With first the Fed, and then the ECB and BOJ, all expected to start reducing their balance sheets over the next two years, a bevy of central bankers has been busy on the jawboning circuit, explaining why this would not have a major impact on either bond yields or stocks. They may be quite wrong, however, according to the latest analysis released overnight by Bank of America's rates strategist Shyam Rajan, who calls the upcoming "trillion dollar mismatch" between Treasury supply and demand over the next five years a "smoking gun" which will trigger an "equity-rate disconnect" due to the gradual phasing out of "price insensitive buyers" and calculates that "either rates need to be 120bp higher or stocks need to be 30% lower to trigger enough demand to match forward UST supply estimates." Needless to say, both outcomes are negative for current record low volatility, and will have a substantial impact on risk-asset prices over the coming years, a vastly different forecast than the one the Fed has been scrambling to convey.
Here is the gist of Rajan's argument:
The bond/equity disconnect vs. UST supply/demand
The trillion dollar mismatch in Treasury supply/demand dynamics over the next five years will likely trigger an equity-rate disconnect correction, in our view. Our analysis suggests either rates need to be 120bp higher or stocks need to be 30% lower to trigger enough demand to match forward UST supply estimates. In this report, we quantify the projected increase in supply, the decline in price insensitive demand and the triggers required for the two main price sensitive sources – pensions (higher rates), mutual funds (lower equities) to step up, to clear the supply-demand mismatch in USTs.
Supply is coming, irrespective of stimulus
There are few things more certain right now than increased Treasury supply, in our view. Whether one believes that tax reform gets done or not, Treasury supply is likely to go up substantially over the next five years. In essence, the US Treasury is underfinanced by anywhere between $2 - $4.5 trillion over the next five years, requiring substantial increase in auction sizes and/or new products. Of this, even conservatively speaking, there is likely to be a $1 trillion shortfall in demand using current trends.
Putting these divergent trends together, Rajan notes that the "smoking gun" shortfall is emerging in a time when price insensitive buyers are gone, "so prices have to move"
Unlike prior cycles, price insensitive sources of demand (reserves, domestic banks and Fed) are no longer present to absorb the substantial increase in UST supply needs. This leaves the duty of absorbing supply to the three main price sensitive channels:
1) Foreign private investors who need a stronger dollar and have bridged the gap in the last two years. However, with both the ECB and the BoJ set to taper, crossover demand from this community is set to decline.
2) Domestic pensions who would need higher rates - Our analysis suggests that rates need to be nearly 120bp higher for this community to account for the entire $1 trillion.
3) Fixed income mutual funds - In our estimate, equities need to be 30% lower for mutual fund to see sufficient inflows into bonds to bridge the gap.
Either way, we believe the peak of the equity-rate disconnect is behind us.
Below we lay out some more details from his critical analysis, first why "supply is going up no matter what":
There are few things more certain right now than the upcoming increase in Treasury supply. As we have written previously (May refunding: eyeing the ultra 27 April 2017), few realize that the Treasury’s supply needs are likely to go up substantially irrespective of the stimulus outlook. The four components that drive future Treasury supply all point towards higher issuance needs:
- Baseline deficit projections from the CBO were projected to trough in 2016 and increase even absent additional stimulus.
- Maturing debt: Given the substantial issuance we had until 2015, the profile of maturing debt increases every year until 2020. This drives gross refinancing needs higher.
- Fed: As the Fed allows Treasury securities to run-off, there is at least one class of maturing debt investors that won’t renew themselves – the SOMA portfolio. Fed run-offs will increase public issuance needs by about $200bn-$250bn a year.
- Tax reform/cuts: depending on the details adds anywhere from $200-$400bn in additional annual issuance needs.
- Recession: If one were to take a pessimistic view of the economy where Fed runoffs or stimulus don’t happen, a recession starting next year would drop federal receipts by nearly $250-$400bn (without accounting for likely increased spending) resulting in an effect almost similar or higher than that of the Fed run/offs.
The easiest way to illustrate this shortfall is Chart 4. Current coupon auction sizes raise the blue horizon line – which is just about enough to cover baseline maturing debt and deficit projections for 2018 but falls substantially short starting from 2019. On top of this, if the Fed were to start portfolio run-offs and/or we get a tax reform package in Congress; the shortfall totals nearly $800bn-$1 trillion a year starting as early as next year1. In a nutshell, we estimate the US Treasury is underfinanced by cumulatively around $3 trillion - $4.5 trillion over the next five years.
BofA notes that the shift in Treasury market dynamics is coming at a time when the "old players - including foreign private - are now weak"
The problem with increasing supply this time around, is that the Treasury market can no longer rely on the kindness of the old players. In Savings glut 2.0 – unwinding the bond market conundrum after peak QE, we detailed how despite the savings glut continuing, there is clear evidence that it is not supporting fixed income assets. More generally,
- Reserve manager demand has declined dramatically. Given that this concept has been well socialized over the last two years, it suffices to say that Chinese demand for USTs has gone from accounting for 40% of net UST supply from 2004-2014 to being negative over the last two years (Chart 5).
- Domestic banks that have purchased nearly ~300bn of USTs in the last five years have net sold duration since the election. The potential for lower regulation has likely stopped the dramatic chase for liquid assets that dominated bank behavior in the last five years. This has led to domestic banks turning into a net seller of nearly $26bn USTS year-to-date.
- Foreign private investors are home bound: While the above two factors have been in place for some time now their effect has been somewhat masked by one factor. The dollar and rate differentials moved far enough to encourage foreign private investors to be the price sensitive buyer that stepped up to cover the gap. In fact, since the end of QE in the US in 2014, crossover demand from Europe or Japan can itself account for nearly $560bn of US demand – nearly matching the peak annual demand from the Fed during QE (Chart 6). Now given that, both ECB and BoJ QE are likely to end next year, the peak of foreign private buying is likely behind us. A cheaper periphery post ECB taper and a steeper JGB curve post BoJ changes next year, will likely result in a greater home bias for these investors
Rajan's calculation boils down to one number: a $1 trillion shortfall over the next 5 years.
