In this study, we evaluate the hypothetical impact of the FOPWL regulation, and we observed a reduction of energy, saturated fat, trans-fat, added sugars and sodium intake, after a simulated replacement of commonly consumed foods in the Mexican diet, with products that would have less WL according to the FOPWL criteria in all the population and by socioeconomics characteristics across the three phases of the norm. The most important reduction was observed for added sugars intake. To our knowledge, this is the first study in Mexico to evaluate through food and beverage replacement the potential impact of the FOPWL norm under the assumption that consumers would replace consumed products with those with a fewer WL or with non-processed food.

The results in these analyses are consistent with similar studies. In a previous study also in Mexico [32], processed food was replaced with food that complied with the Mexican Committee of Nutrition Experts (based on the dietary intake recommendations from the World Health Organization) and Federal Commission for the Protection against Sanitary Risks Nutrition Criteria (based on the cutoff points established by the food and beverage industry) and then compared with the baseline consumption [32]. Similar to our study, a reduction was seen for energy (− 5.4%), saturated fat (− 18.9%), trans-fat (− 20%), total sugar (− 36.8%) and sodium (− 10.7%) intake [32]. The criteria for evaluating critical nutrients were different from the FOPWL used for this analysis, hence the differences in reductions. Besides, an increase in fiber intake was shown; however, we did not evaluate fiber and micronutrients intake since our focus was in the assessment on the critical nutrients that FOWPL regulation involved. However, it would be likely that in our replacement simulation, fiber would also increase since we increased the consumption of foods rich in dietary fiber during the replacement.

In this study, the greatest reduction was observed for added sugars in all the population and by socioeconomic characteristics. A possible explanation is the high intake of sugar-sweetened beverages in the Mexican population. In 2012, added sugars consumption represented 12.5% of total energy and 70% of them came from sugar-sweetened beverages [33]. Due to the high intake, Mexico is the first in the ranking of mortality and morbidity attributable to these beverages [34, 35]. Further, in a recent national survey ENSANUT 2018, adults (85.8%), adolescents (85.7%), school-aged children (85.7%), preschool-aged children (83.3%) reported an important consumption of these beverages [36], despite the efforts to discourage their consumption in the last years.

The impact of Chile’s Law of Food Labeling and Advertising which includes the FOPWL was evaluated on household beverage purchases [16]. Using fixed-effects models they compared the observed volume and calorie sugar content in beverage purchases after the regulation, with a counterfactual based on pre-regulation trends. The volume of high-in beverage purchases decreased by 22.8 ml/capita/day (95% CI − 22.9 to − 22.7; p < 0.001) or 27% after the regulation was implemented. Nonetheless, it is not certain whether the observed changes were due to reformulation or a real change in consumer behavior.

In the Netherlands, in young adults (aged 19–30 years), a substitution of current food intake with those that complied with the Choices Program (an international nutrient profiling system to determine whether foods are eligible to carry a healthier option label) was conducted [37]. After the replacement, they observed that median energy intake was reduced by 16% as well as for nutrients with a maximal intake limit (from 23 for sodium to 62% for trans-fat). Contrary to our study, they showed a greater reduction in trans-fat intake in relation to other nutrients. This may be explained because of the differences in dietary patterns or higher content of this nutrient in Dutch-processed food. Besides, in this study, the trans-fat intake could underestimate since the labeling regulation in force in 2016 presented some deficiencies that allowed to declare 0 g of trans-fats when they were below 0.5 g per portion.

Furthermore, during the third scenario, the cutoff points for trans-fat and saturated fat did not change; if the cutoff points for them would be stricter in the last phase of the norm, there would be more products to replace hence we could observe a greater reduction in these nutrients. Moreover, the trans-fats intake might be already conditioned by other efforts such as the Action Plan to eliminate industrially produced trans fatty acids 2020–2025 and the Replace initiative from the World Health Organization (WHO), where the elimination of trans-fats in industrial food processes is proposed [38]. However, we expect that the reductions shown for all the nutrients in this study will be more important since the food industry will have the opportunity to modify the nutritional content of their products according to the cutoff points of the norm. Thus, even if consumers did not change their dietary choices, in expectation they would consume less amount of energy, saturated fat, trans-fat, added sugars and sodium from the newly reformulated products.

The Mexican Government has implemented different strategies to enhance the dietary quality of the population, such as taxes on junk food and sugary beverages, banning the sale of processed foods and sugary beverages in schools, establishing restrictions for advertising directed to children, implementing physical activity programs and bike lanes in the community, and making public water fountains more available [39,40,41,42,43]. Unlike previous measures, the FOPWL regulation has the advantage that can be monitored and enforced and have the possibility to change the cultural norms that regulate consumer demand, thus food supply over time [18].

