1Although social practice determines the manner of eating and food choices, the range of products (understood here as the range of available products), their prices and characteristics, and the populations’ economic access to these products will also influence food consumption. These issues will be addressed through an examination of supply channels (local production, importation, supply locations) as well as food prices (levels and determinants), and then household consumption and socio-economic disparities in food consumption.

1. Supply channels

2The supply consists of local or imported products, distributed through several supply locations (large and medium-sized supermarkets, convenience stores, etc.), which are used differentially according to the population’s socioeconomic categories.

1.1. Local food production and importation

3Assessment of local products in the overall supply involves analysing the coverage rate, i.e. the proportion of local products as part of the total consumption. For the 2013-2017 average (Odeadom, 2018) it is clear that the coverage rate was very good for eggs in all of the DROM. The best fruit rates were achieved by French Guiana and Mayotte, followed by La Réunion, but they were lower in the Antilles. The Antilles also maintained a low rate for vegetables, while it is high for French Guiana, Mayotte and La Réunion. The lowest results were for animal products, relatively low for all of the DROM (except La Réunion for poultry), especially for dairy products: Mayotte and La Réunion posted low rates, while milk production was virtually absent in the other DROM.

4With regard to changes in the period 2007-2017, the comparison of growth rates of agricultural production and demography shows that the coverage rate tended to increase in French Guiana, decrease in Guadeloupe and Martinique and stagnate in La Réunion. These changes were explained by the dynamic of agricultural areas undergoing plant- and animal diversification, diminishing sharply in the Antilles (minus 1,600 ha on average per year), increasing slightly in la Réunion (plus 146 ha on average per year) and more significantly in French Guiana (plus 1,129 ha per year).

5Given the pattern of current consumption and the agricultural potential, the coverage of meat consumption by local production is not viable20. However, according to the French Overseas Departments Agricultural Economy Development Office (Odeadom), the size of a fruit and vegetable production area needed to cover all the fresh produce requirements is estimated to be only between 11,000 and 26,000 ha, i.e. between 6 and 14 % of the overseas agricultural area (although this calculation does not take into account any need to increase consumption of fruit and vegetables to improve diet quality).

6There were major differences in 2015 in the structure of imports for the five DROM (table 1): for Guadeloupe, French Guiana and Martinique about 40 % of calories were imported in the form of fats, 42 to 47 % in the form of carbohydrates and 14 to 18 % in the form of proteins. In contrast, in Mayotte there was a much higher proportion of carbohydrates whereas the proportion of fats was much lower, which aligns with individual consumption studies that show that the population’s intake of fat was much lower than the recommended levels, whereas they exceeded the recommendations for total carbohydrate intake. La Réunion had a profile in the middle between Mayotte and the other three DROM. In Martinique the imported calories accounted for 90 % of the calories consumed, slightly lower than La Réunion, while a large proportion of local produce was exported to these two overseas regions (sugar, tropical fruits).

7Local food production is concentrated around a few product families. In Martinique local industry mainly supplies the soft drink (fizzy drinks, fruit juices, nectars, etc.) and alcoholic drink (rum) sectors as well as the cured meat, biscuit and dairy product sectors, with regard to which the issue of the nutritional quality of locally produced products compared to imported products has been raised several times. It covers two areas: the first is the segments in which the local industry operates compared to imported products (for example, brine products substantially more salty than the rest of the cured meat products; fruit nectars substantially sweeter than the rest of the soft drinks).

8The second is the nutritional quality of the same types of product: the sugar content of yoghurts (or fizzy drinks) produced locally compared to those produced in mainland France. In relation to this second area studies have revealed the differences between imported products and locally produced products (2011-2013), in particular for local fizzy drinks or dairy products, the sugar levels of which were much greater than their equivalents in mainland France. Following a legislative measure (law of 3 June 2013), manufacturers are expected to have reformulated their products to reduce sugar levels. A systematic evaluation of the discrepancies between Martinique (and the DROM in general) and mainland France in the nutritional quality of the local supply still remains to be made, however. Should discrepancies be found regarding the two areas mentioned previously, it would become necessary to examine ways of supporting the local industry which could help improve the nutritional quality of this local supply.

