Poverty in Canada (2nd Edition)

Introduction

NEXT TO HOUSING, FOOD IS likely to be the biggest expense for low income families. It is a fact, however, that food has never been a better bargain than it is today. The proportion of every family's (including low income families) income spent on food has systematically declined during this century.

The purpose of this chapter is to develop a methodology for determining the essential cost of a nutritious diet. Discussion of food is limited to common, nutritious and palatable diets. This is consistent with the principle elucidated in the previous chapter, that is, basic necessities are those required to maintain long-term physical well-being and must be at least at the standard of quality considered minimum acceptable in the community. Before that is done, however, I wish to critically examine prevailing food budget guides.

Practical guides to food budgeting for Canadian families have typically used the "Food Basket" approach. This involves costing on a periodic basis an extensive list of foods designed to meet nutritional needs. The Montreal Diet Dispensary, the Social Planning Council of Toronto and, more recently, Agriculture Canada (Robbins (1984)), have developed Food Baskets to assist families in selecting a nutritious diet. While in principal the idea of a nutritious food basket is a good one, there are serious problems with each of the above versions.

Agriculture Canada's Nutritious Food Basket makes no claim to be a low or minimum cost way of providing good nutrition and it is well they don't. Included in their Food Basket are such items as bacon, cream cookies, salad dressing, jam, jelly, tea, coffee and condiments—all absolutely inconsistent with either good nutrition or low cost. In the case of processed foods, which make up half the Food Basket, "name" brands have been used exclusively. Many low income shoppers utilize generic products or store brands with no sacrifice of nutrition or palatability while enjoying substantial savings. In addition, the Food Basket does not always select the most economical size of an item. For example milk (2 l.), peanut butter (500 g.), macaroni (500 g.), apples (454 g.), carrots (454 g.), and onions (454 g.) can all be routinely purchased in larger and more cost efficient sizes. Virtually every family could take advantage of a larger sized purchase within the perishability limit.

Related to the issue of purchase size is the "problem" of variety. Agriculture Canada's Nutritious Food Basket implicitly requires families to consume 79 different food items each week. This is physically impossible, even for very large families, because of fairly rigid indivisibilities in purchase sizes of most goods. For example, there are 15 items in the vegetable food group which when purchased in the (relatively small) stated package size will provide about 135 servings of vegetables per week. According to the Canada Food Guide, a family of four requires only 70 servings of vegetables per week to maintain good nutrition. The vegetable list is appropriate only for families of six or more persons. Attempting to achieve this variety each week is clearly uneconomical for virtually all Canadian families. The problem still exists but is less pronounced with the other food groups in the basket. The point here is that while the Nutritious Food Basket reflects average consumption patterns, it is of little use as a practical guide to real Canadian families, whatever their income level. It is impossible for most families to achieve significant variety each week without great expense. Individual families are likely to be selecting food items from a smaller "core" list of foods and substituting when appropriate. The Food Basket is simply not representative of how Canadian households actually consume food.

The most damaging problem with Agriculture Canada's Nutritious Food Basket is that its consumption results in excessive caloric intake. The author provides no explanation of exactly how quantities consumed were to be determined and gives no information on the energy adequacy of the quantities selected. I have elsewhere calculated the Food Basket's caloric intake for various age/sex categories and found that average recommended levels were exceeded in each case. This is true for both the 1984 and 1986 versions of the Food Basket. The consumption of too much food is not only unhealthy but results in an overestimate of the required cost of a nutritious diet.

This critique applies to the 1984 and 1986 versions of the Nutritious Food Basket because these were the prevailing versions during the reference period for this study (1988). In 1989, however, Agriculture Canada developed a "Thrifty Nutritious Food Basket" along with an updated version of their "Nutritious Food Basket." They use the same methodology but somewhat fewer foods than in the earlier versions. This time, unfortunately, they did not include sufficient information to enable researchers to check the energy adequacy of the baskets.

While Agriculture Canada does not claim to determine the minimum cost of a nutritious diet, the Montreal Diet Dispensary (MDD) does. They select a list of relatively low cost foods, use prices from major (Montreal) grocery chains, and use weights which reflect the consumption patterns of the least well off quintile. While these measures undoubtedly result in food costs being lower than they could be, there is nothing at all in the MDD methodology which ensures that the cost of a nutritious diet will be minimized. Indeed there are several rather obvious features of their approach which result in higher than necessary costs.

The prices used in the MDD calculations are the median prices of each item from among three popular Montreal chain food stores. Only the regular price of the item is used, even if it is on sale that week. But surely sale prices ought to be included because they reflect what shoppers actually pay for the particular item while it is on special. In the normal course of shopping some food items are on sale and no shopper, whatever their income level, will insist on paying the higher, regular price. This does not require any deliberate shopping strategy on the part of the consumer. Ignoring sale prices clearly overestimates the essential cost of a nutritious diet.

In addition, weekly food costs are increased by an arbitrary five percent because of "inevitable food wastage." What evidence is there that inevitable wastage is five percent? We should be suspicious of a middle-class bias at work here. There are a great many low income families for whom food wastage is unthinkable and whose actual rate of waste may well be much less than five percent. Even if we accept that food costs are five percent higher because one-twentieth of all food purchased by poor households is wasted, there are surely some offsets which tend to lower food costs. "Stocking-up" strategies and family garden plots are among the more familiar examples.

