| Behind The Store Front
Prepared for the Retail Council of Canada in partnership with Industry Canada By Jacobson Consulting Inc.
The Retail Sector |
The literature on factors impacting consumer expenditure is one of the deepest troves of analysis in the field of economics. Much of the analysis is developed around a model of consumption decisions being made in a household context strongly influenced by factors such as disposable income, interest rates, relative prices and expectations about changes in such factors. The next chart (Figure 18) highlights the broad relationship between income and retail sales showing the average annual growth rate of sales and income over the period 1999-2004.
Generally, regions with above average growth in income also show a similar pattern for retail sales. However, the detailed underlying influences are more complex than simply income. The demographic structure of the economy strongly influences the character of household expenditure. Consumption patterns are expected to be different for older persons than younger ones, not just because of taste but also because of the impact of changing needs (life cycle) and past expenditure history. For example, households with older persons may often be considering replacement purchases of durables, rather than initial purchases. Households with a younger household head might possibly have children emphasizing other differences in the structure and frequency of purchases.
The Survey of Household Spending (SHS) summarizes the complete income and outlay patterns of Canadian households. For this analysis, the data for respondents has been categorized into quintiles by income. Simply put, the data for the households was ranked by income and divided into 5 equally-sized groups, or quintiles.
There are strong relationships between income and household structure. The next chart (Figure 19) shows some of the key demographic values for SHS 2003.
The solid line, plotted on the right axis, shows the number of persons in each household by quintile (20% of the sample, ordered by income). The left-most point is the all-households average. The bars, plotted on the left axis, show the portions of each group having a specific demographic characteristic. Households at the lower end of the spectrum are likely to look much different than those at the upper end.
- Household size rises with income.
- Probability of children increases with income.
- Lowest income households are dominantly comprised of single earners, perhaps seniors.
- More than half of households in the top quintile have two or more full-time wage earners.
Additional demographic details are presented in Table 26 in the statistical appendix. The next chart (Figure 20) summarizes some key relations from the 2003 SHS. Two comparisons are shown. The top bar in the chart shows the values for key spending characteristics as a ratio between the highest and lowest quintiles. The second bar shows the ratio of the values for the top quintile and the average of all households.
The chart shows that households in the top quintile enjoy an income more than 8 times higher than the bottom group. However, the top group spends only slightly more than twice as much on food from stores. This is approximately similar to the difference in household size. A similar relationship is seen for household cleaning supplies.
Obviously, not all households buy all products. The average household spends more than $5000 per year on food but less than $400 on household appliances. The low value for appliances is attributable to the fact that everyone buys food but only a few households replace their appliances each year. Every household buys food. Statistically, this means that the incidence of food expenditure by households is more or less 100 per cent. However, food expenditure does not rise proportionately with income. Average household expenditure for a broad range of goods and services is shown in a table in the appendix.
Not all households buy a car every year. In other words, the incidence of household expenditure on capital items is relatively low. However, it rises with income. The relative average value of expenditure on cars exceeds the similar ratio for household income. However, for most categories, the relationships between the top and bottom for expenditure are not as high as the relationship for income. Details of expenditure incidence are also provided for selected categories in the appendix.
Generally, as income rises, households are more likely to spend money on a specific category. The next chart (Figure 21) shows the incidence of expenditure for three categories: household furnishings, prescription medicine and entertainment electronics such as TVs. Presumably because of the availability of health insurance and because you don't get a prescription unless you need it, the incidence of buying drugs is not strongly tied to income. This contrasts with the other two categories which show the expected relationship.
Because of its capital nature, everyone doesn't buy a TV every year, the incidence for purchasing such equipment is much lower than the incidence of buying some goods for the household.
Retail spending naturally rises with income, but it doesn't rise as quickly. As a result, the lower-income quintiles spend a larger proportion of their income on retail goods than do the higher-income quartiles. This can be seen in Figure 22. The expenditure data used this figure are created by aggregating purchases of retail commodities, essentially goods, for each quintile. The data are presented relative to several key measures:
- Current expenditure — approximately purchases of goods and services,
- Total expenditure includes tax payments and asset accumulation, almost income disposition, and
- Income.
Note first that total expenditure as a share of income falls from the lower quintile to the upper quintile. This is because, as income rises, more of it is devoted to saving. Next, note that retail spending as a share of total expenditure also declines with income. This reflects the fact that spending on services — as opposed to spending on retail goods — rises with income. Thus, retail spending not only fails to keep pace with income but falls as a share of overall spending as income rises. At the lower quintile, this effect is compounded because many low-income households are headed by retirees who are drawing down some of their lifetime savings and so spending in excess of income.
Retail-related expenditure is roughly proportionate to current expenditure for most of the upper quintiles but falls in relation to income because more affluent households devote an increasing share of income to savings and other asset purchases. Current expenditure, spending on goods and services excluding renovations, exceeds household income in the bottom quintile because this group contains many seniors and others who are spending their capital to finance their shelter requirements. This is shown in Table 26, in the statistical appendix, by the negative asset changes for the lower quintiles. Only the upper quintile is able to accumulate significant assets. Total expenditure includes tax payments and asset accumulation and hence approximates income. Essentially, this analysis shows that retail services support a bigger portion of the spending activities of lower-income households than of higher-income ones.
Home renovation activity is a very specific household purchase and is considered an investment under the SHS classification. Only 38 per cent of households in the bottom quintile were likely to own their own home in 2003. In contrast, 93% of households in the top group owned their own dwelling. As would be expected, the next chart shows that both the incidence of the expenditure and the average expenditure rises with income, slightly more than proportionately.
Incidence is plotted as a line on the left axis. The first data point represents the average of all households, showing that almost 23% of households spent money on improvements to their homes. However, the average expenditure, plotted on the right axis, was only slightly more than $1300. The lowest quintile averaged roughly $340 expenditure per year with a very low incidence of only 8.2%. This can be interpreted as meaning that the average expenditure of the 8.2% of households in the lower quintile who actually did renovations was over $4100.8 Similarly at the high end, since only 37.8% of households reported renovation expenditure, the average expenditure of reporting households was actually slightly more than $8000 per household. Note, the top group spent only slightly more than twice the expenditure of the lowest group despite receiving 8 times the income.
The statistical appendix contains data for selected categories of expenditure showing average (Table 30) and median (Table 31) household expenditures. The distribution of household expenditure is slightly skewed for most categories. The median expenditure is slightly lower than the average for most expenditure categories. One of the exceptions is the renovation activity for which the gap between median and average is more significant. At the top end of the income spectrum, median reporting expenditure was $3400 but average reporting expenditure is slightly more than $8000. Essentially this means more than half of the households in top bracket who did any renovations spent less than half of the average amount of all in the class that undertook that expenditure.
8This is calculated by dividing $340 by .082.