Posted September 2000
This is a revised version of a paper presented at the First International Conference on Rural Aging, 7 to 11 June 2000, Charleston, West Virginia, United States. Comments and suggestions received from Jacques du Guerny, Marcela Villarreal, Ida-Eline Engh and Alain Marcoux (Population Programme Service, FAO, Rome) and Jean-Pierre Gonnot (United Nations Population Division, New York) are gratefully acknowledged.
Rural population ageing in developing countries remains inadequately documented and poorly understood. This paper argues that agricultural censuses can be a valuable tool for the analysis of rural ageing and its linkages with agriculture. Based on information from selected national agricultural censuses taken in 1985-1996, examples are presented of ageing-related topics that can be studied using the census data, including overall levels of ageing among persons attached to agricultural holdings, as well as associations between the age structure of agricultural holders and particular techno-economic characteristics of holdings.
While rapid population growth resulting from high fertility combined with lowered mortality has been the key demographic issue during the last decades, population ageing is poised to replace it as a major demographic preoccupation in the twenty-first century. As older people are becoming a larger and more visible part of the social scenery, better statistical information on demographic ageing, its causes and consequences is urgently needed to guide policies and programmes. This is especially true about the ageing of rural populations in developing countries - a process which is all too often overlooked, despite some evidence that in a number of poorer countries rural areas are now disproportionately older than urban ones, primarily because of the age-selective nature of rural-to-urban migration (Marcoux, 1990; Martin and Kinsella, 1994; du Guerny, 1997; Skeldon, 1999; UNFPA, 1999). Unfortunately, not enough is known about rural demographic change in the developing world. In most cases, the basic demographic information is not available from national sources, whereas the United Nations statistical system no longer publishes population estimates for rural and urban areas separately but only for total populations. Such a lack of relevant information represents a serious problem given that most developing countries are still predominantly rural and that in the years to come the absolute numbers of their rural elderly are likely to grow much faster and reach much higher levels than in the more developed regions (UN, 1993).
This paper focuses on censuses of agriculture, which so far have been insufficiently explored as possible data sources for the examination of rural population ageing in developing countries. Although agricultural censuses are primarily designed to describe agricultural holdings and thus can cover only a part of the rural/agricultural population - i.e. holders, members of their households, and agricultural workers employed on holdings - they could be a powerful tool for analysis since they provide an opportunity to link the information on the techno-economic characteristics of agricultural holdings with the demographic attributes of persons attached to them.
The paper reviews and assesses the demographically relevant information contained in agricultural censuses taken between 1985 and 1996, for which a report was published and made available to FAO. The paper consists of four parts: the Introduction is followed by Part 2, which reviews the main methodological aspects of agricultural censuses. In Part 3, examples are presented of subject matters relevant to population ageing that can be studied with the evidence from agricultural censuses. The selection of topics discussed is not intended to be comprehensive but rather to provide food for thought through highlighting issues that appear to be important for agriculture and rural development in poorer countries. Part 4 then concludes the paper by summarizing its main points and suggesting possible ways of advancing the use of agricultural censuses in research and decision making in the field of rural ageing.
The application of agricultural census data for demographic purposes requires good understanding of the census methodology. This section provides a condensed overview of the main methodological issues relevant to population research. For more general information about the production and substantive coverage of agricultural censuses, see FAO (1986, 1992, 1995, 1996, 1997) and Stloukal (1999).
A census of agriculture may be defined as a large-scale, periodic, government-sponsored statistical operation for the collection of quantitative information on agricultural structure, covering in principle the whole of a country within a given agricultural year. The chief concern of an agricultural census is to describe key factors of agricultural production (land and water resources, crops, livestock, agricultural inputs, machinery and equipment), but many agricultural censuses monitor also information on persons attached to agricultural holdings. On the other hand, aspects such as production and yields are commonly excluded from the census scope.
Most countries collect the agricultural census data through an interview of agricultural holders by enumerators who visit individual agricultural holdings. Although the term census implies complete enumeration, an agricultural census may also be conducted through sampling, or a combination of complete enumeration and sampling. Unfortunately, countries that use sampling very rarely report on the sampling techniques adopted and the magnitude of sampling errors.
Since 1950, FAO has been assisting countries in planning and conducting censuses of agriculture. Each of the decennial World Census of Agriculture Programmes prepared by FAO provided methodological guidelines for organizing national agricultural censuses. The six decennial Programmes - centred on 1950, 1960, 1970, 1980, 1990 and 2000 - gradually expanded the census scope while keeping structural aspects of the agricultural production sector as the central theme.
In agricultural censuses the primary unit of enumeration is the agricultural holding, defined as a techno-economic unit comprising all land and livestock used for agricultural purposes and operated under a single management, without regard to title or legal form. For practical reasons, the census enumeration is sometimes limited to holdings that correspond to certain recognized criteria (see Section 2.3).