1.2. Nutritional quality of the food supply

9A good knowledge of the nutritional characteristics of the food supply available in the DROM would be useful for a number of reasons. The first would be to determine whether the foods sold in the DROM, whether imported or produced locally, have a different nutritional composition than the products available in mainland France, and therefore whether the food environment has qualitative characteristics that magnify or lessen health risks compared to mainland France. The second would be to determine the extent to which the nutritional quality of food products varies according to distribution channels and price levels, so as to determine whether or not, in each product category, those of better nutritional quality are positioned at affordable price levels. Finally, a determination would need to be made as to what the relevance and impact on the food supply could be of a public policy which aims to encourage industrial producers to change products’ characteristics so that they have better nutritional content (product reformulation). In general, we have very few documents and publications available that allow us to respond to these questions.

10With regard to product quality, studies conducted in 2011 and 2012 noted differences in the nutritional composition of products available in certain DROM and those sold in mainland France. Following this observation, the law of 3 June 2013 (known as the Lurel Law) aimed to guarantee the quality of the food supply in the French overseas departments and regions. The Decree of 9 May 2016 states that “the added sugar content of foodstuffs […] distributed in the French overseas communities […], but not distributed by the same brands in France, cannot be higher than the highest added sugar content found in similar food products of the same family that are the most frequently distributed in mainland France”. This law applies to the following products: sugary soft cold drinks, dairy and similar products, biscuits and cakes, ice cream and sorbets, breakfast cereals, cereal bars, chocolate and chocolate bars. In response, manufacturers in the drinks and dairy products sectors announced plans to reformulate products and reduce sugar content, but there has been no systematic review to date in the DROM as a whole to assess this law’s impact.

11Studies conducted recently as part of the Nutwind project in Martinique and Guadeloupe do shed particular light on the sugary drinks sector, known to be one of the main vectors of sugar exposure in the Antilles, but the specific nature of these studies require the results to be considered with caution, with generalisations avoided at this stage.

12Nevertheless, available results suggest that there is a tiny fraction of reference products in the sugary drinks market in the Antilles with a sugar content higher than the highest level observed in the same product family in mainland France, while it is also noted that average sugar levels observed in the Antilles in certain product families are significantly higher than those observed in mainland France (fig. 9). Furthermore, the product families with the highest number of reference products are among those with the highest average sugar content, as a result of which the average sugar content per reference product is also significantly higher than that observed in mainland France (table 2, see IV). If we accept that the number of references per family is an indicator of the quantities sold, this means that the pattern of consumption is clearly oriented towards the most sugary product families on average.

13Furthermore, it is noted that there is no correlation between the prices of products and their sugar content – the full range of sugar content is available over the full price range. The least sugary drinks are therefore neither more nor less accessible than the most sugary drinks.

14The results from this Antilles example regarding sugary drinks cannot be generalised at this stage to other product categories and to other DROM - this will have to be done in the future. The example does however underline the importance of considering the qualities of the products on offer in the DROM. The observed quality levels bring into play both consumer demand (the fact that the reference products are the most common of the most sugary product families is a reflection of consumer demand based on preference) and the companies’ decisions regarding product manufacture. Discussions around conditions for improving properties (here a reduction in sugar content) should take into account the interaction between these two aspects.

15These discussions should also include the fact that a large proportion of the food supply is made up of imported products, especially from mainland France. Actions to improve this supply are therefore best dealt with at national level first. But there is also a locally produced food supply which may be significant, in varying degrees according to the DROM in question, for several sectors such as drinks, cured meats, biscuits and certain dairy products. Local producers should also consider reformulation procedures for these product groups, especially in product families with an average sugar content higher than that observed in mainland France.

