Bio-based and Applied Economics 7(3): 249-263, 2018 ISSN 2280-6180 (print) © Firenze University Press ISSN 2280-6172 (online) www.fupress.com/bae Full Research Article DOI: 10.13128/bae-7678 Can menu labeling affect away-from-home-dietary choices? ElEna CastEllari1,*, stéphan MarEttE2, DaniElE Moro3, paolo sCkokai1 1 Dipartimento di economia agroalimentare, Università Cattolica del Sacro Cuore, Piacenza, Italy 2 UMR Économie Publique, INRA, Université Paris-Saclay, F-78850 Grignon, France 3 Dipartimento di scienze economiche e sociali - DISES, Università Cattolica del Sacro Cuore, Piacenza, Italy Date of submission: 2018 28th, September; accepted 2019 23rd, July Abstract. This study aims to evaluate the impact of two menu-labeling formats on changes in dietary choices in an away-from-home meal, specifically in a university cafeteria. A field experiment at a university cafeteria in Italy was conducted providing two different types of nutritional labels. The experiment lasted four days, spread over two weeks during which a total of 930 observations were collected. During each day of the experiment, only in one food line (treated line) a label indicating the healthy options was displayed, while in the other line no label was presented (control line). The paper describes two indexes to measure how the selected food choices for each participant are in line with what suggested by the labels. We define five different class- es of these indexes and we test our hypothesis using an ordered logit model. Results show the labels we provided had no significant impact on changing the tray compo- sition, in accordance with other previous experiments suggesting that adding only nutritional information in a restaurant setting does not necessarily encourage health- ier choices. The paper concludes highlighting the need of a multifaceted approach to design effective public policies enhancing healthier choices in a self-service restau- rant. Specifically, the provision of nutritional information by itself can have zero or low impact unless it synergizes with others instruments such as nutritional education, social norm provision and nudges. In the conclusions, some suggestions on public policies addressing the promotion of healthy food habits are given. Keywords. Menu Labels, Food away-from-home, Healthy food policies, Food labe- ling. JEL Codes. I12, I18, D12. 1. Introduction Food away from home (FAFH) consumption plays an increasing role in the daily diets of many people worldwide. In Italy, the share of FAFH on total food expenditure was 33% in 2015, versus 46% in Spain, 44% in the United Kingdom, 27% in Germany, and 26% in *Corresponding author: elena.castellari@unicatt.it 250 Elena Castellari et alii France (Agrifood Monitor, 2016). In 2015, about 12 million Italians (around 20% of the entire population) had lunch away from home 3-4 times a week (Fipe-Commercio, 2015). These patterns are actually similar for all industrialized and many developing countries (Mottaleb et al., 2017). In the USA, the share of FAFH on total yearly food expenditure rose from 25.9% in 1970 to 43.1% in 2012 (USDA). In the 2007-2008 National Health and Nutrition Survey data 41% of adults said they had consumed foods and/or beverages from fast food-type restaurants during the previous 24 hours, and 27% of them from full-ser- vice restaurants (Seguin et al., 2016). While the rising in FAFH consumption is not a bad habit per se, researchers have found that the frequency of eating FAFH is positively correlated with some unhealthful outcomes, such as overweight and obesity (Binkley et al., 2000; McCrory et al., 1999; Satia et al., 2004, Todd et al., 2010). The link between FAFH and obesity can be explained because people tend to under- estimate calories and fat content when they select their meal in an away-from-home envi- ronment (Backstrand et al., 1997). Indeed, restaurants and cafeterias typically use caloric dense ingredients (butter or dressings) to gain palatability1; yet, it is almost impossible for consumers to detect those “hidden” fats and overall taste remains a major force driving food choices (Glanz et al., 1998). The positive relationship between FAFH expenditure and BMI has also been found in children. According to a study by Bowman et al. (2004), on a typical day when eating at quick-service food, children (aged 4-19) tended to consume more fat (+ 9 g), added sug- ars (+ 26 g), sugar-sweetened beverages (+ 228 g), and less fiber (-1.1 g), milk (-65 g) and fruits and non-starchy vegetables (-45 g), compared to those who did not, leading to 187 extra calories compared to a meal consumed at home. Given the increasing trend in eating away from home, policy makers have consid- ered the urgency of finding policy instruments which can lead to healthier consumption behavior. Labeling2 is among the information-based instruments extensively used to lead consumers towards more informed and possibly healthier choices (Galizzi, 2014; Traill, 2012). We can think that in a FAFH environment providing some nutritional informa- tion may limit the misperception on nutrients’ content when consumers are choosing their meal. However, while the introduction of nutritional information in a restaurant menu is supported by many researchers and health officials, their provision is mainly due to private sector or local government initiatives (Brambilla-Marcias et al., 2011). In fact, the implementation of a mandatory policy in a catering environment would require the 1 Elaborating data from household food consumption surveys conducted by the U.S. Department of Agriculture (USDA) during the period 1977-2008, Biing-Hwan et al. (1999; 2012) have shown a reduction of the share of saturated fat to the overall caloric intake of Americans. However, from their analysis, FAFH is still richer in satu- rated fat than food at home: in 2005-2008, fat contributed to 30.5% and 37.2% of the caloric intake from food at home and from FAFH, respectively (Biing-Hwan et al., 2012; Kozup et al., 2003). Moreover, the FAFH has been found higher in saturated fat, sodium and cholesterol and resulted in lower calcium content and dietary fiber than food at home (BiinHwan et al. 2012). Todd et al. (2010) estimated that in the USA each meal consumed away from home results in 134 additional calories. 