Several food items in the MDD basket are not purchased in the most economical size. Their reference family of four could easily use larger sizes of two percent milk, cheese, peanut butter and rice and achieve some saving. Also, the use of generic and store brands for some two dozen items instead of the more expensive "name" brands would result in additional savings.

The 71 items in the MDD basket provide more variety than most Canadian households can consume during a given week. Like Agriculture Canada, the MDD feels compelled to reflect average consumption patterns while ignoring the feasibility of their diets for real families. While they do attempt to account for the diseconomies (or economies) of family size, their adjustments are entirely arbitrary and without justification. One way of getting around this problem would be to use a shorter list of foods and make substitutions where appropriate.

The MDD uses 1975 guidelines for recommended intake levels. Health and Welfare Canada revised these guidelines in 1983 and in almost all age/gender categories the energy requirements have been scaled down. Yet the MDD continues to use the older, somewhat excessive levels in its calculation of "minimum" food costs. This in itself overestimates the essential cost of a nutritious diet. More importantly, however, like Agriculture Canada, they never demonstrate the energy adequacy of their own Food Basket.

Using the quantities they provide and a calorie counter booklet, it is possible to determine the energy intake of the MDD basket. The calculations show that, for a family of four, food consumption is excessive by about seven percent (using the 1983 intake guidelines). So, as with Agriculture Canada's Nutritious Food Basket, the MDD "minimum" cost diet results in too much food being consumed which is, of course, unhealthy and needlessly expensive. Without question, the Montreal Diet Dispensary has not minimized the cost of a nutritionally adequate diet.

The Social Planning Council of Metropolitan Toronto has a guide for food budgeting which resembles in most ways the Food Baskets of Agriculture Canada and the MDD. The Social Planning Council (SPC) has a list of 106 food items subdivided into some 14 food groupings. An average cost per food grouping is determined by taking a weighted average of the price per unit for each of the foods in the grouping. The weights reflect the actual buying patterns of "sample" families. By multiplying the average cost per food grouping by the suggested weekly quantities (by age/gender), we get the cost of each food grouping for the week which can then be summed to determine total weekly cost. As with the other two approaches, the SPC uses too much variety to be feasible and nowhere demonstrates the caloric adequacy of its diet. In addition, it provides no information about how prices were determined (i.e., one Toronto grocery store? many?), whether "name" brands were used exclusively, or how exactly the suggested quantities were determined. However, the SPC does not claim to provide the least expensive means to a nutritious diet. They state, "It should be emphasized that these costs do not represent a minimum" and go on to offer a number of suggestions to reduce food costs. They observe that families tend to overutilize meat (which is relatively expensive) and underutilize vegetables (relatively less expensive). Among their suggestions to lower food costs are: use larger sizes where feasible, stock up on sale items, buy lower priced store and generic products; and buy powdered milk because it is the nutritional equivalent of its fresh fluid counterpart and is lower priced.

At best, the prevailing Food Basket approaches can be regarded as estimates of how much average Canadian families pay for a nutritious diet. In no way do they represent the essential or minimum costs of a nutritious diet. Indeed, two of the three versions wisely make no such claim.

George Stigler (1945) was one of the first to develop a methodology to determine the minimum cost of a nutritious diet. He began by tabulating the nutrient composition of a long list of common foods. Only the following nine nutrients were examined: calories, protein, calcium, iron, vitamin A, thiamin, riboflavin, niacin and ascorbic acid. Using published urban average food prices, he then calculated the quantity of each nutrient per dollar of expenditure for each food. This enabled him to identify the most economical sources of each nutrient. Using "recommended nutrient allowances" established by the U.S. National Research Council as a constraint, he found, by trial and error, that the least expensive way of providing the recommended nutrients was in a diet of just nine foods. This result was to be expected. By omitting variety and palatability as constraints, cost minimization is achieved with the same number of foods as nutrients. Stigler recognizes this and does not actually recommend his diet to anyone.

Stigler's attempt to apply analytical techniques to the problem of minimizing food costs can be correctly viewed as an interesting academic exercise with no practical value. Variety and palatability are essential components of food consumption and must be included in any useful analysis of food cost determination.

Methodology

The approach used in this study is a variation of the classic linear programming diet problem. The cost of food is to be minimized subject to four considerations—variety, palatability, balance and energy. The first two are relatively subjective concerns and not easily amenable to scientific determination. However, it is inappropriate to ignore them. A reasonable solution to the palatability problem would be to develop a list of highly popular and widely available foods. The issue of variety, however, is more difficult to resolve. There is no consensus among nutritionists as to exactly how many different foods one must consume to obtain the required nutrients. There is certainly no agreement as to the variety required to avoid monotony. This would clearly depend on factors such as ethnic and regional preferences, age, education and income level and perhaps genetics. Their best advice is to eat a variety of foods from within each food group. The list of foods developed below is an attempt to balance off considerations of variety with popularity and regular price data availability. However, the issue of variety is very interesting and important and is examined again later in the chapter.

The final two considerations are formal linear constraints. Balance refers to the number of servings from within each food group required for good nutrition. The Canada Food Guide (1985) is the nutritional standard used. The energy constraint refers to the recommended caloric intake of the individual or family and Health and Welfare Canada's (1983) Recommended Nutrient Intakes for Canadians is the standard used in this case.

