Perhaps the most important economics research paper in several decades was recently published in the Journal of Economic Perspectives, and it has largely gone unnoticed. It reveals that the Japanese government, after thirty years of running huge budget deficits, actually has net financial assets equivalent to only 33% of GDP. In a world where mundane trivia are typically described as ‘astonishing’, expressions of genuine awe have reduced impact. Nonetheless, “Japan’s debt puzzle: sovereign wealth fund from borrowed money” by Chien et al. is more important than they realise, not least because the punchline table contains a simple error, which turns an interesting conclusion into an astonishing one.
At some point it will be conventional wisdom to view Quantitative Easing (QE) and cryptocurrencies as two of the more absurd financial activities humans have hitherto engaged in. One is a massive technological distraction in search of a problem which has hijacked the consciousness and digital ‘wallets’ of a wonderful cross-section of the world, from techno-gold bugs, to the White House, to religious finance bros seeking alternatives to usury. The only people who actually use crypto as an alternative to fiat money (as an actual means of exchange) are those peddling illicit goods on the dark web. And they must curse all the other clowns for creating so much price volatility. Science fiction suffers from an earnestness that underestimates the absurd.
There’s something pleasing about bracketing QE with cryptocurrencies in the lexicon of off-the-wall financial pyrotechnics humans periodically create. Culturally the two could not be further apart: serious economists do QE, brash grifters do crypto. Despite its combination of banality and sophistication, crypto appears easier to understand. In contrast, I would bet that no more than a handful of people on the planet actually understand QE – despite thousands of research papers and ‘serious’ speeches on the subject.
We can prove this statement with a simple observation: if QE was understood there is no way that the editors of the Journal of Economic Perspectives would miss a simple and game-changing error in the punchline table of Chien et al.’s seminal paper (below).
Table 1 presents the Consolidated Japanese Government Balance Sheet, setting out the financial assets and liabilities of the Japanese government. Can you spot the glaring error? (There are two errors, one is glaring).
| Quarter end | 1997Q4 | 2012Q4 | 2024Q2 | Δ 97–24 |
|---|---|---|---|---|
| Assets | ||||
| Deposits | 5.9% | 8.5% | 19.1% | 13.2% |
| Loans | 102.8% | 63.1% | 63.4% | −39.4% |
| Of which by PFIs | 92.2% | 50.1% | 40.0% | −52.2% |
| Domestic Equities | 10.7% | 20.9% | 41.7% | 31.1% |
| Of which by PPFs | 2.4% | 4.4% | 13.5% | 11.1% |
| Of which by BoJ | 0.0% | 0.7% | 12.6% | 12.6% |
| Foreign Securities | 7.5% | 29.7% | 61.8% | 54.3% |
| Of which by PPFs | 1.3% | 7.0% | 27.7% | 26.5% |
| Other Assets | 8.4% | 7.9% | 6.4% | −1.9% |
| Sum (Assets) | 135.2% | 130.1% | 192.4% | 57.2% |
| Liabilities | ||||
| Currency | 10.8% | 18.3% | 20.3% | 9.6% |
| Bank Reserves | 0.6% | 9.5% | 90.8% | 90.2% |
| Bonds & T-Bills | 41.8% | 162.3% | 103.8% | 62.0% |
| Of which by BoJ | −9.6% | −23.2% | −93.6% | −84.0% |
| Loans | 55.1% | 48.5% | 37.8% | −17.4% |
| Deposits FILF | 46.4% | 1.1% | 1.5% | −44.9% |
| BoJ External Debt | 0.0% | 0.1% | 7.1% | 7.1% |
| Other Liabilities | 5.2% | 8.7% | 8.8% | 3.6% |
| Sum (Liabilities) | 159.9% | 248.5% | 270.0% | 110.1% |
| Net Liabilities | 24.7% | 118.4% | 77.6% | 52.9% |
This is a situation where a lack of knowledge can be advantageous. What are ‘assets’ and ‘liabilities’? Simply put, assets are things you own and liabilities are things you owe, sometimes contingently. Financial assets generate an income and/or can be sold for a positive sum of money, and financial liabilities require repayment at some future point or points in time (typically in the form of interest and/or capital), and/or can be repaid/repurchased for a positive sum of money.
