Thread by @michaelxpettis

1/10
Interesting new IMF paper on the extent of industrial policy subsidies to Chinese manufacturers and SOEs (to the extent information is available) and their impact on productivity.
imf.org/en/Publication…

2/10
It measures national-level cash subsidies, tax benefits, subsidized credit, and subsidized land, which collectively amount, it says, to a high 4% of GDP. The authors note that other subsidies exist, including sub-national subsidies, but these are harder to measure.

3/10
I'd include in the latter what is perhaps among the biggest subsidies, which is over-spending on logistical infrastructure. To the extent that these cost more in resources than they create in economic value, they represent a large transfer to the users of the infrastructure.

4/10
In addition, the authors measure credit subsidies against benchmark interest rates, but I would argue that credit subsidies are likely to be much larger if we consider that benchmark interest rates themselves are repressed, sometimes even to highly negative real rates.

5/10
This matters a great deal in a financial system in which nearly all lending is to the investment side of the economy. Repressed interest rates transfer income from net savers, whose saving is effectively "taxed", to net borrowers, whose funding is further subsidized.

6/10
Finally, and perhaps most obviously, China's undervalued exchange rate is a major, but hard to quantify, indirect subsidy to the tradable goods sector, which comes at the expense of net importers (i.e. the household sector).

7/10
These other subsidies are hard to measure, but I suspect they (along with subnational subsidies) are among the highest of all. As it is, the authors of the paper estimate that even a partial accounting of the value of subsidies puts them at 4% of GDP.

8/10
The paper also finds that it is not just SOEs that receive the subsidies but also private-sector manufacturers in industries that are especially favored by Beijing. The whole saga of involution is pretty strong evidence that this is true.
carnegieendowment.org/posts/2025/08/…

9/10
The study calculates that the asset misallocation caused by these subsidies reduces aggregate productivity by about 1.2%, and GDP by up to 2%. I am not sure how much faith I put in these calculations, but they do show the paradox China faces – the point of these...

10/10
subsidies is to keep economic activity high enough to reach politically-determined GDP growth targets, but their effect is to distort investment and lower the sustainable growth rate, leaving China to bridge the growing gap with even more non-productive investment.