Table 5-1 below displays the list of foods used in this study. The foods were selected for the variety of nutrients they provide and for their popularity and wide availability in grocery stores. The foods are categorized according to five as opposed to four food groups because fruits and vegetables are treated separately. The Canada Food Guide recommends four to five servings per day of fruits and vegetables of which at least two must be vegetables. Therefore, there is really a unique constraint for each and that is why they are disjoined. Table 5-1 includes three foods which do not belong to any of the food groups. Margarine and sugar are considered "complementary" to foods in the Breads and Cereals category and would normally be part of the food budget. Salt is the most popular seasoning and a good source of iodine.

Table 5-1: Food Cost Chart — 1988
		Serving amount
Item	Purchase Size	EP*	AP**	Jan. price ($)	Cost per unit ($)	Cost per serving ($)	Cals per serving
Milk—fresh (2%)	4,000 ml.	250	250	2.99	.0007	.1869	128.00
Milk—powdered	25,000 ml.	250	250	11.85	.0005	.1185	90.00
Milk averages					.0006	.1527	109.00
Blade roast	1,000 g.	90	220	3.73	.0037	.8206	221.00
Stew beef	1,000 g.	90	131	5.05	.0050	.6616	203.00
Reg. gr. beef	1,000 g.	90	126	4.39	.0044	.5531	260.00
Liver—beef	1,000 g.	90	113	2.99	.0030	.3379	196.00
Pork roast—PS	1,000 g.	90	248	4.17	.0042	1.0342	220.00
Chicken—whole	1,000 g.	90	225	4.39	.0044	.9878	151.00
Tuna—canned	198 g.	90	90	1.59	.0080	.7227	177.00
Cheddar cheese	1,000 g.	60	60	8.57	.0086	.5142	241.00
Eggs—large	12	2	2	1.09	.0908	.1817	158.00
Peanut butter	1,000 g.	64	64	2.65	.0026	.1696	380.00
Meat averages					.0135	.5983	220.70
Cabbage	1,000 g.	37	37	1.29	.0013	.0477	9.00
Carrots	2,270 g.	58	58	1.49	.0007	.0381	25.00
Celery	1,000 g.	64	64	1.79	.0018	.1146	10.00
Lettuce	750 g.	30	30	1.99	.0027	.0796	5.00
Onions	2,270 g.	85	85	1.49	.0007	.0558	29.00
Potatoes	4540 g.	125	166	1.79	.0004	.0654	137.00
Tomato Juice	1,360 ml	125	125	1.19	.0009	.1094	22.00
Vegetable averages					.0012	.0729	33.86

The second last column gives the cost per serving for each of the foods listed. All food quantities are measured in terms of servings so this value is very important. Its determination is quite straightforward. For each food there is a quantity which constitutes one serving. For example, one serving of milk is 250 ml., one serving of meat is 90 g., one serving of vegetables is 125 ml., and so on. The prices used were those prevailing during the relevant week at only one food store. [The third week in January , 1988, at Food City, a large Ontario grocery chain.] Dividing the price by the container (or purchase) size and multiplying by serving size, we obtain cost per serving. With some foods cooking is required and in those cases the calculations incorporate the distinction between a serving as eaten and a serving as purchased.

In addition to cost per serving, the calories per serving associated with each food is given in table 5-1. Then for each food group, both average cost per serving and average calories per serving have been calculated. It is these average values that are used in the linear program. The implicit assumption is, of course, that within each food group, the various foods are alternated on a roughly equal basis over a one week period. As this happens the averages are accurate representations of costs and caloric intake.

The model

As pointed out above, our problem can be formulated as a linear program. The model, in general form, consists of an objective function (the function to be maximized or minimized) and one or more constraints. Thus,

Minimize

where Z = the total cost of food for the week
C_i = the cost per serving of food i
X_i = the number of servings of food i
n = the number of foods

subject to:

    (1) Balance constraints:

Xi > N_i for all i

where N_i = required minimum number of servings per week of food i

    (2) Energy constraints:

where E_i = calories per serving of food i
and T = total number of calories required for the week

The general model can be applied to any number of specific situations. In this study seven different foods (n=7) are used: milk, meats (including alternatives), vegetables, fruits, breads, margarine and sugar. Salt is excluded because it generates no calories and there is no obvious guideline to its use. Its cost can be added in at the end. The Cis and Eis are calculated in table 5-1 (averages). Once family composition is specified (i.e., ages and gender), the Nis and T are easily determined by consulting respectively the Canada Food Guide and Average Energy Requirements.

The results

Prices for the foods in table 5-1 were collected for six consecutive weeks during the first quarter of 1988. It is worth emphasizing that only actual weekly prices at one store were used. No saving strategies of any kind were employed.

