings rather than fry with more species information; and data included in other reports will not be duplicated.

India has commenced an upgrade of its statistics system. This includes the disaggregation of aquaculture from fisheries data; introduction of data warehousing and electronic dissemination through web services and CDs; increased computerisation of data handling and communication; repackaging of information to suit different categories of clientele at the local, regional, national and international levels; and strengthening of infrastructure (IT).

6.2 FAO aquaculture questionnaire (FISHSTAT AQ)

6.2.1 Describe specific problems in providing to FAO the information requested in the questionnaire, and any reasons for these problems

Country-specific issues included inability to meet reporting time frame due to domestic delays (Thailand), species data being aggregated by group and aquaculture data not being disaggregated from fisheries (India).

Another issue raised was that the reporting forms are Euro-centric; they may need to be amended to suit the Asian situation (Japan).

6.2.2 Comment on the adequacy of the questionnaire instruction sheet

The instruction sheet was deemed to be adequate. However, it was noted that the reporting agencies at national level are not necessarily the primary collectors of data (India).

Annex 2

SUMMARY OF THE NATIONAL REVIEWS ON INFORMATION FOR STATUS AND TRENDS REPORTING OF AQUACULTURE

Annex 3

THE FAO AQUACULTURE QUESTIONNAIRE (FISHSTAT AQ) 1 COMMENTS AND SUGGESTIONS

1.1 Summary comments from Japan:

The present two forms of FISHSTAT AQ seem to have been designed to meet the requirements of EU member states and EU candidate countries. The present two forms for FISHSTAT AQUA are in need of thorough revision so as to suit all countries.

Fishery statistics for both capture fisheries and aquaculture are required for economic analyses as well as biological analyses. To satisfy these two requirements, at least for aquaculture the following statistics are required:

1. Number of aquaculture establishments as economic unit.

2. Number of aquaculture facilities as input item. Out of several items relating to these input items, the area of water in use is considered to be the most important items.

3. Yield (harvest) in quantity and value as output item. However, the value of the yield is the most difficult item to collect.

By having the above three items, dividing the output data by economic unit data and input data will enable a lot of useful analyses in terms of the size of fish farm, productivity per unit area/by aquaculture facility.

When we look at the present two forms:

1. Number of aquaculture establishments is missing.

2. The number of units for aquaculture facilities may be useful for certain aquaculture systems. However, the country has to report the number of fish ponds regardless of its area. There is no country in Asia where the number of ponds is counted.

3. There is no linkage between the two forms. Therefore, analyses as referred in above is not possible.

4. The price/kg is requested for every species. In theory, it is logical. However, such a weighted price by species is not available in most countries.

[Note: In Japan, there is a fish market survey, which is divided into the survey in fish producing area and that for fish consuming area. For both areas, there are several fish markets that are selected and fixed. Based on the sales records of these markets, weighted average price by species are calculated for fish-producing area and fish-consuming area separately. Such a survey cannot be easily done by every country for reasons of resources to implement it.]

Aquaculture production data in value have been reported to FAO since 1984. Subsequently, these value data have been compiled into FAO Fishstat. However, these value data are of little use.

At present, FISHSTAT AQ has two sheets. One refers to aquaculture production methods or mean, and the other to production by species. But there is no linkage between the two sheets in terms of type of aquaculture employed such as carp culture, eel culture, shrimp culture, salmon culture, etc. Normally, aquaculture survey at country level is done separately for such type of aquaculture. Within a same type of aquaculture, the number of establishments, the number of aquaculture facilities, area under culture and harvest in quantity are simultaneously enquired.

The instruction sheet lists only five types of aquaculture facilities. When we look at this item on global basis, there are many more different types of aquaculture facilities in use. There is a fundamental point as to whether statistics on the number of aquaculture facilities are required for all types of aquaculture or not? As an example, the number of fish ponds regardless of size has little meaning. For this reason it may be worthwhile to combine the present two different forms into one standard form as shown in the sample below.

Sample Form for Reporting Statistics on Aquaculture

Country of Area: ______________ Year: _________

1. Fishes

The above will indicate roughly what is suggested, with statistics for Yellowtail Culture of Japan as an example. For the entry of species, spaces for recording 3 alpha codes would be required.

It is assumed that such a new form will be prepared for Fishes, Crustaceans, Molluscs and Seaweeds separately. Within each respective sheet, data on a single type of aquaculture, such as carp culture, eel culture, shrimp culture, oyster culture, etc. are recorded. The advantages from this revision would be as follows:

1. A reporting country will find it easier to record the information into FAO forms, as the survey and compilation of data for aquaculture are normally done for each type of aquaculture separately.

2. FAO will find it easier to compile aquaculture information for international comparison at least for major aquaculture species and systems such as salmon culture, shrimp culture, freshwater fish culture, etc. separately.

3. As referred earlier, international comparison with regard to various productivities will become possible.

For the above purposes, the establishment of national and international classification of aquaculture may have to be considered for Fishes, Crustaceans, Molluscs and Seaweeds separately. In light of above, FAO may wish to consider a tabulation program, by which an FAO Yearbook on Aquaculture is compiled.

The problem of a species name, which is reported at family or genus level, can be solved with the help of a national biologist.

Regarding the aquaculture production in value, FAO may wish to contact the Statistics Department of the UN (UNSD) in New York, requesting to divide "fishery" into "Capture" and "Aquaculture" in the International Standard Classification of All Economic Activities (ISIC). Japan has already done it. There may be some other countries that have done such a division. In those countries, the production data of capture and aquaculture in value are available separately, and have reported them to the UNSD as part of their National Account data.

Another option is for FAO to request national fishery statisticians to contact an office in charge of the National Account to divide the total fishery production in value into "Capture" and "Aquaculture". In this way, the production data in value of both capture and aquaculture will become available simultaneously. Such data may be more comparable and keep some consistency among countries.