Why the case for foreign currency sovereign borrowing is not tenable

According to the current year’s Budget, out of the total proposed borrowings of ₹7 trillion during the current financial year, ₹700 billion may be in foreign currency. To put the proposal in perspective, in what follows, the cost ( risk of loss due to rupee’s unexpected depreciation) benefit ( interest savings due to lower nominal borrowing cost in foreign currency) analysis has been done given the reality of the exchange rate risk being by far the most serious financial risk!

Let us assume that the currency of sovereign borrowing is the US dollar and about $10 billion (assuming exchange rate of ₹70 to a US dollar) is borrowed for 10 years at the current 10 year US Treasury yield of 1.85%, rather than ₹700 billion at the current going 10 year G Sec yield of 6.35 %. Although it is a strong assumption that the Indian government can borrow at the same rate as the US Government, it doesn’t detract from the larger point that has been made in what follows. Since, unlike the private sector, comprising exporters who earn foreign exchange and thus enjoy natural currency hedge on their foreign currency borrowings, and importers who spend their foreign currency borrowing, Government of India will need Indian currency for domestic spending, and will, therefore, have to either sell $10 billion borrowed abroad on an outright basis to get ₹700 billion or , do a swap, like what RBI recently did with banks in India. In the latter case, the fair value per annum forward hedging premium will be the difference between 10 year G Sec yield of 6.35% and 10 US Treasury yield of 1.85%, that is, 4.5%. So the effective cost of such swapped/currency -hedged US dollar borrowing will be 1.85%+ 4.5%, that is, 6.35 % per annum, the same as Government of India would pay to borrow in Indian currency with no benefit of any cost reduction. But, of course, such currency-hedged borrowing , without any benefit of cost reduction, and if anything, at a somewhat higher cost, will still  make sense in the scenario where the domestic private sector does not have access to external borrowings in foreign currency which, if anything, is far from the case currently, given the fact that  as of 31 March 2019 , such foreign currency denominated external debt was about $350 billion , which , incidentally , but significantly, is about $80 billion less than the current level of forex reserves of $430 billion! Borrowing abroad in foreign currency is a veritable problem for the private sector when Current Account Deficit as a percentage of GDP is unusually high,  and forex reserves cover relatively low , as indeed was the case in 2013 when forex reserves as a percentage of GDP were about 18% and Current Account Deficit was about 4.7% of GDP. This amounted to Forex Reserves Coverage Ratio ( FCR) , like the the so-called Liquidity Coverage Ratio (LCR) , of 18/4.7, that is, 3.8 times, quite unlike the FCR for the year 2018-19 of 7.6 times which works out to two times ( 7.6/3.8) the FCR for the year 2012-13! Thus, it is no brainer to see that external borrowing in foreign currency by Government of India to supplement forex reserves kitty in the above scenario would be unexceptionable. But as discussed above, that is no way the case now and, therefore, this borrowing option must rule itself out because excess rupee liquidity created by excess forex reserves will have the downside of sterilisation costs to it. 

Now let us consider the other option of outright sale (as opposed to the currency-hedged one) of US dollar proceeds of external borrowing in exchange for Indian Rupees and leaving the currency exposure unhedged for 10 years until maturity because it is only, and precisely,  then would the benefit of nominally lower  US dollar interest rates accrue but against the risk of rupee depreciation ! It turns out , based on the annualized average (2.20%)and annualized volatility (6.75%) parameters , derived from the historical time series of the daily percentage changes of the $-₹ exchange rate over a fairly long 19 year period from 2001 to 2019, that while the expected exchange rate at the end of 10 years works out to ₹89 to a dollar, there is 99% probability (the closest proxy for a Black Swan event outcome) that the Indian Rupee’s maximum unexpected depreciation will not exceed Rs 140 to a dollar at maturity of the 10 year sovereign borrowing !  Put another way, what this means is that there is only 1% probability that the Indian rupee will depreciate beyond ₹140 to a dollar at the end of 10 years . Equally, based on the same parameters , in the opposite direction, there is 99% probability that the rupee’s maximum unexpected appreciation will not exceed ₹52 to a dollar at the end of 10 years. Since, contextually, it is the maximum unexpected loss , and not maximum unexpected gain, that matters , the rupee depreciation has been considered . Having said that , it doesn’t mean that such a massive depreciation will happen but only that  what happens if, and when , that does happen ! So, if the worst comes to worst , 10 years from now , the Government may have to pay ₹1.4 trillion to redeem the principal amount of  external dollar denominated unhedged debt of $10 billion (₹700 billion) , implying a loss of ₹700 billion ! Of course , the actual loss will be net of interest savings adjusted, again, for the dollar’s unexpected appreciation. Specifically, the annual interest payments at the rate of 1.85% per annum  in US dollars over 10 years are adjusted for the dollar’s unexpected appreciation from ₹70 to a dollar in the beginning to ₹140 at the end of 10 years by compounding at the implied CAGR of appreciation of 7.18% . This gives ₹19.64 as total interest cost in rupees over 10 years  per US dollar of principal amount which , in turn, straightaway gives ₹196.4 billion as total interest cost for $10 billion over 10 years ( Significantly , using the $-₹ exchange rate at the end of 1,2,3...10 years, computed based on the same parameters, gives ₹200 billion which is almost the same as ₹196.4 billion).

To get the net interest savings, all we need to do is subtract ₹196.4 billion from the interest amount of ₹444.5  (0.0635*10*700) billion Government of India would save by not borrowing in the domestic market at 6.35% per annum for 10 years and we get ₹248 billion as net interest saving and which , when added to the loss of ₹700 billion, gives ₹452 ( - ₹700 + ₹ 248 ) billion as the estimated maximum unexpected  net loss on redemption of the foreign currency sovereign borrowing by India at the end of 10 years. Thus, the total cost of foreign currency borrowing adds up to ₹896.4 billion ( loss of ₹700 billion and interest cost of ₹196.4 billion) over 10 years on the principal amount of ₹700 billion , translating into an effective annual simple interest rate of 12.8% compared to 6.35% per annum for domestic sovereign borrowing ! Of course, unexpected maximum net loss on foreign currency sovereign borrowings for shorter maturities like 3, 5 and 7 years will be so much lower.     

And finally , as regards the so called ‘crowding in’ of the domestic private sector to the extent of such external sovereign borrowing of ₹700 billion, it will hardly add up to anything considering the overwhelming ‘crowding out ‘ of the domestic private sector to extent of the balance domestic sovereign borrowing of ₹6.3 trillion! 

So , in conclusion, any which way, the case for foreign currency sovereign borrowing by India is not tenable given the current ground realities. 

(The writer is a retired Executive Director of the Reserve Bank of India. The views expressed are personal.)