Food price inflation for the rich and the poor: Evidence from grocery prices in Latvia
We use unique web-scraped data to study grocery price inflation in Latvia in 2022–2023.
We find that, in percentage terms, prices of cheaper items have increased more than of the expensive items.
Our results thus suggest that low-income households, which tend to buy cheaper products, may have experienced higher food price inflation than the high-income households during the studied period.
The price jump is not the same for all
In 2022 and 2023, prices in many parts of the world went up at a rate not seen for many years. Yet, there were noticeable differences among various goods and services, and among countries.
Looking at the European Union (EU), the Baltics, including Latvia, have faced one of the most dramatic inflation rates (Chart 1).
At the same time, headline numbers for Latvia, like those for most other member states, were largely shaped by skyrocketing prices of energy and food (Chart 2).
Such charts tell us that households may perceive and experience inflation in widely different ways. Aggregating data to produce a single figure for a country or even the whole European Union is useful. However, it masks the fact that each household is unique in what it consumes and how much it costs.
A two-children family living in Latvia and spending most of its income on essential needs like heating, electricity, and food (i.e., a low-income family) was hit by the rising prices much harder than, say, an affluent working-age couple living in Madrid. Even if the latter live in a bigger apartment and spend more on electricity and groceries, this spending likely makes up a smaller portion of their budget, and any increase in such spending is less noticeable. They can also cut back on it by switching from premium brands to regular ones or maintain the previous consumption by using their savings – a freedom that is unlikely for a low-income family.
What does the research say?
Measuring the differences in inflation rates experienced by the rich and the poor is important because it allows us to better understand social inequalities and design targeted support mechanisms. Indeed, it has been attempted many times.
The main problem with this task is data – we would only be able to obtain precise estimates if we could access every household's bank accounts, record all their purchased items, and calculate the price changes. That would be thrilling, although a bit time-consuming. Rough estimates, grouping families into income bins, are luckily easier to obtain.
Charalampakis and his colleagues from the European Central Bank (ECB) analyse the differences between the expenditure weights of the monthly HICP inflation data and those of the Eurostat Household Budget Survey (HBS) data (the survey focusses on households of various income groups and is conducted every five years). They find that, in September 2022, the annual effective inflation rate for the bottom income quintile exceeded that of the top quintile by about 2 percentage points. In simpler terms, the growth rate of prices for the lowest earners was, on average, 2 percentage points higher than the respective rate for the highest earners. This gap was the largest since 2006, contrasting the near-zero difference observed during the 2010s’ low inflation phase. Unsurprisingly, the surge was mostly driven by energy and food prices.
Allocation is not the only factor
It is clear that the rich and the poor differ in terms of their expenditure allocation and we can use that information to calculate the effective inflation. However, they also differ in what they buy for the allocated funds. A well-off family is likelier to buy expensive food, clothes, cars, and travel to exotic places than a family struggling to make ends meet.
Empirically, this has been established by Hallak, who finds that advanced economies import more high-quality goods, and Bems and di Giovanni, who confirm the relationship between income and product choice using shop scanner-level data covering the 2008 financial crisis in Latvia. More recently, Siliverstovs finds the same income-related quality preference in German household scanner data. In other words: when you have more money, you demand more quality, and it tends to cost more.
While this may seem like a statement of the obvious, it actually matters for the issue of inflation. Imagine that across a wide range of prices, for some reason, only prices of the cheapest goods are going up. Then, if we calculate the effective inflation by using different consumption weights from the HBS, and we multiply these weights by a same-for-all inflation index, we will get it wrong.
Although such extremes are unlikely, it is also improbable for prices to change at the same rate across the whole quality range. To determine personal inflation precisely, we need to use both weights and prices of the purchased items.
In this article, we try to discover the relationship between the prices and their inflation rates, focusing on food prices during the current inflation spike. We analyse online prices of two major supermarket chains in Latvia, and we find that cheaper food items have seen bigger price hikes in percentage terms in 2022 and 2023. In other words, the fact that poorer households spend more on food is exacerbated by steeper inflation of lower-priced food items.
Unique data set
During the pandemic, online shopping expanded significantly in Latvia. For monitoring and research, Latvijas Banka started recording online shop prices of two major supermarkets in 2021.
These data do not tell us what each person buys, but they show the daily price changes for many products. To get a grasp of it, Chart 3 presents an example of the price developments for two extra virgin olive oil brands. The products are real and the prices are taken from one of the shops, but the data are anonymised. One product is on the expensive side and the other one is among the cheapest.
Notice how both prices are increasing in a visually similar way. The cheaper olive oil remains cheaper. However, when we calculate the average prices for 2022 and for 2023 and compare the price increases in percentage terms, it appears that inflation of the expensive olive oil is less than half that of the cheap olive oil (Table 1). While the nominal increase in price per litre is comparable, inflation differs greatly, because this increase is compared with very different prices in 2022.
Table 1. Inflation of cheap and expensive olive oil.
|Brand||Average price (EUR/l)||Inflation (%; y-o-y)|
|2022||2023*||2023 vs 2022|
* Data for 2023 include the first 9 months.
