The tail risks that may be generated by Mario Draghi’s monetary policy innovations in the Eurozone include even more intense versions of Andrew Haldane’s “Doom Loops”
The recently adopted QE approach by the ECB, in concert with the negative deposit policy rate (NDPR) introduced last summer, has set off a number of nested disequilibrium dynamics that may unwittingly introduce a material increase in systemic risk for the eurozone, and perhaps beyond. Lord Keynes anticipated what he termed a “euthanasia of the rentiers”, as he expected active monetary policy would be successful in reducing long-term interest rates, and the share of the population living off of bond coupons would eventually just wither away. By way of contrast, if the following assessment is correct, Draghi may have signed a mutually assisted suicide pact with finanzkapital in the eurozone.
The logistics of implementing QE (including questions about adequate bond supply for the ECB to purchase, as well as the related market “liquidity” concerns), or whether or not QE represents what Lord Turner refers to as “open monetary financing”, are not the real problems, or at least not the most compelling ones. Rather, the implementation of QE with a large and increasing share of the bond market displaying negative yields to maturity (NYTM) presents a number of serious challenges to financial stability in the eurozone.
To cut to the chase, the ECB’s QE and NDRP measures may be setting investors up for a discontinuous price event, much like what was experienced in the equity market meltdown back in October 1987. Even if a disruptive yield spike is avoided, or even contained and reversed by ECB heroics, pursuing QE under NYTM market conditions may lead to a significant dampening down of bank and insurance company profitability. In the extreme, the solvency of key eurozone financial institutions could once again come under question. This could further complicate the ECB’s chances of achieving their 2% inflation goal, as it may dampen the bank lending channel as a key transmission mechanism for unconventional monetary policy.
The entire set up, in other words, begins to take on many of the characteristics of Andrew Haldane’s Doom Loops. In this case, however, the ECB may unintentionally be setting off nested Doom Loops that will feed on each other, and thereby magnify systemic risks quicker than investors and policy makers might otherwise imagine possible. Below is a concise sketch of the main elements of the Doom Loop dynamics the ECB may have set in motion.
- NYTM bonds still pay positive coupons, but the price of the bond has been bid so far above the par or face value that over the full life of the bond, the holder of the bond will face a certain loss when the bond matures. NYTM bonds are therefore never likely to be willingly held to maturity by private investors, as they guarantee a loss. No client of institutional investors is likely to ever have a nominal loss as their target return on their assets – especially on assets meant to be managed to cover future nominal liabilities that tend to grow over time, like corporate pension fund liabilities.
- NYTM bonds are therefore held solely for prospective capital gains, via sale of the bond before the maturity of the bond. This requires the holder of NYTM bonds to believe they will be facing a virtually certain Greater Fool (GF) in the fixed income market. This GF must possess ample if not nearly infinite buying power, as well as the motivation to constantly bid up prices of NYTM bonds, or else the existing holder of the NTYM bond faces the risk of not finding a buyer at a higher price, which means they will face a certain nominal loss from having to hold the NYTM bond to maturity.
- Central banks (CB) with fiat or sovereign currencies (that is, money not exchangeable on demand into a fixed number of other currencies or commodities) fit the description of this GF perfectly, as their balance sheet expansion is virtually unrestricted. In the case of the ECB, their purchasing of bonds in GF mode is limited only by the inflation ceiling constraint of 2%. With price deflation spreading across the eurozone, the ECB is clearly highly motivated to play the necessary GF role, and very unconstrained in doing so.
- Buyers of NTYM bonds have to believe CBs will execute QE sufficiently to produce falling yields and rising bond prices. The GF must be committed to delivering risk-adjusted returns that are compelling enough to private investors to take the risk of holding NYTM bonds. Unless the GF can credibly and continually create the expectation of higher prices and hence capital gains to holders of NYTM bonds, private investors will want to dump their entire holdings of NYTM bonds if the GF appears to be backing away from the designated role.
- Since the CB is the GF, and since existing holders and new buyers of NTYM bonds require the GF to continually be prepared to bid NYTM bond prices higher, any indication the CB will terminate, or taper, or even pause on its originally indicated path of QE purchases introduces a significant risk. The stock demand for NYTM bonds, not just the flow demand of private investors, then shifts against NYTM bonds en masse. There is a likely revulsion of investors, a la Charles Kindleberger, with NYTM holdings, since the market set up with NYTM bonds is not unlike the game of Musical Chairs.