2 In the United States, under provisions of the Affordable Care Act of 2010, restaurant chains with twenty or more locations operating under the same brand are required to provide detailed nutritional information to consumers and to display calories on their menus. In the European Union (EU), with Regulation no. 1169/2011, new rules regarding nutritional information for food, both pre-packed and non-pre-packed, have been introduced. However, this regulation does not impose stringent rules for restaurants, unless differently required by each member state. 251Can menu labeling affect away-from-home-dietary choices? capacity of standardizing ingredients and portions, which is not a trivial task especially for smaller size and not-chained restaurants. As a consequence, requiring stringent adop- tions of nutritional labels in a catering environment can have the side effect of pushing smaller, non-chain business out of market, and can reduce options for consumers (Maz- zocchi et al., 2009). The aim of this study is to analyze if the provision of some nutritional information in a university cafeteria has an effect on the composition of the meal chosen. Specifically, it can be expected that, by providing some nutritional information using labels, consumers might be facilitated to reduce the bias of “hidden calories” and consequently to identify healthier options. At this end, the authors conducted a field experiment in a university cafeteria in Italy, where two types of informative labels were alternatively provided. This article proceeds as follows: after a literature background, first the experimental design and the indicators used to evaluate the quality of the meal are described; then the model and the empirical results are presented, followed by some discussion and policy implications. 2. Background Previous literature showed the provision of nutritional information can lead to mixed findings. In a systematic review, Mazzocchi and Trail (2005) evaluated the effect of food label in portion size consumption and they found varying impacts, from increasing, to decreas- ing or no effect. However, none of the studies examined found an effect on reducing energy- dense foods (Mazzocchi and Trail, 2005). Similarly, a literature review by Swartz et al. (2011) and another by Kiszko et al. (2014) have shown the provision of caloric labels had none effect on the caloric intake of the food ordered and consumed. Further, Harnack and French (2008) concluded that, even if some studies support the evidence of a relation between the provision of caloric labeling and food choices, these effects are weak or inconsistent. Similarly, empirical studies have shown mixed results. Some have found the provision of nutritional information in a restaurant menu helps reducing the caloric intake (Rob- erto et al., 2010, Wisdom et al., 2010), others have measured no significant effect (Elbel et al., 2009, Finkelstein et al., 2011). Ellison et al. (2014) showed that numeric labels alone (i.e. labels where nutrients content was shown as grams or mg per 100 grams of prod- ucts or as percentage) have no influence on food choices, unless reinforced by traffic light symbols. In fact, traffic light labels (i.e. labels where some nutrient contents are classified with colors red, orange or green based on some thresholds with respect to dietary recom- mendations) may lead restaurant patrons to introduce in the menu lower-calorie options. Marette et al. (2019) showed that the appearance of traffic light labels significantly impacts the Willingness to Pay of products offered in the experiment. On the other side, an experimental study conducted by Seward et al. (2016), where traffic labels where provided in a university cafeteria setting, has shown that, while stu- dents reported to use the traffic light regularly and support their use, the intervention had no effect in improving dietary quality. Vasiljevic et al. (2015) have shown that, on select- ing different snacks, emotion labels (such as smiling faces) yields stronger effect on the perception of the healthfulness of the snack than colored label; and overall frowning labels are more effective than smiling ones. 252 Elena Castellari et alii Using a random control trial, Oliveira et al. (2018) find the provision of a menu labe- ling displaying different food information being positively associated with healthy food choices. Elbel et al. (2009) find that the provision of caloric labels on fast food menus in New York had no effect on the caloric content of the purchased meal. In general, the literature has found paternalistic interventions (nudges), eventu- ally combined with information provision, being more effective in producing behavioral changes (Downs et al., 2009; Thapa and Lyford, 2014; Thunström and Nordtröm, 2013; Castellari and Berning, 2016). Other studies have shown the importance of providing