The clarity of having a personalized index crystallized all the steps needed. It was a matter of hours to create a list of our own regular purchases with specific quantities of each that we consumed in a month. My list ended up with around 100 products covering staples, personal hygiene, household cleaning, cosmetics, fruits-and-vegetables and packaged food. This looked like a fairly good list to begin tracking.
The next step was to find a source for each of these products to get regular prices from. A few trials on sites we regularly use showed that no single site was adequate to cover the 100 odd products in my list. But there was one (which shall remain unnamed to keep my scripts going!) which could give me prices for 80 products. That looked like a very viable option.
The easy part next – a script to pick up prices for those 80 products every day early in the morning. A few tweaks to make sure we get the data correctly and can handle the errors that can get thrown and the tool was ready to get going. The first day when I collected a full set of data for the 80 products was a special day, in more days than one. Feb 29th, the Leap Day made for a perfect reference point. The total price of the basket became my Index.
The cost of my basket of 80 goods, at their respective quantities, on Leap Day 2020 was ₹15,928.23. The first Real Price Index.
Leap Day 2020, 29TH FEBRUARY 2020
This became the benchmark for measuring the change every day. Not every product price could be picked up every day. Either due to availability issues, or more likely due to the website rejecting our crawl request. The index calculation evolved to use the most recent available price, if the day’s price was unavailable. A freshness parameter was associated with an index to indicate how current it was. A few missing prices wouldn’t really invalidate the index.
The final piece was in putting together a front-end to view this data. realpi.tanksali.com was born.
With the quest begun, it was now a matter of trying to connect the dots.
I read all about inflation from my daughter’s economics books, read many results that Google presented and ended up no wiser. The theory was simple enough. Something that used to cost X at a given time costs (x + I%*x) when inflation is ‘I‘. Find the base price at a given time and its current price and elementary math will give you the inflation.
I hunted around and found a table of “somethings” that is used to calculate Consumer Price Index (CPI) in India. This table is constructed from a list of everything that people in the country spend on – food, housing, transport, clothing, education, fuel and lots more including even intoxicants. There are tables, organized by state to account for differences in consumption patterns. Data is collected from over 2300 urban and rural locations across the country. Every month. This makes it look pretty sound and a robust and defensible metric. But, I wondered, does it represent something real – something that I could relate to in my own expenses or is it like the proverbial data of the average person with 2.2 children who no one can ever meet?
Looking around at how other countries arrived at inflation did not yield much satisfaction. They all followed similar patterns with similar high level categories. With data collected at a similar periodicity.
This called for a deeper dive. Going beyond the categories to each individual item resulted in a list 300 items long, an eminently manageable list. Each item looked like we could find the price for it online, in real time, from a suitable e-commerce provider. Food was easy, clothing with a bit of stretch, things like fuel also were manageable, but education, housing and transportation were trickier.
That started the challenge of finding a online source of price for each item in the list. Two thirds were relatively easy to manage and the rest required some calisthenics to reach. In this process of mapping the CPI index to online sources, a bigger question arose. The list of items were such that, in my opinion, there wouldn’t be a single family anywhere in India that would consume that specific basket of goods. If we did proceed with obtaining appropriate prices for all the CPI components, would it still reflect actual inflation observed by any single person anywhere? Would it relate to anything that I can experience?
The answers to both these questions clearly showed that a changed approach is needed. A more personalized one. What’s needed was a personalized Price Index.
This should show, in very concrete terms, that I will actually be paying X% more (or less) than the previous month for my needs. It should be a measure of an impact on my wallet that I can correlate directly with this index. It should be as close as I (or anyone) can get to having inflation index that you can almost touch and feel. Something true. Something Real.
I am not talking about the Big Bang or the Inflationary Theory of Cosmology but the prosaic, far more mundane, almost boring inflation that Economists talk about.
In 2018 the RBI, in its bi-monthly meeting, decided that inflation was higher than it should be and they should increase the interest rate to rein it in. This meant prices are rising faster than desirable. However, nearly everything I could buy was cheaper than at any time in the previous few years. This was driven largely by the boom in digital payments, nearly all of which involved some form of cashback, discount or offer that made prices really low. This divergence between my personal experience and the view that central bank carried made me start looking at prices more closely. I began looking at personal household expenses to see if I could correlate my experiences to the macro data that was reported periodically. To my disappointment I found that I could not discern any direct correlation at all. Obviously, inflation was a very well established metric and calculated by experts at every country’s central banker/government etc. But then why couldn’t I relate it to my personal experiences? Perhaps it was just that my methods were too informal. Or perhaps it was to do with the averaging effect over a billion++ people.
Feb 2016: CPI was 123.8
Feb 2019: It was 138.6
an inflation of 11.9% relative to an year ago
FEB 2020: 147.7 – 6.5% over the year
Govt of India statistics
When seen on the macro scale of my life, inflation was so clearly obvious that it did not need math. Salaries of engineers fresh out of college had moved from a few thousand Rupees when I started working to the few tens of thousand – a 10x increase for what was (arguably) the exact same type of work. School fees that my dad paid for my schooling was approximately a thousand times lower than that I paid for my daughter’s schooling. Key groceries like sugar were 20-30 times higher now than in my childhood. But why couldn’t I observe this factor playing a role in my daily life? Too small a change on timescales of months or even an year? And yet, there seemed to be a very clear expectation that changing interest rates by 0.5% will change people’s behaviour.
Inflation of 5% – 8% were the norm in the last decade. That meant nearly everything should be twice as expensive as ten years ago. Rentals in my area were very similar, perhaps 15% more than they were 10 years ago. Staples were more expensive but I felt that they were not 2x – though the lack of personal records meant that I could not really tell. Milk did seem to be 2x from memory while domestic help seemed to be at about 4x.
Going further back, I recollect a company, who we used to supply our products to, which had this unusual practice of linking compensation directly to inflation. Each employee would get a performance rating at the end of every year. This would translate into a salary raise as rating+a base (which was equal to the country’s inflation rate). This, when I first heard of it, was sort of an eye opener. I tried to apply this in my organization and it fell flat – a 5-6% base raise (which was the inflation then) for a software engineer at that time would have elicited either convulsions of laughter or a quick departure to a saner company. But my interest was piqued.
Then there were the doomsday predictors – the insurance salesmen – who would whip out sheets which showed that nothing you saved could help once you stopped earning, as everything would be super expensive by then. Medical expenses for the merest consultation would cost more than a year’s income after retirement, children’s education would take the arm and leg it does now and one will need to trade in a kidney as well. The source of these enormous projected costs? Inflation.
Amidst all this were economists who would, in more recent times towards the end of 2019, be confidently predicting that inflation was falling and people will spend more because interest rates were reducing and yet my housing loan EMI stayed where it was, with such a minor change in the tenure that I would need a microscope to see its movement.
So what really is inflation that is relevant to me and, more importantly, what role does it play in my life? That’s the question that started this quest.
