By Suyash Choudhary, Head – Fixed Income, IDFC AMC
While we remain in steady state for now, it is stating the obvious that some sort of central bank normalization cycle is ahead over the next couple of years. Data releases on the margin tweak expectations around this (recent CPI being one example) but are unlikely to change the broader picture; unless data changes start turning extreme. Given this, this is as reasonable a time as any to reiterate and analyse the bond framework that we are currently operating with and which is informing our broad approach to markets.
The Three Pillars
There are 3 pillars that define our current framework. These are as under:
- The current effective policy rate is at emergency levels: At the end of the day, India remains a growth economy with steady state nominal GDP growth rates of 9 – 10%. The current policy rate of 3.35% has been justifiably arrived at in the midst of considerable shocks to growth, enduring tail risks and substantial modelling uncertainties. However, once these large shocks and modelling uncertainties abate the continuation of this rate can no longer be argued for merely on growth-inflation dynamics. To elaborate, sustaining support to growth while allowing for more gradual climb-down of inflation will definitely involve continued accommodation but just not emergency levels of it. Thus irrespective of near term data, it is prudent to build for a steady normalization path ahead for monetary policy. To be fair, to say this is also stating the obvious since current market curves are pricing a lift off from 3.35% sooner rather than later. Also to be fair, the RBI has already embarked upon first steps in the process (stepping down on bond market intervention, increasing variable rate reverse repo amounts) and these should logically get followed up with the commencement of narrowing of the policy corridor via gentle hikes to the reverse repo rate at the latest by early next year.
- Peak policy rates in this cycle are likely to be lower than in the last: While the normalization cycle may have already started and the first reverse repo rate hikes may get delivered within the next 2 quarters, we also think the process is likely to be gradual. Apart from RBI not wanting a sudden and unwarranted tightening in financial conditions, there may be another reason to expect this: the distance to be covered in terms of total overnight rate adjustment may be less than 200 bps. If we are correct on this, it implies that peak rates in this cycle will be substantially lower than in the last. This follows from a view that there is likely a significant aspect of long period scarring which may have had an impact on the potential growth rate of the country. To be sure this can heal and growth can come back strongly, but it is unlikely that a full restoration happens within the next couple of years. In fact the argument for lower peak rates is valid for developed economies as well and indeed markets seem pricing for it. However, there the logic may be less to do with permanent scarring and more to do with a short period intense fiscal burst largely aimed at revenue spending (low enduring growth multiplier effects) that then substantially fades out in the years ahead.
- Bond demand versus supply dynamics may remain a longer term issue: As is well known, RBI has played a pivotal role in maintaining bond demand-supply balance (meaning supply absorption in a relatively narrowly defined range on yields) since the start of the pandemic. This period, and the period immediately preceding the pandemic, was also consistent with a turn in the interest rate cycle which naturally buffered the appetite for bonds from market participants as well. While going forward the gross supply of duration will start coming down consistent with a gradual return to fiscal compression glidepath, this is expected to be a relatively slow process. The government sector (including public sector entities) has had to shoulder a larger part of the growth burden over the past few years as the private sector started to slow down. This has been sound economic sense but has nevertheless eroded some fiscal flexibilities. Again while judicious management has restored some of these flexibilities, the government sector nevertheless may not be able to reduce its participation dramatically in the years ahead. While we have an export led recovery in place now after a long time, and one can hope that this heralds an investment revival and from there a boost to incomes and hence consumption, it is also true that developed markets are most likely going to slow down appreciably over the next couple of years thereby mitigating an important tailwind to emerging market exports at the current juncture. Thus even as headline deficit starts to reduce, gross supply of duration in relation to demand for it may still likely pose a tricky balance in the time ahead. It is to be noted that India’s global bond index inclusion and associated foreign portfolio flows into our local bonds being talked about for next year onwards cannot be looked at keeping all other factors constant. Thus it is likely that these bond flows will lead to more forex purchases by RBI and the central bank will correspondingly dial back on its own purchases of government bonds. Also, with a revival in the economic cycle some of the additional bond appetite from local market participants may naturally start to diminish. Hence, the bond index inclusion most likely replaces one set of buyers with another (loosely speaking) rather than materially alters the demand – supply equation for bonds in our view.