social descriptive norms to encourage change in food choices (Burger et al. 2010). This work evaluates the effect of nutritional labels’ provision on the menu items selec- tions, rather than the caloric content of the meal choices. Nutritional information may have little effect on the caloric content of the overall meal but might impact its compo- sition inducing a shift from ‘worse’ to ‘better’ choices3. Other studies have found only a small portion of consumers (between 16% and 29%) have responded to nutritional labels changing their menu selections (Balfour et al. 1996, Yamamoto et al. 2005). We evaluate two different label intervention: (1) a label where the green color is matched with a posi- tive emotion (smile) to identify the item within the same food group (first, second, side dish, fruit and dessert) that has the lowest caloric intake among the available options; (2) a label which ranks within the same food group (first, second, side dish, fruit-dessert) the options available based on their caloric composition using a medal (gold, silver and bronze). 3. Methods 3.1 Experimental design The hypothesis behind the experiment is that displaying some nutritional labels (i.e. indicating either a partial or a complete ranking of dishes in terms of their caloric con- tent) in a self-service restaurant may influence consumer when selecting food options. We expect the presence of the label would help consumers to identify the hidden calories and thus the less caloric options. At this end, we collected data at a university cafeteria located in Piacenza, Italy; the experiment lasted four days, spread over two weeks (with a four-week break between them) between March and April 2016. The cafeteria is a self-service caterer, presenting two lines, each one providing identical food choices. The cafeteria meal has a fixed price and it allows to select one option within each menu category: first dish; second dish; side dish; fruit-dessert.. To test our hypothesis we provided (in separate settings) two different types of labels: 1) Less Caloric Labels (LCL): within each menu category (first dish; second dish; side dish; fruit-dessert.) the label indicates the option with the lowest level of calories4 per portion (Fig. 1, left panel); 3 Within each food category (first dish, side dish, second dish, fruit-dessert), we rank food choices based on their caloric content from best (less caloric content) to worse (higher caloric content). 4 The canteen staff provided us the recipes of the dishes and, with their supervision, we used the website http:// www.myfitnesspal.com to rank every dish in each category, from the less to the most caloric. 253Can menu labeling affect away-from-home-dietary choices? 2) Calories Ranking Labels (CRL): within each meal category (first dish; second dish; side dish; fruit-dessert.) the labels indicate a ranking among options based on the lev- el of calories from the least (gold medal, 1st place) to the most (bronze medal, 3rd place) caloric (Fig. 1, right panel). The labels were chosen together with the canteen managers. We proposed different types of labeling selected from previous studies. During the first week (1st and 2nd day) the effect of providing a LCL was tested, whereas the CRL was used in the second week (3rd and 4th day). During each day of the experiment only one food line (treated line) displayed a label while in the other line no label was present (control line). It is assumed people randomly choose between the two lines, although to account for a possible self-selection bias the treated and the control lines from day one to day two (LCL) and from day three to four (CRL) were switched. Participants were not aware to be part of the experiment before selecting the food choices. The first contact with the labels took place at the beginning of the treatment line, where a flier explained the meaning of the label (LCL in day 1 and 2: CRL in day 3 and 4 as in Fig. 1). Individuals taking the control line did not receive any nutritional infor- mation during the meal selection. The recruitment of participants to the experiment took place at the end of the lines (both control and treatment), where, with the support of a flier, two recruiters explained to users how to take part to the experiment. If they accept- ed, they were asked to take a picture of their tray using their smartphone before starting to eat and to share it using a digital platform. Moreover, after lunch, participants were asked to complete a survey including both demographic and behavioral questions. All partici- pants were rewarded with a coupon redeemable at the university coffee shop. We collected 459 observations during the first week (1st and 2nd day) and 471 during the second week (3rd and 4th day). The final dataset contains 930 observations recording tray composition, --------------------------------------------------------------------------------------- These symbols identify caloric content of dishes as: the lowest: ; the medium: ; the highest: relatively to the group they belong to: First dish, Second dish, Side dish, dessert. --------------------------------------------------------------------------------------- Example: Dessert Ice cream A slice of Sacker cake Fruit Salad --------------------------------------------------------------------------------------- Figure 1. Explanation of LCL (left panel) and CRL (right panel). 254 Elena Castellari et alii demographics and behavioral characteristics for each individual. The final sample is most- ly composed by university students (around 84% of the sample), and in small percentage by university faculty and staff. For a detailed description of the participants, please refer to the model and empirical results sections. 