The sentence in Spanish means "we do not want goals, we want beans", and it has been painted on a Mexico stadium during the football Mundial in 1986. In some way, it is the most concise statement of the dilemma that sport has to cope with in the developing world. Sport events and sport shows cannot make people forget underdevelopment, poverty and hunger. Should the governments in the Third World provide the population with sport events or with grass root sport practice? Should they attract more (usually foreign) money to organise big sport events or should they allocate a greater share of state and municipality budgets to the development of a "sport for all" or a school sport practice?
Any economic analysis of the relationships between sport and economic development can hardly avoid to start with a very sad look at the empirical evidence which shows that the least a country is economically developed, the least sport practice, sport performance, sport facilities, and sport finance are developed. Although data are missing to calculate an exact correlation between economic underdevelopment and sporting underdevelopment - this lack of available data in itself exhibits a low level of both economic and sporting development -, a number of real facts converge towards a strong relationship between the level of economic and sporting development. In such a context, the great bulk of Third World countries are somewhat dependent for their sporting activities on foreign aid, money flowing from abroad, import of sporting goods and equipments, multinational sponsors, and sometimes the relocation of industrial production by multinational corporations (Andreff 1988), while they cannot hedge against a dramatic "muscle drain" of their most talented athletes abroad (in the same sense as we usually speak of a brain drain of the elites from the educative and research system of these countries).
Then, the long lasting question to be addressed is: what is to be done for improving sporting development in Third World countries? The problem is that economic development itself is a good part of the recipe after the last forty years which can be coined the "decades of economic development failure", except in a handful of South-Asian and Latin American newly industrialized and emerging market economies. So, we will confine to discussing only a few suggestions, in the second part of the paper, well aware that they do not fit to solve but simply to alleviate some causes of the joint economic and sporting underdevelopment in most developing countries.
1. Low-level economic development triggers low-level sporting development
"The gap between developed and underdeveloped countries has become so wide that the latter have lost any hope to close it as regards either sport performance or the organisation of big sport events" (Ezziani, Kaach 1984). Such a statement remains basically true in the nineties, up to 2000, but it must be differentiated according to the level of economic development in various Third World countries. It is quite usual to distinguish, among developing countries, the least developed countries (LDCs) with a GNP per capita lower than $ 600 per year, the middle-income developing countries (MICs) with a GNP per capita between $600 and $ 2,000, and the newly industrialised or emerging market countries (EMCs) with a GNP per capita between $ 2,000 and $ 9,000 - even though some of them are not yet market economies (Cuba, Belarus). The distinction relies on some other economic variables that we would not comment on here. Applying the previous criteria, we have selected 48 LDCs, 53 MICs and 41 EMCs (See Tables 1a, 1b and 1c in Annex). The issue of sporting underdevelopment is of course the most serious in LDCs and more serious in MICs than in EMCs, and we discuss it much more here in reference to the two first groups of countries.
1.1. No available statistical data
Even from Tables 1a, 1b and 1c, we can yet see that statistical data availability is lower for economic and education figures, the lower the level of economic development. The same evidence pertains to sporting development data. "Precise data about money spent on sports and games in India is not available" (Bhatty 1989). It is just not possible to draw a statistical profile of the people who are participating in various sports and games in today India, for example. It is quite impossible to collect economic data related to sporting activities in LDCs and MICs and still quite difficult to get it in EMCs. One of the rare data source we can rely on is the result of a questionnaire sent by UNESCO to 32 African LDCs: the response rate has been 50% (Souchaud 1995a).
1.2. Not enough physical education andsport for all practice
A) Physical education: school enrolment and exposure
Children and teenagers first must attend school programmes in order to benefit from physical education at school. Let us define the net enrolment ratio in the primary school as the ratio between the number of children enrolled and the whole relevant age group. Whereas this net enrolment ratio is in the range of 95%-100% in all developed countries, it is simply not exactly known in many developing countries in 1998 (Tables 1a and 1b).
Among the EMCs for which information is available, the ratio varies from 76% in Lebanon and 81% in Botswana to 100% in Mexico. In the MICs, the interval of variation is slightly wider, from 74% in Morocco and 78% in Nicaragua to 100% in Philippines and China. The worst situation is observed in the LDCs among which the enrolment ratio falls down to 30% in Eritrea, 28% in Ethiopia and Mali, and 25% in Niger. In other words, a child in Niger has an ex ante probability to be exposed to physical education in the primary school which is only a quarter of the probability for a child in Mexico, Philippines and China (or France).
Now, once exposed to physical education, the duration of exposure has a deep influence on developing sport practice among children and teenagers. In the 16 LDCs surveyed by UNESCO (Souchaud 1995a), one country had no physical education scheduled in the primary school in 1995, three countries had scheduled one hour per week and the other countries between two and three hours per week (Table 2); in the secondary school, physical education was scheduled two to four hours per week in all countries.
The problem is that the hours supposedly devoted to physical education are practically never fulfilled, at least at 80% of the time schedule, and in several countries the sport time schedule is simply not accomplished at all (response: no, in table 2). This is mainly due to a shortage of sport teachers, to the number of pupils (40 to 100 per one sport teacher) in the primary school, and, in addition in the secondary school, to a lack of sport facilities and equipements, a shortage of qualified teachers and no vocational training to improve their qualification.
Most sport teachers qualify abroad in bilateral co-operation programmes with European countries, after having got a grant. International cooperation is badly needed for the training of sport teachers from developing countries. Nevertheless, we observe a contrasted capacity to train sport teachers between EMCs and MICs like Argentina, Cuba, Egypt, Syria or Thailand (Table 3) on the one hand, and LDCs such as Benin, Niger, Guinea or Mozambique on the other hand.
The number of sport teachers in Tunisia has increased from 3,606 in 1990 to 4,419 in 1994; but the population in school age is roughly 2.5 million people, so that the ratio is still one teacher for roughly 565 sport pupils on average (Khiari 1999). The number of newly hired sport teachers has fallen from 268 in 1991 to 190 in 1994. Tunisia is quite representative of the school sport situation in the MICs and lower income EMCs, although it is far better in several top EMCs.
B) Sport for all: does it mean anything in developing countries?
When we want to see which part of the population in a country has a sport practice, we have to define a participation ratio. Such a definition is not unique. The simplest definition is to compare the number of affiliated members to sports federations with the number of inhabitants. Although it is not a satisfying definition, we can stick to it in developing countries where accurate data is so rare.
In developed countries, we can use a frequency (once a week) or a duration (one hour per week) criterion to define more exactly how many are the sport participants, but a data collection process by a national institute for statistics is then required. A broader definition is to leave the participants self-define and quantify their sporting practice in their response to an ad hoc inquiry or questionnaire.
The great bulk of participants is concentrated on a few sporting practices, determined by available sport facilities; in Tunisia, in 1994, there were over 2,000 participants in only seven sports: football, handball, basketball, track and field athletics, karate, tennis and volleyball. Even though less precise information is available, the main reasons for the low level of participation to sports in India are assessed to be: inadequate sport facilities, too little attention paid and too little money allocated to sports and the development of physical culture, the general absence of positive attitudes towards sport, society and the fact that parents rarely encourage young people to take physical exercise (Bhatty 1989).
In African LDCs, sport participants were even fewer in 1995 with, on average, one affiliated member to sports federations per 800 inhabitants (participation ratio: 0.0013%) in the 9 LDCs surveyed by UNESCO (Table 5). The participation ratio is between 1.4% (1/71) in Cap Verde and 0.0007% (1/1500) in Gambia, compared with a participation ratio in the range of 20-25% (1/5 to 1/4) of the population in European countries in 1990.