1.3. Description of supply locations

16Depending on their socio-economic position, people in the DROM do not obtain food and drink in the same way from hypermarkets, supermarkets, convenience stores, markets and informal channels (self-consumption, donations).

17As in mainland France, hypermarkets and supermarkets are where the majority of people obtain food and drink, ahead of small shops which are seeing a decrease in their sales (except for bakeries) (table 3, see IV). Alongside hypermarkets and supermarkets, convenience stores and specialist shops are developing new distribution points with commercial catering (fast food), mass catering, street traders and food trucks. Informal channels remain difficult to document.

18In Martinique and La Réunion as of 2015 nearly eight out of ten purchases are made in supermarkets and hypermarkets (Iedom, 2017; DAAF de La Réunion, 2015; CCI de La Réunion, 2016), more than in mainland France (Brasset et al., 2015; Cossou and Forgeot, 2010, Kranklader, 2014). La Réunion inhabitants buy half of their fresh fruit and vegetables in hypermarkets and supermarkets, with markets still a major supply location. Although hypermarkets and supermarkets are the main shopping locations in Guadeloupe, regardless of level of education, employment status and income level, they are visited more by people with jobs, those with medium or high incomes and those with a level of education better than the high school diploma or A levels, compared to less-advantaged categories (Brasset et al., 2015; Orsag, s.d b). In Guadeloupe the number of people who visit hypermarkets and supermarkets increases with the rise in the level of education and among working people.

19Nearly half of the superstores in Martinique and La Réunion are hypermarkets, a higher proportion than in mainland France (Iedom, 2017; CCI de Guadeloupe, 2015; CCI de La Réunion, 2016; Kranklader, 2014), whereas hard discount stores in Martinique and Guadeloupe enjoy a greater market share compared to mainland France and La Réunion. As regards relationships between the supply location and socio-economic position (Brasset et al., 2015; Orsag, 2010c), supermarkets and hypermarkets are the main shopping locations in Guadeloupe for all categories, but they are visited more by people belonging to the most advantaged socio-economic categories. In La Réunion also, all households shop in supermarkets and hypermarkets regardless of income. Hard discount stores are visited more by people with low SEP in Guadeloupe and La Réunion, as are convenience stores. The inhabitants of Mayotte shop in large supermarkets (at Mamoudzou), as well as in small village stores and a little less in specialist stores (especially for fresh produce) (Martin, 2010; Iedom, 2017). Nearly four in ten products come from self-production, donations and informal supplies (Guyot, 2013). Furthermore, many poor Mayotte inhabitants and asylum seekers turn to food aid as a permanent source of supply (Aboudou et al., 2019). In French Guiana, the prices of basic food products are high (Iedom, 2018a; CCI de Guyane, 2014), and for many shopping is done in supermarkets and hypermarkets alike (mainly in Cayenne).

20Supply through self-consumption (personal production or donations) is significant in La Réunion, Martinique and Guadeloupe, especially for couples with children and the least advantaged populations. In Guadeloupe, two out of three people benefit in this way (Orsag, 2010c). In La Réunion, it helps increase the consumption of fruit, eggs, milk and cheese, while in Mayotte it is significant with regard to the consumption of bananas and fishery products.

21In conclusion, the significant and growing proportion of purchases in hypermarkets and supermarkets in the DROM, as in some provincial cities in mainland France, already exceeds 80 %. The Antilles stand out as a result of their greater proportion of hard discount stores. Apart from the hypermarkets and supermarkets, purchases are also made in bakeries and, for fresh produce and fish, at markets. Informal and roadside purchases are not very quantifiable, but they are not insignificant. The proportion of self-consumption and local production is still sizeable for vegetables, starchy foods, root crops and fruit. In French Guiana the proportion of local production and self-consumption, as well as informal channels, remains poorly documented – the purchasing criteria, their typology and their development are little known due to the scarcity of data on the subject. And there are no data available on purchasing influences according to presentations, contents labels, or Nutriscore in the DROM.

