As the Fed moves towards monetary tightening, financial markets have started to not only consider the implications of higher policy rates but also the impact that quantitative tightening (QT) may have on financial markets, especially further out in US rates curves. It is our view that the impact of QT will be more keenly felt by risky assets, rather than via a significant increase in longer-end UST yields as purported by some commentators.
The Mechanics (QE/QT)
Central bank (CB) balance sheet expansion (QE) is a sectoral asset swap where CBs create CB reserves for other assets, usually government bonds owned by other sectors of the economy. A stylized QE takes the form as shown below with the CB purchasing assets from the private sector (pension fund as shown in the example by the BoE) via a financial intermediary (commercial bank). The CB finances its government bond purchase by creating reserves and subsequently crediting the reserves to the financial intermediary representing the private entity. The financial intermediary credits the private sector entity with deposits in exchange for the bonds they held. QE swaps interest-bearing, longer duration government bonds for a zero-interest, zero duration asset.
The Fed’s balance sheet size is currently approximately US$9tril (end Jan). Its securities portfolio stands around US$8.3tril.
QT works in reverse. In QT, the Fed allows its asset holdings of USTs and MBS to decline, resulting in lower reserve & overnight reverse repo (ON RRP) levels on the liability side. For instance, consider the likely scenario that the Fed ceases to reinvest some of its UST maturities during QT. As the USTs mature, the Treasury delivers the par amount to the Fed; which in turn reduces the TGA by that amount. In order to replenish the TGA, the Treasury issues debt to the private sector; the bank of the new security buyer will have its reserve account debited (unless, the bank buys the USTs which will equate to an asset swap from reserves to USTs on the bank’s balance sheet), with a corresponding credit to the TGA. Ex-post accounts will show an unchanged TGA, while both Fed holdings of USTs and CB reserves will be lower (i.e. balance sheet shrinkage).
Hence the two main implications arising from QT: 1) private ownership of USTs and/or MBS will rise, and 2) a reduction of cash/reserves vs. an increase in collateral in private lending markets.
Below is a stylized accounting of QT implementation (where $10 of QT is partially ‘absorbed’ by MMFs, banks, and other private sector participants):
The extent/magnitude to which each of the accounting identities above occur depends on numerous dynamic interactions (beyond the pace/scale of QT).
We note a few examples below:
- The structure of treasury issuance. E.g., An increase in the proportion of bills vs. coupon issuance would drain the O/N RRP relatively quicker as MMFs can shift allocation from the O/N RRP to bills.
- How the TGA evolves. E.g., A rebuild of TGA cash would drain other aspects of Fed’s liability-side (reserves, O/N RRP) quicker.
- How the private sector adjusts its asset mix. E.g., Commercial banks may increase/reduce its asset allocation to loans depending on the economic outlook, assessment of credit risks, and profitability vs. capital considerations (i.e. RoRWA), thereby holding more/fewer securities/USTs
- The regulatory landscape. E.g., Regulatory exclusion of USTs in SLR requirements for banks introduced during the pandemic incentivizes dealer market-making in USTs markets.
- Foreign buyers. E.g. The accumulation of FX reserves (in the form of USD assets) by foreign central banks – dependent on factors such as BoP dynamics and the extent of FX intervention.
The above illustrates how complex/unpredictable/incalculable such dynamic decision-making/interactions among market participants and policy-makers can be, making any approach to determine the identities of future ‘net buyers’ of USTs a futile exercise. As argued previously, the nature of flows into USTs – the global risk-free asset – is unique vis-à-vis other asset classes, given its role in absorbing a significant portion of global ‘excess’ liquidity and currency transaction flows. Direct buyers of USTs partly reflects where excess liquidity sits, depending on how global financial flows are directed/recycled (what matters for price/yields is how growth-conducive/sustainable such flows are, rather than the eventual balance sheet accounts).
Additional points to note (considering market narratives):
- The choice between reducing holdings of UST and MBS have indirect consequences beyond the 1st-degree transfer of ownership. For example, focusing on MBS reductions can affect the inherent convexity within the rates market.
- While 1st-degree accounting may suggest that banking system deposits may decline during QT implementation, actual deposit growth within the banking system has a very limited relationship with QE/QT implementation. Bank deposit growth is contingent on many other drivers (loans growth, economic trajectory, wealth velocity, regulatory landscape etc.). Private agents rebalance their balance sheets to adjust to the desired level of deposits.
- The standby repo facility introduced by the Fed in July 2021 serves to reduce the tail risk in funding market squeezes as reserves decline (note repo-funding stress in September 2019).