3.2 Indexes of meal composition To summarize the food selections made by participant i at day t we computed two different indicators of the tray’s meal composition. The purpose of this index is to meas- ure how close the composition of the meal is to an “optimal meal”, which in the case of LCL would correspond to a tray with all green labeled choices, while in the case of CRL to a tray with all gold medals. Two different indexes for both the treatment and the con- trol subsamples were computed: (a) a uniform index (UI) where we attributed the same weight to each of the dish selections; (b) a weighted index (WI) where we attributed differ- ent weights to dishes of different categories (first dish, second dish, side dish and dessert). The UI was computed as follows: UIi,t = j=1 Ni ,t ∑S jit 1 Nit (1) where Nit is the total number of dishes composing the tray of individual i at day t while Sjit is the score, which in the case of the LCL would be equal to one if individual i at day t made a choice j labeled as healthy (green label), and zero otherwise. In the case of the CRL Sjit has a value equal to 1 if the choice j made by individual i at day t was labeled as gold, equal to 0.5 if choice j was labeled as silver, and zero if it was labeled as bronze. Similarly, to compute the WI we used the following: WIi,t = j=1 Ni ,t ∑S jit Pjit j=1 Ni ,t∑ Pjit (2) where Pjit is the weight attributed to each dish selected by individual i at day t The weight Pjit depends on the meals’ category. Specifically, a weight of 0.35 was attributed to the first and second dishes, since they are typically more caloric, and a weight of 0.15 to side dish and dessert. Both indexes (UI and WI) range from one, when an “optimal tray” was cho- sen, to zero, when all choices are not the one “suggested” by the labels. Fig. 2 shows the distribution of the two indexes (WI and UI) under both label treatments (LCL and CRL). The index computed using the uniform approach is more concentrated around some spe- cific values. Moreover, the distribution of all indexes is concentrated around zero: for the LCL indexes the zeros account for more than 80% of the observations, while for the CRL indexes this share reduces to around 60%. 255Can menu labeling affect away-from-home-dietary choices? 4. Results To test whether the label provision had an effect on the food selections, we generated a variable indicating the propensity to select an “healthy option” (PHO), using the UI and the WI. Specifically, based on the index values, we compute the PHO as an ordinal varia- ble with five possible outcomes as described in Table 1. The probability of being in a PHO class (k), is given by: Pr(PHO =k | Z) = Φ (βk + [Z]`β) – Φ (βk+1 + [Z]`β) (3) where Φ (.) is the standard logistic density function (CDF), k= [0,…,4], β0 = - ∞ and β5 = + ∞; Z is a set of covariates influencing PHO and β is a conformable set of parameters. Specifically, Z includes the following variables: a) T is a dummy variable equal to one if participant i in day t belongs to the treated sample; b) Xi is a set of demographics and behavioral variables collected for each person i, as described in Table 2. Summary statistics are presented in Table 3. In both weeks, the sample is almost equally split between treated and non-treated observations. Students are the largest share Figure 2. Distribution of the Weighted Index (WI) and Uniform Index (UI) for the Less Caloric Label (LCL) and Calories Ranking Labels (CRL). 0 2 4 6 8 10 D en si ty 0 .2 .4 .6 .8 1 Uniform Index (UI) Less Caloric Labels (LCL) 0 2 4 6 8 10 D en si ty 0 .2 .4 .6 .8 1 Weighted Index (WI) Less Caloric Labels (LCL) 0 2 4 6 8 10 D en si ty 0 .2 .4 .6 .8 1 Uniform Index (UI) Calories Ranking Labels (CRL) 0 2 4 6 8 10 D en si ty 0 .2 .4 .6 .8 1 Weighted Index (WI) Calories Ranking Labels (CRL) Source: Own data elaboration. 256 Elena Castellari et alii of the sample (around 84%) and females are around half of the sample. Around 40% of the sample is commuting from nearby areas and almost 90% of the participants use the cafeteria at least three times a week. A large share of the sample declared to usually pay attention to the labels of the food they purchase (around 80%), to prepare its own meal often or sometimes (around 70%), to practice regular physical activity at least once a week (around 70%), and to not substitute water with other drinks during a meal (around 70%). More than 30% of the sample experienced some weight gain in the last six months and more than 20% sometimes visited a nutritionist to receive diet advises. Surprisingly, only 2.6% of the whole sample reported to consume at least five portions of fruit and vegeta- bles daily. Given the nature of the dependent variable, model (3) was estimated in STATA using an ordered logit model. Equation (3) was estimated for both LCL and CRL samples, using both uniform and weighted indexes. Results are reported in Table 4. All parameters on the treatment line variable (T) are not statistically significant, sug- gesting that all our specifications fail to identify any significant positive effect of the label Table 1. Definition of the classes of PHO. Outcomes Classes PHO=0 UI or WI = 0 PHO=1 0 < UI or WI ≤ 0.25 PHO=2 0.25 < UI or WI ≤ 0.50 PHO=3 0.50 < UI or WI ≤ 0.75 PHO=4 0.75 < UI or WI ≤ 1 Table 2. Demographic and behavioral variables. Variable name Variable Description Student One if student, zero otherwise Female One if female, zero otherwise Commuter One if commuter, zero otherwise Frequent User One if he/she eats at the cafeteria at least 3 times a week, zero otherwise Cook One if he/she prepares his/her own dishes often or sometime, zero if rarely or never Label One if he/she reads the label of the food consumed often or sometime, zero if rarely or never FV5 One if he/she consumes at least 5 portions of Fruit or vegetables per day, zero otherwise Water One if during the meal he/she never or rarely substitutes water with other drinks, zero if often or always Weight One if in the last six months he/she had a weight increase, zero otherwise Nutritionist One if he/she ever visits a nutritionist for a diet, zero otherwise Active One if he/she practices physical activity at least once-twice a week, zero otherwise Source: Own data collection. 