2. Food prices and socio-economic determinants of food

22Lower incomes in the DROM compared to mainland France, combined with higher food prices, raise questions about the influence of the economic aspects on household food choices.

2.1. Levels and determinants of food prices

2.1.1. Comparison of food basket costs

23Overall, the food prices assessed by the Insee in the DROM are much higher than the food prices in mainland France. As the average food choices are different according to the DROM and in mainland France, the price comparison conducted by Insee involves the value in each DROM of a basket of food21 in mainland France (column 2 of table 4), and the value in mainland France of a basket in each of the DROM (column 3). The average price difference between the DROM and mainland France can be derived from this (column 4).

24It is clear that in 2015 food prices in the DROM were higher than in mainland France, particularly in Martinique, Guadeloupe and French Guiana. Price differences tended to increase between 2010 and 2015, except in French Guiana. It is also noticeable that the differences are more significant when measured for a basket sold in mainland France than evaluated for a local basket in each DROM. Thus in 2015 these differences vary from 37 % (La Réunion) to 48 % (Martinique). The price differences between the DROM and mainland France for the DROM basket are around 20 % (- 23 % to - 18 %) for the four most populated DROM. In the case of Mayotte, while the price difference of the basket sold in mainland France is 42 %, the price difference is zero for the food basket consumed in Mayotte.

25This means that the choices made by consumers in the DROM relate to products the price differences of which are lower compared to those of mainland France. However, this overall information does not help to determine whether or not these different choices promote a diet of good nutritional quality and to what extent the structure of food prices promotes such a diet. Specific analyses on this question were carried out as part of the expert appraisal and are presented in the paragraph “Food expenditure and diet quality”.

2.1.2. Role of dock dues and competitive mechanisms

26There are multiple factors which contribute to higher prices in the DROM compared to mainland France and they have been the subject of several assessments by the Competition Authority (Autorité de la Concurrence, 2009):

• the narrowness of the markets and therefore the additional production costs;
• the cost of sea freight;
• the weak competitiveness of local production, which seldom profits from tax exemptions which are favourable to it;
• specific local taxation with internal and external dock dues (one third of commune and regional financing in 2007);
• the weakness of competitive mechanisms due to the concentration of distribution (dominated by national brands) and the vertical diversification of distribution groups. Several distributors are producers of certain food products, but they also act as their own retail outlet suppliers as wholesaler-importers. It is therefore complicated for new retail players to establish themselves, hence there is less competition in the DROM. At the same time, wholesaler-importers achieve relatively high profit margins which can be to the detriment of the consumer, who ends up paying a higher price.

27Therefore, the lack of competitive mechanisms at several stages of the supply chain, local taxation and local production which is uncompetitive and which does not sufficiently reap the benefits of tax exemptions to which it is entitled, are all factors that contribute to higher price levels than in mainland France. Local production can contribute to lower-priced supplies, but for certain categories of food products the import barriers (dock dues, cost of freight, middleman margins) and the concentration on local markets diminish this possible effect on prices.

28Products imported or produced in the DROM are taxed for the benefit of local authorities via dock dues. The “main” dock dues include an external component (taxes on imported products) and an internal component (on local products). The regional dock dues are also divided into an external rate and an internal rate. These taxes represent one third of the operating revenue for the communes and around 30 % for the overseas regions, except in La Réunion where they only contribute 17 %.

29The internal dock dues can play a protectionist role for local production if their rate is lower than the taxation of imported products (external dock dues). The modulation of the dock dues rates according to the nutritional quality of the products is one of the possible ways of promoting higher quality or local production.

30As far as the external dock dues are concerned, the rates vary according to the products (table 5, see IV). Across all of the DROM less processed products are taxed at a lower rate than processed products; the lowest rate is for meat; cereals and milled grain products are taxed more than meat, medium-level taxes are for animal oils and fats; fish and shellfish are highly taxed.