Although agricultural holding is the recommended unit of analysis, demographic information is in fact obtained through investigating another unit: the holders' household, defined on the basis of the arrangements made by persons for providing themselves with food or other essentials for living. When using agricultural census data, one has to remember that in some contexts it is common that members of one household operate separate holdings, or that holdings are operated by two or more persons belonging to different households.
The demographic data collected in a census of agriculture refers only to persons attached to agricultural holdings. A holding may consist of the holder, other persons belonging to the holders' household, and hired workers who either permanently or occasionally work on the holding. Other categories of the agricultural population - members of landless hired workers' households, persons engaged in hunting, forestry and fishery activities or agricultural services - are by definition excluded. On the other hand, the census may include people who belong to an agricultural household but are not dependent on agriculture (artisans, etc.). The number of members of the holders' households may thus be either smaller or larger than the agricultural population (defined as all persons depending for livelihood on agriculture). Likewise, the data produced by agricultural censuses are inconsistent with the concept of rural population as applied in demographic statistics (i.e. people living in rural areas).
In each of FAO's Programmes, countries were cautioned against overloading the census with too broad a substantive coverage as this could complicate collection and processing of data. Ultimately, however, it is up to the national authorities to choose the statistical topics to be monitored and the classifications to be used. Any serious analysis of agricultural census information should therefore begin with a careful examination of the basic concepts used and applied in practice during data collection.
The agricultural census is supposed to cover all agricultural activities in a country. There are, however, many exceptions which may affect the completeness of data. The main ones are summarized below, while more details for individual countries can be found in national reports:
In spite of the widely recognized importance of the agricultural census for the management and study of rural issues, a number of countries do not organize their national censuses according to FAO recommendations. A full-scale agricultural census is an expensive operation that some developing countries may find difficult to support. Besides, many countries have alternative sources for some of the information they need on persons working in or dependent on agriculture (for instance, population census or sample survey evidence). Therefore, they may be inclined to collect only structural data on physical and technical characteristics of agricultural holdings which represent the core of the agricultural census, and reduce or omit items related to the population of the holders' households.
With respect to the above discussion, what aspects of rural ageing can be examined with the evidence from agricultural censuses? This section singles out several such topics, with the objective to draw attention to issues that could be important in relation to agricultural production, food security and rural development. The discussion is based on the premise that, while the 'static' census data do not provide answers to all relevant questions, they can nevertheless offer interesting insights into factors influencing age-specific patterns of social organization in the agricultural sector, and highlight the need for further investigation.
Like most other population datasets that contain the information on peoples' ages, the data obtained through a census of agriculture can be used to assess the overall levels of ageing of the population attached to agricultural holdings. The added advantage of an agricultural census is that it can portray rural age structures in two different but complementary ways: (a) for the population of members of holders' households, and (b) for the sub-group of agricultural holders. The ageing parameters of these two entities are likely to be dissimilar, with distinctive research and policy implications.
Table 1 provides an overview of basic statistics of demographic ageing that can be derived from agricultural censuses conducted in developing countries around the year 1990; Table 2 then summarizes the basic methodological information on these censuses. This information gives a very crude indication of the availability and analytical potential of census data for individual countries. In this context it is worth noting that during the period in question, some of the world's leading population 'heavy-weights' (China, India, Indonesia, Pakistan, Nigeria, to name just a few) either did not conduct a census of agriculture or did not design it in a way to permit the study of rural population ageing.
The census-based indicators in Table 1 confirm the considerable variability in levels of rural ageing between the countries concerned. To some extent this variation echoes the diversity of national practices in the census design: differences in definitions and enumeration criteria used, completeness of coverage of various agricultural activities, use of sampling, quality of age reporting, etc. Unfortunately, most national census reports do not provide enough information on data collection approaches applied. However, the variation in the country indices of ageing undoubtedly indicates also genuine differences in the demographic structures of the populations of agricultural holdings. These can result from local patterns of rural-urban migration, fertility and mortality levels, age-related aspects of land tenure, etc. At the outset it should be recognized that the census data by themselves do not allow exploring the root causes of differences in rural ageing between countries, but they can nevertheless greatly facilitate the formulation of assumptions about these causes. For instance, it is possible that some of the relatively high proportions of elderly agricultural holders in parts of sub-Saharan Africa result from the practice of communal use of natural resources and the corresponding form of social organization in which the management of land and decisional authority lies primarily in the hands of the elderly chiefs. In fact, it can be hypothesized that whereas the ageing levels among members of agricultural households are determined by the interplay of the principal demographic processes (fertility, mortality and migration), the proportions of elderly agricultural holders reflect also non-demographic factors (socio-economic, cultural and political conditions). This hypothesis appears to be consistent with the finding obtainable from Table 1, namely that a given level of ageing in the population of agricultural households can be associated with very different proportions of elderly holders, and vice versa.
