Central banks are making large losses, resulting from money creation and asset purchases over many years. This matters to all of us. Why? Because ultimately we, as taxpayers, pay. Because the politics of losses might mess up the institutional independence of monetary policy. And because losses affect the cost benefit analysis of quantitative easing.
This issue is, however, a complicated subject and too much for one newsletter. So today I will look at some of the accounting and institutional issues, followed next week by an assessment of how much it matters.
Why do QE losses arise?
Quantitative easing is the process whereby a central bank creates money and buys assets. The vast majority of money created globally has bought government bonds from outside the banking system. In doing so, the central bank earns the coupon from the bond it has purchased and the money used to purchase the bonds ends up as private deposits in commercial banks. These banks have an excess of these deposits and park them overnight at the central bank where they are remunerated at the policy rate.
When interest rates are low or zero, the coupons exceed the overnight interest rate and the central bank makes a profit, which is generally passed back to government. But if interest rates rise, as they have, the overnight rate exceeds the return on assets, generating a net interest loss (or negative carry). In addition, as interest rates rise, the value of the bond falls and when it is redeemed, it is generally worth less than the amount that was paid, although this depends on the coupon on the bond and the price paid.
So, as interest rates rise, central banks tend to make both a net interest loss and capital loss on redemption.
How should we think about this?
Ultimately, payments between central banks and governments do not matter — they are both part of the consolidated public sector, with the central bank just another arm of government.
The best way to think about it, as helpfully outlined in a recent paper by Stephen Cecchetti and Jens Hilscher, is as a form of fiscal policy at the public sector level. Essentially QE is just an exercise in increasing the amount of short-dated debt remunerated at the overnight policy rate in exchange for reduced amounts of long-dated government bonds.
The government could do this itself, buying up long-dated bonds in issuance and swapping them for short-dated debt — there is nothing special about it happening in an entity called the central bank.
How do countries account for this in deficits?
This is where the conceptual purity meets the hard reality of individual country institutions, different accounting practices and a variety of fiscal rules. It is unavoidably messy.
The UK is an example of a country that accounts for QE and quantitative tightening well. Its headline public finance measures are at the public sector level, consolidating the central bank into the government sector.
It means that, as the chart from the UK’s fiscal watchdog below shows, the QE process made significant profits in the 2010s when overnight interest rates were close to zero but this ended in 2022 as rates rose. With the Bank of England engaged in active sales of bonds, sometimes bought at very high prices and sold at low ones, the Office for Budget Responsibility expects significant capital losses in the years ahead and a decreasing interest loss as QT decreases the central bank’s balance sheet. The OBR assumes a rapid pace of QT (probably wrongly) and the period of losses is essentially over by the early 2030s.
But this is not the important point — the faster QT is, the bigger are the valuation losses and smaller the interest losses, so the net present value are likely to be similar.
The BoE has an indemnity arrangement with the UK government, covering it for losses, but again, that is also not the important thing here. These are real losses, with the government sector losing and the private sector gaining. In the UK, the net interest losses are incurred and accounted for when they happen and it is right that they should show up in the UK public finances as an item that contributes to the public deficit. The treatment is no different to the UK government having issued much shorter-term debt and now paying more interest as short-term rates have risen.
If the UK is a bit of a paragon of virtue in accounting for losses properly, the US is the opposite. It measures its public deficit by default at the government level with the central bank sitting outside. This means money flows between the central bank and the US Treasury matter.
When QE makes profits, the treatment is in effect the same as in the UK because the Federal Reserve pays profits to the US Treasury by law. But when it makes losses, instead of a symmetrical money flow from the US government to the central bank, they sit in the Fed’s accounts as an accumulated “deferred asset”, which will be reduced in future once the Fed starts making profits again. The chart below show’s the Fed’s accounting, which moves from a flow when it is making profits to a growing stock of a deferred asset when it is making a loss.
People in the US and the Fed think this is normal, but it is far from that. It is simply the US consolidated government borrowing more than it reports and saying it will account for it later using an off-balance sheet vehicle. The US is not Greece, but hiding debts off balance sheet did not ultimately prove a boon in the early 2000s for the Mediterranean nation.
As Willem Buiter has recently written, this is a “gibberish” form of accounting and the Fed “must be honest about possibly deeply negative conventional equity or net worth”. Most of the Euro system adopts the same approach as the Fed, although without the “deferred asset” naming convention.
What about public debt?