The problem was solved initially for the "standard" family of four consisting of two males aged 36 and 13 and two females aged 33 and 9. According to the Canada Food Guide, the minimum required number of servings per week in this case for milk, meat, vegetables, fruit, and breads are 70, 56, 70, 56 and 112 respectively. [In all cases where a range of servings is given in the Food Guide (i.e., milk for children up to 11 years = 2-3 servings), the midpoint of the range was used as the relevant value.] The quantities of margarine and sugar are assumed to be constant at 140 servings and 70 servings per week respectively. [For the purpose of this research, consumption of margarine is assumed to be 5 servings per person per day. This amount is roughly related to the average consumption of bread and to a lesser extent potatoes. One serving of margarine is 5 ml. Similarly, the consumption of sugar is assumed to be 2.5 servings per person per day, roughly related to consumption of cereals. One serving of sugar is 5 ml.] The required weekly energy intake for this family is 65,100 calories. The linear program was solved for these values and the results are summarized in table 5-2.

Table 5-2: Minimum weekly food costs for a family of four, January/February, 1988
	Serving quantities per week		Costs
Food Category	Food Guide minimum	Optimal solution	Week 1 ($)	Week 2 ($)	Week 3 ($)	Week 4 ($)	Week 5 ($)	Week 6 ($)
Milk	70	70	$10.69	$10.69	$10.69	$10.69	$10.69	$10.69
Meat	56	56	33.50	36.79	31.33	34.27	35.03	32.47
Vegetables	70	70	5.10	4.84	5.06	5.06	4.78	5.06
Fruit	56	56	7.90	7.55	8.25	8.36	7.71	8.36
Bread	112	380.77	14.47	14.17	13.86	13.86	12.91	13.33
Margarine	—	140	3.07	3.07	3.07	3.07	3.07	3.07
Sugar	—	70	.24	.24	.24	.24	.18	.18
			74.97	77.35	72.50	75.55	74.37	73.16

The optimal solution could have been anticipated. Since we wish to minimize the cost of nutritious eating, it should be intuitive that we fulfill the minimum quantities of the relatively more expensive foods (milk, meat, vegetables and fruit) and load up on the relatively inexpensive food (breads) until caloric requirements are met. Indeed, I believe most low income families understand this point without having to formally optimize and their consumption would generally reflect this pattern. The real problem is likely to be a failure to consistently meet the minimums in the non-bread categories.

On the basis of the six week experiment, the minimum cost of a nutritious diet for a family of four during the first quarter of 1988 is a bit under $75 per week. This comes to about $3,882 for the year. [Excluding the cost of salt. The consumption of 1 kilogram of salt per person per year at the 1988 prevailing price of .99/kg would keep the cost of eating nutritiously under $3,886 for a family of 4.]

In order to catch the effects of possible seasonal price changes on costs, the six week experiment was repeated during both the second and third quarters of 1988. The summary results are contained below in table 5-3.

Click here to view Table 5-3: Minimum weekly food costs for a family of four, 1988

The average weekly cost during the second quarter period was $75.52, only slightly above the first quarter value. The rise in many food prices during the summer of 1988 is reflected in our July/August period average weekly cost of $78.85—about four percent higher than the first quarter value.

Finally, the model was applied to other family structures commonly cited in poverty literature. Nis and Ts were determined and optimal solutions calculated for the following types:

a)single female parent under 35 with two children, ages two and three

b)elderly couple, both over 65

c)single elderly female

The results just for the July/August 1988 period are presented below in table 5-4.

Table 5-4: Minimum weekly food costs (various family types), July/August, 1988
	Family type
Food	Family of 3 (single mother)		Family of 2 (elderly couple)		Family of 1 (single elderly female)
	Servings	Cost ($)	Servings	Cost ($)	Servings	Cost ($)
Milk	49.00	$7.82	28.00	$4.47	14.00	$2.23
Meat	42.00	26.51	28.00	17.68	14.00	8.84
Vegetables	52.50	4.00	35.00	2.67	17.50	1.33
Fruit	42.00	6.34	28.00	4.23	14.00	2.11
Bread	87.18	3.23	154.24	5.72	56.63	2.10
Margarine	105.00	2.37	70.00	1.58	35.00	.79
Sugar	52.50	00.19	35.00	00.13	17.50	00.06
Total calories	31,500.00		28,700.00		12,600.00
Total cost		$50.47		$36.47		$17.48

For comparison purposes, the following list contains our cost estimates along with those of each of the "food baskets" for the reference family of four for the period of July/August 1988:

	Annual Cost ($)
Sarlo	4,052
Agriculture Canada	6,075
Montreal Diet Dispensary	5,360
Social Planning Council*	7,000
*This is the 1986 SPC estimate updated to 1988 using the "food" component of the CPI.

There is a 50 percent difference [For a family of 4 in the city of Toronto during April/May of 1988.] between my estimate of the cost of a nutritious diet and that of Agriculture Canada. This is not surprising. The purpose and methodologies are fundamentally different. There can be a wide variety of costs consistent with good nutrition. This study is unique in examining the lower bound of that range while incorporating consideration of variety, palatability, balance and energy into the problem. Because no shopping strategies, substitutions, coupons, etc. have been used in the analysis, it is reasonable to conclude that it would be relatively easy for the typical family to achieve the determined minimum costs. No effort is required to economize. [Clearly many low income families do make an effort to economize—because they have to. While the assumption made here is not realistic, it is necessary as a starting point in determining essential food costs. It would be an interesting and useful exercise to examine the effect on food costs of relaxing this assumption. This is done later in the chapter.]