Because the data collection approaches that countries use in their agricultural censuses vary significantly, the applicability of census data for inter-country comparisons will always be limited. At a country level, however, agricultural censuses could be of much value in specifying levels of, and possibly trends in, demographic ageing in agricultural areas; and in highlighting the linkages through which the ageing process affects the agricultural sector. For instance, increases in the proportion of older people in the farm population usually mean that the supply of agricultural labour and the flexibility of the agricultural labour market are declining. Although older workers may not necessarily be less productive (this would depend on the nature of their work and external circumstances, such as the level of technological advancement, and also on their actual age), there seems little doubt that they have less incentive to invest in new skills, with less time expected in the agricultural workforce to recoup that investment. In other words, changes in the age structure of the farm population are likely to be accompanied by important shifts in the character of economic activity in agriculture. Furthermore, in conditions of advanced demographic ageing, the farm population may not be able to adequately respond to agricultural intensification and technological change. Thus, the agricultural census may highlight the need for specific strategies to ensure that an older farm population can cope with the challenges of technological change, but also that the agricultural labour market as a whole is equipped for these challenges and that agricultural development is not impeded unnecessarily by labour shortages.
Agricultural census data may also indicate, in an indirect way, a growing demand for adequate services to complement or replace traditional support networks for older people. Healthy and well-resourced aged people are likely to demand various distribution services and formal as well as informal support, while frail aged persons will need a wide array of care services. In particular, the need for appropriate health care including hospitalization may increase dramatically as a result of rural ageing because in many developing countries the rural elderly have a lifetime of exposure to health risks and are on average less fit than their urban counterparts. Some of the required services will be highly skilled (for example, health services) but others will be low skilled (such as transport). However, almost all these services will be labour-intensive. Thus, demographic ageing is likely to considerably alter both the supply of labour and the demand for various kinds of workers in the future.
As with farm populations in general, the ageing of agricultural holders has the potential to considerably influence the socio-economic situation in rural areas. As the indicators in Table 1 prove, agricultural holders tend to be older than the population of members of agricultural households. This is because holders typically execute important legal and managerial functions that in many societies tend to be associated with more advanced stages of adulthood. Besides, despite a worldwide trend away from employment in agriculture, work in this sector continues to be one of the most important sources of economic activity for the elderly in developing countries (Martin and Kinsella, 1994). It needs to be stressed, however, that the data on elderly holders captured in agricultural censuses cannot be taken at face value. In many developing countries, particularly in Africa, the holder will very often be identified, by tradition, as the oldest male in the household. In addition, over-reporting of age is known to be common in many developing countries, especially in cultures where old age is respected and esteemed. Moreover, as in household surveys of older persons everywhere, in the censuses based on sampling there is likely to be a bias towards the relatively healthy and well-off elderly. Without a deeper knowledge of a given socio-cultural context and the concepts and definitions used in collecting information on ages of holders, the data cannot be meaningfully compared between countries.
The reality of elevated numbers of aged agricultural holders can obviously have significant implications for the agricultural sector. For instance, older holders may be inclined to shift to less labour-intensive forms of production or to renting land out in order to receive income without having to work the land. The ageing of agricultural holders can thus trigger important changes in agricultural production practices and land use patterns. Elderly holders (especially the poorer ones) may also be less able to adapt to technological change and less willing to experiment with new modes of production, which in turn could slow down agricultural modernization. On the other hand, under certain conditions (particularly in settings where farming generates sufficient profit or where adequate social security systems exist) older holders may possess considerably more property than younger farm managers, have less farm debt, and generally be in a more favourable financial position to intensify production and thus be a resource for agriculture. Thus, high proportions of the elderly among agricultural holders can be both an obstacle and an advantage for development; the outcome will very much depend on local socio-economic and political circumstances.
But the ageing of agricultural holders can also be viewed in a dynamic perspective. Compared to younger agricultural producers, elderly holders are much more likely to retreat from farming, due either to retirement , ill health, or death. In other words, elevated proportions of holders in the age categories above 60 years suggest that a substantial number of agricultural holdings will be without the present manager in a near future. What can happen to these holdings? In principle, they can either be transferred or sold to other farmers and perhaps integrated with other agricultural holdings; or, if the pool of persons with sufficient capacity to take care of such holdings is limited, they can be left unused and exposed to environmental degradation. Evidently, in each case the implications for agricultural production and food security would be very different.
A related point is that when a relatively high proportion of agricultural holdings is found in the census to be managed by elderly holders, this could be of much importance from a policy viewpoint. For instance, it may suggest an optimum timing for an intervention (such as land or agrarian reform, which could be more effective if coinciding with transfers of land induced by demographic ageing of the holders); or, conversely, indicate that a policy action (say, introduction of a technological innovation) is not appropriate at a given time because many holders are older and thus less able to make the necessary adjustments. In any case, the age distribution of agricultural holders is a factor that could significantly influence responses of the agricultural sector to policies and programmes, and to social change in general.
