India joins Global Bond Indexes: What it means and What it doesn't

Indian bonds will be added to global bond indexes from 2024, says JP Morgan. An index they build, called the GBI-EM Global Diversified Index, will add India's government bonds to their composition. The weight will reach 10% eventually, starting June 28, 2024 with 1% weight and increasing by 1% every month.

The JP Morgan note (no public link) says that this index has around $213 billion linked to it. This will mean that if India reaches 10% of it, we will see inflows of $21 billion, or around Rs. 170,000 crores.

Sounds brilliant, but is it? It's an excellent first step, but it won't do much about the following things:

It will not materially impact the rupee versus the dollar
It will not dramatically increase the demand for the Indian bonds
It is not going to magically fix the current account deficit; we get large surplus inflows anyhow.

We will explore.

Why is India not already in bond indexes?

If there are bond indexes, why aren't Indian bonds already on them? In the past there were many issues. The RBI did not allow foreign investors to own shorter term bonds (maturing within 3 years of purchase), and there were limits on how much out of the total issuance of a government bond could be owned by FPIs. These restrictions were a problem if Indian bonds were added to an index. Because indexes, per se, require the freedom to be able to invest in the underlying securities without dealing with restrictions.

Therefore, these restrictions were removed in 2020, by a new concept called "Fully Accessible Route" (FAR) for certain bonds in 2020 by RBI. (Geek note: Read circular)

FAR bonds are like those Politician types that break all sorts of queues. There are restrictions, but they don't apply for these bonds.

Now, FAR bonds are huge - over 25 lakh crore ($330 billion) of issuance has happened, with 31 bonds qualifying, maturities between 2024 and 2051. Given this, adding Indian bonds to the indexes makes complete sense, and they are now easier to invest into. India will join a whole lot of other emerging economies in the bonds, like China, Indonesia, Thailand, Hungary and even Egypt.

What about the tax thing?

India charges tax on bond income make by foreign investors in two ways:

Capital gains tax: when they sell bonds, FPIs are taxed capital gains taxes according to Indian laws
Interest tax: there is a "withholding" tax of 5% on interest (or "coupon") paid to foreign investors

This is different from most other foreign countries that will charge an interest tax but no capital gains tax. But note that the taxes they charge can be considerably higher than India: (Source)

The Czech Republic charges as much as 35% tax on interest! Brazil too charges 23%, and Indonesia 10%. This is much higher than India's 5%. And the Indian capital gains tax only applies on sale and is very low (10%) if you hold a bond for one year or more. Relatively speaking, this is not as much a problem, but the issue is really in estimating returns. An index fund that sees outflows may have to sell certain bonds and could incur short or long term capital gains, and this will not be known in advance, so their tracking error will fluctuate. This can be an issue but by and large, the volatility should even out in the longer term.

The tax thing wasn't so much an issue, though it was debated for a long time. Finally, with the size of issuance available, India became a lot more easy to invest in.

Does this mean foreign funds will buy all Indian debt?

In Marathi, there is a word called "Thamba", meaning "hold your horses". So, please Thamba.

India has 98 lakh crores - around $1.2 trillion - of government debt outstanding.

Out of this, some $20 billion will be bought, $2 billion at a time, between 2024 and 2025. That's less than 2% of total outstanding debt.

Every year, the government issues new bonds to finance their deficits, and some bonds mature too. This year they will issue Rs. 15 lakh crore of new bonds (gross) - but of this around 3 lakh crores will mature, so net issuance is about 12 lakh crores. Let's consider the gross amount - 15 lakh crores.

The entire bond index based buying will be $21 billion, or Rs. 170,000 crores. That's only 11% of the gross issuance of bonds. And then, that's only for one year - after these inflows, there won't be so much more inflow.

Will bond yields cool off?

To some extent, yes, because those flows will come in over a year. So about 16,000 cr. per month can be assumed to be bought by FPIs who follow these indexes. That's not a lot - the government issues over Rs. 130,000 cr. worth bonds every month. So they will be a small part of overall issuance.

Secondly, does this attract new investors into Indian bonds? The answer lies in a peculiar concept called "spreads". What's the interest rate difference between the US long term bond and the Indian long term bond? That spread has fallen to multi decade lows now, at just 2.6%!

Usually this spread indicates that the US interest rates are very low and Indian rates are higher. In another way, it reflects inflation expectations in either country.

What seems to be happening now is - India's at about 7.15% for the 10 year, and the US is 4.5%. If the US 10 year yield keep rising, then this spread is even smaller. India isn't that attractive or developed to be able to draw lots of debt inflows when the difference is this small. So don't expect huge participation by active funds into Indian government debt.

Therefore, price demand will probably remain good only from index funds, only for the period they invest.

Does it help the rupee?

Note that all the bonds in this index will be Indian rupee denominated bonds. Therefore, the government (which issues the bonds) will not have to worry about the rupee's devaluation against the USD. Why is that a problem?

Imagine you're the Indian government borrowing, say, 100,000 cr. rupees (heck, why not, if you're imagining) at 7%. Your cost is Rs. 7,000 cr. per year.

If you borrow in rupees, what you have to pay back after say 10 years is 100,000 cr. (principal) and 7,000 cr. every year (interest). Finished, that's all.

If you borrow in USD where rupee depreciates:

Say USDINR is Rs. 80 at the start. So you borrow $12.5 billion (that's a 100,000 cr. in rupees) issuing a 10 year bond
Say the interest is just 5% per year
Say the USDINR falls to say 100, steadily over 10 years. (Rs. 2 per year)
In 10 years, you will have to return Rs. 125,000 crores
And if you add the increasing cost of interest in rupees, since the coupon is in dollars, the effective interest rate is 7.11% for you.