The typical family described in this study makes one or two trips per week to one grocery store. However, the family is not buying every item every week. For example for the family of four a bag of potatoes lasts about three weeks, a bag of rice two weeks, a bag of apples less than two weeks and a bag of sugar about two months. Only bread and some fresh meats may need to be purchased more frequently than once a week. While many urban families live within a short distance of a grocery store there will be some for whom transportation costs start to become meaningful. However, it would be inappropriate to include these costs in with food costs.

The "serving" is the base unit in this study, however, most foods are not purchased by the serving. The purchase sizes are given in table 5-1. In most cases it is quite easy to subdivide the purchase size into serving sizes. For some foods such as roasts, it may be necessary to have a small freezer (within a refrigerator) to store portions for a short time. Nevertheless, the distinction between the purchase size and serving size presents no obvious barrier to cost minimization for the family of four. It may, however, present a problem for the other family types. The purchase sizes of some vegetables and fruits may be too large for smaller families to consume within the perishability limit. Staggering [Staggered use is the most reasonable solution. The same variety is achieved but over several weeks instead of each week.] the use of these vegetables and fruits or splitting the purchase size among two or more families are possible solutions. The other alternative is to purchase a smaller size and typically pay an increased cost per serving. It is worth noting, however, that some vegetables simply do not come in smaller sizes, i.e., a head of cabbage, a head of lettuce, a head of celery. This problem is a persistent one for smaller families and is not unique to this study.

At first glance, the bread group aggregates appear to be excessive. However, when the totals are broken down on a per person per day basis the quantities do not seem to be unreasonable. For the family of four, on average, each person consumes 13.5 servings from the bread group per day. [Somewhat more for males, somewhat less for females.] This might consist of a cup of cooked cereal (or 1.5 cups of corn flakes) and two slices of toast at breakfast, two sandwiches at lunch and 1.5 cups of cooked rice or macaroni and two slices of bread at dinner. The single elderly female consumes just eight servings of breads and cereals each day.

It should be relatively easy to visualize a series of menus that can be generated from the foods listed in table 5-1. These basic foods permit a wide range of standard and ethnic fare. However, for all its appeal, developing and costing menus is not a rigorous approach to the problem! [I have elsewhere (Sarlo, 1987), utilized the menu approach to examine the issue of the essential cost of good nutrition. In spite of the difference in methodologies between that study and this one, the results are remarkably consistent. The (low) cost of a nutritious diet for a family of four was found to be approximately $3,750 in mid 1987.] A greater degree of subjectivity is introduced (including the likelihood of cultural bias in the selection process) and trial and error is required to search for minimum costs. The methodology used in this study with the explicit incorporation of an optimization procedure is far more elegant and precise.

It can be argued that the list of foods in this study is too narrow. There may well be sufficient variety to ensure good nutrition, but there may be insufficient variety to avoid monotony, at least for some. Would expansion of the food list significantly raise the cost of a nutritious diet? To examine this question, an additional 22 foods were selected from Agriculture Canada's Nutritious Food Basket list along with their typical mid-summer prices. Only four of the items were at their special, "in season" prices. The additional food list along with package sizes, prices, calories per serving and costs per serving are given below in table 5-5. The additional foods were merged in with the original group and the linear program re-run with the revised average values. The new weekly minimum cost of eating nutritiously for a family of four is $75.40—about four percent less than the original result. This experiment demonstrates that the variety of the diet can be significantly expanded without adversely influencing costs. Clearly, new foods cannot be added indefinitely without costs rising. It is equally clear that a wide range of foods give us consistent cost results.

Table 5-5: Additional food list
Item	Serving Size	AP*	Calories per serving	Jul-Aug 1988 price ($)	Cost per serving ($)
Ice cream	125 ml.	125 ml.	142	2.32/2 l.	.1450
Wieners	2	2	236	1.90/12	.3167
Cooked ham	90 g.	90 g.	163	5.83/kg	.5247
Baloney	90 g.	90 g.	286	2.95/kg	.2655
Turkey	90 g.	225 g.	148	4.25/kg	.9563
Fish sticks	3 sticks	3 sticks	158	4.49/28	.4811
Sardines	90 g.	90 g.	183	.79/100 g.	.7110
White beans	250 ml.	125 ml.	230	1.53/ 907 g.	.1603
Sausages	3 links	3 links	177	3.29/16	.6169
Average			198		.4843
Tomatoes	123 g.	123 g.	23	.62/kg	.0763
Spinach	30 g.	30 g.	7	3.10/kg	.0930
Frozen fr. fries	50 g.	50 g.	111	1.49/kg	.0745
Canned peas	125 ml.	125 ml.	62	.49/ 398 ml.	.1539
Canned beans	125 ml.	125 ml.	15	.49/ 398 ml.	.1539
Average			44		.1103
Canned peaches	125 ml.	125 ml.	100	1.79/ 796 ml.	.2811
Canned pears	125 ml.	125 ml.	100	1.79/ 796 ml.	.2811
Grapes	85 g.	85 g.	60	1.96/kg	.1666
Plums	66 g.	66 g.	36	2.07 kg	.1476
Strawberries	125 ml.	125 ml.	24	2.00/kg.	.1600
Average			64		.2073
Whole wheat bread	1 slice	1 slice	63	.99/24	.0413
Flour	40 g.	40 g.	146	2.59/ 2.5 kg	.0414
Wiener bun	1/2 bun	1/2 bun	75	.68 / 8	.0431
Average			95		.0419
*AP = as purchased

In 1986, the average family of four (two adults, two children) in Canada spent $6,049 on food, of which 21 percent was spent in restaurants. The estimates provided here, without resort to any economizing strategies, show that the same family of four can eat a nutritious, palatable diet for about $4,000. To say that Canadians are spending more than they need to on food is merely to say that Canadians eat more red meat than they need to and fewer fruits, vegetables, and whole grains than they should. Nutritionists have long advised that a healthier diet is a less expensive diet.