Does this happen to all products? At this point, we cannot tell – this is just one example we handpicked on purpose. To be able to generalise, we can build upon findings in the literature and compile a shopping list for the poor and the rich by selecting the cheapest and the most expensive items in each category. Then, we can see how inflation affects both lists.
We retrieve data from two different sources, and we have to make sure the items match up. To do this, we group the items into 68 individual food product categories defined by the Latvian Central Statistical Bureau. These categories are easy to understand, like milk, wine, and butter.
The cheaper the item, the higher its price inflation
We focus on the highest and lowest 10% of prices in each category every month. For example, out of 100 prices, the 10th price is our "cheap" item benchmark and the 90th price is our "expensive" benchmark. We then turn these prices into an index, with the starting price as our base and inflation averaged across categories (Chart 4). We use the first month of data, March 2021, as a starting point of the ‘race’, which coincides nicely with early signs of accelerating inflation, as we show earlier in Charts 1 and 2.
What do we find? First, both the cheap and expensive items became substantially and gradually more expensive throughout 2022, with the prices rising by at least 30% compared with the base. The prices of cheap items peaked in March 2023 and those of expensive items – in January of the same year. Second, initial prices matter for inflation. By March 2023, the expensive items had risen by 33%, but the cheap ones had gone up by 53% – that is a 20 percentage point difference. Interestingly, after that highest point, the prices of the cheaper products started to come down, while those of the pricier ones remained broadly unchanged. This makes March 2023 a distinct peak of the mountain for cheaper products in Chart 4.
Our evidence confirms what was already reported by Which?, a UK magazine running their own price tracker. Simmonds, a researcher for the magazine, wrote in April this year that the own-label budget supermarket items were almost 25% more expensive than a year ago, compared with 14% for branded goods and premium own-brand ranges, suggesting a gap of 11 percentage points in the UK (March 2023 data).
Why did this happen?
One reason could be that people who buy pricier items have more options. If their favourite expensive item becomes too costly, they can just switch to something cheaper. But those who are already buying the cheapest option are stuck with it. This means that sellers can raise prices for cheaper products more confidently, knowing that people will still buy them.
Another reason may have to do with the size of profit margins. If stores enjoy higher margins when selling expensive items, they need not raise prices as much to preserve absolute profit when their costs go up. This explanation would still require retailers to sacrifice some of the relative profit (margin), which could be the case given the remarkable rate of inflation and price competition between the supermarkets. Moreover, a significant sacrifice in the upscale segment could explain why expensive items have not seen price drops with a sort of "margin recovery".
While our data cannot tell us exactly why this is happening, we believe it is still important to highlight the difference in how inflation affects cheap and expensive products. People with lower income might be hit twice: they spend more on food, and the prices of the food they buy are growing faster.
Are we confident in our findings?
Our method is somewhat simplistic. We are making a few assumptions to compare how inflation may affect households with different levels of income. We are assuming that households with more money buy pricier versions of all products, like wine, butter, or milk. In reality, however, a rich family may buy regular milk or maybe no milk at all. Further, we assume that every category has equal weight in grocery spending; it is unlikely to be the case. Plus, people may start buying cheaper items if their usual choices become too expensive.
Nonetheless, we construct a comparison of inflation experienced by average low-income versus high-income households, and we even out some of the differences on purpose. We have made sure our results are not thrown off by just one or two categories with massive inflation gaps. Since we do not know from the data how much of each product people buy, we believe it is reasonable to assume that every category has equal importance. The option to switch to more affordable products is only available to those not buying the lower-end products in the first place. Thus, this could only exacerbate the inflation gap.
The web-scraped data reveal that food prices may be going up quicker for those earning less compared to those earning more. As less affluent households already allocate a larger share of their income to food, they are double-squeezed – both by their spending allocation and by prices. Charalampakis and his colleagues estimated last year that the gap in effective inflation between the lowest and highest income quintiles is at a multi-year high; our findings imply it might be even higher.
For academics, our results provide another piece of evidence that officially reported HICP inflation figures may mask important differences in spending patterns and the realised changes in costs of living faced by households with different income levels, especially during the periods of high inflation. For policy makers, they highlight the importance of targeted support measures. Indeed, it appears that those having no choice but to accept the higher inflation are in a different position than those having the alternative of switching gears.
 Charalampakis, E., Fagandini, B., Henkel, L., & Osbat, C. (2022). The impact of the recent rise in inflation on low-income households. European Central Bank.
 Harmonised indexes of consumer prices (HICP) are calculated using the same methodology in every EU member state, improving comparability between countries.
 Hallak, J.C. (2006). Product quality and the direction of trade. Journal of International Economics 68(1), 238-265.
 Bems, R., & di Giovanni, J. (2016). Income-induced expenditure switching. American Economic Review, 106(12), 3898-3931.
 Siliverstovs, B. (2023). You are what you eat and You eat what you earn: Evidence from the German Household Scanner Data Panel. Unpublished manuscript.
 Latvian Central Statistical Bureau [CSB]. PCC010. Average retail prices of selected commodity (euro per 1 kg, if other - specified).
 Simmonds, E. (2023, April 18). Everyday food prices soar as supermarket inflation hits 17.2%. Which?