31However, despite these overall trends, the external dock dues rates vary. On average, the lowest dock dues are observed in La Réunion (6 %), the highest in French Guiana (14 %) and Mayotte (14 %) and medium rates in Guadeloupe (10 %) and Martinique (10 %).

2.1.3. Relative price structure

32In addition to higher average price effects than in mainland France, the DROM also have to contend with the issue of the relative prices of various food items compared to one another. For example, the price ratio between 1 kg of poultry and 1 kg of beef can be different depending on the DROM and between the DROM and mainland France. In this case the price difference can have an influence on consumer choices (besides the fact that average prices are higher in the DROM than in mainland France), leading to the question as to whether price differences of food products in relation to each other might drive consumers to over-consume certain goods and under-consume others. If so, are the choices biased in favour of or against the nutritional quality of diets?

33We have limited reliable information on the prices of various food products, and while studies in this regard are underway, for example in the ANR Nutwind project, the results are not yet available. Specific price comparison with regard to a drive in Martinique and a drive by the same distributor in mainland France shows that, for many product categories the prices in Martinique are about one and a half times higher than in mainland France. This ratio is the lowest for fruit, starchy foods, fish, red meat and, to a lesser extent, aperitif products, so in comparison to mainland France the difference in this case encourages the consumption of these products. Conversely, the ratio is the highest for soft, sugary drinks, salty snacks and fast food, poultry/offal/eggs, vegetables and, to a lesser extent, oleaginous products and ready-to-eat meals. In comparison to mainland France, this discourages the consumption of these products. It is therefore difficult to form a conclusion regarding the positive or negative effect of price ratios on the quality of the diet (compared to mainland France), since, for example, the price structure would encourage fruit consumption but discourage the consumption of vegetables.

2.2. Household consumption and socio-economic disparities

2.2.1. Food budget and expenditure per level of income

34With higher prices and lower income than in mainland France, overseas households devote a greater proportion of their budget to food. This proportion rises even as income declines, however overall food expenditure is still on average lower than in mainland France because the income levels are also lower (except in Martinique).

35The expenditure per food group is also different between the DROM and mainland France (Brasset, 2014a) (table 6, see IV): all DROM populations spend less money on meat, milk, cheese and eggs than in mainland France (except in Mayotte for meat). The inhabitants of Guadeloupe, Martinique and Mayotte buy proportionately much more fish and seafood than those of French Guiana, La Réunion and continental France. Compared to mainland France, the overseas populations spend less on fruit, but more on vegetables. The purchase of soft drinks appears double that of mainland France, while the purchase of alcoholic drinks is slightly lower. It appears very likely that consumers steer their purchasing towards lower-quality products.

36The lower the income, the higher the percentage of the budget allocated to food; therefore the percentage of the budget allocated to food decreases when the income increases. In the DROM, food is the largest item of expenditure, as it is in mainland France, for people on low incomes, except in Guadeloupe (2^nd largest after housing) (Brasset, 2014a; Jeanne-Rose and Clarence, 2015; Orsag, 2010c). In the overseas regions, as in mainland France, the budget constraint for food does not have the same weighting for households with different standards of living. By way of an example, in Mayotte for food consumption as a whole, and taking into account the size of the household, the 20 % most affluent households spend on average 213 euros per month and per consumption unit, i.e. 4.3 times more than the 20 % lowest-income households, who spend 49 euros per month on average. But these food expenses account for a greater share of their total budget; food represented 33 % of the lowest income households’ budget compared to 16 % for the most affluent. According to the literature, as an unhealthy diet (energy-dense, low in essential nutrients) is cheaper than a balanced diet, this would imply that social nutritional inequality may sometimes be due to budgetary constraints on food choices.