Again, the UK accounts for this properly. Public sector debt rises annually by the net interest loss (the deficit) and the realised losses once a bond matures or is sold. It is consolidated at the public sector level. This is what you would expect because at the point a bond matures or is sold, QE is over. Money has been created, used to purchase an asset and destroyed. If there is a loss on that transaction, it should be added to public debt as a financial transaction.
The Fed and the ECB again are living on the never never. Eventually, once the central bank is made whole, the public debt figures will be the same as those in the UK — there is ultimately little difference in the mechanisms — but because losses are kept in central banks they do not show up as public debt until some future moment when they have repaired their balance sheets.
Even though QE is long over, the effects will linger. The numbers here can be of reasonable size. On the Fed’s balance sheet, for example, there is a $1tn mark-to-market loss on the assets it currently holds, amounting to one-seventh of the fair value of its assets. It is about 3 per cent of GDP.
What this accounting means for fiscal policy
The important thing to remember is that however a country accounts for QE and QT, the ultimate effect is the same. It is likely to come out in the wash. The UK is taking the pain upfront on its accounts at the time the transactions take place, while others will take their hits later once QE is long over.
The UK has some remarkably silly interactions between QE and QT policies and its fiscal rules, which a new government should sort out. These, however, stem from badly written rules, which neither fully consolidate at the public sector level nor abstract from QE altogether, rather than bad accounting practices for QE.
In the US, there is little talk about the QE hangover that will depress the public finances for some time. The Fed’s latest estimate is that the deferred asset will be paid off by mid-2027, but many others think this is wildly optimistic. The US, in any case, is not wildly keen on fiscal discipline at the moment, so QE is not the biggest issue in its fiscal list of horrors.
Europe, too, will have a lingering legacy of QE that will not be fully on the books for many years. Let us hope this will not come to bite it.
What the accounting means for monetary policy
Almost everyone agrees that QE losses do not interfere with monetary policy because central banks can set interest rates without a problem even with a hole in their balance sheets.
But there are some weird incentives created. In the UK, for example, the Bank of England’s choices interact with the government’s public finance rules on debt in a deeply unhelpful way.
In the US and Eurozone, the accumulated losses on their balance sheets leave money created in the system not backed by assets. I am not suggesting this is monetary financing of their government — far from it because it will eventually be paid down — but monetary financing would look the same on central banks’ balance sheets. This is also far from ideal.
Who is happiest?
There is no doubt about this. Very clearly, the central banks and accounting systems that sweep looses under a giant rug labelled “tomorrow’s problem” are happier.
Losses are a political issue in the current UK election and also in Sweden, where the costs are upfront. Kicking the can might not be sensible or transparent, but it is the easiest thing to do.
What I’ve been reading and watching
-
The Fed held interest rates in a range between 5.25 and 5.5 per cent on Wednesday, signalling between one and two rate cuts this year. Claire Jones interpreted the data expertly in this article
-
Newly installed in Paris, Olivier Blanchard, former IMF chief economist, chatted about inflation with Soumaya Keynes on her podcast. He gave an expert defence of standard macroeconomics, accepted the public wants a low inflation target (new for him) and said that needed to be combined with active fiscal policy in a downturn
-
If you want some cheery news Shamaila Kahn, of UBS Asset Management, finds plenty of it with emerging economies being “particularly quick to control inflation”
-
If you like something completely different to Adam Posen’s view that credible central banks beat inflation, a paper for the European parliament by Jens Van’t Klooster and Isabella Weber suggests committees to give early warnings of inflation, buffer stocks to ease supply shocks and price caps to prevent price rises and wage price spirals. It’s radical stuff and pretty difficult to implement. If Posen is right, it is also utterly unnecessary. I will come back to this
A chart that matters
There is a golden rule when you suggest someone else has “slipped up”: don’t mess up yourself. I flouted that rule in the newsletter last week, confusing Eurozone “negotiated wages” when I meant to write “compensation per employee” in a section saying Christine Lagarde had slipped up.
The upshot is that Eurozone compensation per employee in the first quarter was close to the staff forecast.
In even better news for the Eurozone, unit profits — the part of the GDP deflator that can be attributed to corporate profits — was negative for the first time since Russia invaded Ukraine. As Philip Lane, ECB chief economist, told an Irish audience last week, “the net impact of labour cost increases on prices is being buffered by a lower contribution from profits”.
He attributed this to the transmission of higher interest rates, preventing companies passing on cost or wage increases and said it would keep disinflation on track even as wages played catch-up. Compared with disappointing inflation numbers for May, it is certainly a more encouraging picture.
