Defining an ideal interest rate regime: Financial Express August 6th 2015

What should be the ideal policy interest rate, or in our case repo rate? We keep saying it should be lower when we speak from the investor’s perspective, and argue for higher rates when we represent the savings community. A useful barometer is CPI inflation number as we have decided to adjust the repo rate according to this. It is helpful as we know what should be the anchor, which was uncertain earlier as economists tended to play safe with the cliched tradeoff between growth and inflation.
With the direction being clear, what should be the ideal policy rate which is linked with inflation? The real rate of interest can be calculated as being the difference between the policy rate and repo rate. Hence, can this ideal be defined? RBI is speaking of a rate between 1.5% and 2%. While policy-makers target inflation, the real interest rate is not overtly the objective of RBI. To get a fix of this number, we can look at what the global trend is in the difference between the policy rate and inflation. As expected, there is no clear picture and there are countries with high and low real rates, as table 1 shows.
The table highlights real interest rates for countries with highest and lowest levels as of a specific date—inflation numbers would change for all the countries and hence this is only illustrative. The real rate for India, at 1.85%, looks like being at the median for countries with higher levels. The right panel gives the countries with lowest real rates after excluding countries with negative inflation. It is hard to put one’s finger on the ideal rate.
Another way is to carry out a regression on data for a cross-section of 42 countries to investigate the relationship between repo rate and inflation rates. Extreme values of inflation have been excluded as they distort the results. This is a useful way to conjecture as to what should be the ideal repo rate for a given inflation rate based on a global statistical relationship. The results are very powerful as the equation explains 78% of the repo rate across the countries (called coefficient of determination). The coefficient for inflation rate is 1.162 with an intercept of 0.41. Broadly speaking, it says that an inflation rate would have a multiple effect on repo rate by 1.162 times, to which we add the intercept term. Using these results, the ideal repo rate comes to 6.69%, which can be rounded off to 6.75% as central banks normally change rates in multiples of 25 bps.
This result suggests that if inflation remains at 5.4%, which is the case now, the policy rate should be 6.75% based on the defined norm. The implication is that there is scope for reduction of 50 bps in repo rate to bring it to global levels, provided inflation remains at this level. But if we extrapolate RBI governor’s assumed target of inflation of 6% for January into this model, then it will suggest a repo rate of 7.4%, which can be rounded off to 7.5%. The limitation of this exercise is that the regression has been performed on single data point on a particular date. If the same can be done for a period of a quarter or year for both inflation and policy rate, a clearer picture would emerge for fine-tuning the exercise.
Another issue is the spread between the policy rate and the 10-years G-Sec. The latter is globally accepted as being the benchmark for all market rates and hence is indicative of the level of all interest rates in the market. The premium for a set of countries at the higher and lower ends has been presented in table 2. Countries which carry negative premium or discount to policy rates have been excluded.
The differential between repo rate and the G-Sec rate appears to be on the lower side for India, which is not surprising because, based on time value of money, the 10-year yield should be significantly higher than the repo rate. Further, a regression analysis has been carried out in an analogous manner based on cross-sectional data for 42 countries. The coefficient of determination is lower at 68%, but the coefficient of G-Sec rate at 0.89 is significant with an intercept of 1.60. Substituting the repo rate of 7.25%, the idealised 10-year G-Sec rate should be 8.03%, around 20 bps higher than current market rate. The current yield of 7.8% would be consistent with a repo rate of 7%, while it would be 7.6% for a policy rate of 6.75%.

The conclusion is, first, the repo rate is a bit on the higher side for inflation at the current levels. However, as monetary policy should be forward-looking, the expected inflation of 6% can warrant an increase in interest rates. RBI has probably pitched between the two as while inflation is stable today, there are chances of increasing in future. Hence the stance of status quo appears to be in order.
Second, the G-Sec yields are probably still on the lower side given the repo rate. But if the market is efficient and is moving on the basis of expected repo rate changes, then maybe it is at the right level. However, as stated in the beginning, this exercise is based on single data points and an average across a time period would probably make the picture sharper.
This, in fact, can be a template worth experimenting with to arrive at ideal policy rates.