257Can menu labeling affect away-from-home-dietary choices? provision on the level of the index (UI and WI)5. These results are in line with several other studies which found information based policies are effective on improving consum- er awareness but not necessarily to significantly impact behavior (Galizzi, 2014). Results show students tend to be more reluctant to change their food selections (for all models coefficients are negative and significant). In line with previous studies ( i.e. Krieger et al., 2013), this paper also finds women have a different attitude towards menu labeling, with specifications (3) and (4) of table 4 showing positive and significant coefficients. Frequent users of the canteen service do not seem to respond differently than less fre- quent users (i.e. coefficients are not significant). This study also finds people who some- times or often cook their own meal tend to have a higher index under the LCL approach, while for the CRL the difference is not significant. Variables associated with more attention to the diet, as the attitude on reading food labels, or consuming more fruit and vegetables, are significantly correlated with higher PHO under the CRL scheme, but not under the LCL. Similarly, people who declare to never substitute water with other drinks, or having required the opinion of a nutritionist, tend to have higher PHO, with a positive improvement of the index, only under the LCL scheme. Furthermore, results show variables such as having gained weight in the previous six months, or practicing sport at least once a week, are not associated with different PHO. 5 This study considers only a selected sample of a university cafeteria in Italy, for regulatory purpose and policy interventions an extended study with a more representative sample need to be consider. Table 3. Summary Statistics. Variable Less Caloric Label (LCL) N=459 Calories Ranking Label (CRL) N=471 Mean Std. Dev. Mean Std. Dev. Uniform Index (UI) 0.279 0.289 0.286 0.262 Weighted Index (WI) 0.245 0.279 0.256 0.261 PHO ( from UI) 1.251 1.243 1.314 1.122 PHO (from WI) 1.203 1.278 1.306 1.167 Treated line (T) 0.468 0.500 0.482 0.500 Student 0.843 0.364 0.851 0.356 Female 0.525 0.500 0.501 0.501 Commuter 0.397 0.490 0.372 0.484 Frequent User 0.854 0.353 0.868 0.338 Cook 0.786 0.410 0.769 0.422 Label 0.806 0.396 0.794 0.405 FV5 0.026 0.160 0.030 0.170 Water 0.778 0.416 0.726 0.446 Weight 0.327 0.470 0.344 0.476 Nutritionist 0.255 0.436 0.225 0.418 Active 0.691 0.463 0.705 0.457 Source: Own data elaboration. 258 Elena Castellari et alii 5. Discussion The main objective of this research was to analyze whether the provision of nutrition- al information influenced the meal composition in an away-from-home environment. To this end, we conducted a field experiment at a university cafeteria. Table 4. Results - Ordered Logit model. VARIABLES (1) (UI LCL) (2) (WI LCL) (3) (UI CRL) (4) (WI CRL) T 0.019 -0.121 0.260 0.146 (0.175) (0.175) (0.171) (0.170) Student -1.024*** -1.263*** -1.387*** -1.366*** (0.232) (0.235) (0.251) (0.246) Female 0.008 -0.046 0.310* 0.368** (0.179) (0.178) (0.182) (0.183) Commuter -0.139 -0.189 -0.786*** -0.769*** (0.184) (0.183) (0.186) (0.185) Frequent User -0.025 0.091 0.021 0.076 (0.264) (0.262) (0.269) (0.268) Cook 0.500** 0.514** 0.216 0.252 (0.222) (0.222) (0.214) (0.212) Label 0.044 0.024 0.399* 0.430** (0.229) (0.229) (0.216) (0.216) FV5 0.544 0.671 1.042** 1.188** (0.568) (0.559) (0.482) (0.475) Water 0.550** 0.493** 0.305 0.280 (0.220) (0.219) (0.204) (0.201) Weight 0.120 0.041 -0.018 -0.073 (0.190) (0.189) (0.180) (0.179) Nutritionist 0.396* 0.469** 0.345 0.199 (0.203) (0.202) (0.211) (0.209) Active -0.052 -0.066 -0.003 0.175 (0.192) (0.190) (0.201) (0.199) Constant cut1 -0.254 -0.554 -1.222*** -1.064** (0.461) (0.470) (0.456) (0.450) Constant cut2 0.195 0.245 -0.200 0.151 (0.461) (0.469) (0.452) (0.448) Constant cut3 1.753*** 1.503*** 1.734*** 1.745*** (0.470) (0.476) (0.460) (0.457) Constant cut4 3.333*** 2.534*** 3.306*** 2.963*** (0.512) (0.495) (0.508) (0.484) Observations 459 459 471 471 Standard errors in parentheses: *** p<0.01, ** p<0.05, * p<0.1. 