From the above it should be clear that the information obtainable from agricultural censuses on overall levels of ageing within the population attached to agricultural holdings can be very useful for both researchers and policy-makers. Although in most developing countries the census evidence does not capture all the relevant facets of demographic ageing, many agricultural censuses do provide data that can easily be used to assess the quantitative aspects of ageing among persons attached to agricultural holdings and identify some of the relevant developmental and policy implications. In addition, more detailed agricultural censuses could help to predict the possible outcomes of the ageing process (for instance, a comparison of elderly holders with younger holders may suggest replacements in the future).
Rural areas are not homogeneous but differ from one another, reflecting dissimilarities in environmental conditions, social organization, cultural patterns, economic relationships and agricultural production practices. Demographic structures, too, can vary with the local situation. It is therefore highly unlikely that the same levels of rural ageing will be found within distinctive regional situations. Spatial differences in levels of ageing have thus to be taken into account as one of the key issues for describing and understanding the process of rural population ageing.
There is no uniformity in how rural areas are subdivided for purposes of national agricultural statistics. Some countries pay rather limited attention to geographic differences, but most collect agricultural census information in ways that enable regional comparisons. Clearly, analyses of spatial differences in rural ageing will produce different results depending on whether the unit of analysis is at the county, district, or sub-district level. For instance, an analysis that uses larger regions as the unit of analysis may disguise variation within smaller geographic regions. Of course, the opportunities for geographic analysis of census data are often predetermined by the census methodology. Censuses that use sampling techniques are generally less suitable for this kind of analysis than censuses conducted through complete enumeration.
To demonstrate the usefulness of regionally subdivided data, Table 3 shows the indicators of demographic ageing based on the 1990 agricultural census in the Republic of Korea. The Korean census was carried out as complete enumeration and therefore it is possible to directly compare the results for the country as a whole, its 15 regions and 258 administrative districts. The various measures of variation confirm that the ageing of the agricultural population in South Korea is a regionally differentiated phenomenon, and that smaller geographic units (districts) have a wider variation in levels of demographic ageing than the larger ones (regions). In some districts the proportion of elderly within the population of agricultural households was less than one-half of the national average but in some it was more than 1.4-times above the national figure. Interestingly, territorial differences in ageing were more pronounced for agricultural holders than for the population of agricultural households taken together, and for instance in seven districts elderly holders managed as much as 40 per cent of agricultural holdings or more. Although increases in statistical variability can always be expected as one moves from larger to smaller units of analysis, the main merit of data such as those given in Table 3 is that they capture the exact quantitative dimensions of territorial variation in demographic ageing.
The above example also illustrates that analysts should be cautious when interpreting broad descriptions based on aggregate statistics. Even in countries where geographic differences in agricultural production are relatively small (such as in South Korea), national averages may hide much intra-country variation because rural ageing can have very different impacts across space. Because individual rural regions can face different issues based on economic, demographic and locational realities, they may have different policy needs. If these differences are not taken into account, policies and strategies can have unanticipated or even unintended effects on rural areas because of their distinctive characteristics. In this regard, agricultural census information for smaller geographic areas could be very useful for policy design, implementation and outcomes monitoring.
Demographic ageing is by definition a time-related process, hence it always makes sense to look at how the indices of ageing change with the passage of time. Findings obtained in such a way are not only interesting per se but may also help to specify some of the mechanisms responsible for changes in the age structure of the farm population.
As an illustration, Table 4 displays the data on elderly holders and members of agricultural households aged 60 years and above as recorded in two agricultural censuses in the Republic of Korea in 1980 and 1990, respectively. Both censuses were conducted as complete enumeration and used identical criteria, therefore their results are fully comparable. The absolute figures indicate that whereas the overall numbers of holders and persons belonging to agricultural households declined, numbers of the elderly increased. As a result, in 1990 the elderly accounted for a significantly higher share of the population attached to agricultural holdings than they did in 1980. Several processes could account for this change. First and perhaps foremost, younger members of the farm population may have moved out of agriculture, thus causing a numerical reduction at the 'bottom' of the farm population's age pyramid. Second, demographic ageing in agriculture may have been accelerated by reductions in rural fertility levels. Third, mortality rates at older ages may have declined in agricultural areas, allowing more elderly members of the farm population to stay alive. Fourth, some of those who had once moved to cities may have migrated back to rural areas upon retiring from the urban workforce. Overall, however, the table tells the story of a population that became older primarily due to an outflow of younger persons (relative ageing) rather than to increases in the age-specific numbers of older persons (absolute ageing). On the basis of this evidence, one can draw preliminary conclusions and ask further questions about the nature of demographic and social change in agricultural areas of the Republic of Korea. Furthermore, trends in the evolution of the older population, especially increases of holders aged over 60 and the increasing gender-imbalance among the oldest members of the farm population, are important sociological factors to be considered in future planning for agricultural areas.