We think all the above contain limited informational value (beyond signaling tightening intentions) regarding future UST price/yields. Many tout the notion that higher net supply (net of Fed purchases) causes yields to rise. Yet, while we are not keen to extrapolate past episodes, treasury yields have not had the expected correlation to net supply during QE/QT implementation (yields have in fact risen during periods of QE and vice versa).
Another seemingly intuitive argument is that the transfer of UST ownership from the Fed to the purportedly more ‘price-sensitive’ private sector will force yields higher. However, this implies two important assumptions: a) potential QT cannot be priced ahead by markets despite prior Fed communication (i.e. yields only can adjust when ownership of USTs actually change hands), and b) the ‘fair/true’ yields of USTs are significantly suppressed by Fed purchases. We disregard the importance of assumption (a) given that USTs is such a deep and liquid asset class.
To determine whether assumption (b) holds true at each point in time will require a more perspicuous segmentation and reading of the curves. Analysis needs to establish hurdle rates across the short vs. long-ends, real vs. nominal, term-premia vs. risk-neutral, spot vs. forward curves, and the path (pace and scale towards terminal) vs. destination (actual terminal & market neutral) of each rate cycle. We articulated this earlier (and will do so with an updated, follow-up piece) – the key point being that UST yields are primarily determined by expectations over its term structure and term premia. Implicit in the levels and shape of rates are the market assumptions over the Fed’s reaction function, its “neutral” and terminal rates, the “neutral” rates of growth and inflation considering the economy’s super-structure (socio-economic model, institutional framework, sectoral balance sheet profile, income/wealth distribution), and the expected dominant regime, going forward. Only after (i) thorough consideration of the factors above, and (ii) credibly determining that, if not for Fed purchases, segments of the curve will trade closer to higher ‘fair’/’justified’ levels, can one argue that QT directly lead to higher UST yields.
What’s that lurking in the dark?
Rather, the predominant impact of QE and the Fed’s ZIRP (which the former enhances), is its impact on psychology and the change in market behaviour (elements of both signalling and portfolio re-balancing) amid an evolution of the financial market structure. The prices of assets are elastic, predominantly reflecting the asset allocation preferences of asset holders. Such preferences are heavily affected by both materialism and perceived reality (the material environment of financial markets limits/dictates available options which together with beliefs of market participants shape the consolidated asset allocation of markets).
The presence of the Fed’s ZIRP and QE, with its implicit central bank put, coupled with the evolution of the market structure – democratization of market access, rise of financial leverage (vis-à-vis real economy lending), levered-based (options-based long positions, delta hedging, repo-based lending by NBFIs)/ esoteric investment strategies (Alts, longshot investments), and the shift in investment strategies (buy the dip, momentum, vis-à-vis fundamentals) have led to a upward reflexive market – ZIRP + TINA + FOMO + YOLO + “Robinhoodisation” + Low Cost Financial Leverage + Fed Put + BTD – in this investment cycle. Not surprisingly, this has led to an outsized aggregate allocation to risky assets vis-à-vis history.
Thus, the biggest risk from QT and the end of ZIRP, is the reversal of this process – negative reflexivity of markets. As interest rates increase and liquidity is withdrawn, the price of liquidity will rise, in effect, dampening the effect of TINA. Financial leverage will not be as cheap and readily available, making levered investments less appealing to both lenders and borrowers. The downward repricing of the Fed put makes “BTD” less attractive, with losses more likely to trigger forced closing of positions or the need to put up higher margins. ‘Democratization’ of markets makes trading easier, both on the way up and down, but the difference being a lack of bids in selling markets leads to illiquidity and price gaps. FOMO turns into fears of losses and insolvencies. Long-shot, loss-making investments now faces rising opportunity costs as zero profits and cash burn faces the challenge of rising risk-free rates. Financial markets could fall into a negative downward spiral, on more aggressive monetary tightening amid rising recessionary concerns.
Hence the key risks for financial markets are not that of higher UST yields from QT or the Fed’s Lift-off but its impact on broader risk markets, with assets of risk profiles furthest from the risk-free rate (think of PIMCO’s concentric circles) the likeliest to face heavier drawdowns.
Secondarily, market functioning may be disrupted by QT as reserves, deposit holdings, and access to liquidity are not equally shared across market participants. Large commercial banks are the main holders of such liquidity vis-à-vis smaller banks, NBFIs and other market participants. Tighter liquidity and rising interest rates may lead to disruptions in lending markets, especially for levered entities which have depended on leverage and financial engineering to juice up returns.
To conclude, we believe that the implications of QT will be more keenly felt in other riskier assets vis-à-vis rates which may at worst see some negative passing effect before more important drivers forces rates to converge to its “equilibrium” level and shape.