259Can menu labeling affect away-from-home-dietary choices? The use of information based policies is among the most debated instruments when policy makers look for solutions to promote behavioral changes towards healthier and more sustainable food choices. Yet, effects of these information-based policies on actual behavioral changes are mixed. While some previous studies have found some potential benefits of menu labeling in a restaurant setting, in terms of calorie intake reduction and healthier food choices (Oliveira et al., 2018; Ellison et al., 2013; Roberto. et al., 2010), this paper did not find any statistically significant effects of caloric labeling on food selections, in accordance with several other studies (Elbel. et al., 2009; Swartz et al., 2011; Downs et al. 2009; Mazzocchi and Trail, 2005; Swartz et al. 2011; Kiszko et al., 2014; Harnack and French, 2008). While these results can also be driven by the experimental settings, they suggest that compulsory nutritional labeling in a dining-out environment may not be effective per se. First the effect of a label on dietary choices depends on many unobservable or not-record- ed factors, such as the environment characteristics, the sample composition, the way the labels have been explained and communicated, the type of labels, and many behavioral characteristics. Most of the studies are referred to relatively small sample and to selected group (such as university students), so it becomes difficult to generalize the results from this type of studies to the whole population, as well to find ad hoc recipe valid for all set- tings. Further, even if a strong link between nutritional label and caloric intake reduction as well as food environment improvement would be found, there would still be the need to consider the final outcomes of this policy interventions on health and BMI (Jaime and Lock, 2009). Bonanno et al. (2018), using a quantile regression approach, have shown that the rela- tionship between reading food labels and BMI highly differs among demographics groups. Krieger et al. (2013) have measured the effect of calories posting in fifty restaurants, and after eighteen months, have found a decrease of menu calories only in some sites and in women, but not in men. These previous studies highlight the difficulties to find a “best for all” policy. Thus, some ad-hoc interventions are needed to set up eating environments where healthy food choices are enhanced. In this sense, the synergies among different actions can be valuable to reach broader demographic groups. However, in general, especially in cafeterias linked to educational or working environments, a sure action that need to be reinforced is the setup of common protocols to monitor the nutritional quality of the service and to measure the effect of any in-site healthy initiative. Only continuously and carefully monitoring the nutritional quality of food options and the effects of interventions can ensure their effectiveness on enhancing healthier behavioral changes. 6. Conclusions and policy implications The provision of nutritional labels in a food canteen have many practical difficul- ties. First, recipes need to be standardized and carefully followed; second, dish sizes need also to be standardized, with additional burdens on the food preparation process. How- ever, asking to provide nutritional labels without enforcing the use of standard procedures on food preparation might lead to misleading information signaling, while, at the same time, enforcing this standardization might push out of business small no-chained restau- 260 Elena Castellari et alii rants (Mazzocchi et al., 2009). Given these practical issues, applying the requirements for the labels only to chain restaurants, as experimented in the USA, is probably the easiest option to be applied in Europe. Moreover, chain restaurants are usually chosen to dine out by people driven by time and price constraints, which represent a population group most likely to be targeted by policy makers. However, even if nutritional labels alone will not be the solution to the obesity problem, their provision can increase consumer awareness and lead to some beneficial spillover effects, such as encouraging restaurants to offer healthier food and meal “reformulation” (Schulman, 2010). As nutritional information is presented to consumers, restaurants might find incen- tives to offer lower calorie and healthier options, as observed by Ellison et al. (2014). In this scenario, if the final goal of these policies is to improve the healthiness of food choices, our results, together with the existing literature, suggest the need of continuous monitoring of behaviors in order to design effective policies. However, we can also think of label policies for only their information value “per se”, independently from their effect on final food choices and health outcomes. In this sense, Marette et al. (2019) have found a traffic light label significantly impacts the willingness to pay for the different types of products offered in an experiment, showing that consumer positively evaluate the provi- sion of an easily readable label. An analysis sizing the cost and the benefits of implement- ing a labeling policy could be valuable to understand to what extent this policy is eco- nomically feasible and if it can be potentially sustained under a voluntary scheme. Howev- er, at this end, it is also important to consider that the literature has previously mentioned that an overload of information reduce the marginal effect related to it (Keller and Stae- lin,1989), at the point that consumers can even lose any interest, which is a big challenge for regulators. In accordance with previous findings, we believe that, in order to encourage behav- ioral changes in an away-from-home food environment, public policies need to rely on a multifaceted approach, where the provision of nutritional information synergizes with other instruments such as nutritional education, social norms provision and nudges (Downs et al., 2009; Thapa and Lyford, 2014; Thunström and Nordtröm, 2013, Burger et al. 2010, Storcksdieck genannt Bonsmann and Wills 2012, Castellari and Berning, 2016). Moreover, the discussion highlights the importance of reinforcing common protocols to ensure the nutritional quality of the food options in cafeterias, and to constantly moni- tor any intervention promoting healthy food styles. Moreover, further research needs to evaluate if the implementation of a “health related” intervention in a cafeteria, such as the introduction of nutritional label, has an effect on the sustainability of the food environ- ment and on the produced waste. Overall, it is important for regulators to follow a multi- disciplinary and systemic approach to the food system where all possible spillovers from the demand and supply side are evaluated in order to promote a more sustainable and healthy food environment. 