The comparability of individual country data between consecutive censuses is normally easier as countries tend to use the same concepts and definitions. Nevertheless, when studying the time series one can sometimes detect unexpected irregularities which may be difficult to account for. Sometimes difference between population figures may be explained by differences in definitions of technical (agricultural) subjects applied, rather then by 'real' changes over time. Data users having limited experience with agricultural censuses should therefore be warned when undertaking comparison among countries, or even among consecutive censuses in the same country.
An agricultural census is capable of providing fairly detailed information on demographic attributes of agricultural households enumerated in the census. To a researcher interested in population ageing, the census can thus be a source of information about the associations between ages of holders and sizes of their households. Such associations can shed light on the social and economic conditions of the rural elderly and hence improve the understanding of the implications of demographic ageing in the agricultural sector.
Figure 1 shows the distribution of agricultural holdings by household size and age of the holder as captured in agricultural censuses in Cape Verde in 1988 and Thailand in 1993, respectively. Although each country used a different categorization of household sizes, the findings are nevertheless roughly comparable. The two graphs portray two very different social realities.
Distribution of agricultural holdings by household size and age of holder: Cape Verde (1988) and Thailand (1993)
In Cape Verde in 1988, the average size of agricultural households was relatively high (5.5 persons) and a large proportion of agricultural holdings (34 percent) reported having a holder aged 60 years or more. According to the census data, households with up to six persons were overwhelmingly headed by elderly holders, and only in households with seven or more persons were the oldest holders outnumbered by those aged 50-59 years. Even one-person households were mostly headed by older persons, which seems to challenge the traditional assumption that older people are necessarily dependent for support on coresidence with younger family members. In addition, the data for agricultural households with 2-3 and 4-6 members followed an U-shaped distribution implying that these holdings were more often managed by persons aged either over 60 or between 20 and 39 years than by persons aged 40-49 years. This lends some support to the thesis proposed by Iaquinta et al. (1999) that transfers of land management entitlements may sometime move directly from the elderly to the adult grandchildren, thus bypassing the parent generation.
In Thailand in 1993, the average size of agricultural household was lower (4.4 persons), probably reflecting a higher fertility level than in Cape Verde, but much less holdings (19 percent) were managed by elderly holders. A significant proportion of holdings was held by holders aged 30-49 years, especially if the holding comprised four to five persons. The percentage of holdings with a holder aged over 60 was certainly not unimportant but in no size-category of holdings were the elderly holders as prominent in Thailand as they were in Cape Verde. Part of the reason for such a situation could be the pattern of older parents living with adult children - a well known feature of many Asian countries, which is typically reciprocal in nature and could thus facilitate fairly smooth transfers of property and management rights from the elderly to the younger generation.
These contrasts between Cape Verde and Thailand are hardly explainable in demographic terms because overall levels of population ageing were very similar in both countries at the time (according to the 1999 UN population estimates, the proportion of persons aged 60 and more was 7.0 percent in Cape Verde and 6.7 percent in Thailand in 1990). Although the agricultural census data by themselves cannot provide the required explanation, it seems very likely that above discussed country differences in household size and age of holder are at least partly related to agricultural issues: age-specific aspects of land tenure, intra-household distribution of management responsibilities, allocation of economic roles within agricultural households, patterns of intergenerational transfers of land, etc. Some of these factors could be dictated by tradition (such as the reporting of the oldest household member as the holder, which may be more common in Cape Verde than in Thailand), but others may result from local environmental conditions, agricultural practices, socio-economic policies or other externalities, many of which are known to have quite distinct age-specific implications for members of agricultural households. The evidence presented in Figure 1 thus indicates that the linkages between the holder's age and the size of his/her household are complex and potentially important for appreciating the demographic underpinnings of the agricultural sector in developing countries. In particular, data about elderly holders and their households have implications for household living arrangements, intergenerational relationships, care-provision and social support systems in rural areas. Such information, and therefore also the data source on which it is based, should be of interest to both agricultural professionals and specialists dealing with rural ageing.
In some agricultural censuses, holdings are classified also according to the main type of farming activity performed. FAO does not provide any specific recommendations in this regard and country practices differ to a large extent. Although the collection of this type of information presents certain difficulties, especially in regions where mixed farming is common, the resulting evidence could nevertheless illustrate important interactions between demographic and agricultural factors.
Using census information for the Republic of Korea, Cyprus and Tanzania, Figure 2 shows the structure of agricultural holdings by main types of farming and ages of holders. The Korean data portray a very homogeneous agricultural system: a decisive majority of holdings enumerated are paddy farms while other types of farming are rare though not negligible. However, different farming types are associated with different age distributions of holders. For holdings engaged in paddy farming or producing vegetables or upland field crops, the distribution is heavily skewed towards older holders, whereas holdings specializing in livestock or fruit production display a more regular age distribution of holders. In Cyprus, holdings located in the vine and mountain zones are much more likely to be managed by elderly holders than those in the remaining agro-economic regions. In Tanzania, the three farming types distinguished in the agricultural census do not differ much from one another as far as the age structure of holders is concerned, and none of the age group of holders stands out, except for those aged 30-39 years who hold a large proportion of crop-producing farms.
