The foolishness of inflation measurements

A lot of garbage statistics is used to collect inflation data and to calculate stuff like CPI (Consumer price index). Your mortgage and other interest rates and sometimes your rent are adjusted to this number; and this statistical trickery of taking data to find out the value of money and how it varies over time has a bearing on your everyday life.

Problems with how inflation is measured

Can you give a precise value for money (dollars) by measuring the value of all other things the way the Index with the inflation basket does? I don't think so.

1. It does not consider the statistical deviation in prices. Whenever an inflation data is published, I would like to know the standard deviation (variance) of the data. No mention is made; to hide the fact that the prices vary all the time! For the basket used to measure this data, what is the standard deviation of the price of the basket? The Index covers prices consumers pay for services from medical visits to airline fares, movie tickets and rents-all kinds of variable things, and you just can't add the means and come up with a value without talking about the variability or standard deviation of these things! Note that variances of sums of variables add, and the overall variance of a sum is larger than the individual variances of the variables being added. Since prices of the inflation basket constituents are quite variable, an inflation basket derived by summing together these prices will be even more variable (the co-variance terms being zero or very small).

2. The choice of basket is arbitrary, and even if the data for the basket was statistically reliable, one can't say that the basket can be used to measure the general value of money over all things for all human beings (thousands of things and millions of human beings consuming them in different ways, plus what a person consumes this year is often different from what they will consume 5 years from now). It is a measure of money for that basket only, and any conclusions about the general spending habits of people in so many other ways they spend money from that basket is outright silly.  The inflation data, strictly speaking, is valid just for that basket. Even that basket is constantly changing, sometimes because products in the baskets change (better packaging, more durable, more vitamins in the products, etc.) and sometimes because the people who measure inflation numbers (BLS) change the basket! Talk about fixing data to suit your model...

The very fact that different baskets give different values for inflation shows that there's nothing real behind the inflation number, it is all noise. If there were a real inflation number, it would be independent of the basket design, would be more robust than only a tiny basket determining it's value. The basket is not representative of anything-it is noise being sold as representative data.

3. Measuring the value of a basket in money terms, or a bunch of items in money terms, does not give you the value of money. The demand of money varies just like of other things (potatoes, electricity, whatever) and you cannot just average it out over a bunch of things and come up with "one value" of money, as a CPI calculation does. The natural variation of prices of everything, and of money (cash) in itself, is not considered at all.  It also assumes that the reason the value of money is what it is is because of increase or decrease in supply of money; and does not consider that the demand and supply of money, just like of any other commodity (if you could treat it that way), is itself variable.

To explain this further-let's say we started measuring things in potatoes, and we could control the production of potatoes so that every year we would increase the potato supply by 2%. Can this be captured by measuring everything else in potatoes and coming up with something like "the supply of potatoes increased 2% in 1 year"? I believe it is impossible. With all things valued in potatoes, and because their value is itself very variable depending on the supply and demand, you will end up with a very noisy idea of the value of things in terms of potatoes. It is very unlikely that you will hit the mark of 2%. So even in a controlled environment, when you simply increase the supply of potatoes by 2% in a year, you will not be able to come up with the exact value by your statistical data taking of market prices. The same holds if you measure it in dollars or pesos.

The basic idea of measuring things to get the value of money is not bad-you can get an idea, a very rough idea about inflation, if EVERYTHING in your basket goes up in price by a significant amount (at least 10%). For small variation in prices, less than 5%, you cannot conclude that the price of things increased because you printed more money, or because of simply that the demand of money went down in relation to other things, etc.

If there was something real in measuring the value of money using things in the inflation basket-ALL items would go up with a very similar amount e.g. 5%. Then you can say with confidence that it is the price of money which is changing and it is not the statistical variation in it which you are capturing. But because the individual components of the basket are quite volatile, the correlation coefficient (if measured, of all these prices going up together) will still be small, because the standard deviations enter into the denominator of the calculation of the correlation coefficient. Note that high co-variance (high correlation) does not necessarily imply causality.

Note that most transactions in almost all developed and developing countries are electronic-by checks, credit cards, etc. The printed currency is a small part of the overall money in the system and is responsible for a small dollar value of transactions. The biggest transactions are almost always 100% electronic- by direct bank transfers, by checks (checks are electronic in the sense that you don't need a wad of bills to pay, you write checks using your checkbook, and the depositor normally gets the money in their account electronically), by credit cards (purely electronic), etc. The idea of cash and printing money to cause increase in money supply has become obsolete in the last 50 years because of these developments.

4. Inflation results (that it is generally a positive number, and that prices are going up with time) are contrary to what every human being tries to do in life-save money, get a better deal and a constant push towards finding the same product at a cheaper price.

