Thursday, June 05, 2008

India needs an `inflation report'

All countries experience shocks to inflation. The countries with badly structured monetary policy regimes are those where these shocks get embedded into the structure of expectations, where inflation is more persistent, where the economic cost of wringing inflation out of the system is higher. Ila Patnaik had an article in Indian Express about care in thinking about fluctuations of inflation, and the urgent need for an `inflation report'.

Here are examples of some well written inflation reports - UK, Turkey. A good inflation report is one that is intellectually sound: it is rooted in a conceptually correct understanding of the macroeconomy. The people writing the report have to be clear-headed thinkers, grounded in modern macroeconomics, who understand what is actually going on in the country. The authors of the inflation report should have no conflicts of interest other than honestly talking about inflation: there should not be an effort at covering up for mistakes or upholding a broken monetary policy framework. And, it must be written in plain English, not mumbo jumbo, so as to influence the thinking of millions of individuals in the economy.

What shapes our outlook for inflation? The INR/USD has depreciated in recent weeks, which will fuel inflation in coming months. The small rise in WPI Fuel will also worsen inflation in coming months. These factors need to be properly understood by the authors of an inflation report. On the subject of exchange rate pass through, while one can do a little work, even elementary eyeballing of a graph is quite revealing.


  1. But wouldn't the first step (before writing inflation reports), be to build a clear and consistent measure of inflation itself? India uses the WPI in contrast to world practices, and has no unified measure of CPI across the nation.
    Moreover, on inflation targeting, determining what to target itself seems to be a question - the WPI, an aggregated form of CPI or core inflation? Critics argue that the central bank must target core inflation as it cannot control noise in headline inflation and may signal the markets wrongly.. I'd be interested to know your thoughts on this.

  2. Yes, measuring inflation better is very important. We should certainly get to it. It is shockingly cheap to measure inflation properly.

    Whatever is measured today: for better or for worse, it's being used for policy analysis and policy formulation. The question is: Can this downstream process of analysing and responding to inflation be done better? I think the answer is Yes.

  3. If it's so cheap, why isn't a private econometric firm doing it? Surely RBI and MoF doesn't seems to be listening or care, although they have access to vast amounts of data. I am sure there is a market for the inflation report and analysis that goes it.

    For example, I think, in US, a private economic analysis company declares if US is or has been in a recession and for how long - after the fact, of course.

  4. I do think there are many private (international, to the best of my knowledge) firms that do such analysis. But their data is often exclusive. Consensus Forecasts is one example. Going back to what Ajay said, I completely agree that there needs to be better presentation/dissemination of what the RBI thinks/wants to happen to inflation - beginning with a clearer measure of it, moving to a specific forecast that is published in a regular format that can be compared across different reports, and presenting the foreseen impact of this on growth and employment.

  5. Inflation measurement, like much of the statistical system, suffers from the `public goods' problem. Knowledge is non-rival and non-excludable. Once a fact - this month's CPI - is out, it costlessly flows from one person to another. It is hard to restrict one additional person from knowing the CPI.

    An agency like CMIE would hence find it difficult to recoup the costs of producing a CPI. When large databases are involved - e.g. firm level data - there is a good degree of excludability and it's possible to have the private sector build a statistical system. But when it comes to the prominent macroeconomic time-series like CPI or unemployment or consumer confidence or IIP, this is hard.


Please note: Comments are moderated. Only civilised conversation is permitted on this blog. Criticising me is perfectly okay; uncivilised language is not. I delete any comment which is spam, has personal attacks against anyone, or uses foul language. I delete any comment which does not contribute to the intellectual discussion about the blog article in question.

Please note: LaTeX mathematics works. This means that if you want to say $10 you have to say \$10.