What effect will Fed balance sheet normalization have on Treasury yields?
That’s quickly becoming something of an obsession for market participants and rightfully so.
After all, no one (least of all the Fed itself) has any idea how this is going to work or how it will impact markets.
This situation is complicated immeasurably by concurrent tightening from other DM central banks and from factors outside of developed markets including, most notably, the ambiguity created by the reversal of EM reserve accumulation.
That latter dynamic (the end of the “great accumulation” as Deutsche Bank dubbed it some two years ago) became the subject of heated debate in 2015 as GCC countries weighed how to plug budget gaps and defend currency pegs in the face of falling oil prices and China liquidated its reserves to control the pace of the yuan deval.
If the post-crisis era has been characterized by voracious demand for DM debt by price insensitive buyers, then the rolling back of crisis-era policies will be characterized by the opposite. It will, simply put, fall to price sensitive buyers to absorb supply. So it isn’t just that tapering will by its very nature put upward pressure on yields, it’s that the folks who step in to replace central banks on the demand side of the equation will be rational investors where “rational” means their propensity to absorb supply will depend on the attractiveness of the investment. This represents a complete shift in the dynamic.
In short: Fed balance sheet normalization won’t happen in a vacuum.
This is something Citi has discussed at length and it’s illustrated simply in the following chart:
On Tuesday, Citi’s Harvinder Sian is out with an in-depth look at what’s in store and his analysis is pretty interesting.
First, Sian notes that central banks have effectively taken half the burden of absorbing budget deficits off the private sector’s shoulders:
[Important note from Citi on the EMU chart: "The Euro area looks very different with the supply of AAA bonds falling, via rating downgrades, meaning that the sovereign debt crisis has huge legacy issues for the structure of the market. This means, that even with ECB PSPP buying of AAA bonds at 1/3 of the total purchases, the supply-demand imbalance can be profound.”]
Sian goes on to give the obligatory nod to the projected “benign” QE unwind scenario which, in sum, is this:
Our assessment of 1% GDP Fed balance sheet shrinkage lifting yields by just 6bp is consistent with Yellen’s numbers of +60bp yield compression unwind if we scale up to the Fed’s planned $ 2tn shrinkage (which is, c10% of GDP).
But as noted above, when we think about all of this we have to think holistically about what a coordinated unwind from DMs would mean in terms of a supply shock in safe haven securities and the attendant burden that would place on price sensitive buyers.
What does US flow-of-funds say about Treasury price sensitivity? Before looking at global QE (demand) against global supply metrics it is worth taking a quick detour to assess why Treasury yields have proven less sensitive to various bond bearish shocks than many investors may have believed. Figure 17 and Figure 18 breaks down the holders of debt into three categories.
- ‘Regulatory & Fettered’ holders are those investors coerced to buy or with low price sensitivity. From the flow-of-funds categories, this includes Private Pensions and Insurance, Banks, Money Market Funds, and State and Local holders of Federal debt, and GSE, ABS issuers and Broker Dealers.
- ‘Policymakers’ who are typically also not price sensitive. This category includes Fed holdings and reserve holdings.
- ‘Price Sensitive’ holders are the remainder and include Households, Mutual Funds, ETFs, Non-Financial Corporates, and Foreign Private purchases.
There will be some overlap in the categories for some investors groups. For instance, Foreign Private investors are classed as price-sensitive but that price may well be the level of JGB and Bund yields making the outright Treasury yield far less relevant than in recent years.
With that proviso in mind a striking feature is that despite the acceleration in US debt volumes the bulk of the increase whether measured as a portion of the market or as a % of GDP has been absorbed by regulatory and official sector flows.
The important point here is that while the market is rightly focused on how the private sector will absorb the unwinding official sector flow it is also important to keep an eye on other moving factors.
We have written previously (U.S.) Peak Regulation and for instance expect weaker LCR and CCAR rules may see a drop in bank demand for Treasuries. That is another form of safe bond supply shock.
Again, this issue of who is buying and whether they are price sensitive is critical and is an additional (and potentially aggravating) factor that must be taken into consideration when we think about the coming supply shock.
With that out of the way, here’s Citi’s Sian on supply/demand:
Measuring global supply of safe bonds and global QE offsets
Given the risk that the US QE unwind should be followed in 2018 by ECB tapering and possibly a higher JGB yield target as relative policy divergence triggers JPY weakness – there is a need to look at cross-market influences.
In the following we look at the supply and absorption of safe bonds. We define safe government bonds as Treasuries, core EMU plus France and Belgium (which we label EMU core+), Gilts, Swedish government bonds, Australian government bonds and Canadian government bonds. We exclude JGBs from the calculations on the grounds of ratings and largely domestic ownership, and so we exclude BoJ QE effects on demand.
Figure 19 through to Figure 21 shows some of the data that we are working with, in the context of rising debt levels into and post-crisis, and then QE volumes from Advanced Economies (AEs) that have been absorbing some of this debt increase.
Figure 21 shows the value of global (ex-gold) reserve accumulation measured against the IMF’s Advanced Economy GDP. The rationale, as we discussed earlier, is simply that the reserve accumulation should be viewed as QE by another name in terms of its low price sensitivity and its large influence on bond markets. For instance, reserves rose by some $ 5tn from 2009-14 dwarfing all QE programmes in volume but probably not in 10yr equivalent terms.
To look at the global supply and global QE effects we convert the data to a common $PPP metric. That would avoid currency effects that would alternatively show, for instance, that Euro safe bond supply fell markedly in the crisis period and especially amid the drop in EURUSD from 1.40 to 1.10 region over the QE expectations and delivery period.
The aggregation in stock terms is shown in Figure 22, which in this case includes reserves accumulation. The stock effects and the flow effect are linked but we think that markets focus far more on the flow effect. The economic literature speaks to stocks and portfolio balance frictions but we see the market’s experience with QE as a series of stock level shocks, so perhaps a half-way house between investors and academic viewpoints.
Looking at the flow effect, Figure 23 shows the 4Q change in net supply of safe government bonds when offset by QE flows.
Figure 24 shows net safe government bond supply flow but this time including the offset of QE and now also FX reserves. The pattern is broadly similar until 2015 when FX reserve drawdowns added to the global safe bond supply.
Now Citi’s conclusions after all this aren’t entirely dire and to a certain extent the methodology becomes a bit tortured, but what’s important to remember about all of the above is that in many ways “QT” (quantitative tightening) is far more important than incremental rate hikes. Because we’re talking about actual liquidity here – who’s providing it and how willingly (i.e. at what cost).
In any event, you should take a holistic approach to thinking about the end of the QE regime – hopefully this helps in that regard.