If we go through the Japanese government’s assets everything looks fine. The government (or governmental agencies) have made loans to companies – these are government assets because the government receives income and a capital repayment. The government also owns equities. These are assets because the government receives dividends and can sell these holdings for a positive sum of money.
Now let’s turn to the liabilities. First up, ‘currency’. What? Or even wtf? Japan’s outstanding currency – notes and coins – is equivalent to 20.3% of GDP, apparently. Thankfully, we defined a ‘liability’. Chien et al., consistent with standard practice among ‘serious’ thinkers on the subject, simply transcribed the category, and in doing so made a category error. It’s obvious that currency isn’t a liability. Nothing is owed for currency and nothing gets repaid. It’s money, a bit like bitcoin on the dark web, in fact. It’s called the ‘means of exchange’. I won’t go through chapter and verse on why convention has it that accountants just drop currency into the ‘liability’ category – I’ve bored someone with this before. But it matters. Currency is not a liability – so Chien’s estimate of Japan’s net debt is already overstated by 20% of GDP.
The error doesn’t end there. ‘Bank reserves’ are up next. Again, if you have never looked at a central bank balance sheet before you’re at an advantage. Bank reserves are simply digital currency. If you don’t believe me, try this rambling podcast with Mervyn King, where he states this definitional truth. Now if currency is obviously not a liability, why would its digital version become one? Now someone with a dangerously small amount of knowledge will point out that we pay ‘interest’ on reserves, so this transforms them from non-interest-bearing physical cash into a digital liability. Here’s where Japan is interesting. We used to think that in order to control money market interest rates central banks had to pay interest on reserves – but the Bank of Japan has proved otherwise. By tiering reserves, it is now clear that there is no need to remunerate aggregate reserves in order to control money market rates. A positive interest on reserves (IOR) is actually a helicopter drop of cash to holders of reserves (banks) – that’s a means of voluntary gift-giving, not the creation of a liability. If in doubt check the share price of bank stocks since IOR started rising (reminder to self: stay long).
So bank reserves drop out of the liability category. Now what happens to the maths? Apparently the government of Japan has net financial assets equivalent to 33% of GDP, despite decades of the highest budget deficits in the developed world. That’s astonishing. How is this possible? There are two reasons: money creation in a deflationary shock is a free lunch to the state, and the state can issue liabilities at a lower cost than the return on assets it can purchase or create. In the last three decades, these activities dominated the budget deficit in Japan.
What does this have to do with QE and cryptocurrencies? Not a lot. It’s far more important. Well, QE is relevant, because by engaging in QE central banks were in fact retiring government debt – so most of the developed world simply doesn’t have a public sector debt problem. For the same reason, QT is an act of inane financial destruction – it is needless creation of public sector debt.
There are profound and multiple consequences from this paper:
- The Mayor of Greater Manchester was right: the bond market is ad hoc. The bear market in global bonds that began in 2020 has nothing to do with public sector debt levels – it primarily reflects the fact that bond yields are interest rate expectations, and rates have risen far more than the bond market expected in 2020. There has also been a structural break in Japan’s inflation dynamic, which is causing contagion and a positive bond-equity correlation in recent shocks.
- QE is the creation of money (an asset to the holder, and power of the issuer) in exchange for an asset. When you buy an asset you are investing. QE from 2020 onwards in particular was a huge operation with the state’s balance sheet to buy assets/retire debt – and other central banks would have accumulated substantial wealth had they invested in equities, like the Bank of Japan. Most of their arguments for not doing so are spurious (‘equities are risky’ – equities are not as risky as long-dated bonds yielding zero).
- In future, central banks should invest like a sovereign wealth fund when they need to create reserves. Part of the role of the state is to provide insurance when the private sector is in recession – this holds as true for asset purchases as it does for counter-cyclical fiscal policy.
Finally, the paper by Chien et al. is genuinely excellent, so much so that in a world where truth is defined by how shocking something is, I’m claiming to have predicted it. Seeing as this is my first Substack, and intellectual flexibility is a virtue, I have changed my mind on crypto while writing this. More on that soon.
Leave a Reply