The Coming Curve Flattening
With impending policy normalization, it is natural to expect the yield curve to start flattening ahead; especially given the start point of such an extraordinary steepness. However, there is scope for considerable nuance and analysis in this observation which in turn is needed to decided which points on the curve offer the best risk versus reward trade-off. This is attempted below, in light of the 3 pillars of our framework described above.
The broad counters of the government yield curve are as under (approximate values as on 15th September 2021; source : IDFC MF)
1 year (treasury bill): 3.6% (note: equivalent government bond yields would be somewhat higher)
5 year: 5.6%
10 year: 6.16%
14 year: 6.66%
(we have ignored the longest end for lack of consistent market volumes and for the fact that its inclusion here won’t change the basic analysis)
Broadly speaking, there is approximately 200 bps spread between 5 year and 1 year, and a smaller 50 – 55 bps spread each between 10 year to 5 year and between 14 year to 10 year.
It is obvious that the maximum curve flattening ahead will be between 5 year and 1 year. This is both owing to the extraordinary starting steepness, as well as due to the nature of things: as RBI starts to lift the overnight rate, the shortest segments may bear the maximum brunt of this and so on (It is to be noted that there are various points on the curve between 1 year and 5 years as well as below 1 year. However, the general observations here may be extrapolated to those points as well). Thus, and to state the obvious, the most susceptible rate in this example is that on the 1 year. This draws from pillar 1 above and encompasses 2 aspects: One, the level of this rate is unnaturally low. Two, it is bound to have limited shelf life. The analysis thus far, and basis only the yield curve points highlighted above, seems to suggest that the 5 year is a much safer point on the curve. To further buttress this point, the 200 bps point gap between the 5 year and 1 year translates into an approximately 60 – 65 bps yield rise protection at the current 5 year point (which rolls down to 4 years in 1 year) over 1 year, mathematically speaking. That is to say, that the holder of the bond can suffer that much of a yield rise over the course of the year and still be no worse off than having held the current 1 year instrument over this period. This seems a very reasonable buffer especially when one considers pillar 2 of our framework: given our view of a relatively shallow normalization cycle and lower peak policy rates, the current 5 year bond yield may not have to rise much over the next couple of years once one factors in the maturity roll down effect as well over the cycle.
But what of longer maturities? That is to say, isn’t one safer in 10 year and 14 year bonds? Given what we assessed under pillar 3 of our current framework, we don’t expect the demand-supply balance for bonds to sustainably turn comfortable for the market for a long time to come (note this doesn’t rule out tactical plays in longer duration). Thus the steepness available at longer duration point is not necessarily playable in our view. Remember that a basic assumption embedded in this whole framework is that yields are still going to go up over a period of time; it’s just that at points on the curve the steepness provides more than enough cushion to be able to absorb these rises. However, by definition the mark-to-market loss in longer duration will be higher if yields were to rise. Also, it is hard to play roll-down in longer maturities (there may be no appreciable roll down benefit if a 10 year bond rolls down to being 9 years or a 14 year becomes 13 years). Hence, a longer duration call will have to be basis a view on duration performance and not necessarily to exploit yield curve steepness alone. Save for tactical considerations, we find it difficult to argue for this case so long as our current view remains on bond demand-supply dynamics.
Putting It All Together
The above framework and analysis clearly favours intermediate maturity points (anchored around 5 years) as the best way to play the likely evolution of policy and the expected bond demand-supply dynamics going forward. Within this one can add extra nuances. For example, corporate bond spreads at intermediate maturity points are thin and illiquidity spreads on off-the-run government bonds have shrunk. Thus there is very little justification left for running semi-liquid portfolios anymore, broadly speaking. Further, if there is indeed demand rotation forthcoming in favour of foreign buyers then the so-called ‘FAR’ (fully accessible route) securities should probably find favour in portfolio construction.
An additional point that guides us in this phase is the distinction between holding period return (HPR) and yield to maturity (YTM). When the shape of the yield curve as well as relative spreads on securities are likely to start jumping around a bit, the former concept may turn out to be of more relevance than the latter (which is more relevant in steady state markets or if one intends to hold investments to maturity). HPR is a slightly more dynamic way of looking at things when one makes portfolio choices. It encompasses aspects like willingness to give away some YTM temporarily by not chasing spread assets when one expects such spreads may rise, bar-belling when needed even if YTM gets impacted for a while, and so on.
Finally, what has been presented here reflects our current thinking. As always this can change at any point in time.