To be sure, I've made a multitude of contentious arguments over the past six months or so. Of those arguments, one that has to rank near the top of the list in terms of its being especially controversial is the idea that central banks can -- despite their command over the printing press -- go broke.
I began to explore this issue in several articles (here and here notably) after it became clear over the course of 2012 that the ECB was set to poison its balance sheet by buying up risky Spanish and Italian bonds and by generally accepting stodgy "assets" as collateral for loans to institutions, which were themselves teetering on the brink of insolvency (i.e., Spanish banks). This problem was compounded by the ECB's steadfast refusal to write-down Greek debt, which had already been written-down in the private sector.
As I noted previously, central banks can unquestionably find themselves in a situation where their capital is negative; that is, a situation wherein the value of the assets they have purchased falls below the amount listed as "liabilities." In the case of the ECB, my argument was simply that purchasing trillions of euros worth of periphery bonds was (and is) extraordinarily dangerous under the prevailing circumstances.
As we learned last July, yields on periphery bonds are at risk of rising without warning due to any sudden deterioration in Europe's still-precarious financial situation. Because the price of those bonds falls as yields rise, holding trillions worth on the books could quickly turn into a disaster. Let's be clear: if it's not a disaster in terms of the central bank's ability to function, then it might still be fairly classified as a disaster politically or in the context of the central bank's legitimacy.
It occurred to me at the time that moving from the merely theoretical example of the ECB to a more concrete case might help explain this dynamic to those who may still find it confusing. As Bruce Krasting noted Saturday, the Federal Reserve Board recently provided a perfect opportunity for reflection on this issue by showing just what will happen to its own balance sheet and the "profits" it funnels to the Treasury department as interest rates rise.
One argument you might hear quite often when it comes to the Fed goes something like this: "If their balance sheet is in danger, then why did they make some $90 billion in profits for the Treasury last year?" The first question that no doubt needs to be answered here is not why, but rather how? How does the Fed "turn a profit?"
The "profits" the Fed makes are not mark-to-market gains on its books, although it is sitting on some $200 billion in theoretical paper profits. The $90 billion remitted to the Treasury comes from the difference in the coupons on the Fed's Securities Open Market Account (SOMA) holdings and the interest the Fed pays on excess reserves. As nice as this is (the $90 billion figure is a record, and it certainly helps with the deficit), one has to wonder what would happen if interest rates rose, causing the $200 billion in mark-to-market paper gains to become several hundred billion in mark-to-market paper losses. This is an issue I have raised before when I noted that the Fed's DV01 was skyrocketing.
So what happens if interest rates start to rise? As illustrated by both the Fed itself and Bruce Krasting in the article cited above, based on the current projected trajectory of interest rates going forward, the $200 billion or so in paper profits would evaporate, and depending on the pace of future asset purchases, would become somewhere between $100 billion and $350 billion in mark-to-market losses by 2016:
Next, observe what happens to the Treasury remittances under each of the scenarios posited in the graph above:
If you don't stop to reflect on this for a moment, it might seem strange. That is, since the remittances to Treasury are made from the difference between the interest income on the SOMA holdings and the interest paid on excess reserves, and because the Fed doesn't have to mark its book to market, then it might not be immediately apparent why spiraling mark-to-market paper losses due to rising rates should necessarily have such a dramatic impact on the remittances to Treasury. The answer is that the coupons on the SOMA portfolio are fixed, whereas the rate that must be paid on excess reserves rises with the Fed funds rate isn't:
SOMA interest income remains similar to the baseline because the securities in the SOMA portfolio have already been purchased and their coupons are fixed. However, interest expense becomes greater once the federal funds rate lifts off from the lower bound because of the higher interest rate path. In addition, because sales of MBS occur when longer‐term interest rates are higher than in the baseline, realized capital losses are somewhat greater.
This theoretically will lead to capital losses at the Fed between 2016 and 2021, and will thus eliminate the Treasury remittances:
Obviously, this would seem to be evidence that the Fed's "business model," so to speak, is not working. Can you imagine if your favorite bank realized capital losses of between $10 billion and $50 billion every year for six years?
The trick for the Fed is the use of a "deferred asset." This serves to ensure that the Fed's capital isn't depleted in the event it suffers a loss. Later on down the line, when the Fed eventually turns a profit, it reduces the value of the "asset." This item is recorded on the liabilities side of the balance sheet.
Of course, as Bruce Krasting notes, this an extraordinarily suspect way to operate that can be somewhat equated with a tax loss carryforward. However, deferred taxes are not a qualified asset under Basel III, and I will leave it to readers to determine the legitimacy of this maneuver.
Perhaps more importantly however, consider that when the Fed begins to finally wind-down its portfolio, it will be selling into an unfriendly market in terms of where rates will be when the assets are sold as opposed to where they were when the assets were acquired. The Fed does book gains and losses upon asset sales of course, so you can bet that the pain will be severe when the party finally stops and the FOMC attempts to normalize its balance sheet.
Additionally, the effect will likely be exacerbated by the fact that as rates rise, the stream of interest income from the spread between the SOMA portfolio and the rate paid on excess reserves (which will still be elevated no doubt) will be close to zero at best, or deeply negative at worst. As Stone & McCarthy put it recently:
Ultimately, the Fed will be faced with booking losses on asset sales, while also suffering from declining net interest income as the interest rate on reserves is increased to effect a higher Fed funds target.
There are two takeaways here. First, the Fed (and other central banks) can go broke or become insolvent, even if only for a period, and even if that insolvency is covered with accounting gimmicks. Second, the still-enthusiastic primary dealers and direct bidders notwithstanding, there are at least some folks at the Fed who are concerned with exactly what kind of pain will be associated with the theoretical end of QE. If you own Treasury bonds (TLT), you should be concerned, too.