It is important to stress that food shoppers require no optimization model nor do they need to use any cost saving strategies in order to achieve the food cost levels determined here. They need merely shop at a grocery store on a regular basis, select a variety of tasty, nutritious foods, avoiding junk food and only the few top-end items within each food group. Costs can, of course, be further reduced by practicing a bit of home economy. During 1990 and 1991 Canada's major newspaper services (Toronto Star, Southam, CP, etc.) regularly published articles advising food shoppers on various ways to stretch their food dollars. However, none of that is assumed here.

People who have lived most of their lives in the comfort of a middle class living standard have a tendency to lose sight of basic costs of living. There are now and have always been people forced by circumstance to make do with tight budgets. There are students, people on welfare, low income immigrant families, etc., who manage to eat nutritiously for considerably less than that spent by most middle class families of the same size.

One important limitation of the study is that no allowance has been made for either storage or "plate" waste. [The loss due to cooking as well as bone and fat waste has, of course, been accounted for in the calculations.] While it is reasonable to expect low income families to minimize waste, some of it is unavoidable. The amounts on average may not be insignificant. Surprisingly, food science and home economics literature is devoid of any reliable estimates of these losses. Since any attempt to account for storage or plate waste would amount to guessing, they have been implicitly set to zero for this study. The effect on family food costs of such considerations as plate and storage waste and regional price differentials on the one hand, and home gardens, shopping strategies, and gifts of food on the other hand requires serious examination.

Techniques for calculating the least costly way of achieving some objective are widely used in economics. There is no reason why these techniques should not be applied to very human activities such as the purchase and consumption of food. [Indeed, economic analysis has been useful, in recent years, in helping us understand human behaviour and decision making in such areas as marriage, procreation, and religion.] Food is the fundamental necessity and we should certainly be able to determine the minimum cost of supplying basic nourishment and nutrition to families. This study provides an objective and easy-to-use methodology for such determination. This methodology can be adapted to any family type including families with special nutritional requirements. Indeed, it can readily be modified to incorporate assumptions other than the ones I have made, such as the use of saving strategies and the inclusion of additional nutritious foods. The methodology, therefore, is at least as important as the results.

Recent statements by well known politicians illustrate why it is important to accurately establish essential food costs. In July, 1987, former Liberal leader John Turner commented on a proposed sales tax on food and its effect on low income Canadians by stating, "Just to meet basic nutritional requirements, a family of four must spend $5,668 a year on food." [Toronto Star, July 4, 1987.] Mr. Turner is using the Nutritious Food Basket estimate (Canadian average) as if it were a minimum requirement. This distortion was either intentional or else Mr. Turner did not understand what the Nutritious Food Basket represents. In any case, the statement is incorrect. This study, using 1988 (first quarter) prices, has shown that a nutritious diet for a family of four costs about $3,900, a difference of 45 percent. In September, 1986, a report on Ontario nursing homes revealed that the average expenditure on food for elderly residents was $2.70 per day. Bob Rae, Ontario NDP leader, called this (low) amount "a scandal" and said it was time somebody blew the whistle. [Toronto Star, October 22, 1986.] Again, this study using 1988 prices demonstrates that an elderly couple and a single elderly female, shopping at the retail level would have no problem eating nutritiously for less than $2.70 per person per day.

There is great value in establishing the minimum costs of providing good nutrition to families and individuals. This information would be of practical value to poverty researchers, among others. It is too important to be determined by ad hoc estimates. This research is an attempt to provide a reliable approach to the problem.

Shopping strategies

In determining the minimum cost of a nutritious diet, no saving practices have been employed. Families are assumed to make regular trips to one grocery store and purchase a list of food items in given quantities at the prices prevailing that week. Absolutely no strategic behavior involving stocking up on sale items, substitutions, purchase of marked down items that are day old, end of day or slightly damaged, or couponing was permitted. While this is a good starting point for the study of necessary food costs, it is unrealistic to leave it at that. Many low income families practice shopping strategies because they have to. They recognize that they can enjoy non-trivial increases in living standards by employing some quite simple procedures.

The gains from such shopping strategies can be easily estimated once our optimization procedure is in place. Using our reference family of four, we examined two common practices and measured the savings in food costs as each was employed.

The first strategy involved modest "stocking up" of items on special. Specifically, we assume that the consumer buys two weeks' worth of sale priced perishables (such as fruits, milk, meat and bread) and four weeks' worth of sale priced non-perishables (canned, dry, or frozen items). To take full advantage of savings due to stocking up on perishables, the use of a freezer or, more likely, making two rather than just one weekly trip to the grocery store would be necessary.

