He gets three half answers.
Drezner’s first question: “Just how much third-party holdings of U.S. debt does Saudi Arabia have?”
Wish I knew. The custodial data doesn’t really help us out much. $117 billion—around 20 percent of reserves—certainly seems too low. So it is likely that the ultimate beneficiaries of some of the Treasuries custodied in places like London, Luxembourg or even Switzerland (Swiss holdings are bit higher than can be explained by the Swiss National Bank’s large reserves) are in Saudi Arabia or elsewhere in the Gulf.
The Saudi Arabian Monetary Agency (SAMA) is generally thought to be a bit of a hybrid between a pure central bank reserve manager (which invests mostly in liquid assets, typically government bonds) and a sovereign wealth fund (which invests in a broader range of assets, including illiquid assets). So there is no reason to think that all of SAMA’s assets are in Treasuries.
There are a couple of benchmarks though that might help.
If you sum the Treasury holdings of China and Belgium in the TIC data (Belgium is pretty clearly China, not the Gulf) and compare that total with China’s reserves, Treasuries now look to be around 40 percent of China’s total reserves. Other countries have moved back into agencies, so Treasury holdings aren’t a pure proxy for a country’s holdings of liquid dollar bonds. But this still set out a benchmark of sorts.
And if you look at the IMF’s global reserves data (sadly less useful than it once was, as the data for emerging economies is no longer broken out separately), central banks globally hold about 65 percent of their reserves in dollars. This also sets out a benchmark. Countries that manage their currencies tightly against the dollar would normally be expected to hold a higher share of their reserves in dollars than the global average, though this imperative dissipates a bit when a country’s reserves far exceeds its short-term needs.