Our assumption regarding product substitution is in line with existing empirical results in Chile. Specifically, evidence from Chile shows that WL mainly encourage product substitution within the same food category [44]. Moreover, some studies provide evidence of the underlying mechanisms behind the purchase adjustments after the WL implementation in Chile, accounting for the simultaneous response to this policy by consumers and producers. For example, Barahona et al., in the case of breakfast cereals in Chile, showed product substitution between labeled and unlabeled cereals due to the interplay between updates in healthfulness beliefs by consumers because of the information conveyed by WL and price adjustments by producers in response to the cost linked to product reformulation [45]. Meanwhile, Pachali et al. found, likewise for breakfast cereals in Chile, that reductions in the nutrients of concern resulted from consumers substituting labeled for unlabeled cereals [46]. Moreover, Pachali et al. showed this substitution was reinforced by a price decrease across unlabeled cereals [46]. The authors pointed out that this price decrease was a producers’ response to the new composition of consumers demanding unlabeled cereals that were more price-sensitive than previous consumers before the WL implementation [46]. However, it is worth noting that the expected WL effect is likely to be food group specific in light of the findings by Araya et al., who showed no clear change in purchases of chocolates and cookies in Chile [47]. Evidence in Mexico in the context of the sugar-sweetened-beverage tax showed that sugar content is a demand driver, aside from prices and products’ brand, so consumers are likely to decrease their demand when producers reformulate their products in case this reformulation leads to a sweetness reduction (i.e., sugar reduction with no substitution for non-nutritive sweeteners) [48]. Thus, if WL encouraged reformulation in Mexico, consumers would likely adjust their demand downwards, at least for sugar-sweetened beverages. Finally, aside from reformulation, producers might respond to WL by removing/launching products. However, evidence from South Africa shows that this update in producers’ product portfolios has a minor contribution to the overall purchase changes linked to sugar-density taxes on sugar-sweetened beverages, which encourage reformulation as WL does [49]. In light of all these results, we consider our product substitution assumption consistent with existing evidence. However, we acknowledge our estimate could be conservative by overlooking producers' response to WL (i.e., reformulation and price adjustments) and how consumers react to this response. When the FOPWL regulation would be finished, there will be more food products options, thus more variability to select food products for the replacement; unfortunately, this was not possible due to the scope of this work and the phase of the FOPWL regulation.

Previous studies have proved that modeling is useful to accurately estimate the nutrient intakes when processed food is replaced using healthier alternatives [26,27,28]. These studies are frequently used to anticipate dietary changes before they are implemented in populations like in Mexico. Thus, the data could be turned into nutrition policies and help decision-makers to design policies based in evidence that would help to decrease diet and non-communicable diseases.

Healthy foods are often unaffordable or unavailable, the increase in urbanization causes population to change their dietary habits, to move away from producing and cooking their own food to prepared foods (high in sugar, salt and fat) [50]. Besides, marketing influences food choices, and unhealthy foods are an important target, especially among children [51]. Therefore, legislation and policies should focus on regulating prices, production, processing and distribution to make healthy foods more accessible as well as to restrict marketing to reduce the promotion of unhealthy food. FOPWL regulation is an important strategy to help population to identify unhealthy food in a simple way [52], along with public health interventions that can emphasize the importance of balanced diets on maintaining good health. The community health programs can educate children, parents, teachers and community leaders about the impact of their nutrition choices on health [53, 54].

Furthermore, through this simulation, we can estimate the potential impact of the FOPWL explained uniquely by change in consumer’s behavior. In Mexico, when the implementation of the regulation is completed, it will be difficult to determine if intake changes are because of consumer’s behavior, product reformulation, changes in marketing, other policies or the interaction of all of them.

This study presented some limitations. First, it is uncertain if the Mexican population will replace products with less WL or with non-processed products. This is possible since the population choices and intakes are determined by an array of factors, and not only on the presence or absence of WL [55]. Moreover, ENSANUT 2016 is a cross-sectional observational dataset; for this analysis, we used self-reported intake data which may be affected by measurement error. Further research is necessary to know if this error is randomly distributed through the different subgroups or whether it affects systematically certain subpopulations. However, in the Mexican population, this database is the most comprehensive nationally representative data to study dietary intake. Another limitation may be the availability of products presented in 2017 from the INFORMAS dataset [31]. Further, despite the results observed, we were confronted an important limitation to replace sugar-sweetened beverages, which was the low variability of the number of warning labels in this group of drinks between 0 and 2 WL. During the second and third scenarios, the replacement was using the same drink; thus, the reduction would be greater in the third scenario if more options were available. Finally, our results will need to be corroborated by future observational studies when the implementation of the regulation will be finished.