37A special feature of the DROM is the high proportion of self-consumption which reduces the differences in food consumption between the socio-economic categories. In fact, this practice is more common for households with the lowest income, enabling them to increase their food consumption by 54 % in Mayotte and 12 % in La Réunion, in particular for fruit (37 %) and eggs, milk and cheese (29 %). In Guadeloupe only the self-consumption of starchy foods is greater among individuals with a low SEP.

2.2.2. Socio-economic disparities in food consumption

38There are many causal pathways that explain social inequalities in health, especially with regard to chronic diseases (Inserm, 2014). Among the many factors which determine these diseases, food appears to be a major modifiable factor – for example, food consumption would explain up to 66 % of social inequality in terms of mortality (Méjean et al., 2013; Stringhini et al., 2010). Documenting the socio-economic disparities in food consumption is vital to understanding the underlying mechanisms behind the major social inequalities in health in the DROM and better identifying target populations for efforts to improve their nutritional status without increasing, indeed by reducing, socio-economic inequalities related to health. Within this context the relationship to the SEP will now be considered successively according to the various dietary profiles and then according to food groups.

39Overall the studies show a very clear association between the dietary profiles and the SEP, regardless of region. The “healthy” and “modern” profiles are associated in opposition to the level of education and income respectively, as is traditionally shown by the literature. However, the “traditional” profile is associated with a low SEP in Guadeloupe, Martinique, Mayotte and La Réunion, while an association with the SEP is not found to operate in mainland France in any systematic way (Bertin et al., 2016; Kesse-Guyot et al., 2009). This “traditional” profile is probably better characterised by the SEP indicators in these territories than in mainland France because the culture-specific dishes and foods which make up the DROM diets are far more identifiable and commonly consumed by the populations. Other common features were identified between these DROM regarding “in transition” profiles, in line with profound changes experienced over the last few decades from a diet based on starchy foods, pulses and roots to one with foods that are rich in fat and simple carbohydrates (especially sugary products and drinks). This profile is more prevalent in the intermediate socio-economic categories and conforms to the results of positive associations between the consumption of sugary products, red meat (in La Réunion only) and cured meats and income, or the inverse relationships between the consumption of pulses and oil and income which, for their part, are not wholly in line with expectations compared to results described in the literature. All of this shows that nutritional transition, although advanced, has not yet been completed and may result in a future continuation of rising prevalence in obesity, diabetes and high blood pressure. Furthermore, by exploring the differences in results between the early 2000s and 2014 in the Antilles one may wonder whether the socio-economic disparities in food consumption which have appeared more recently there might be linked to an on-going nutritional transition or to the economic crisis of the late 2000s instead. In the latter regard the financial pressure on households and rising food prices may have caused the least advantaged consumers to replace their more expensive purchases of fruit and vegetables with inexpensive starchy foods. It is for this reason that it is necessary regularly to update data regarding the consumption variations according to the SEP so as to analyse the possible effects of economic crises on the socio-economic disparities in diet.

40Regional variations will be considered according to food category. Although there are wide variations for both fruit and vegetables according to the SEP in mainland France (Inserm, 2014), the SEP has a greater effect on fruit consumption than vegetable consumption in La Réunion, Mayotte, French Guiana and Martinique, whereas in Guadeloupe it is the consumption of vegetables that is more socially distinguished.

41Rice and pulses are consumed more by low-income households in Mayotte and La Réunion. Among the higher socio-economic classes these products are more often replaced by refined cereal products (pasta, bread, etc.), unlike in mainland France (Brasset et al., 2015; Brasset, 2014b). In the Antilles the consumption of starchy foods and pulses is inversely correlated to the level of education and professional category in Martinique and Guadeloupe respectively, and to income for both territories. In French Guiana, the percentage of the budget devoted to rice and pulses decreases with income (Brasset, 2014a).