7. Acknowledgements We thank Giorgia Ciulli for the help during the data collection, Stefano Longo and Elena Barbieri and EDUCATT for the fruitful collaboration. The authors only are respon- sible for any errors or omissions. 261Can menu labeling affect away-from-home-dietary choices? 8. Funding This work was supported by the Daniel & Nina Carasso Foundation through the FP7 SUSFOOD ERA-Net  Research Project “SUSDIET - Implementing sustainable diets in Europe” (Scientific Coordinator: Louis-George Soler). 9. References Agrifood Monitor (2016). http://www.agrifoodmonitor.it/en/food-consumption . Accessed February 3, 2017. Biing-Hwan, L., Frazão, E. and Guthrie J. (1999). Away-From-Home Foods Increasingly Important to Quality of American Diet. Agriculture Information Bulletin No. 749, Food and Rural Economics Division, Economic Research Service, U.S. Department of Agriculture, Washington, DC. Biing-Hwan, L. and Guthrie J. (2012). Nutritional Quality of Food Prepared at Home and Away From Home, 1977-2008. Economic Information Bulletin No. 105, U.S. Department of Agriculture, Economic Research Service, Washington, DC. Backstrand, J.R, Wootan, M.G., Young, L.R. and Hurley J. (1997). Fat Chance: A Survey of Dieticians’ Knowledge of the Calories and Fat in Restaurant Meals. Center for Sci- ence in the Public Interest, Washington, DC. Balfour, D., Moody, R., Wise, A. and Brown K. (1996). Food Choice in Response to Com- puter-Generated Nutrition Information Provided about Meal Selection in Work- Place Restaurants. Journal of Human Nutrition and Dietetics 9(3): 231–37. Binkley, K., Eals, J. and Jekanowski M. (2000). The Relation Between Dietary Change and Rising US Obesity. International Journal of Obesity 24(8): 1032-1039. Bonanno, A., Bimbo, F., Cleary, R. and Castellari E. (2018). Food Labels and Adult BMI in Italy – An Unconditional Quantile Regression Approach. Food Policy 74: 199-211. Bowman, S.A., Gortmaker, S.L., Ebbeling, C.B., Pereira, M.A. and Ludwig D.S. (2004). Effects of Fast-Food Consumption on Energy Intake and Diet Quality Among Chil- dren in A National Household Study. Pediatrics 113(1): 112–118. Brambilla-Macias, J., Shankar, B., Capacci, S., Mazzocchi, M., Perez-Cueto, F.J.A.,Verbeke, W. and Trail W.B. (2011). Policy Interventions to Promote Healthy Eating: A Review of What Works, What Does Not, and What is Promising. Food and Nutrition Bul- letin 32(4): 365-375. Burger, J.M., Bell, H., Harvey, K., Johnson, J., Stewart, C., Dorian, K. and Swedroe M. (2010). Nutritious or Delicious? The Effect of Descriptive Norm Information on Food Choice. Journal of Social and Clinical Psychology 29(2): 228-242. Castellari, E. and Berning, J.P. (2016). can Providing A Morning Healthy Snack Help to Reduce Hunger During School Time? Experimental Evidence from an Elementary School in Connecticut. Appetite 106: 70-77. Downs, J.S., Loewenstein, G. and Wisdom, J. (2009). The Psychology of Food Consump- tion. Strategies for Promoting Healthier Food Choices. American Economic Review: Papers & Proceedings 99(2): 159-164. Elbel, B., Kersh, R., Brescoll, V.L. and Dixon, L.B. (2009). Calorie Labeling and Food Choices: A First Look at The Effects on Low-Income People In New York City. Health Affairs 28(6): 1110-1121. 262 Elena Castellari et alii Ellison, B., Lusk, J.L. and Davis, D. (2013). Looking at the Label and Beyond: The Effects of Calorie Labels, Health Consciousness, and Demographics on caloric intake in restaurants. Journal of Behavioral Nutrition and Physical Activity 21: 10-21. Ellison, B., Lusk, J. L. and Davis, D. (2014). The Impact of Restaurant Calorie Labels on Food Choices: Results from a Field Experiment. Economic Inquiry 52(2): 666-681. Federazione Italiana Pubblici Esercizi – Confcommercio (2015). Ristorazione: Rapporto Annuale 2015. Finkelstein, E.A., Strombotne, K.L., Chan, N.L. and Krieger J. (2011). Mandatory Menu Labeling in One Fast-Food Chain in King County, Washington. American Journal of Preventive Medicine 40(2): 122-127. Galizzi, M.M. (2014). What is Really Behavioral in Behavioral Health Policy? And Does it Work? Applied Economic Perspectives and Policy 36 (1): 25-60. Glanz, K., Basil, M., Maibach, E., Goldberg, J. and Snyder, D. (1998) Why Americans Eat What they Do: Taste, Nutrition, Cost, Convenience, and Weight Control as Influ- ences on Food Consumption. Journal of the American Dietetic Association 98(10): 1118–1126. Harnack, L. and French, A.S. (2008). Effect of Point-of-Purchase Calorie Labelling on Res- taurant and Cafeteria Food Choices: A Review of the Literature. International Jour- nal of Behavioral Nutrition and Physical Activity 5: 51 Jaime, P.C. and Lock K. (2009). Do School Based Food and Nutrition Policies Improve Diet and Reduce Obesity? Preventive Medicine 48(1): 45-53. Keller, K. L. and Staelin, R. (1989). Assessing Biases in Measuring Decision Effectiveness and Information Overload. Journal of Consumer Research 