Energy and nutrient composition data for frequently eaten foods in Mexico (n = 151) came from the 2012 round of national health and nutrition survey. The list included corn and flour tortillas; grains; meats and fish; snacks (including sweets and nuts); milk and dairy; beans and eggs; fruits and vegetables; sugars, fats and oils; and sugar sweetened beverages (16).

The Nutrient Rich Food Index has two components: NR9 and LIM3. The positive NR9 component is the sum of percent daily values for nine nutrients to encourage ( N u t _ i n c i ): protein, fiber, vitamin C, vitamin A, calcium, iron, potassium, magnesium, and vitamin D. The negative component (LIM₃) is based on the sum of percent daily values for saturated fat, added sugar, and sodium (14). Percent daily values are capped at 100%. The NRF_9.3 score is calculated as follows:

Nutrient standards were based on the Mexico national institute of public health (INSP) and the Codex Alimentarius (18). The reference daily values (DV_i) were protein (50 g), fiber (25 g), vitamin A (800 RAE), vitamin C (90 mg), vitamin E (9 mg), Ca (1,000 mg), Fe (18 mg), K (3,500 mg), and Mg (310 mg). The maximum recommended values (MRV_j) were saturated fat (20 g), added sugar (50 g) and sodium (2,000 mg). The NRF_9.3 index scores, split into tertiles of nutrient density, were then used to calculate the nutrient rich food price index (NRFPI).

Average prices (per kg or L) for the 151 foods, coinciding with the ENSANUT survey period and denominated in Mexican Pesos (MXN), were obtained from the National Institute of Geography and Statistics (INEGI). Data collection procedures have been published (19). Food prices were corrected for preparation and waste and were expressed in MXN/100 g and MXN/100 kcal, edible portion. Monthly inflation rates for each food, also available from INEGI, were used to establish prices between June 2012 and March 2022. The annualized monthly inflation rates ( π t ) were calculated using average food prices ( p t ) in a given month as follows π t = ( p t p t − 12 ) − 1 , where p t − 12 is the food price twelve months before. Affordability was the ratio of NRF_9.3 score points and cost (MXN/100 kcal):

The present adaptation of the Laspeyres price index (1, 2) used market baskets based on nutrient density tertiles instead of commodities. The NRFPI_9.3,t price index score was computed as the average of m food price indexes weighted by the respective food NRF_9.3 score, which is held fixed over time. Using the NRF_9.3 as the main nutrient density indicator and metric of scale (15), the NRFPI_9.3,t attached projected food prices (per 100 g or 100 kcal) to individual NRF_9.3 scores starting in June 2011. The estimation of average food prices for the whole sample period employed the monthly growth of individual price indexes to project every price at time t. The present version or the NRFPI was related to the NRF9.3 nutrient density score and can be called NRFPI_9.3,t. The NRF9.3 score was the time-invariant weight attached to individual food prices as follows:

where P i , t are the average prices of foods i = 1 , ... , m in time t . The nutrient scores for each food i , e.g., NRF_9.3,1,…, NRF_9.3,m, described above, serve as time-invariant weights for each price index. The NRFPI_9.3,t was constructed using variable prices and fixed nutrient scores as weights. This NRFPI_9.3,t keeps nutrient density constant while allowing prices to vary over time. The base month t=0, where prices and nutrient scores are fixed, was May 2012, the end month of ENSANUT 2012 data collection.

Descriptive analyses examined the association between nutrient density and nutrient cost. First, we examined energy density, food prices and affordability by NRF _9.3 tertiles. Second, the relation between NRF_9.3 scores, energy density, and prices per 100 g and 100 kcal was examined using scatter plots. Next, the time series properties of food prices were analyzed to assess dynamics both on MXN/100 kcal and MXN/100 g over the sample period. Mean food prices in May 2012 constituted the base period. A battery of tests assessed linear trends and seasonality using dummy variables, t-tests, analysis of variance and F-tests. The association between price levels and inflation with the degree of nutrient quality is tested using Wald tests.