Distribution of agricultural holdings by age of holder and type of farming: Republic of Korea (1990), Cyprus (1994) and Tanzania (1994-95)
These findings are open to a range of interpretations; for instance, they could reflect historical trends in agriculture (older holders dominate traditional farming types but are under-represented in more modern specializations) or they could correspond to differences in the intensity of labour required for certain types of farming (activities such as fruit and livestock production are more labour-intensive and therefore less attractive for older holders). Age-selective outmigration from specific rural areas can also play a role, given that in most countries certain farming types tend to be concentrated in certain geographic locations and can thus be exposed to different migration regimes. But the age structure of holders is likewise an important predictor for future developments within individual farming types. For instance, it can be speculated that the agricultural sectors where older holders prevail will react differently to market influences, agricultural policies and other interventions, or follow a different path of agricultural intensification than those with younger holders. This, in turn, can foster further demographic, social and economic discrepancies between different farming types.
Previous examples have been based on data about absolute numbers of agricultural holdings. However, some agricultural censuses make it possible to examine also the area of holdings. In combination with the information about the age structure of holders, these two perspectives are likely to produce different results, as the data from the 1991-92 agricultural census in Nepal reveal (Figure 3).
Distribution of agricultural holdings by age of holder, number and area of holdings: Nepal (1991-92)
Although the two distributions depicted in the graph are not radically different, they nevertheless show that the amount of agricultural land associated with specific age-categories of holders is different from what one might have deduced from the information on absolute numbers of holdings. The graph provides evidence that the share of older holders is more important in terms of area than in terms of number of holdings; for instance, male holders aged over 55 years hold 21 percent of agricultural holdings, but this corresponds to 26 percent of agricultural land. Thus, one may deduce that older farmers have on average larger holdings than their younger counterparts. Findings of this kind highlight the importance of a factor that is discussed in more detail in the following section: the association between the holder's age and the average size of the holding.
Given that characteristics of production are generally omitted from agricultural censuses as most countries consider them unsuitable for monitoring through the census approach, the size of the holding is often the only census item that can be used as a proxy of the holding's overall productive capacity. Demographically speaking, size of the holding can have direct bearing on demand for agricultural labour and therefore on household size and structure; migration of household members to and from the holding; hiring of agricultural workers; and various kinds of relationships between those living on the holding and their distant relatives. A larger size of the holding is typically associated with an increased capacity to support additional children, accumulate savings and rent out cultivable land, which in turn could increase a holder's quality of living and the probability of support during his/her old age. Moreover, in traditional agricultural societies the size of the land held is closely connected with social status: the larger is the holding size, the higher is the social status of the household members. In reality, of course, the holding size is the outcome of several interconnected factors: environmental circumstances, existing agricultural practices, possibilities of access to land (i.e. conditions for land tenure) and, last but not least, the holder's previous demographic and economic experience. Considering this, it is clear that the size of holding can vary significantly with respect to the holder's age.
Most countries follow in their agricultural censuses the FAO definition of the total area of agricultural holding. However, there are two major factors that make it difficult to compare data on area of holdings between different countries: (a) agricultural practice factor, which is related to climatic conditions and refers to the fact that in some countries there may be two or more cropping seasons while in other countries there is only one season; and (b) land tenure factor, which refers to the fact that in many countries some categories of land use, such as pastures, are not a part of the agricultural holdings but belong to the community.
In fact, the average size of holdings calculated from the data of the 1990 round of agricultural censuses varies very much from country to country, from around 1 hectare or less in countries such as the Democratic Republic of Congo, Nepal and the Republic of Korea, to 78 hectares in Paraguay and 469 hectares in Argentina (FAO, 1997). Once again, however, national averages are likely to hide considerable variation by the holder's age. Unfortunately, only a few countries publish the crosstabulation required to analyse this relationship.
For Thailand and the Republic of Korea, such data is available and the results are indeed quite interesting (Figure 4). Both countries display a similar pattern: within younger age groups the size of holding tends to increase with the holder's age; the averages reach maximum values for holders aged 45-49 years in Korea and 55-59 years in Thailand; but in older age categories the size of holding declines with each added unit of age. Although the sizes of holdings and the intensity of the relationship between holder's age and size of holding are far from identical in the two countries, in both cases the age of the holder appears to be a significant determinant of access to land and, presumably, of agricultural productivity as well. In both countries, farmers between 45 and 60 years of age emerge as proprietors of the largest agricultural holdings, and one can speculate that many of them will be among the key agricultural producers. Interestingly, the sex-specific data for Thailand reveal that the age differentiation follows a similar distribution for both male and female holders. In other words, in the Thai case age is more important than sex when management of agricultural land is concerned. A comparison of this data with other country evidence could be helpful in investigating the extent of gender (in)equality in agricultural areas of developing countries.