15(4): 504-508. Kiszko, K.M., Martinez, B.S., Abrams, C. and Elbel, B. (2014). The Influence of Calorie Labeling on Food Orders and Consumption: A Review of the Literature. Journal of Community Health 39(6): 1248-1269. Kozup, J.C., Creyer, E.H., and Burton, S. (2003). Making Healthful Food Choices: The Influence of Health Claims and Nutrition Information on Consumers’ Evaluations of Packaged Food Products and Restaurant Menu Items. Journal of Marketing 67(2): 19-34. Krieger, J.W., Chan, N.L., Saelens, B.E., Ta, M.L., Solet, D. and Fleming D.W. (2013). Menu Labeling Regulations and Calories Purchased at Chain Restaurants. American Jour- nal of Preventive Medicine 44(6): 595-604. Marette, S., Nabec, L. and Durieux, F. (2019). Improving Nutritional Quality of Consum- ers’ Food Purchases with Traffic-Lights Labels: An Experimental Analysis. Journal of Consumer Policy, forthcoming. doi.org/10.1007/s10603-019-09420-5 Mazzocchi, M. and Trail W.B. (2005). Nutrition Health and Economic Policies in Europe. Acta Agriculturae Scandinavica Section C 2(3-4): 138-149. Mazzocchi, M., Trail, W.B. and Shogren J.F. (2009). Fat Economics: Nutrition, Health and Economic Policy. New York: Oxford University Press. McCrory, M.A., Fuss, P.J., Hays, N.P., Vinken, A.G., Greenberg, A.S. and Roberts, S.B. (1999). Overeating in America: Association Between Restaurant Food Consumption and Body Fatness in Healthy Adult Men and Women Ages 19–80. Obesity Research 7(6): 564-571. Mottaleb, K.A.,  Rahut, D.B. and Mishra, A.K. (2017). Consumption of Food Away from 263Can menu labeling affect away-from-home-dietary choices? Home in Bangladesh: Do Rich Households Spend More? Appetite https://www.ncbi. nlm.nih.gov/pubmed/28347779119: 54-63. Oliviera, R.C., Fernandes, A.C., da Costa Proença, R.P., Hartwell, H., Rodrigues, V.M., Flemming Colussi, C. and Fiates G.M.R. (2018). Menu Labelling and Healthy Food Choices: A Randomised Controlled Trial. British Food Journal 120(4): 788-801. Roberto, C.A., Larsen, P.D., Agnew, H., Baik, J. and Brownell, K.D. (2010). Evaluating the Impact of Menu Labeling on Food Choices and Intake. American Journal of Public Health 100(2): 312-318. Satia, J.A., Galanko, J.A. and Siega-Riz ,A.M. (2004). Eating at Fast-Food Restaurants is Associated with Dietary Intake, Demographic, Psychosocial and Behavioural Factors Among African Americans in North Carolina. Public Health Nutrition 7(8):1089– 1096. Schulman, T. (2010). Menu Labeling: Knowledge for a Healthier America. Harvard Jour- nal on Legislation 47(2): 587-610. Seguin, R.A., Aggarwal, A., Vermeylen, F. and Drewnowski, A. (2016). Consumption Frequency of Foods Away from Home Linked with Higher Body Mass Index and Lower Fruit and Vegetable Intake among Adults: A Cross-Sectional Study. Journal of Environmental and Public Health 2016. doi.org/10.1155/2016/3074241. Seward, M.W., Block, J.P. and Chatterjee, A. (2016). A Traffic-Light Label Intervention and Dietary Choices in Colleges Cafeterias. American Journal of Public Health 106(10): 1808-1814. Storcksdieck genannt Bonsmann, S. and Wills, J.M. (2012). Nutrition Labeling to Prevent Obesity: Reviewing the Evidence from Europe. Current Obesity Reports 1(3):134-140. Swartz, J.J., Braxton, D. and Viera, A.J. (2011). Calorie Menu Labeling on Quick-Service Restaurant Menus: An Updated Systematic Review of the Literature. International Journal of Behavioral Nutrition and Physical Activity 8: 135. Thunström, L. and Nordtröm, J. (2013). The Impact of Meal Attributes and Nudging on Healthy Meal Consumption- Evidence from a lunch Restaurant Field Experiment. Modern Economy 4: 1-8. Todd, J.E., Mancino, L. and Biing-Hwan, L. (2010). The Impact of Food Away From Home on Adult Diet Quality. ERR-90, U.S. Department of Agriculture, Economic Research Service, Washington, DC. Trail, B.W. (2012). Presidential Address: Economic Perspectives on Nutrition Policy Evalu- ation. Journal of Agricultural Economics 63(3): 505-527. Thapa, J.R. and Lyford, C.P. (2014). Behavioral Economics in the School Lunchroom: Can it Affect Food Supplier Decision? A Systematic Review. International Food and Agri- business Management Review 17(A): 187-208. Vasiljevic, M., Pechey, R. and Marteau, T.M. (2015). Making Food Labels Social: The Impact of Colour of Nutritional Labels and Injunctive Norms on Perceptions and Choice of Snack of Foods. Appetite 91: 56:63. Wisdom, J., Downs, J.S. and Loewenstein, G. (2010). Promoting Healthy Choices: Informa- tion versus Convenience. American Economic Journal: Applied Economics 2: 164-178. Yamamoto, J.A., Yamamoto, J.B., Yamamoto, B.E. and Yamamoto, L G. (2005). Adoles- cent Fast Food and Restaurant Ordering Behavior with and without Calorie and Fat Content Menu Information. Journal of Adolescent Health 37(5): 397-402. Positive Mathematical Programming and Risk Analysis Quirino Paris The hedonic contents of italian super premium extra-virgin olive oils Luca Cacchiarelli1,*, Anna Carbone2, Tiziana Laureti1, Alessandro Sorrentino1 Corporate R&D and the performance of food-processing firms: Evidence from Europe, Japan and North America Heinrich Hockmann1, Pedro Andres Garzon Delvaux2,*, Peter Voigt3, Pavel Ciaian2, Sergio Gomez y Paloma2 Can menu labeling affect away-from-home-dietary choices? Elena Castellari1,*, Stéphan Marette2, Daniele Moro3, Paolo Sckokai1 A preliminary test on risk and ambiguity attitudes, and time preferences in decisions under uncertainty: towards a better explanation of participation in crop insurance schemes Attilio Coletta1, Elisa Giampietri2, Fabio Gaetano Santeramo3,*, Simone Severini1, Samuele Trestini2