Hello! You’re receiving this newsletter as one of 6,764 subscribers (+69 thank you to everyone who singed up, welcome.)

We have some more really interesting videos coming about the economic impact of the Iran war. Be on the lookout this weekend.

You’ll find the “Share the Newsletter” feature at the end of the current issue. Today’s newsletter contains 1,021 words, and reading it in full takes about 6 and a half minutes.

Think about how much trust we place in a number like GDP growth or the inflation rate. Central banks use them to set interest rates. Investors move billions of dollars based on them. Governments use them to justify spending decisions. 

And most of us assume that somebody, somewhere, has done the maths properly. Which they have. Sort of.

The problem is that most of the systems we still use to collect economic data were built for an era when you sent people door to door with clipboards, ran phone surveys, and hoped for honest answers.

The Bureau of Labor Statistics and the Census Bureau still rely heavily on exactly those methods, which should give you some pause.

So in 2025, when they suspended portions of CPI data collection due to a federal hiring freeze, the inflation number that mortgage lenders, retirees, and central banks rely on had literal gaps in it. 

And it doesn't take a budget crisis to break these numbers. Sometimes all it takes is how you collect the data.

In 2023, the University of Michigan’s consumer sentiment survey — one of the most closely watched indicators of American economic mood — switched from phone interviews to online ones. That single change caused the consumer sentiment index to drop by nearly 9 points. Nothing had changed except how they were asking.

Now, all of that would be manageable if countries were at least being honest with the data they do collect. Which, as you guessed, isn’t always the case.

Argentina faked its inflation numbers for nine years. Not once. Nine years. From 2007 to 2015, the government demoted the official who was correctly reporting surging prices and replaced her figures with ones that kept borrowing costs artificially low.

When reality caught up, as it always does, the fallout landed on ordinary people: destroyed purchasing power, junk credit ratings, and a mistrust of official data that lingers to this day.

Greece faked its budget deficit to qualify for the Eurozone, and that particular piece of creative accounting helped trigger a sovereign debt crisis that cost the continent hundreds of billions of euros.

And then there’s China. In early 2025, Beijing reported GDP growth of 5.4% — impressive numbers for an economy simultaneously dealing with a property market collapse, deflation, and falling exports.

Independent researchers at the Rhodium Group weren't buying it, putting the real figure closer to 2.8%.

And it's not hard to see why the gap exists. Beijing sets GDP targets and local officials are promoted or demoted based on whether they hit them, making the incentive to keep reporting good ones essentially irresistible.

China's own government even publicly acknowledged what it called "the biggest corruption in the statistical sphere."  What it hasn't acknowledged is that it built the incentive system that made it inevitable.

Now, to make things even more complicated, the research that's supposed to help us make sense of all this can't always be trusted either.

A lot of that comes down to something called the h-index, a score that ranks researchers by how often their work gets cited. The idea is to measure influence, but what it actually measures is output.

So academics game it by publishing more papers, citing each other's work, and in some cases buying authorship through businesses that manufacture fake peer-reviewed research for paying clients, which we covered in a previous newsletter. The motivation is straightforward: your career depends on it.

And now AI has added a new layer to this, generating references to studies that don't exist, authors who were never real, and invented journals.

Fraudulent citations in academic papers have increased sixfold since 2023, from one in every 2,828 papers to one in every 458. Those fake references enter the academic record, get cited by other papers, and the contamination spreads.

And the truly uncomfortable part is that AI systems are trained on published research. If that research is itself increasingly AI-generated and hallucinated, future models learn from corrupted foundations, creating a degenerative feedback loop that researchers call model collapse and that has no obvious off switch.

Now, none of this means the whole system is worthless. The best data we have is still better than guesswork, and a lot of people work very hard to produce it honestly.

But the next time you see a GDP figure, an unemployment rate, or a growth forecast, it’s worth asking: how was that measured, by whom, and what did they have to gain from the answer?

Nobody wins a war. Well, except the countries that aren't fighting one.

While the conflict in Iran has sent oil prices surging, disrupted global shipping, and left Europe scrambling for energy alternatives, two economies have been quietly watching it unfold with something close to satisfaction — Russia and China, neither of which fired a single shot.

For Russia, higher oil prices are a lifeline. Sanctions enforcement has relaxed as western governments prioritise keeping the lights on at home. And global attention that was fixed on Ukraine is now pointed squarely at the Middle East.

China's calculation is different and more ambitious. Developing economies already wary of American financial pressure are suddenly more open to alternatives. And soaring oil prices are driving exactly the kind of demand for solar panels, batteries, and electric vehicles that China has spent a decade positioning itself to supply.

In our latest video, we look at what each country actually needed going into this conflict, how the war has served those interests, and why the countries with the most influence to end it may also be the ones with the least incentive to try.