42No variation according to the SEP could be measured regarding meat, poultry, fish and eggs in Guadeloupe. However, in the Antilles fresh fish is replaced by frozen or tinned fish among the higher income categories while fresh fish and canned fish are consumed less in high-income households in Mayotte and fish consumption, regardless of type, does not vary according to income in La Réunion. Unlike the Indian Ocean, where fresh fish remains a part of the traditional diet, it is likely that risks associated with chlordecone contamination in the Antilles motivate the most affluent classes to choose imported processed fish, which is more expensive but seen as less risky than fish caught locally. With regard to meat, in Martinique the consumption of poultry is decreasing, whereas that of cured meats increases with the level of education (Kannari). In French Guiana the percentage of the budget spent on meat decreases when income increases, whereas the percentage for seafood products increases (Brasset, 2014a). In Mayotte, similarly to the consumption of fish and seafood, poultry consumption decreases as income increases and it is replaced by red meat and cured meats instead. In La Réunion the consumption of meat, cured meats, seafood and eggs generally appears to be less socially distinguished, or else possesses a social status contrary to that in mainland France. Poultry is actually eaten a lot less when the SEP decreases in mainland France, while red meat, cured meats and fresh fish are consumed a lot more among disadvantaged groups (Brasset, 2014b). Conversely, in La Réunion consumption of meat, including red meat and poultry, is higher among low-income households than among wealthy households, probably because all meat (red meat, poultry and pork) are more valued in La Réunion and Mayotte, especially in the low-income categories.

43Among dairy products, cheese and yoghurt are consumed more when income rises in the five DROM, whereas there is little variation for milk. In French Guiana, Mayotte and La Réunion consumption of dairy products increases with the SEP, whereas the relationship is either reversed or insignificant in the Antilles. It is likely that, as the population of the Antilles is particularly aged in relation to the other territories, it maintains a more traditional diet with low consumption of dairy products.

44Unlike mainland France, sugary products (confectionery, jam, chocolate products, ice cream, etc.) appear to be consumed more in Guadeloupe, Mayotte and La Réunion by people with a high SEP (Inserm, 2014; Méjean et al., 2011), as are alcoholic drinks in Guadeloupe. The proportion of the budget for sugary drinks also decreases with income, except in La Réunion, and more markedly so than in mainland France – especially in Mayotte.

45With regard to added fats, especially oils, their consumption decreases significantly as income rises, whereas oil is favoured by people with high incomes in mainland France.

46To conclude, the overall quality of the diet (compliance score and healthy profiles) of adults is positively associated with their socio-economic position, confirming the results described in both international and national literature (Inserm, 2014). On the other hand, the diet of adults in a disadvantaged socio-economic position is notably characterised by a lower intake of healthy foods, such as fruit and vegetables, wholegrain cereal products, fish and yoghurt. However, studies concluding an association between the SEP and individual consumption of unhealthy food groups in adults remain too few in each of the DROM to allow for identification of social groups which might be more at risk than others. As a result, the findings related to these food groups require further study. Similarly, although the link between the socio-economic environment created by parents and their children’s health-related behaviour is well established in international literature, the results of the studies available in the DROM also remain too inconclusive for proper utilisation, probably due to differing levels of consumption within this population, but also due to the small sample size in the Kannari study.

47Nevertheless, in studies which rely on several indicators in the Antilles, slight differences can be distinguished: although income and education are positively associated with a healthy diet, it is not automatically the same regarding other indicators, especially employment status and socio-professional category. Studies in other territories are limited to the use of a single indicator, income, to assess socio-economic position, while the different indicators mentioned reflect different social and dietary concepts and could even have cumulative effects. The income level reflects the financial affordability of a high-quality diet. People with more financial resources can therefore purchase more expensive food, such as fruit and vegetables, the nutritional quality of which is superior (Galobardes et al., 2001). Furthermore, the level of education affects an individual’s ability to understand and incorporate nutritional information in a critical manner and then establish healthy eating habits. This effect of education independent of income shows that individuals with a high level of education can develop strategies to overcome the barrier represented by the price of certain foods.

