Refining natural interest rate estimation for India

Natural interest rates (NIRs) are expected to exceed those in advanced economies in emerging market economies (EMEs) because their higher growth is thought to require more savings. But the mature economy homogeneous agent models in which the real interest rate equilibrates savings to investment are inappropriate. Aggregate savings rise with income, as productivity and employment expand, while rates of interest largely affect allocation of savings and investment. The latter is often the real constraint. The Reserve Bank's official estimation of the NIR follows the standard framework and finds the NIR to rise with growth. We replicate their method, for a more recent dataset, and use it as a benchmark to see how the estimated NIR varies when features relevant for EMEs are modelled. The first variant introduces the dualistic structure of the economy, with pervasive informality and still large numbers employed in agriculture. The second has an alternative data-based decomposition of the trend from the cycle. The third uses one year ahead realized inflation as a proxy for expected inflation, since using the past trend misses the impact of inaccurate forecasts on the real interest rate. We also try alternative priors. In each case we find the estimated NIR is lower than in the benchmark, which itself is about unity for 2024, compared to the official estimate of 1.4-1.9. Allowing correlation between the cycle and the trend, which is a feature of EME growth, lowers the NIR substantially.