If everyone is given 10% more money: the salaries of wage-earners go up by 10%, the capital returns of all capital owners go up by 10%, and the rents which landlords collect also go up by 10%), it cannot be said that the prices of all goods will go up. We hate to pay higher prices for the same product-and many times will simply not buy the product at all if prices are raised. The inflation proponents in this case will say that part of the raise in salaries (or capital gains and rents) will go up to raise prices (e.g. 3% of the gains will go towards raising prices, while the rest of 7% will go towards finding better deals, etc. or just saving the money, lending it out at interest).

A raise in a salary will not cause you to run out and pay a higher price for the exact same product (only a fool would do that!)-and the natural desire of human beings to find better deals from several producers is always pushing prices lower for the exact same product. A corollary: buyers will pay higher prices for higher quality products in the same category.

Similarly, if by just having more money in your pocket you pay more for the same goods (and cause inflation), rich people are causing inflation all the time! They have more money..therefore they will pay more for the same goods than poor people? I don't think so. Rich people are not fools; and just because they have more money doesn't mean that they will overbid for stuff. Inflation caused by more money in the system assumes this behavior.

When you get more money and go to a supermarket, you don't pay more for the same things. What you do do is to buy more things-in other words, you increase the volume of your purchases, but do not push up prices. The richest buy more volume of things; but do not pay more per unit for those things. Being richer really means that the shopping cart has more items in it; and it does not mean that you have paid more for the same items.

This essential thing about volume is lost in many other fields of life as well, not just in measuring inflation. When oil prices rise people who are long the oil market celebrate-not realizing that the net revenue of the oil producer is price multiplied by volume, and if you reduce the volume too much on increasing the price, your revenues are reduced. Or, similarly, many merchants celebrate when they can increase prices; not realizing that they are digging their own grave, because increasing prices leaves competitors a chance to get in, because they can offer goods at lower prices.

One must note that it is buyers who decide the final price in any transaction; sellers only make offers. Buyers enter into the transaction voluntarily (no one forces them to buy something, they can just sit with the money, or buy something else with the money) and are the final deciders of what all goods are priced at. Since buyers are always trying to lower prices, it follows that there is a constant pressure on goods to go lower in money price, or a constant tendency of deflation. Every man wants to buy the cheapest, and deflation is the norm in that sense. If sellers were setting prices it would not be the case; but it is the buyers who set prices. Buying is a voluntary act, and they can choose not to buy, with nooone forcing them to go ahead with the transaction. Since most goods except  basic foods are in reality luxuries, it follows that buyers are always lowering prices of goods, shopping between different producers to find a better deal on the same or similar products.

However, buyers will pay higher prices if the product improves; which is really the best use of their increased wages (or capital gains or rents). This explains why over long periods of time (several decades), when you can see that wages do go up, why goods seem to go up in price. Many goods are now better quality than before, and comparing even simple products at two different points in time, the points separated by several decades-will make you realize how much their quality has improved. The quality, includes durability and packaging of vegetables in much better today than 50 years ago. The price of carrots seems to go up in 50 years; but if you consider how much the quality of carrots has gone up, you will realize that the gain has been nothing more than an adequate compensation for better quality, better packaged, cleaned, ready to eat and nice looking carrots rather than the same carrots being sold at higher price. In real life at a certain point in time, when we go to a supermarket, we do pay slightly more for a better product (you will pay more for cleaned, ready-to-eat carrots than unwashed carrots with their leaves attached to them). Accumulate this preference over several decades and you will see why carrots today are much better than carrots 50 years ago. I chose carrots as an example of a simple good; the real quality improvement in complicated machinery, electronics goods, etc. is much larger over a period of decades; $1000 would buy you a 2MHz Processor, 200KB RAM computer 20 years ago, today it gets you a 2GHz processor, 5GB RAM. Or in other words, the computer of 20 years ago will cost you pennies-because of the improvement in the production technology of computers. I remember paying $500 for a 1.2M pixel camera 15 years ago, today a 10M pixel camera is a standard, free feature in all smartphones which cost $200. Improvement in digital camera technology has been very rapid, and the costs of the same product have dropped down exponentially. It is the slow accumulation of higher prices for better products by consumers which caused the elimination of the 1.2M pixel camera. Let me explain this further.

Fifteen years ago, when the 1.2M pixel camera was launched, it cost $500. In one year after that, the price of 1.2M pixel camera was $400, but a 1.5M pixel camera was $500. I chose to pay $500 for the 1.5M pixel camera. Do this successively for a few years and you will realize how it was consumer preferences together with better (cheaper) technology which caused the elimination of the 1.2M pixel camera from the market.

Inflation baskets and measuring inflation using them do not consider well the improving quality of products, i.e. they do not control for product quality. While some exact products might really increase in price with time, for most products, there is constant improvement going on in quality. Most price increases you see in life on products is because of a real improvement in the quality of these products, which is not captured by inflation measurements. In the United States, the cheapest car might have cost US $5000 in 1965 and today in 2016 it costs US $15000, but the increase in price is not because of inflation or more money in the system; it is a real improvement in the quality of the car in this period. Wage increases are also completely real-and inflation doesn't "eat away" into wage increases as often quoted.

