When prices change, how that impacts people depends crucially on which prices increase and what goods and services people are consuming. Across the western world, price inflation–the rate at which prices increase–has been relatively slow for over a decade. Central bankers have consistently undershot their inflation targets despite their careful implentation of complex monetary policy. 

The supposed dearth of inflation might seem like small comfort–or a cruel joke–to the Californian hipster paying $15 for a smoothie bowl, the German renter whose rents are increasing at a stunning rate or the London young professional shoveling out £5 for an unimpressive lunch sandwich. The larger the diversity in consumption patterns, the less appropriate it is to aggregate price changes into a general price index such as CPI or PCE statistics.

One reason for the dissonance between official figures and real-world experience is the weight that statisticians place on various items when constructing a consumer price index (e.g. the Bank of England’s CPI; the ECB’s HICP; the Fed’s PCE). For instance, in the price index used by the European Central Bank, housing costs make up only 17% of the index, whereas the Federal Reserve places a 24% weight on housing expenses. That divergence turns a 25% increase in housing costs–with all other prices and consumption patterns held constant–into a 4.25% overall inflation in the Eurozone but a 6% inflation in the U.S.

While policymakers are aware of those data limitations and we have standardized statistical ways to adjust for quality improvements, these problems can still cause headaches. One illustrative example is the impact of iPhone prices on Sweden’s price index; Martin Enlund, FX strategist at Nordea, estimates that the quality adjustment of iPhones alone reduced the reported price increase by 0.1 percentage points every year for the last 5 years.

That minor detail has some implication for our modern world, considering that the Riksbank’s interest rate decisions have turned on such small margins before. Looking at these differences in consumption bundles and quality adjustments over longer historical periods, they quickly become astronomical. In a famous paper, Nobel Laureate William Nordhaus surveyed “lumens”–a unit for light–emitted by various sources throughout the centuries. Nordhaus estimated the price of light, the essential service its originators provide us with, to have fallen by 99.97% between 1800 and 1992.

Over decades or centuries, even small differences can result in very large adjustments when we evaluate past incomes. For instance, how much better is a computer as a calculating tool than an abacus? Is a keyboard and word processor ten, fifty or a hundred times better than quills, ink, and bulky, slowly decaying paper?

A recent study by Vincent Geloso and Peter Lindert makes a big deal out of consumption bundles. By disaggregating purchases by working classes and upper classes, they make a revolutionary discovery: beginning earlier than we used to believe, the poor’s standards of living improved faster than those of the rich. Contrary to the tired claim that capitalism involves the rich getting richer while the poor get poorer, it seems that during the 19th century the opposite was true.

The authors reach this conclusion by using different consumption bundles for two different income segments. People’s standards of living depend on what they themselves consume, not on what they could buy if they had the rich’s consumption patterns:

 “[T]he contrasts that matter are contrasts in individuals’ abilities to buy what they care to buy, or need to buy, and not the (nominal) inequality in their ability to buy the same common bundle as some other class could buy.”

The components that drove this extraordinary reduction in cost of living, argue Geloso and Lindert, were falling prices of grain-based foods and a rise in the relative price of services that the poorer classes supplied (mostly wage rates for common labor).

The American rise in inequality over the nineteenth century, using both top-1% / bottom-99% and top-10% / bottom-40% metrics, is much less pronounced than previously believed. The authors conclude: 

“[T]he ‘nineteenth-century’ period 1815–1914 brought a clearly egalitarian shift in the price structure for all four countries—England, Canada, the USA, and post-1850 Australia. The net change over these 100 years is unmistakable.”

A century before Paul Ehrlich would predict imminent starvation in the entire world (specifically in what he thought was a remarkably backwards India), the world surplus of grains had enriched the poor–even in the “dark Satanic mills” of Britain. The lower relative price of grains mitigated and partly reversed the economic inequality we tend to associate with the nineteenth century.

The exact bundles used to measure consumption matter greatly for understanding prosperity, today as well as in the past.