This study used intercept dummy variables to test for linear trends and seasonality of NRFPI_9.3,t monthly nutrient costs and annualized monthly rates of inflation. The model adds time trends that were jointly tested employing individual t-tests and Wald tests in the parameter estimates of the following regression:

Where N R F P I 9.3 , t are the nutrient price indexes of each tertile group per 100 kcal (MXN$ /100 kcal) or 100 g (MXN$ / 100 g), respectively; D m are the m=[1,…,11] seasonal dummy variables with one for every month and 0 otherwise. The linear time-trend variable is t with a gradient parameter (θ). The average shift impact of each seasonal month to the conditional mean is captured individually by β m , while θ is the average impact of each passing month on N R F P I 9.3 , t scores. Linear OLS estimations methods assume that u t ~ i . i . N . ( 0 , σ 2 ) . As in previous studies (20), these tests are based on robust standard errors to account for potential violations of residuals' independence.

Table 1 shows energy density (kcal/100 g), NRF_9.3 values, food prices (per 100 kcal and 100 g) and affordability metrics by NRF9.3 tertiles. Foods in the bottom NRF9.3 tertile were characterized by the highest energy density (287.41 kcal/100g), lowest cost (MXN$ 3.10/100 kcal) and low nutrient density to price ratio (−5.52). Foods in the top tertile of NRF 9.3 scores had lower energy density and were more expensive (MXN$ 5.40/100 kcal) but had a better nutrient density to price ratio (27.21) A similar pattern is observed with nutrient affordability measured per 100 g.

Figure 1A shows the time series behavior of the nutrient rich food price index (NRFPI_9.3,t) between June 2011 and March 2022. For foods in the top tertile of NRF_9.3 scores, the NRFPI index rose from 100 base period points to 206.32 score points in March 2022 (an annualized average monthly increase of 7.13%). For foods in the middle tertile of NRF scores, the NRFPI rose from 100 points to 183.07 points (a monthly increase of 6.21%); while for the foods in the bottom tertile of the NRFPI rose from 100 to 164.56 points (a monthly increase of 5.21%). Figure 1A further shows that foods in the top tertile of NRF_9.3 scores not only cost more but also showed more price spikes and more price volatility than foods in the bottom tertile. Foods in the top tertile of nutrient density scores included fresh produce.

While the price gap (% difference) between NRFPI price indexes was set to zero in May 2012, by March 2022 the cost gap grew to 12.70% compared to mid NRF9.3 tertiles and by 25.38% compared to low NRF9.3 tertiles. Figure 1B shows annualized monthly inflation rates by tertiles of NRF_9.3 scores. Inflation rates were higher for the more nutrient-rich foods as compared to foods in the bottom NRF tertile. What is more, price increments for the most nutrient rich foods were highly erratic. Figure 1B also shows a significant increase in costs for all foods starting in 2021, likely associated with the global increases in food prices from mid-2021. Figure 2 shows NRFPI average price levels and annualized monthly inflation rates by tertiles of NRF_9.3 nutrient density.

Table 2 shows mean difference in NRFPI, inflation rates and seasonality trends by tertiles of NRF_9.3. Higher-scoring foods were associated with higher inflation rates. The trend gradient (θ) also rose from 0.58 for foods in the bottom tertile of NRF_9.3 to 0.64 for the middle tertile and 0.84 for the top tertile. The historical monthly price increases were higher for the more nutrient rich foods. Finally, seasonality tests based on dummy variables for NRFPI levels showed differential increases in monthly costs by NRF_9.3 tertiles (see Figure 3). The analysis employing annualized monthly inflation rates also confirms the strong presence of seasonality effects. While the source of seasonality effects still needs to be determined, the monthly price and inflation variations may be related to weather effects, harvesting seasons, and local and international trade patterns (20). The pattern of prices and inflation rates for low and middle NRF9.3 tertiles reveals a more stable seasonal variation of nutrient prices compared to the irregular seasonal behavior of prices observed in highly nutritious foods. The relatively high spikes of seasonality in the prices of highly nutrient foods in particular months suggest potential trade anomalies and may have an impact on diet quality.

Table 3 assesses the behavior of NRFPI scores and annualized inflation rates of foods with different nutritional values during the COVID-19 pandemic. The most significant variation of prices and inflation was observed in the last year of the sample, from March 2021 to 2022, when foods with the highest nutrient quality changed from an average score of 176.16 score points in 2021 to 207.50 points in 2022, an annual growth of 17.79%. The prices of foods in the low and mid-levels of nutrition experienced a lower rise: 6.69% and 12.96%, respectively, during the same period. The most nutrient rich foods did not only cost more per 100 kcal but also experienced a higher inflation growth at a time of health, social and economic crisis.