Distribution of agricultural holdings by average size of holding and age of holder: Thailand (1993) and Republic of Korea (1990)
To summarize, the data presented in Figure 4 seem to reflect important linkages between demographic and socio-economic phenomena. Several possible interpretations may be proposed; for instance, one can use the life-cycle perspective. In many Asian countries, farmers aged between 50 and 60 years are at the peak as agricultural producers; have at their disposal considerable resources; and as members of high-fertility cohorts are often heads of larger households - all of which could create the conditions favourable for large-scale farming. In contrast, older holders who on average are less physically capable and closer to retirement are more likely to intentionally reduce the size of their holdings through transfers of land to younger family members or through sale or rental of land to non-relatives. Small average sizes of farms managed by young farmers may be the result of deliberate choice on part of these farmers, reflecting the lack of interest in full-time farming among younger generations; or they could be the outcome of the prevailing economic conditions which may not allow younger farmers to obtain more land; or they may simply mean that new land is rare or non-existent. It is clear that the relationship between the size of the holding and the holder's age can only partially be understood with data from agricultural censuses. Conclusive evidence on this association may only come from appropriately designed surveys. However, the analysis of agricultural censuses could be the first step towards uncovering the possible determinants involved in this interaction.
In several developing countries agricultural censuses could be used to identify important associations between the age structure of the farm population and certain aspects of the agricultural system. Most importantly, agricultural censuses provide a unique opportunity to link information on the techno-economic features of the agricultural holding on the one hand and the demographic and other characteristics of the holders and their households on the other. Since agricultural censuses do not specifically target the older population, they make it possible to apply a more holistic perspective and study the situation of the elderly within the broader demographic, socio-economic and institutional context of the agricultural sector. In some countries the evidence needed for this analytical approach will not be obtainable from other sources.
Given the various limitations of agricultural censuses as sources of demographic data (concepts and enumeration criteria applied, data collection and data publishing practices followed), these censuses cannot be expected to provide perfect information on the rural elderly and associated factors. In fact, the agricultural census is not well suited for soliciting information on a number of important topics (e.g. living arrangements and health conditions, agricultural productivity, income and wealth of older persons, support systems available to the elderly etc.). Moreover, given that the main aim of the agricultural census is to provide data on economic characteristics of agricultural holdings, information on older, economically unactive persons may not be fully reliable. On the other hand, the census data are readily available in a number of developing countries and provide a rich reservoir of evidence that can be analysed with little extra cost. In addition, most agricultural censuses cover a range of socio-economic and environmental contexts and thus permit the kind of disaggregated analysis which is often impossible with in-depth surveys.
The following recommendations suggest ways of advancing the use of agricultural censuses, which can in turn lead to findings that challenge some of the prevailing misconceptions about the role of rural elderly in developing countries:
A great deal of the data on elderly persons in agricultural censuses is rarely analysed and hardly ever reaches users, in part because of restricted public access to such data. This is all the more unfortunate since analyses based on agricultural censuses have a significant potential to expand the already existing understanding of demographic change in the agricultural sector and to address significant policy issues. In particular, the inferences that could emerge from agricultural censuses can be crucial in identifying interconnections between demographic ageing and issues that are often of vital importance at the national level: food production and food security, supply of agricultural labour, poverty and access to resources. Insights derived from agricultural censuses can then be used to guide policies and programmes aimed at managing the social and economic challenges brought about by the ageing of rural communities. Agricultural and population specialists interested in developing countries, including policy-makers dealing with various aspects of rural demographic ageing, should therefore be alerted to the information potential of agricultural censuses. At the same time, more attention needs to be paid to using agricultural census statistics in decision-making, and for monitoring government policy and performance in the area of rural ageing.
1 However, in many developing countries, stopping work due to age is less of an event and more of a process of gradual withdrawal from various activities, so the age pattern of retirement may not be so marked. This is especially true in the traditional agricultural sector where mandatory retirement ages have little or no effect.
2 An excess of women over men at older ages is typically viewed as problematic since it reflects higher levels of widowhood and because older women, especially those without spouses, are thought to suffer a greater disadvantage than elderly men.
3Of course, in most settings the productive potential will depend not only on the acreage of the holding as such, but also on the location of the holding, the quality of land and its utility within a given farming system.
4 According to FAO (1997), total area of agricultural holding is the combined area of all the holding's parcels. Land owned by the holder but rented to others should not be included in total holding area. The holding area includes farmyard and land occupied by farm buildings. The total area of a holding practising shifting cultivation should include area under crops during the reference period and area prepared for cultivation but not sown or planted at the time of enumeration. Holders having access to communal grazing land should not include their estimated share of such land in their total holding area. Total holding area should be equal to total land area under various land use classes.
5An example of such a combined dataset may be found in the 1996 Agriculture-Population Linkage Database developed by Statistics Canada.
