As the output gap closes, wages will start rising in rural areas as well. This will put pressure on food inflation from both the demand and supply side
When headline and core inflation diverge so sharply—as is the case in India at the moment—the key question is which eventually converges to which? Does headline eventually converge to core? Or does core converge to headline? With India’s headline inflation below 3% and core inflation at almost 6%, the directionality of convergence matters crucially and will determine the outlook of inflation and monetary policy in 2019. There are good theoretical reasons, a priori, for the directionality to run either way. If core inflation were to begin asymptoting towards that, markets will start pricing in an easing cycle. In contrast, if headline inflation starts converging to core inflation, talk of rate hikes will be back on the table. The purpose of this piece is not necessarily to ask why food inflation has decelerated so much. Or what sticky core inflation is telling us about output gaps. It is simply to use India’s inflation history to ask a narrower, more technical, question. What does past data tell us? Does headline inflation eventually converge to core? Or does core revert to headline?
To answer this question, the standard approach followed by Cecchetti and Moessner (2008) and Anand, Ding and Tulin (2014) is utilised and monthly CPI inflation data (year-on-year growth rates) commencing in 2002 is used. In the baseline data, CPI-IW (industrial workers) is used because that enables a longer time series. Then, running the same analysis using the new CPI combined series starting in 2011, the results don’t change. For CPI-IW, core inflation is defined as headline adjusted for food and fuel. A finer distinction is not available due to data constraints. In the case of CPI combined, core is defined as headline adjusted for food, fuel, housing, and transport or communication.
The tests that have been run suggest that, in the 2002-2012 period, shocks to food and fuel were either persistent and/or had meaningful second round effects, thereby impacting core inflation. The transmission mechanism was likely through inflation expectations, which subsequently affected wage and price behaviour and eventually impacted core prices. Over time, therefore, core converged to headline. The important role that shocks to food and fuel played in influencing household inflation expectations, and thereby subsequently seeping into core inflation, underpinned the Urjit Patel Committee’s choice of headline inflation as the inflation target. Interestingly, however, these results completely turn on their head over the last few years. In other words, over the last 5 years, headline inflation has converged to core, though not completely. To check that convergence has indeed changed over time—and is not an artifact of the chosen time periods—a series of rolling regressions for both equations across sub-samples was run. The results depict that, a decade ago, it was core that used to converge to headline, but over time, headline has progressively converged to core.
The results are both decisive and robust to (i) the choice of CPI series (CPI-IW versus CPI combined) and (ii) different definition of core inflation. But how should one make sense of these results? Prima facie, it points to the evolving role of inflation expectations. To the extent that inflation expectations were elevated and unanchored—as was the case between 2009 and 2014—it is unsurprising that shocks to food and fuel would quickly translate into a more generalised inflation through wage and price-setting behaviour. In other words, second-round effects were large, inducing core to converge to headline. However, given the adaptive nature of inflation expectations, the sharp disinflation in recent years has likely caused expectations to both moderate (visible in RBI’s household surveys) and get more anchored. The implication is that second round effects are much more muted, and shocks to food and fuel prices do not propagate as strongly into core inflation. Correspondingly, core inflation is relatively more responsive to slack. In other words, transitory shocks do not get generalised as completely or as quickly. Against this backdrop, it is understandable why headline eventually converges to core inflation, ultimately reflecting underlying slack. This also perhaps explains why food inflation has seen such a sharp deceleration in the last few years and yet core has remained relatively sticky.
What these results suggest is the headline inflation—expected to be in the 3% handle in the near future—will eventually start converging, over a 12-month period, towards core inflation which is currently running above 5%. If this were to come to pass, space for any monetary policy easing cycle—notwithstanding a one-off cute in February or April this year—would virtually evaporate. How would the economic transmission occur, though? It would have to be through wage dynamics. Specifically, as output gaps close, wages would be expected to rise, eventually also translating into higher rural wages, which would be expected to pressure food inflation from both the demand and supply side. Real rural wages have already begun to tick-up, though largely reflecting the disinflation in the rural economy. Expansive fiscal policy—that both expedites the closing of output gaps and targets stimulus to the rural economy to push up wages and food prices—would simply reinforce and accelerate these dynamics. If these results hold, any monetary easing in early 2019, is likely to be transient and shallow.
It is also possible that, one can look at these results and draw the wrong conclusion. If headline inflation converges to core, should RBI target core inflation instead, as developed market central banks do? Why target headline? Reaching this conclusion, however, is fraught with risk. As has been previously found, inflation expectations in India are both adaptive and rational. Consequently, the transition to the inflation targeting framework has also played a role in anchoring expectations and, therefore, de facto contributed to shocks to food and fuel having less pronounced second round effects. In effect, therefore, headline inflation gradually asymptoting to core—and not the other way around—is perhaps the best testimony to the growing efficacy of flexible inflation targeting in India. Any backsliding from there runs the risk of causing expectations to get unhinged again, and for transitory shocks to get more generalised again.
Chinoy is chief India economist and Jain is an economist at JP Morgan Chase