Madan Sabnavis: The cost of monetary policy

Should the RBI be spending the equivalent of 75 paise to help save the economy Re 1?

Conducting monetary policy has often been compared to driving a car in inclement weather with a cracked windscreen and a fogged rear-view mirror, with one foot on the accelerator and the other on the brake. Actually, it is more complex "" like halting at a crossroad on a narrow Mumbai road with three arrows signalling the directions, but with space for only two vehicles. You have to choose a lane but which one "" you can be fined for lane cutting.

This is the predicament of the Reserve Bank of India (RBI) governor who walks the tightrope with criticism from all sides: the "CRR cut is meaningless"; the "repo rate should have been lowered"; "sterilisation of dollars is a must"; but "where are the securities?" and so on. The question to be answered is: all this for what?

The textbook says that monetary policy serves two purposes: GDP growth and stability, or inflation. Of late, monitoring the exchange rate is yet another covert objective of the RBI. With India's real GDP at around Rs 13,00,000 crore, it can be said that the RBI is trying to affect, say, a 1 per cent change in output of prices, or simply Rs 13,000 crore of the value of output or price rise (it does not seem likely that monetary policy could affect more than this level of GDP and inflation).

To do so, it uses different policy tools to influence the schedule of money supply in the country. But in the process, the RBI necessarily incurs a cost that is actually hidden because, unlike you or I, it can always print notes to cover the cost. So, it does not really matter. Let us see what this cost could be since, say, March 31.

Looking at the composition of money supply and the instruments used by the RBI, there are essentially four costs the RBI incurs while trying to control the growth in money supply and achieve its other objectives. The first is the growth in forex reserves. So far, dollars worth close to Rs 80,000 crore have flowed into the system, which the RBI has been buying. This is being done ostensibly to stabilise the rupee and ensure market discipline.

At best, these dollars are invested in US treasuries, which carry a return of 1.5 to 2 per cent. The alternative would be to invest the rupees in, say, a 10-year government paper where the yield is 5 to 5.2 per cent. Thus, there is an implicit opportunity cost of not investing in government paper and preferring US bonds "" around Rs 2,800 crore a year.

Additionally, the RBI gets banks to keep money with it in the form of cash reserve ratio (CRR) to control growth in credit creation and, hence, money supply. The payment of interest on these balances is in two parts: the first 3 per cent earns no interest, while the RBI pays the bank rate (6 per cent) for the balance 1.5 per cent (at present, the CRR is 4.5 per cent).

Total deposits would soon be in the region of Rs 14,00,000 crore and the cost incurred on the CRR will be Rs 1,260 crore a year. Further, with the RBI contemplating a standing deposit facility for surplus commercial banks' funds, this component could be expected to move up further.

The third cost of monetary policy is incurred on repos. In order to maintain stability in the money market, the RBI holds daily repo auctions to draw out liquidity from the system and ensure that call money rates do not crash. This year so far, the RBI has been drawing out close to Rs 25,000 crore on an average daily basis through one-day repos.

This means an average annual cost of around Rs 1,200 crore (the repo rate was 5 per cent for the first six months of the year and has subsequently come down to 4.5 per cent). What is interesting here is that the RBI is actually interfering with the market forces by artificially keeping call rates up and not letting them find their true level. This is a distortion that should be done away with.

Lastly, the RBI has conducted open market operations (OMOs) to permanently draw out money from the system (as against repos where there is a buyback) by selling government paper. By mid-November 2003, it had a lower balance of Rs 56,000 crore of government paper. The loss would be the interest on these securities.

Assuming an average cost of 8 per cent on such paper, the cost in terms of loss of income would be around Rs 4,500 crore. The total cost of all these policies is around Rs 9,700 to Rs 10,000 crore. The benefit derived could, at best, be a 1 per cent saving in inflation or a 1 per cent gain in real output.

The assumptions made here could, of course, be open to subjectivity, but the issue is whether the RBI should be spending around Rs 10,000 crore to generate (GDP) or save (inflation) Rs 13,000 crore. This is tantamount to passing a value judgement of whether it is worth printing currency to gain Rs 13,000 crore. Simply put, is a cost of 75 paise justified to save Re 1?

There are two views here. The first justifies the expense of 75 paise a rupee. That's because the gains to society in the form of higher GDP growth and lower inflation vindicate the cost. Besides, the RBI is not really affected by these transactions; in fact, they help discipline the markets. Thus, there is no harm in such measures.

The contrarian view would be that even though there is no real net cost for the country, we could get more efficient. There is no need to keep accumulating dollars when we can go in for capital account convertibility and the dollars could be better invested in remunerative, though risky, avenues.

The CRR need not earn any interest, which is the global norm. Repos are artificially distorting the call market and are unnecessarily propping up returns in this market. The same holds true for OMO sales where banks should be forced to lend more rather than seek safer investment havens. As is the case in economics, all such subjects have arguments on both sides. Which one would you choose?

(The writer is chief economist, National Commodity and Derivatives Exchange. These views are personal)