Basic indicators of rural demographic ageing for developing countries that took a census of agriculture in the period 1986-95
Coverage of agricultural censuses and application of sampling for developing countries that took a census in the period 1986-95 and monitored information on the age structure of the population attached to agricultural holdings
Indicators of regional variation in ageing of the population of agricultural households (proportion of persons aged 60 years and more): Republic of Korea, 1990
Population of agricultural households aged 60 years and more: Republic of Korea, 1980 and 1990
Argentina, 1989. Censo Nacional Agropecuario 1988: Resultados Generales, Vol. 26: Total del Pais. Instituto Nacional de Estadistica y Censos: Buenos Aires.
Bahamas. 1995. Second Census of Agriculture 1994: Final Results. Ministry of Agriculture and Fisheries - Department of Agriculture, Department of Statistics: Nassau.
Botswana. 1995. 1993 Botswana Agricultural Census Report. Division of Agricultural Planning and Statistics, Ministry of Agriculture - Central Statistics Office, Ministry of Finance and Development Planning: Gaborone.
Cape Verde. 1990. Recensamento Agricola 1988, Vol. 1: Dados Globais. Ministerio do Desenvolvimento Rural e Pescas - Gabinete de Estudos e Planeamento: Praia.
Colombia. 1990. Primera Encuesta Nacional Agropecuaria, Vol. 1: Total Departamentos. Ministerio di Agricultura: Bogota.
Cyprus. 1997. Census of Agriculture 1994. Department of Statistics and Research - Ministry of Finance: Nicosia.
du Guerny, J. 1997. The rural elderly and the ageing of rural populations. FAO Sustainable Development Department.
FAO. 1978. Collecting Statistics on Agricultural Population and Employment. FAO Economic and Social Development Paper No. 7. FAO: Rome.
FAO. 1986. Programme for the 1990 World Census of Agriculture. FAO Statistical Development Series No. 2. FAO: Rome.
FAO. 1992. 1980 World Census of Agriculture: Methodological Review. FAO Statistics Series No. 107. FAO: Rome.
FAO. 1995. Programme for the World Census of Agriculture 2000. FAO Statistical Development Series No. 5. FAO: Rome.
FAO. 1996. Conducting agricultural censuses and surveys. FAO Statistical Development Series No. 6. FAO: Rome.
FAO. 1997. Report on the 1990 World Census of Agriculture. FAO Statistical Development Series No. 9. FAO: Rome.
Grenada. 1996. Final Report on 1995 Grenada Agricultural Census. Ministry of Agriculture, Planning Unit: St. George's.
Guinea Bissau. 1990. Resultats du Recensement National de l'Agriculture 1988-89, Vol. 1: Resultats Nationaux et Regionaux. Ministere du Developpement Rural et de l'Agriculture - Cabinet de Planification: Bissau.
Iaquinta, D., du Guerny, J. and Stloukal, L. 1999. Linkages between rural population ageing, intergenerational transfers of land and agricultural production: are they important? FAO Sustainable Development Department.
Korea, Republic of. 1982. 1980 Agricultural Census, Vol. 10: National Edition. Ministry of Agriculture and Fisheries: Seoul.
Korea, Republic of. 1991. 1990 Agricultural Census, Vol. 1: Whole Country. Ministry of Agriculture, Forestry and Fisheries: Seoul.
Marcoux, A. 1990. Ageing rural populations in the developing countries: patterns, causes and implications. United Nations International Conference on Aging Populations in the Context of the Family, Kitakyushu, October 1990. In: Ageing and the Family, United Nations: New York, 1994, pp. 144-148.
Martin, L.G. and Kinsella, K. 1994. Research on the demography of aging in developing countries. Demography of Aging, National Academy Press: Washington, pp. 356-403.
Nepal. 1993. National Sample Census of Agriculture 1991/92, Vol. 1: Nepal. National Planning Commission Secretariat - Central Bureau of Statistics: Kathmandu.
Paraguay. 1994. Censo Agropecuario Nacional 1991, Vol. 6: Sector Agricola. Ministerio de Agricultura y Ganaderia: Asuncion.
Skeldon, R. 1999. Ageing of Rural Populations in South-East and East Asia. FAO Sustainable Development Department.
Stloukal, L. 1999. Using agricultural census data for demographic purposes: some pros and cons. FAO Sustainable Development Department.
Tanzania. 1996. National Sample Census of Agriculture 1994/95, Vol. II: Tanzania Mainland. Statistics Unit, Ministry of Agriculture - Bureau of Statistics, Planning Commission: Dar es Salaam.
Thailand. 1993. 1993 Agricultural Census, Vol. 1-5. National Statistical Office - Office of the Prime Minister: Bangkok.
UN. 1993. Urban and Rural Areas by Sex and Age: The 1992 Revision. UN Population Division: New York.
UN. 1999. World Population Prospects: The 1998 Revision, Vol. I: Comprehensive Tables. UN Population Division: New York.
UNFPA. 1999. Population Ageing: Challenges for Policies and Programmes in Developed and Developing Countries. Edited by R. Cliquet and M. Nizamuddin. UNFPA & CBGS: New York / Brussels.