2.2.3. Food expenditure and diet quality

48To establish a relationship between the level of food-related expenditure and diet quality, we have used the data from the Kannari survey and a specific project comparing prices in a drive in Martinique and in a drive, by the same distributor, in mainland France, which allows for the determination of the cost of specific diets. Two approaches were followed: the first, expanded upon in the “Consumption typology” paragraph (Part II, 1.2) consisted of classifying the various diets observed in Martinique into four typologies and determining the minimum costs (established from the lowest prices observed for each type of food) and average costs (based on average prices observed for each type of food) of each one. The second approach consisted of classifying the diets observed according to four categories depending on their nutritional quality (score for compliance with the PNNS recommendations) and determining the minimum and average costs of the diets for each one. The cost of each diet22 is calculated using the quantities consumed for each of the average diets of the four typologies and the four categories, as well as the minimum or average prices observed in Martinique or mainland France for a total of more than 1,000 food products. These costs, therefore, correspond to the food expenditure of the average consumer for each typology or each category if they were buying the food at a minimum or average price in Martinique or mainland France.

49The following table (table 7) shows the minimum cost thus estimated for each typology and category, providing an indication of the accessibility of each of the diets in question.

50Typology 1: “healthy” typology, characterised by a high consumption of fruit, vegetables, wholegrain products and yoghurt and a low consumption of sugary drinks, alcoholic drinks, snacks and fast food, resulting in a high score for compliance with the recommendations.

51Typology 2: “in transition” typology with a high consumption of cured meats, red meat, offal, sauces and alcoholic- and sugary drinks, but also a high consumption of local tubers and pulses and a low consumption of biscuits and cakes, resulting in a medium score for compliance with the recommendations. Typology 3: “traditional” typology, with a high consumption of fruit, vegetables, local tubers, rice, fish and seafood, a low consumption of red meat, cured meats, alcoholic drinks, snacks and fast food products and of which the majority of the dishes consumed are specialities of the Antilles, with a high score for compliance with the recommendations.

52Typology 4: “modern” typology with a high consumption of sugary drinks, snacks and fast food products, biscuits, cakes and poultry, and a low consumption of fruit, vegetables, fish and seafood products, resulting in a low score for compliance with the recommendations.

53As has already been pointed out, the costs of the Martinique diets are substantially higher (around 40 %) in Martinique than in mainland France, but otherwise there is no clear link between the minimum cost and nutritional quality of the diets under consideration. When we look at the typologies, the best diet from a nutritional point of view ( “healthy” typology) and the worst diet ( “modern” typology) have a similar minimum cost (especially when they are evaluated using the current prices in Martinique). They thus show the same level of accessibility, with their cost being much lower than the cost of the other two diet typologies.

54The price structure in Martinique (compared with that of mainland France) tends to reduce the difference in cost between the Typology 1 diet ( “healthy”) and the Typology 4 diet ( “modern”). In other words, with the current price structure in Martinique it is possible to have a diet of good nutritional quality (Typology 1, “healthy”) at a minimum cost similar to that of a diet of poor nutritional quality (Typology 4, “modern”).

55Comparing the costs of diets according to their nutritional quality also shows that the minimum cost of the diets of worse nutritional quality (category 1) is higher than that of better-quality diets. Overall, the minimum cost decreases with the nutritional quality.

56These results suggest that in, the case of Martinique at least, eating properly or badly costs substantially more than in mainland France, but that the current price system does not necessarily make diets of good nutritional quality, such as those observed within the population, less accessible than diets of poor nutritional quality. This indicates that the adoption of diets of poor nutritional quality is most probably not primarily the result of the price system but that the quality of the diet is influenced more by eating habits, and therefore consumer preferences. All the same, it should be noted that these price comparisons relate to a single drive and cannot be considered representative of the general situation in either Martinique or mainland France – these results remain to be confirmed through further studies based on more comprehensive price comparisons to be extended to all of the DROM.