- The minute the CB is even vaguely suspected of pausing in its purchase program (the equivalent of reaching for the needle on the record in the Musical Chair analogy), there will be a simultaneous desire to dump NYTM bonds, which could also have repercussions for the creditors and owners of financial institutions trying to dump NTYM bonds. To the extent that the “smart money” (aka wise guys in hedge fund land) recognize there is a distinct first mover advantage in the game of Musical Chairs, the mere hint the CB is preparing to pause may be sufficient to set off a stampede of private investors out of NYTM bonds.
- When portfolio preferences adjust against holding NTYM bonds in this sharp and simultaneous fashion, bond pricing may go discontinuous (huge gaps in prices, or even no bid), as the stock supply offered up in the market will tend to overwhelm the natural flow demand for NYTM bonds. When pricing goes discontinuous, quantitative driven models (as well as “algos”) tend to blow up, since they presume continuous pricing at all times. Measures of volatility will also tend to surge, which means risk budgets will get gobbled up quickly. As risk budgets get eaten up, forced sales of risky assets tend to spread across institutional investors. In the extreme, investor sell large blocks of financial assets that are still liquid, even if they are unrelated to the original asset class under distress, as a hedge against the holdings they wish to sell, but are facing no bid market conditions.
- In this fashion, feedback loops with adverse consequences can be set into motion, with attempts to sell NYTM bonds leading to more attempts to sell the same; leading to more discontinuous pricing, and so on, as we witnessed in 1987 with the portfolio insurance mechanism in the US equity market. It is not unlikely, as in 1987, that forced or panic selling could spread to more liquid assets in the bond market, or even to more liquid assets in the equity and other markets
- In theory, this vicious cycle/Doom Loop process would supposedly end when yields are once again high enough to represent a good risk/return proposition to private investors. Yet there are likely to be large capital losses in the portfolios of holders of NTYM who could not find buyers once the music stops and the GF/CB steps back. Their risk budgets will also be very restrictive at that point. So-called value buyers may be few and far between under these circumstances. After all, it is harder for private investors to reach for yield when they are facing a risk budget that has been exceeded, and sitting on large unrealized capital losses).
- There are several reasons why the GF, in this case the ECB may elect to pause or taper early. The financial media is beginning to recognize some of the reasons for the game of Musical Chairs to end earlier than September 2016, which marks the currently planned termination of the PSPP. These include: (a) Supply constraints may complicate the ECB buying program in future months – after all, the ECB has been massively front run by investors flocking to government subsidized and virtually government guaranteed capital gains, and the supply of high quality bonds may not be large enough given ECB portfolio risk constraints; (b) Eurozone macroeconomic data may continue to surprise on upside, and inflation may revive sooner once oil and other commodity prices find a bottom (note that Draghi has already declared victory in turning eurozone economy around, just by announcing QE); (c) Bank net interest margins and insurance company profitability may be sufficiently eroded by negative deposit rates (which are acting as a tax on excess reserves held by banks at the ECB), by flattened yield curves (which are reducing their various carry trades), and by an ever increasing share of the bond market exhibiting NYTM (which means they effectively are off limits to purchase by the banks and insurance companies), that major financial institutions begin to push back, such that political opposition could build within ECB (probably through German financial institutions getting the Bundesbank to lead the charge) to curtail or even reverse QE early.
To conclude, what we appear to have in play here is a game of Musical Chairs, but it may prove to be a game with remarkably high stakes. Once the music stops - or even the hint of the music stopping filters out across institutional investors - everyone will race for a chair (i.e. try to sell NTYM bonds). The absence of a GF at that point in time means there will be no buyer, at least not until yields spike sufficiently to warrant taking on the perceived risk, which will probably be quite high under those circumstances, as longer maturity bonds at low nominal yields have high durations – that is, bond prices react a lot to small changes in nominal yields.
The Draghi Doom Loops traced above are probably features of the ECB’s current QE and NDRP policy that either have not been recognized by investors, or have not been thought through by ECB, or both. Indeed, one could argue institutional investors are quite happy to have government guaranteed returns (that is, government subsidies in the form of capital gains for wealthy bondholders as the ECB bids up bond prices in its PSPP operations) from the GF/ ECB all the way out to September 2016 termination of QE. But to mash up metaphors, these institutional investors may be picking up pennies in front of a steamroller, and caught (unwittingly, and in part by the demands of the quarterly relative performance objectives that largely dictate their behavior) in a game of Musical Chairs that is likely to end sooner than they think - and end very badly, at that.
We will elaborate further on these and other Draghi Doom Loops in future essays, as there are further complicating dynamics that have been set in motion, but still do not appear to be widely recognized yet. Keynes’ forecast of the fate of rentiers may prove to have been far too kind: It would be a great irony if Draghi’s QE and NDRP initiatives have forged a mutual assisted suicide pact with finanzkapital in the eurozone.