It’s always fun reading a Jason Hickel paper. You know he’s gone wrong somewhere because, well, Hickel, see? So the task becomes well, why has he gone wrong?

Here he’s claiming that, acshully, poverty increased when China stopped with the Maoist idiocy and, acshully, poverty in China now is about the same as it was under Maoism. Which, you know, but Hickel.

This one saying that poverty has increased in China since the 1980s is also interesting.

Indeed, this one is:

Drawing on the BNPL data, we show that between 1981 and 1990, China’s poverty rate was considerably lower than in capitalist economies of similar size and income. On average, during this period only 5.6 per cent of the Chinese population lived in extreme poverty, compared to 51 per cent in India, 36.5 per cent in Indonesia, and 29.5 per cent in Brazil. Furthermore, we find that during the capitalist reforms of the 1990s, China’s poverty rate increased dramatically, reaching a peak of 68 per cent in 1995. While China’s poverty rate gradually recovered during the 2000s, rough estimates for 2018 indicate that it remains at roughly the same level as during the 1980s.

That is, you know, fun. There wasn’t any noticeable economic growth in China between the 60s and late 70s. In the 60s tens of millions starved to death. We’d therefore sorta assume quite a level of poverty. But, the claim is:

This problem becomes particularly acute when comparing socialist and capitalist economies, and when assessing transitions from socialist to capitalist systems. Socialist policies of public provisioning and price controls may keep the cost of meeting basic needs quite low compared to capitalist contexts characterised by high levels of commodification and privatisation. This means that any given level of broad-gauge PPP income would have a greater welfare purchasing power – in terms of basic needs – under socialism than under capitalism; people would have better access to the key goods, such as food and housing, that are necessary for escaping extreme poverty.

Hmm, well, mebbe. If things aren’t distributed by price then measuring prices might not be all that good a measure of distribution. Seems a reasonable enough point.

Hickel then goes on to point out the real issue he’s investigating:

The World Bank’s poverty line (WBPL) suffers from a significant empirical limitation, however, which scholars have recognised now for more than a decade (Reddy and Pogge Citation2010). The purchasing power parity (PPP) exchange rates that undergird the WBPL are calculated on the basis of prices across the entire economy – including commercial airfares, sports cars, and meals at high-end restaurants – rather than the prices of goods that people need in order to meet basic needs, such as food and shelter (Moatsos Citation2016, Citation2021, Allen Citation2017, Citation2020). When it comes to measuring poverty, what matters is not income as such but rather what that income can buy in terms of access to essential goods; in other words, what matters is the welfare purchasing power of income. Allen (Citation2017) analyses commodity prices around the world in 2011 and finds that the cost of meeting basic needs, measured in PPP terms, changes depending upon the price of food and shelter relative to prices across the rest of the economy. In Zimbabwe a person’s subsistence needs can be met with $1.74, PPP (ibid., p. 3713). But purchasing a similar basket would cost $3.19 in Egypt, and $4.02 in France (ibid., p. 3713). Because the WBPL does not account for the variable cost of meeting basic needs in different countries, it cannot be used to establish meaningful estimates of poverty.

And, well, ok. If we’re in a subsistence economy then sure, the price of air tickets isn’t really relevant to the price or availability of a bowl of rice. So, if we work out those Purchasing Power Parity exchange rates - the ones we use to determine poverty and relative incomes across countries - using air tickets in our basket of goods then sure, we might be misled.

Hickel’s whole explodin’ the World Bank because they’re wrong proof depends upon this. That we’re measuring poverty in subsistence economies by including air tickets and thus PPP exchange rates are wrong and therefore bugger you capitalists!

And, well, sorta:

Drawing on the BNPL data, we show that between 1981 and 1990, China’s poverty rate was considerably lower than in capitalist economies of similar size and income. On average, during this period only 5.6 per cent of the Chinese population lived in extreme poverty, compared to 51 per cent in India, 36.5 per cent in Indonesia, and 29.5 per cent in Brazil.

But, of course that same problem afflicts - in these time periods at least - at least both India and Indonesia. So, you know, maybe not.

We could even mutter something about it doesn’t really matter what the price of a bowl of rice was in China if there were no bowls - at those low, low, socialist prices - of rice to be had. As there weren’t in the 1960s.

But is there anything else? Say, this point about using air tickets in our price comparison basket? Well, yes, there is:

ICP PPPs are published for GDP and selected expenditure components (see Table 1 in the technical note and the ICP Classification of Expenditures). Many applications and indicators use PPPs estimated at the level of GDP or at the level of Households and Nonprofit Institutions Serving Households (NPISHs) Final Consumption Expenditure (otherwise known as private consumption), which reflect the average difference in prices between countries across the whole economy or for those goods and services that fall under private consumption, respectively.

Which leads us to:

Economies participating in the ICP are expected to estimate their GDP for the reference year in line with the SNA and to disaggregate this GDP estimate into the relevant expenditure components identified in the ICP expenditure classification.

And thus to:

And so Hickel is right, eh? Expensive things - expensive things not relevant in a subsistence economy - like restaurant meals and air tickets are included in PPP weights and thus we’re going to have significant problems using PPP FX rates to measure poverty in subsistence economies.

Ta Da!

Except. Except - always read the footnotes.

b: “Actual” expenditure components in italics reflect the sum of household consumption expenditure and expenditures by nonprofit institutions serving households (NPISHs) and government on goods and services (within the given category) actually consumed by households.

Ah. The World Bank are not blithering idiots then. Or they might be but we should not use this instance as our proof. For what’s happening is that yes, sure, expensive stuff is included in the PPP calculation. But only in proportion to how much of that expensive stuff is happening in that economy. That’s what the “actual” means. If, say, eating out is 8% of US household expenditure - something near right - then we’ll apply a weighting of 8% to eating out in our calculations of PPP. Seems fair. If, as in 80s China, eating out is 0% of household expenditure - or eating out means gulping down a raw salamander found in the ditch - then restaurant prices get a 0% weighting in our PPP. Raw salamander is one of those non-cash transactions that Hickel so greatly prefers, of course.

The same is true of air tickets and so on. The influence of these expensive things on PPP exchange rates is only by the amount they actually happen in the economy under discussion. Therefore the Hickel objection to PPP exchange rates fails.

So now we know what’s wrong with Hickel here - he didn’t read the footnotes. Which gives us our lesson for the day, don’t be like Jase, always read the footnotes.