| What was the reasoning behind omitting outliers for the purpose of the press release? I appreciate the transparency, but why leave out the 8 countries where your figures resulted in more than 220 hours of labor a month (and still ranking the “worst” countries just below that point)? https://www.thepricer.org/hours-to-afford-essentials-best-an... I don’t know much about methodology for this kind of data. So forgive me if these are silly questions. But something feels off: I’m just not convinced that everyone in Mexico works 11 hours a day, 7 days a week to afford their basics. And not to pick on Mexico, but OECD’s hours worked index [1] seems to put Mexican workers at ~2200 hours a year—rather less than the 3840 your method suggests. Where would you imagine the confounding factors might lie? Would it be silly to imagine some of the differences might involve * large portions of consumer trade happening in the informal economy, with only relatively legible (thus probably higher-income) consumption factoring into the price indices? * related: elements of subsistence lifestyles confounding the consumption figures? For example, if I want to rent or build a formal place in Lagos, it would cost a certain amount, but 60% of people there live in informal dwellings whose cost wouldn’t be captured in housing price statistics [0]. So perhaps the price index captures what it costs to live relatively richly there, but the net pay captures a broader range of the income spectrum? * related: what’s in this constant basket that we’re comparing across vastly different nations? Do the items in your “basket” reflecting typical behavior rather than subsistence minima per se—so reflecting appetitive consumer preferences in some economies? In the sense that the “basket” we’re pricing in the US involves a 4,500 square feet single family dwelling and 3 SUVs, while in Ghana (which you call out for cheap rent) it’s “a roof over your head”? I bet I could survive on big ol’ sacks of rice and beans for a long time if I needed to, but that’s not what people buy where I live: would the “food” in your basket be “big ol bag of rice and beans” or “what a normal family buys there”? * it looks like the price data is per capita and the wage data is per formally-taxed worker after tax. Do these figures then reflect different patterns of labor force participation or household composition between economies? That is, are these household-level hours? If my whole 12-person extended family live together, and I’m the sole breadwinner, do we get dinged for 12 rent bills and 12 transportation bills but only count one formal paycheck? * OECD net pay—that’s net of taxes, yeah? Sounds from the abstract like they factor in cash transfers from social programs as if it were earned as pay—so those “hours” aren’t actually worked—but granted I can’t imagine that’s too big of a number. In the same spirit though what proportion of the “fixed basket” gets paid for through those taxes rather than out of a household’s net pay (in, for example, the socialist countries you call out)? * could low-paying jobs be paradoxically overrepresented in formal data, since “multinational mining concern” might keep more formal books than “neighborhood merchant who prefers that you pay in cash”? [0] https://punchng.com/60-lagos-residents-live-in-informal-sett... [1] https://www.oecd.org/en/data/indicators/hours-worked.html |
Why the >220h outliers were excluded from the press note: We used a trimmed summary to avoid letting a handful of fragile cases (thin/biased price data, capital-city rent proxies, wage series with poor coverage, currency quirks) dominate the headline. In the full write-up, I’ll show the entire distribution, the eight omitted cases, and winsorized vs. raw comparisons so readers can audit the effect.
Where confounding can creep in (you called several):
Informal economy: Formal price series + formal wages can overstate hours where many transact or earn informally.
Housing formality: If rent stats miss informal dwellings, the “solo renter” basket skews too expensive vs. typical living arrangements.
Basket design: Ours is a minimal, non-tradable basket (modest 1-BR rent, basic utilities, ~2,100–2,400 kcal groceries, local transit). No cars, no tuition, no luxury goods. I’ll add a shared-housing variant and a city vs. national sensitivity.
Household composition: The metric is per worker, not per household. A 3-adult household sharing one rent bill will show fewer hours per person than a single renter.
Net pay & transfers: We use net/typical pay (after taxes, including standard cash benefits where the source does). Publicly funded services (health care, schooling) are intentionally excluded from the out-of-pocket basket—this is a cash-flow lens, not a full welfare measure.
Mexico/“11 hours a day” example: That interpretation mixes hours of pay needed with hours actually worked. If the basket costs more than one typical monthly take-home, the ratio can exceed ~173–184 hours/month—even though people don’t work 11h/day; they share housing, live outside the cost center, buy informally, or pool incomes.
What I’ll add to make this clearer (and falsifiable):
A methods box with the exact basket, wage series, and the full, untrimmed table (including the eight outliers).
Two companion views: Shared-housing hours and Discretionary hours = paid hours – hours to essentials.
A city edition (e.g., SF/NYC/Lagos vs. national) so price-level heterogeneity isn’t hidden.
Happy to be corrected—if you have a better wage or rent series for any of the questionable cases, point me to it and I’ll rerun that row and show the delta.