Each month, the Swedish government agency Statistics Sweden (Statistiska centralbyrån or SCB) collects data on goods and services consumed by the Swedish population to produce official statistics.
Analysing and comparing price changes over time across different sectors of the economy (food, housing, energy, transportation etc.) represents the basic principle behind the consumer price index (CPI) measurement. Goods and services are included in a representative consumption basket (which remains somewhat stable over time) reflecting what an average person consumes compared to a reference or “base year” (1980 for SCB). This helps identify which aspects of living costs are increasing more rapidly than others.
The CPI is also a key measure of inflation, which is often the talk of the town, although not always easy to grasp for the common people. Discussions about inflation not only happen within the walls of the Riksbank, but it often becomes a tool for election campaigns, as we have recently seen during the 2024 US presidential election, leading to oversimplification and misunderstanding of the context.
Taking a closer look at how the CPI is measured in Sweden is important to understand whether representative the official inflation rate is, as measured by the CPI. Not all households have the same weight: richer ones have greater weights than poor ones in a ‘plutocratic price index’ that is the one used by SCB.
“We work in a plutocratic way, which is a governmental decision, although there is another way to do it, that is the ‘democratic way’ where you try to collect the data more at a household level”, Filip Hellberg, statistician at SCB, says.
Since producing a separate price index for each household is not doable, statistical agencies construct an average to measure the effects of price changes that do not always allow us to see who is affected the most by a high inflation.
They use ‘regression models’ and ‘supported judgmental methods’ that deliberately exclude items with the lowest expenditures. “We update these regression models each year, but this is a very hot topic in the statistics business, whether to do this or not”. When talking with Filip Hellberg, he stresses the fact that the plutocratic method they use is hardly representative of the entire population’s spending habits: “If people buy more Volvos even though it is a quite expensive car, than Kia, Volvos get a bigger impact on the CPI. We look at where the consumption is, we do not investigate how much a certain societal group consumes. The top 10 % in Sweden has a bigger impact on the CPI than the bottom 10%”.
Due to what is called the plutocratic gap, the CPI can contribute to underestimating the impact of inflation on inequalities.
When asked about what is the most common issue for people who want to understand these figures, Filip Hellberg explains “If you go to the store to buy an iPhone, you think it’s very expensive, right? It’s around 10,000 kr and maybe it was 8,000 kr five years ago, but the phone today is so much better. So we adjust the price for that, which makes it cheaper. This is a discussed topic because the public pension is dependent on the CPI. Elderly people sometimes call us and say ‘I get less money because you say that the phones are better, but I don’t even buy these good phones, and still I get less money because of these phones’. This is a very hard topic, and we get criticized a lot”.
What consumers buy changes over time and once a year, Statistics Sweden updates which goods and services are to be included in the measurement. Moreover, CPI series are not always comparable because of the differences between items and the usage of quality adjustment methods.
One of the most glaring examples is revealed by the CPI for printed newspapers and periodicals that seems to have seen the most significant price surge. While it is common knowledge that the media market has been hit hard by rising paper and energy costs, and that media outlets such as Dagens Nyheter are raising the cost of the newsprint subscription for its audiences, this record does not represent the reason why people are switching to digital subscriptions, nor it makes printed newspapers the most costly item in the basket. It does, but it is not comparable.
“You can’t really compare the CPI rates of newspapers with the ones of computers, for example. Newspapers are about 20 times more expensive today than it was in 1980. Computers are only 0.34”.
That is because with computers they use quality adjustment methods, while printed newspapers do not change over time in terms of quality and content although we find both items in the same series.
However, having 1980 as the base year is going to change soon. This year SCB is going to change it to 2020. We asked Filip Hellberg the reason behind this choice.
“After 40 years the numbers become a bit hard to interpret: it’s easier to state that something went up by 10% from 1980 if you’re in 1985, but now if you say that from 1980 something went up by 9 times, it becomes harder for the normal population to interpret it”.
Food was at the center of the discussion because prices have been increasing so much. “Since 1980 it has gone up by 4 times so it’s 409 in 2023, but the amount of money that people spend on food has decreased heavily”. In 1980 people may have spent 30% of their income on food, while today they spend between 5% and 10%.