Operation Twist: New twists and turns for RBI’s playbook

By Andre de Silva & Himanshu Malik

Back in May 2020, we stated that RBI had outlined a more realistic borrowing plan to increase the full-year issuance of Gsec for FY20/21 by Rs 4.2trn to Rs 12trn. However, since then the growth outlook has turned even weaker (GDP 2Q2020: -23.9% y-o-y, consensus: -18.0% y-o-y) and the centre’s tax collection is running far below the February budget numbers, suggesting that the additional borrowings announced in May will largely cover the tax revenue shortfall. And, it is not just the central government, but state governments that also face increasing pressure to borrow more. The central government has proposed that state governments could cover the estimated Rs 2.35trn shortfall in GST compensation, either through a partial borrowing of Rs 970bn via RBI’s special window or through market borrowings.

Moreover, fiscal support is likely to be needed, given the rising growth headwinds worsening the fiscal outlook. Consequently, we raise our estimate of combined net supply of Gsec and State Development Loans (SDLs) to Rs 19trn (Gsec: Rs 11.7trn, SDL: Rs 7.3trn) from Rs 15.9trn (Gsec: Rs 9.6trn, SDL: Rs 6.3trn). We expect full-year gross Gsec borrowings to be increased by another Rs 2trn, from Rs 12trn to Rs 14trn. RBI will announce the borrowing calendar for the H2FY20/21 (October 1 2020 to 31 March 2021) by end September.

For state governments, gross issuance will depend on the proportion of SDL repayments met by the Consolidated Sinking Fund (CSF). State government gross borrowings in H2FY20/21 will reach Rs 4.85trn, and states will need to issue at least Rs 200bn of SDLs every week in H2FY20/21 vs Rs 120-160bn now.

To some relief, Gsec issuance has been quite aggressively front-loaded in the first half of the year with regular exercise of green-shoe options for a total amount Rs 720bn. As such, Rs 7.06trn of Gsec have already been issued and with two remaining auctions in September, the gross Gsec issuance will reach at least Rs 7.66trn by the end of September. This means that gross Gsec issuance of Rs 6.34trn will be needed in the second half of the fiscal year and supply at weekly Gsec auctions will fall to around Rs 250bn (vs Rs 300-340bn now) in H2FY20/21 even if the additional Gsec borrowing of Rs 2trn is announced.

However, we note that demand dynamics for Gsecs are likely to be more challenging in the second half of the year. We also note that increasing share of SDLs in combined net supply in H2FY20/21 is likely to lead to a widening of the spread between SDL and Gsec yields unless RBI announces additional measures to reduce the SDL supply at auctions.

Where is the demand for bonds?
We estimate that at least Rs 12.5trn of Rs 19trn combined net supply in FY20/21 will need to be absorbed by commercial banks and RBI. Banks have already net added Rs 3.8trn to their government debt holding (i.e. SDL and Gsecs) in H1FY20/21 so far. RBI bought sizeable amount of Treasury bills in April and May while it purchased only Rs 470bn of Gsec through five operation twist. RBI’s Gsec holdings are going to further increase by another Rs 100bn with upcoming operation twist on September 17. But this still means that additional net combined supply of at least Rs 8.1trn will need to be absorbed by RBI and banks during the rest of FY20/21.

The central bank will continue to do so through ‘Operation Twist’, but any outright purchases of Gsec by RBI will require the liquidity to normalise first or concerns around inflation to recede. What is needed is better management of duration supply while banks can further scale-up their SLR holdings with investments in short-dated bonds. The central bank can also to take out duration from the markets through operation twist auctions and increase its holdings of Gsecs, but such operations are constrained by RBI’s treasury bill holdings.

SLR holdings of banks have reached 29% of Net Demand and Time Liabilities (NDTL) on the back of aggressive buying of government debt by banks. It is not unreasonable to assume that further aggressive increase in SLR holdings of banks up to 32% of NDTL can create additional net demand of up to Rs 7trn for government debt from banks. The remaining supply-demand gap of Rs 1.1trn will need to be bridged by either the central bank or by other sets of investors that may be willing to scale up their investments with higher bond yields.

We also note that banks will be reluctant to take duration risks with interest rates close to bottoming out and policy backdrop becoming more uncertain with rising inflation and fiscal headwinds. This implies the need for a certain shift in issuance strategy.

Higher short-dated issuance and more operation twist
Banks are now allowed to hold more SLR securities (which include Gsecs and SDLs) on their held-to-maturity (HTM) book. Previously, banks could hold investments in excess of HTM cap (i.e. 25% of investments) provided that excess investments are in SLR securities not exceeding the overall limit of 19.5% of NDTL. But, this limit has been increased to 22% from 19.5% effective from September 1, which is equivalent to Rs 3.5trn.

In H2FY20/21, we think that the proportion of bonds (both Gsec and SDLs) in less than 5-year maturity tenors will need to be increased to 30-40% so that such supply can be easily absorbed by banks. In comparison, over 26% of issuance at Gsec auctions and 25% of SDL issuance in H1FY20/21 so far has been done in maturity tenors of 5-year or less. Issuance of floating-rate bonds can also be increased but more crucially, the central bank will need to take out duration supply of at least Rs 3-4trn from markets in H2FY20/21 through ‘Operation Twist’ auctions or other measures.

We, however, note that current ‘Operation twist’ policy is constrained by RBI’s holdings of Treasury bills. This may therefore require some policy innovation from the central bank that could include the issuance of Market Stabilisation Bonds (MSBs) by the central bank at the short-end of the curve and purchasing Gsecs at the long-end of the curve. This will have a similar effect as ‘Operation twist’ and will help the central bank to reduce duration without increasing the liquidity.

In terms of PV01 terms, the Gsec supply at weekly auctions has increased from an average of Rs 13-14bn in FY19/20 to Rs 24-30bn in H1FY20/21. Yet, the term premium in 10yr Gsec yields has declined from a peak of over 142bps in March 2020 to only 120bps currently. Besides, a host of liquidity easing measures by RBI as well as aggressive policy rate cuts, the compression in term premium has been led by RBI’s OMOs.

However, we find that it was the signalling effect of the announcement of the OMOs rather than the actual OMOs that had a meaningful impact on the compression of the term premium.

RBI also provided strong guidance that it is not comfortable with higher Gsec yields by rejecting all bids at the auction of 10yr Gsec on August 28, while the central bank also bought 10-year Gsec at 4-7bps below the market yields at operation twist auction on August 27. Despite this, the decline in term premium did not sustain this time with rising expectations that such measures will not be enough.

The latest Gsec auction on September 11, also resulted in devolvement of almost full 10-year Gsec supply primary dealers as RBI rejected bids for higher yields. While RBI is likely to continue to provide a soft cap on 10yr Gsec yields, we believe that paying pressure in ND OIS rates is likely to persist at least in the near term with increasing supply strains in bond markets and especially with central bank likely to maintain a policy pause for the rest of 2020.

With policy rate on hold and increasing macro uncertainty around fiscal dynamics and inflation, we see scope for higher risk premium in ND OIS rates in the near term and expect 5yr ND OIS rates to reach 5.70%.

Edited excerpts from HSBC Global Research’s INR rates: New twists and turns (dated September 14)

Silva is head of global EM rates research and Malik is Asia-Pacific rates strategist. Views are personal