There’s a fight happening in the intellectual circles around inflation, and you shouldn’t be surprised. It involves the terms “model” and “regression” and stuff that involves figments of people’s imagination, but in the end tries to answer the question most important to the RBI: Should we even bother about RBI when we look at inflation? Is Monetary Policy useful at all?
Before my opinion muddles the issue further, let’s see what happened. In a column titled “Where Monetary Policy is Irrevant” (Indian Express, Sep 13, 2014) Surjit Singh Bhalla wrote about how “Tight monetary policy has as much to do with Indian inflation as it has to do with India’s magnificent loss to England in the recently concluded Test series.”. Essentially, he said, monetary policy doesn’t influence inflation. That RBI raises rates or reduces them has no impact on the very thing it is raised or reduced for, that is, to control inflation.
He “proves” this by using a model that maps past data. Bhalla takes inflation as measured by the Consumer Price Index (Industrial Workers) or the CPI (IW), between 1978 and 2004 and attempts to correlate it with various parameters – fiscal deficit, money supply changes, repo rates etc. He comes down to exactly one thing: Minimum Support Prices (MSP). And only that. (Why stop at 2004? Because you have to use the formula you discover up to 2004 to test if the results are correct after 2004. )
The Indian government buys food from farmers at a certain preannounced price every year. These prices – the MSP – is set so that farmers can grow corps knowing there is a buyer at this particular price (they are free to sell to others at higher). MSP changes every year, and Bhalla observes, historically, that inflation is close to 5.2% + 1/3rd of MSP changes (1 year lagged).
This, he argues, is the only thing that determines inflation. RBI can take its monetary policy and throw it out the window, because that doesn’t impact inflation rates at all. He challenges the RBI to rebut this.
At one level, this makes sense. 50% of the CPI (IW) is based on food. And food prices will move according to MSP changes. Consequently, a good part of inflation must be because of MSP changes.
A professor of economics, Amartya Lahiri, then chimes in favouring the RBI. (“Don’t blame MSP for Inflation”, Indian Express, Oct 07, 2014). Lahiri does more complex math; he decides that MSP changes are driven by inflation that’s already in the system, and then something else. The former helps predict inflation, the latter (“something else”) does not. In an argument that is already twisted enough to cause aneurysms, you find him saying: past inflation is responsible for future inflation.
To stoke the fire further, Bhalla responds saying the lack of high inflation in September was anyhow predicted by his model. (Which wasn’t true for the last two months; the data has just “fit”)
To the person who can’t care enough for the math, one guy’s telling you that inflation (which is largely about food) is impacted by food price increases (which is what MSP changes cause).
Another guy tells you that the food price increases are because there was inflation in the past, which therefore causes future inflation.
It is difficult to understand why they’re on different sides; to the lay observer, this looks like the bleeding obvious.
The point they disagree on is whether monetary policy is relevant. But it would be difficult to say this with quantitative analysis, given the relative lack of proper monetary data history. Repo rates, for instance, have only about a 15 year history from today. How can you even prove a correlation (or the lack of it), especially since monetary policy changes are lagged as long as 18 months in other countries? Even then, Bhalla’s data shows massive differences from his own model – in four of the 10 years that his paper “predicts” inflation, the model is wrong by more than 1.5%, which is huge when inflation itself is around 7-9%. In two of them, it even gets the direction wrong (falling versus rising or flat inflation). If his model is wrong, then it’s probably not considering some other factor, and then we can add all sorts of monetary and mathematical magic to fill the gap.
This hasn’t been addressed by either person but it is likely that high rates can be used to control inflation; at some level, even if MSP increases don’t cause inflation, interest rates can be used to stop them from spiralling out of control. In effect, when your car’s brake fails, would it not be obvious that you slow the car using the gears instead, instead of arguing that the gears and the brakes aren’t related?
To those looking to monetary theory as a saviour, it’s been consistently wrong for many years. Printing money causes inflation, they tell us. And yet, we have a Japan, a US and a Europe that are printing like it’s going out of fashion, and continue to beg the world for inflation. High interest rates in Brazil and India haven’t really managed to quell the inflationary beast. The base theory needs a better answer than “wait, it will happen”.
At some level, statistics is like a bikini – it hides more than it reveals. You can tweak data to make your argument sound good, if you have enough complex data to hide behind. You can argue that Per Capita Consumption of Cheese in the US depends on the number of people that died of being tangled in their bedsheets. (No, really)
The RBI’s real case is that MSP increases do cause food inflation, and the RBI chief has provided data and insights on how we can change things. But let that not disturb a theoretical argument.
Subscribe to Capital Mind:
To subscribe to new posts by email, once a day, delivered to your Inbox:
Also, do check out Capital Mind Premium, where we provide high
quality analysis on macro, fixed income and stocks. Also see our
portfolio which has given stellar returns in our year, trade by trade
as we progress. Take a 30-day trial: