As a long-term investor in Consumer Alpha™ products and companies, Infusive researches the macro drivers of aggregate demand to anticipate where Consumer Alpha™ products and companies are likely to be found now and in the future.
To understand the purchase behavior of individual consumers over long periods of time, Infusive Intelligence is analyzing the US Bureau of Labor Statistics (BLS) Consumer Expenditure Survey, the most comprehen sive data set on buying habits, incomes and consumer characteristics worldwide. The BLS has run this survey annually since the early 1980s, generating billions of data points on hundreds of product categories. The survey informs the calculation of the US price index.
The BLS interviews tens of thousands of “consumer units” – ie households, which might be a large family or a single person – nationwide on quarterly expenditures. It collects several thousand records of all purchases made by a household during a two-week period. Our medium-term research objectives are threefold:
- Identify stable consumption patterns and preferences that connect specific demographic groups with specific product categories.
- Understand how consumer behavior changes when economic conditions change.
- Predict household-level and aggregate consumption over time and across dozens of product categories.
The complexity of both data and topic demands thatour research effort proceeds iteratively so that we build longer-term forecasts on a solid descriptive understanding of the US consumer landscape. As our research continues over coming months, Infusive Intelligence will share key illustrative findings with FUSE readers. We start this series with initial results that demonstrate the detailed insights that can be gained from this effort.
Relationships Between Households and Spending Patterns
Unlike cyclical factors that are difficult to forecast accurately over the long term, structural factors like disposable income and particularly demographics have a unique quality: they are predictable years in advance (with a high degree of precision). Stable relationships between the characteristics of a household and its spending patterns are key to long-term consumption forecasts.
As a starting point, we want to learn how a household’s purchase behavior depends on its income and age profile. To this end, we calculated product expenditure across age and income brackets and identified preferred product categories for given demographic groups.
We plotted the data on heat maps, and we include one series on this page. These heat maps show the preferences of each age and income group. For each one we show the percent deviation of a product category’s (like tobacco’s) share of total expenditure by the group in question from that category’s average share of total expenditure for all age and income groups. In other words, we show relative spending preferences adjusted for differences in absolute income. If a category accounts for 15% of a segment’s total expenditure whereas it only accounts for 10% of all consumer expenditures, the corresponding figure would be 50%.
Figure 1: Heat Maps of US Tobacco Consumption by Age and Income 1993 – 2013
Consumption of Tobacco: We observe a strong preference by households with heads who are between 45 and 64 years of age earning between $5,000 and $20,000. In other words tobacco is favored by lower-income, middle-aged consumers in the US. While this group has been the core demographic of tobacco consumption for the 20 years in question, we also observe a slight weakening of the youngest age group’s preference for buying tobacco. This may jeopardize future sales of tobacco if the weakness in demand persists, because younger consumers will age in to the tobacco industry’s target segment in a decade or two.
Consumption of Food Away From Home: We observe a strong preference by younger, lower-income consumers. Households headed by younger people at the lower end of the income scale buy more of their food out of the home – as a share of their total consumption – than any other age or income group.
Consumption of Entertainment: Two key groups show an outsized preference for spending on entertainment: upper income groups of all ages and lower income groups with household heads under 25. With basic needs taken care of, wealthier people of all ages spend more on movies, the theater, music concerts, and other forms of entertainment.
Case Study: Age Hotspots in Apparel Spending
Young adults – namely, 20 year-olds – spend an outsized portion of their income on apparel compared to all other age groups. This spending pattern, clearly visible in the graphs on the next page, is one preliminary insight of our research.
Our finding is especially notable because it holds for both men and women and extends across many apparel categories: men’s and women’s footwear, women’s dresses, men’s shirts, and many other types of apparel (we show two representative examples). Because the data extend across two decades, the pattern appears to hold across time and shifting generational tastes in clothing. One hypothesis for this pattern involves identity. Teenagers and young adults are often focused on finding their identity, and apparel is a material manifestation of that. Clothes are an expressive experiment in identity and are therefore worth the high cost (relative to income) to these young consumers. Consumers who are 45 to 50 years old spend the absolute most on apparel, although their spending on apparel relative to their total income is lower than the overall population’s.
Figure 2: Expenditure on Women’s Dresses by Age Group
Figure 2: Expenditure on Men’s Shirts by Age Group
Our Methodology and Scope of Research
Our analysis of youth spending on apparel is a small portion of the research we are conducting. We are working with a very large dataset. We estimate that the BLS Public Use Microdata Consumer Expenditure dataset contains over 2 billion individual data points. Significant time and effort is required to understand its structure and components – and process and arrange them into a workable form before analyzing the data and drawing conclusions. The main goal of our research using this dataset is to understand at the granular, individual level how consumers behave over time. This means analyzing spending on broad product categories such as food, housing, apparel, and entertainment as well as delving into Consumer Alpha™ subcategories such as coffee, confectionery, and alcoholic drinks.
Our research also entails analyzing consumer groups by age, income, wealth, location, gender, marital/familial status, race, and education, among other characteristics. Synthesizing this data on demographic groups, consumer products, and spending behavior will give us granular insights into what a wide range of consumers are buying. It will also demonstrate to what extent purchase behavior depends on group characteristics as opposed to individual preferences.
Because the data spans 1980 to 2013, we see the evolution of consumption across time.
The example of youths’ spending on clothing is an easy-to-understand result of analyzing complex matrices covering age, income, and products across time. In that example, spending on one product among people of a specific age held steady not only across socioeconomic groups and both genders but also across time. Infant underwear is another example of a product where spending patterns are stable over time. As can be seen in the chart on the next page, normalized spending on the category is more or less constant across age groups from 1997 to 2012. This makes sense: demand for infant underwear is defined less by in dividual preferences and more by necessity when people become parents, which tends to occur within the age range of 20 to 45.
But our research often yields more complex results. These results give us opportunities for multi-faceted insights. In cigarette consumption, for example, we observe several features that do not hold steady across time. It appears that a group of consumers in a particular generation had an outsized preference for cigarettes as a portion of their total spending. As time passed this group carried its preferences with them as it aged, causing the spikes in relative spending at different intervals over the years. This trend is evident in Figure 5.
Figure 4: Expenditure on Infant Underwear by Age Group
Figure 5: Expenditure on Cigarettes by Age Group
The Impact on Investment Structural shifts in demographics and income affect consumer spending, which in turn directly influences future industry revenues and earnings. Foreknowledge of these long-term shifts, which are often not priced in by financial markets despite being understood by some players in the market, has been shown to cause excess returns on investment.
Infusive is a resolutely long-term investor focused on the performance of consumer-facing companies over a horizon of 20 or more years.
Infusive Intelligence positions Infusive to take account of long-term structural factors mispriced by markets that are dominated by shorter-term investors. When Infusive analyzes companies from the top down, assessing their earnings and prospects in the context of industry peers, Infusive Intelligence provides a complementary perspective from the ground up. It assesses the demographic groups and product categories that are positioned to grow or decline. This helps assess the competitive strength of a company’s product portfolio and geographical exposure.
As Intelligence continues its research of BLS data over the next few months, we will share selected insights with FUSE readers.