When the economy tanks, as it has done during the COVID-19 pandemic, people typically cut back or reallocate spending on food, clothing, and other consumer goods. Shrinking income and wealth force consumers to revisit where and how they shop. Consumers switch from grocery stores to discount stores or warehouse outlets, rethink decisions about food and nonfood items, and move away from national brands toward private-label products, according to research.
But the combined effects of these shifts have been poorly understood, according to Chicago Booth’s Pradeep K. Chintagunta and Sanjay K. Dhar and Booth PhD candidate Shirsho Biswas. They argue that focusing on the individual elements of store-outlet choices, product-category trade-offs, and brand decisions can lead to misleading conclusions. In reality, consumers are playing three-dimensional chess, sometimes with surprising results.
This matters because merchants looking at one or two dimensions can make bad purchasing and pricing decisions, the researchers write. For example, they demonstrate that as consumer housing wealth declined during the Great Recession, the joint share of purchases of private-label goods at warehouse-club stores went down even as combined expenditure share of private-label items at all stores rose. Choice of shopping venue altered consumers’ decisions on brands, the researchers suggest. If a warehouse-club manager expected demand for private labels to rise amid the downturn and consequently introduced more such products at higher prices, the manager might “have made a potentially incorrect decision,” they write.
Biswas, Chintagunta, and Dhar set out to develop a better understanding of how the three dimensions play out in consumer decision-making on stores, product categories, and brand levels. They used data from 2004 to 2014―before, during, and after the Great Recession—and tapped the Nielsen Datasets at Booth’s Kilts Center for Marketing. They matched household shopping trips with income and other demographic data using the Nielsen Homescan Consumer Panel, covering 150,000 households’ purchases of goods with more than 3.2 million universal product codes from 1,075 product categories, including health and beauty, dry grocery, frozen food, fresh produce, nonfood, and general merchandise. The Zillow Home Value Index enabled them to measure housing values by month and zip code as an indicator of household wealth.
The analytical models the researchers used to study macroeconomic effects on micro behavior will be important for marketers and researchers in a variety of settings.
To study store choice, the researchers tracked household-level share of purchases by store format, by product for food and nonfood categories, and by brand choice for private-label goods and national brands. They assessed household decisions on each of the three dimensions in isolation and in a cross-dimensional analysis.
The researchers find significant differences between the single-dimensional and multidimensional analyses during the recession. For example, food purchases rose as a share of spending, but the proportion of spending at grocery stores declined as consumers changed where they shopped. At the same time, people spent more buying private-label items at grocery stores.
They also argue that it’s not possible to use results from a single-dimensional analysis to predict how levels and trends change when it comes to different, joint combinations of stores, categories, and brands—whether the changes were caused by the Great Recession or are more generally due to income and wealth changes. This may be because consumer choices about budget allocation between stores, categories, and brands are not independent of each other, or because companies undertake marketing activities that affect consumers’ store, category, and brand decisions simultaneously.
The researchers also document that there can be large trends over time in the shares of different stores, categories, and brands. The share of money spent at grocery stores declined from 2004 to 2014, while the share spent at discount stores and warehouse clubs increased, and the shares that went to food and private-label products also rose. The impact of income and wealth changes also varied by store, category, and brand type and were large in magnitude relative to the year-on-year change in share, as captured by an annual trend in a few specific cases. For instance, the researchers find that a 10 percent drop in household wealth caused an increase in expenditure share in discount stores of about 25 percent of the magnitude of the annual trend. On the other hand, the corresponding change in spending on private-labels (versus national brands), foods (versus nonfoods), or grocery stores and warehouse clubs among store types was only about 15 percent of the change captured by an annual trend in these shares, the researchers report.
The researchers analyzed the effects of declining wealth on various combinations of store choice, product category, and brand level. The same finding as described above for the single-dimensional analysis holds. Changes in share caused by, say, a 10 percent shock to income or wealth varied for different combinations of stores, brands, and categories, but were generally small (less than 20 percent in magnitude) compared with annual trends in most cases.
The analytical models the researchers used to study macroeconomic effects on micro behavior will be important for marketers and researchers in a variety of settings, Biswas, Chintagunta, and Dhar suggest. These include measuring the impact of climate change on consumer travel behavior, the influence of landﬁll development on packaging choices by consumers and businesses, and the effect of rising inequality on individuals’ charitable contributions.