This is assuming only a 25% depreciation in 10 years. If the rupee depreciates by 4% a year, the effective interest rate increases to 8.7% and so on.

This is what killed the likes of Malaysia and Indonesia during the East Asian crisis of the late 1990s, where debt was denominated in foreign currency.

India's bond index additions are not going to be vulnerable that way - the entire currency risk is borne by the person buying bonds. In the East Asian crisis, every devaluation of the local currency hurt the debt heavily, so the central banks then had lesser ammunition to fight a speculator from hurting the economy. In effect, the exchange rate impacted the solvency of a country, when it should be the other way around.

In that context, the rupee is somewhat helped because the rupee fluctuations against the dollar don't impact the bonds that will be added to the index. Usually, index additions are unhedged against currency fluctuations.

Another smaller term way it will help is that inflows will help the rupee temporarily when the money comes in. How much will come in? Around $2 billion a month. This is not significant, so at best it's a small help during 2024-25. Don't count on it for saving any runs on the rupee; that's still an RBI job.

Expectations of high rupee depreciation are absent

If you think that oh my goodness, the rupee will fall and keep falling, the markets don't seem to reflect that. As of Sep 29, the one year future for the USDINR contract on the NSE trades at Rs. 84.6, while you can buy the Spot USDINR at 83.07. That's less than 2% depreciation expected one year ahead.

In fact, right now, the rupee forward premium is the lowest in a decade. It's like 1.7%!

A low forward premium means we aren't expecting to see much of a depreciation anyhow. Even with adding our bonds to the index, this hasn't changed that much.

RBI protects against Foreign Inflows even with a Current Account Deficit

The current account deficit has fluctuated quite a bit, but the important point to note is that it is usually financed very heavily by a combination of:

Portfolio investments from abroad (FPI) into debt and equity markets
Direct investments (FDI) - startups, Indian subsidiaries etc.
NRI deposits (tiny amounts now, but was prevalent earlier)

In the 1990s it was NRI Deposits that helped bridge our current account deficit.

In the 2000s it was FPI. Foreign investors flocked to Indian markets.

In the 2010s all the way till end 2021, it was FDI that was dominant. Since then we've seen FDI slip as startups struggle to raise capital, but FPI has come back strongly in the last two quarters.

The additional money that bond index inclusion will bring is relatively small compared to the FPI or FDI flowing in, and it will only be income for a year. The inflow is $5 billion a quarter, for about a year. Which is nice, but not a game changer.

Read: What is the Balance of Payments?

Case in point: the Current Account Deficit for the April-June-2023 quarter was $9 billion, but FPI (mostly in equity) was $15.7 billion by itself. There was an additional $5 billion in FDI, and $2 billion in NRI deposits. (What happens to the extra? The RBI gobbles it up and bolsters reserves, or the rupee will appreciate to compensate)

In 2022, there was a massive exodus of FPI investors in equity, and that still didn't hurt the USDINR much because the RBI sold dollars to compensate.

In short, we don't have a massive CAD problem, and if there are outflows, the RBI has the strength to keep the exchange rate less volatile. RBI forex reserves are over $600 billion again (if you consider their forward exposure too) and the only reason that we have seen a rise in USDINR Is because the RBI is buying like there's no tomorrow:

RBI Reserves:

So, to put it mildly, getting $20 billion in a year - $5bn in a quarter - is not meaningful for us. The current account is already getting bridged, and in all probability, the RBI will buy any new incoming dollars to not let the rupee appreciate.

(I don't agree with the idea of RBI controlling our exchange rate so much, and eventually they'll have to let the rupee trade freely. I just hope I'm alive when they do)

Does this mean the government will borrow more money?

Governments will take money any which way they get. And interest payments are the biggest cost that the government faces every year. A drop in yields would help reduce that, but it looks only likely if the western countries reduce rates - India's rates are at the lowest spreads compared to US debt. If the yields don't fall, our government's budget deficit will only widen if they go bonkers on borrowing, so the shorter answer is: no.

More importantly, of course, consider the fact that:

this bond index inclusion will bring new demand of Rs. 170,000 cr.
If our bank deposits go up 12% (that's about how much they've grown in the last year), we get an additional 22 lakh crores in deposits. Roughly 18% of that has to be in government bonds
That means just bank SLR requirements due to local deposit increases will add Rs. 400,000 cr. in demand from banks
That is much larger than the bond index inclusion impact.

The TLDR version

We'll repeat this, about India getting added to bond indexes.

It will not materially impact the rupee versus the dollar even next year, as the flows are small for India
It will not dramatically increase the demand for the Indian bonds. It's only 1/10th of gross issuance, and Indian deposit growth will do 3x more demand.
It is not going to magically fix the current account deficit; we get enough inflows already, and the RBI will likely buy to avoid the rupee from appreciating.

It's a nice thing to hear about, and that's about it. We should be in foreign bond indexes so that eventually we can internationalize the rupee properly. It's a good first step, that's all. In fact, we shouldn't even expect it to happen - too many geopolitical issues can happen and derail the geo-economics of the whole thing. I'd never get too patriotic about stuff like this; this is not about India having arrived and all that. We're late. So late that it's like walking in to a dinner with a swagger when everyone else is already having dessert.

Expecting too much of what are relatively tiny flows can be just as harmful as, erm, going through insane traffic to a Trevor Noah show in Bangalore.