The second shopping strategy utilizes some substitutions. Specifically, the three least expensive items in the fruit and meat entree categories are double weighted each week as the family is assumed to increase use due to their relative inexpensiveness. It is important to note that for practical purposes, vegetables have been excluded from both shopping strategies. Normally, in one week, a family of four consumes about two-thirds of a head of celery and about half of each of a cabbage and a lettuce. They could not feasibly stock up or substitute into these items within the perishability limits. The same problem applies, to a lesser extent, to carrots, onions, and potatoes. Therefore, for most families, there is limited scope for shopping strategies in fresh vegetables.

The results of re-running our linear program with costs adjusted to account for shopping strategies for three separate time periods between January, 1988 and March, 1989 show that, on average, a family of four that consistently employed a stocking up strategy could reduce their food expenditures by about four percent or about $160 per annum. To the extent that they were able to utilize the substitution strategy in combination with "stocking-up," the saving would be about seven and a half percent or about $300 per annum.

Clearly smaller families (size three or less) will be less able to use these saving practices and therefore their percentage gains will be smaller. In addition, rural families will typically be unable to fully utilize these strategies. However, it is equally clear that many poor families can and do use other savings practices such as couponing, comparison shopping and garden plots [It might be assumed that the poor could make very limited use of gardens. However, as we will see in the profile of the poor, fully one-third of poor families are homeowners. Garden plots are not an insignificant consideration.] which could further reduce food costs.

The minimum cost of a nutritious diet has, so far, been estimated only for Ontario. The fact is that food prices vary, sometimes substantially, in different regions of the country. Additional information is required to estimate essential food costs across Canada.

Statistics Canada is the only source of easily available and reliable food price data. They publish quarterly the "city average retail prices" of some 50 food items for 25 major cities in every province and region of this country. [Statistics Canada, Consumer Prices and Price Indexes, cat. 62-010.] While the list of foods in table 5-1 substantially overlaps with the Statistics Canada list, direct comparability is limited by two considerations: the use by Statistics Canada of just "name" brands for a number of items and their use of inappropriate sizes of some items.

Statistics Canada does not price store brands or generic equivalents of items such as cheese, peanut butter, tomato juice, orange juice, apple juice, macaroni and cornflakes. Our list made no such exclusion and permitted the low income shopper to select the least expensive equivalent. In addition, Statistics Canada has selected uneconomical sizes of certain items, such as milk, cheese, peanut butter and macaroni. In the case of milk, the most perishable of the items listed, the purchase of just one litre at a time is expensive and clearly unnecessary. Most families, whatever their income, have been purchasing fresh milk in the four litre plastic bags throughout the 1980s. Even a single person can consume a four litre bag of milk within the usual perishability limit if they derive most of their dairy intake from this source.

In spite of substantial mismatching, there remains 13 key foods which are directly comparable with the Statistics Canada list. Table 5-6 below lists these 13 foods with size of purchase in parenthesis and displays the July 1988 prices of each food for the 25 cities that Statistics Canada surveys. At the bottom of the table are the corresponding prices at Food City, the Ontario grocery chain from which this study's prices have been drawn. While the prices of many individual items varied substantially, the overall cost of the aggregate "basket" showed much less variation between cities. The cost of this basket (sum total of the 13 foods) is a fair representation of food costs in each community as it includes popular foods from each of the food groups except dairy. The total cost of the bottom row is referred to as the Ontario base.