The combination of increasing supply and decreasing traditional demand leaves the UST market in search of new price sensitive buyers. While we are not usually alarmists about supply/demand mismatches, even taking a conservative estimate this time around worries us. Using a baseline of our Fed portfolio runoff schedule and assuming no stimulus, the Treasury faces a shortfall of ~$2 trillion over the next five years (primarily starting in 2019). We assume that the Treasury will finance nearly 50% of this in sectors with heavy demand – bills through 5s. On top of this we will also assume that the current auction sizes have more than sufficient demand. This leaves the Treasury to net increase 5y-30y auction sizes by approximately $4bn each. We only assume that this additional increase in auction sizes will need a new sustained buyer, leaving a demand shortfall of ~$1 trillion over five years.
But with conventional buyers phasing out, who will buy? The answer: mutual and pension funds. And this is where the existing equilibrium math gets tricky.
Currently, the projected pension obligations of the top 100 corporate defined benefit plans stand at $1.71 trillion with a funded ratio close to 85% (~$250bn deficit). More importantly, the rate sensitivity of pension liabilities more than dwarfs the equity sensitivity of pension assets. The empirical sensitivity of the combined funded ratio to rates stands at roughly at $2.1bn improvement for every 1bp increase in 30y rates. On the other hand, it exhibits negligible sensitivity to a move in equities – for example in 2012 and 2016, despite equities being up >10% and rates unchanged, funded ratios barely budged.
So simplistically speaking, a 120bp increase in 30y rates will see the top 100 corporate pension plans return to 100% funded ratios.
What would it need for pensions to account for $1trillion demand? Currently, asset allocations for these pensions stand at ~ 35% equities, 45% fixed income (with 20% in other assets). Ignoring costs and funded status volatility considerations, it is reasonable to assume that as pensions reach a 100% funded ratio, they will likely move out of equities to a fully fixed income portfolio (as companies don’t benefit from overfunding). So in the most simplistic case, a 120bp increase in rates, will lead to a $600bn outflow from equities into bonds (35% of $1.7 trillion) from the top 100 plans. Scaling this up to the entire universe of DB plans ($2.9 trillion) would suggest that a 120bp increase in rates, will lead to a total of ~$1 trillion in demand from this community for fixed income assets.
In other words, one possible solution to the "smoking gun", i.e., the $1 trillion supply/demand gap to be completely met by pensions, would require the clearing level in markets would have to see interest rates nearly 120bp higher. Such a move would have dramatic consequences for the entire treasury curve, as it would imply much higher longer-date inflation expectations, not to mention vastly steeper funding costs.
What about mutual funds?
The second source of price sensitive demand is the domestic mutual fund community. Here, while some of the inflows are steady given an aging demographic, a substantial increase in demand from this community would need a risk-asset shock that motivates outflows from equity funds into bond funds. To isolate the potential demand in a riskoff shock from this community:
- First, we believe the entire mutual fund and household sector holdings of corporate equities are susceptible to reallocation during risk-off shocks. There is roughly $11 trillion of equities held by mutual funds and about $15.8 trillion by the household sector according to flow of funds (total $27tn).
- Second, in order to quantify the potential outflow from this community we use EPFR flow data during risk-off shocks. Specifically, we isolate recent 5% correction episodes in the SP 500 and see net outflows experienced by equity funds. We then scale up this number given that total net assets represented by EPFR is about $8 trillion or a third of the universe that is susceptible to reallocation.
Can mutual funds account for $1 trillion demand? According to BofA, the average outflow in the two most recent prominent 5% risk-off corrections (Q1 2016, Q3 2015 China deval) has been ~$47bn. Scaled up to the entire universe this would suggest ~$160bn in equity outflows for a 5% correction in the equity market.
This means one would need roughly a ~30% correction in equity markets to result in a shift of nearly $1 trillion from equity funds into bond funds.
* * *
This brings us to BofA's concering conclusion, according to which "the trifecta of increasing UST supply, declining traditional sources of demand and the need for either pensions or mutual funds to step-up will trigger an equity-rate convergence trade in our view."
Why now?: The combination of increasing Treasury supply and decreasing traditional demand has left the onus of absorbing UST supply to price sensitive buyers. While this was met over the last two years by the foreign private community, the end of ECB/BoJ QE will likely require the $1trillion supply/demand gap to be met by domestic pensions and mutual funds.
- For pension funds to fill in this gap completely, the clearing level for rates needs to be nearly 120bp higher.
- For mutual funds to have enough inflows to justify this demand, equities would have to be nearly 30% lower.
Rajan leaves off with this caveat: "This is clearly a simplistic framework given its assumptions around linearity. For example, pensions could use nonlinear glide paths – adopting a very risky strategy if underfunded and substantially lower risk allocations as funded ratios move past 90%. Similarly, investor outflows from mutual funds during risk-off shocks of 10% will likely more than simply double a 5% correction episode. Nevertheless this provides a first order framework to evaluate what “price sensitivity” these particular investor bases would require to account for a trillion dollars of additional bond demand."
While the analysis may be simplistic, it is accurate, and brings up numerous questions about the current dyseqilibrium between bonds and stocks. It also provides a fascinating discussion topic for the next Yellen press conference, when we hope at least one journalist will ask just how the Fed hopes to cross this particular $1 trillion "supply demand imbalance chasm" without major market disruptions, something the Fed has not even once acknowledged yet is a possibility.