Table 5-6: Comparative food prices in various Canadian cities ($ cost per unit of weight) — July, 1988
	(kg)	(kg)	(kg)	(kg)	(doz.)	(kg)	(kg)	(4.54 kg)	(kg)	(kg)	(kg)	(675 g)	(2 kg)
Cities	Blade roast	Stew beef	Ground beef	Chicken	Eggs	Cabbage	Celery	Potato	Lettuce	Banana	Orange	Bread	Sugar	Total   Cost
St. John’s	5.07	6.25	3.93	4.05	1.69	.87	1.49	2.89	1.23	1.46	1.78	.78	1.39	32.88
Charlottetown	4.18	6.03	3.72	3.35	1.52	.92	1.89	1.89	1.52	1.16	1.55	1.11	1.25	30.09
Sydney	4.10	6.48	3.89	3.65	1.60	.91	2.26	2.22	1.74	1.48	1.63	1.11	.98	32.05
Halifax	3.66	5.99	3.62	3.20	1.48	.99	1.77	2.05	2.10	1.51	1.74	.98	1.15	30.24
Moncton	4.21	6.44	3.42	3.42	1.58	.86	1.58	1.85	1.35	.95	1.77	1.05	1.10	29.58
Saint John	3.90	6.59	3.35	3.09	1.57	.98	1.58	2.11	1.19	1.12	1.55	.97	1.08	29.08
Chicoutimi	5.87	7.74	4.16	4.39	1.37	1.29	1.70	2.76	.98	1.74	1.94	.99	1.66	36.59
Quebec	6.95	7.00	3.32	3.73	1.37	.95	1.21	2.13	1.60	1.30	1.96	1.18	1.49	34.19
Trois-Rivieres	4.74	6.90	2.74	3.73	1.52	1.00	1.15	2.43	1.12	1.38	1.95	1.18	1.79	31.63
Sherbrooke	3.93	7.13	3.67	3.81	1.54	.89	1.32	1.64	1.61	1.46	2.05	1.14	1.84	32.03
Montreal	4.68	6.26	3.76	3.98	1.55	.94	1.22	2.41	1.42	1.32	2.01	1.27	1.94	32.76
Hull	4.85	4.70	3.03	2.28	1.13	1.07	1.20	2.88	1.08	1.03	1.96	.83	.89	26.93
Ottawa	4.71	5.00	3.16	2.91	1.09	.88	1.24	2.64	1.53	1.08	1.94	.85	.89	27.92
Toronto	4.61	5.79	4.02	3.89	1.26	.83	1.11	3.34	1.33	1.36	1.85	.94	1.42	31.75
Hamilton	4.42	5.80	3.86	4.09	1.05	.97	1.22	3.54	1.29	1.22	1.69	.80	1.00	30.95
London	4.62	5.95	4.21	4.05	1.19	.81	1.25	3.14	1.68	1.49	1.97	1.01	1.23	32.60
Sudbury	4.46	5.21	3.74	3.29	1.16	.82	1.29	2.69	1.27	1.36	1.95	.88	1.35	29.47
Thunder Bay	4.91	6.50	3.41	3.18	1.43	.92	.99	3.60	1.38	1.48	1.95	1.24	2.03	33.02
Winnipeg	4.26	5.08	2.09	3.74	1.02	.78	.88	2.44	1.44	.80	1.59	1.32	1.40	26.84
Regina	3.02	5.31	2.06	3.72	1.16	.78	.98	3.99	1.32	1.00	1.73	1.41	1.55	28.03
Saskatoon	2.65	5.41	2.01	3.95	1.07	.94	1.23	3.94	1.36	.93	1.67	1.34	1.55	28.05
Edmonton	2.82	6.10	2.64	3.91	1.17	.90	1.63	3.09	1.91	.95	1.97	1.27	1.52	29.88
Calgary	3.48	5.74	2.77	4.41	1.44	1.14	2.00	2.94	1.77	1.48	2.04	1.14	1.46	31.81
Vancouver	5.32	6.29	2.82	2.93	1.37	.62	1.03	3.07	.90	1.42	1.59	1.46	1.72	30.54
Victoria	5.52	6.13	2.96	3.35	1.40	.69	1.16	2.63	1.08	1.46	1.55	1.45	1.75	31.13
Ontario (Food City)	5.59	6.59	3.73	2.18	1.09	.99	.99	3.99	1.29	1.52	2.33	.69	1.69	32.67
Sources: Statistics Canada, Consumer Prices and Price Indexes, cat. 62 – 010, July-September 1988; Food City, North Bay, Ontario, July 2, 1988.

Using food prices derived from our one Ontario grocery chain, we calculated essential food costs for families of size one to six. The results, contained in table 5-7, are referred to as "base" food costs. It is in relation to these costs that essential food costs in other cities will be determined. A variety of age-gender compositions were considered but in each case the most expensive result was selected. Since the typical composition of a poor family is unknown, it is best to err on the high side.

Essential food costs by family size for the major Canadian cities were estimated as a proportion of the corresponding "base" costs. The proportions used represent the ratio of the cost of the basket of comparable foods table 5-6) in the major city to the cost of the same basket in the Ontario base. Table 5-8 presents these results and as well, the corresponding ratio (R) for each community. The values in the table represent the minimum cost of a nutritious (and palatable) diet for the major metropolitan areas in all regions of Canada (excluding Yukon and N.W.T.) by family size. It is noteworthy that these metropolitan areas comprised slightly more than 60 percent of the population of Canada in 1986.

Click here to view Table 5-8: Essential annual food costs—1988 ($)

Some critics will argue that there is a much simpler way to determine food costs relevant to low income families. Since Statistics Canada conducts family expenditure surveys every four years, why not just use the food expenditure results of, say, the lowest quintile, as representative of required expenditure by the poor. The fundamental problem with this approach is that we have no way of knowing whether average food expenditures of this (or any) quintile will result in a nutritious diet.

Evidence that a significant proportion of school children, from a variety of backgrounds, are malnourished [A 1988 study for the Waterloo Region Separate School Board found that less than a third of elementary school children are properly nourished and that lifestyle is more at fault than poverty. See Toronto Star, September 8, 1988.] suggests that income and a healthy diet are not related. It would be inappropriate to use actual food expenditures to represent the essential costs of providing a nutritious diet.

Is it possible that those with low incomes shop more frequently at convenience and corner stores and thereby need to pay significantly more to eat than their more well-off counterparts? The food expenditure surveys conducted by Statistics Canada reveal no evidence of this. Overall, only 5 percent of unprepared (i.e., non restaurant) food purchases in Canada are bought from convenience stores. [Statistics Canada, Family Food Expenditure in Canada 1986, cat. 62-554, table 3, p. 22 ("Average Weekly Food Expenditure, 1986, All Families and Unattached Individuals").Note] Those in the lowest quintile of the income distribution spend a slightly higher proportion (8 percent) but a lower absolute amount of their non restaurant food dollars at convenience stores. [Ibid, table 6, p. 30.] There is, in addition, no evidence that when they do shop at convenience stores, they do so because they have to.