What happened to the $5000 car of 1965, why is it not available in 2016? That's because there has been a real improvement in the US (and other countries) in these years, and people earn enough money to not have to buy the $5000 1965 type car anymore. In poorer countries, you will find a car for $5000; because society there has not improved to an extent that everyone can afford a $15000 car. A $5000 car today in a poor country is much inferior in quality than a $15000 car in the US, as anyone can plainly see. It is normally much smaller in size, power, sturdiness, etc.

An improvement in the quality of products is the reason their price (as a category) may up in time, and inflation measurements by baskets do not consider these improvement in quality, attributing all the increase to an increasing supply of money.

Comparison to wage data-average or median wage doesn't tell you much

Expanding on point 1. above, just like the mean or median wage of a country doesn't tell you much, an average inflation or a basket doesn't either. Worse still, at least the wage is one number; the inflation basket is averaged over hundreds of products, and that makes it even more variable. No one made policy decisions on the basis of an increase of 2% in the median wage (or mean wage), because we all know that the distribution of wages is highly variable, and the mean or median does not account for what's happening for most people in real life. Strangely, we are made to believe that there is such a thing as an average inflation rate, and that that's representative for all the different ways consumers of a country spend their money.

We need to think of inflation like a wage data. Your wages are not my wages, and your inflation is not my inflation. Neither of us has anything to do with the basked based inflation data published by a Government agency, which is the inflation of another imaginary human being who consumes the inflation basket goods, in the exact proportion the basket says.

The case of Panama and Ecuador

The futility of inflation measurements is best illustrated by countries like Panama and Ecuador, which use the US dollar as their currency. Since they have no influence or control over how US dollars, a foreign currency there, is valued, or is brought into their country, their inflation measurements are nothing but noise-in the sense that there can't be regular pattern to it. Amazingly, they do have inflation measurement departments...where dozens of statisticians and bureaucrats are employed to come up with this data. Link for Panama here   and Here for Ecuador. Since they can't do anything if the inflation is high or low, why measure it at all!

In Ecuador, the reported annual cumulative inflation was 3.38% in 2015, and 3.67% in 2014. The geniuses in Ecuador and Panama who are reporting these numbers, massage the data to make it look like a nice number, instead of saying it is statistical noise. They publish this number to keep the banking community and financial guys happy. As long as it is between 0 and 5%, everyone is happy. Data which will make it go beyond these is readily discarded.

Digression 1-As a general comment, I think most people, including Scientists, will make up or polish data to prove what they want to prove-because their jobs, their funding, etc. depends on it. In the end, for most people, it is just a job; and a bit of data jugglery is a part of their job, just as it is for people who work in Sales and Marketing-who will make up data to sell their products. Not more than 5% will publish the data as it is, and agree that it might just be noise (and this is in all walks of life, not just Economics, Medicine and Global Warming; subjects I have covered in this blog to show you how a lot of what you see is noise disguised or polished as data).

Digression 2-As I grow older, I realize that a lot of data we have around is is just noise, and a large amount of foolishness of humanity is to confuse this random data as relevant. This is how stuff like the existence of God and astrology was born. Even though Science has advanced greatly in the times since we invented the idea of God and planetary movement affecting our daily lives, I still find that a lot of humanity still confuses noise with data. Wherever people see patterns, I generally see noise. And as Nassim Taleb said so well in one of his books-good data shouts at you, you don't need to look for it or justify it by long arguments.
 

There are a lot of holes in the basket measurement of inflation. Any number which is less than 5% and is quoted as inflation is but noise-and they are careful to hide the noise by not publishing the standard deviation of data.

Other holes in the measurement of inflation: a) they invented something called "core inflation" to remove the volatile parts of the inflation basket, food and energy. A real statistician never discards any data, here it is being discarded on purpose to remove the deviations and noise in the data, to make the inflation numbers "smooth" and b) a small change in the composition of the inflation basket will lead to an entirely different number for inflation. If the real value of money was being measured, a change in just the composition of the inflation basket should make no difference.

Inflation measurements which are increasing steadily at say 2 or 4% per year are in reality a clever way of defrauding the persons who borrow-who almost always see their interest payment go up with time, because they measure the index in a peculiar way to make it increase slightly almost every year. The noise of the measurements is removed carefully to always come up with a slightly positive number  between 0% and 5%, to give the banks a license to increase interest payments on the borrowers. Or rent payments, which are also adjusted to inflation in most big cities. The inflation data publishers meet the needs of the market well-the massage the data to come up with a number of between 0 and 5%, and get to keep their jobs. But the real value of all this data is zero, for the reasons mentioned above.

Inflation feedback loops

The most ridiculous use of inflation data is by utilities  (electricity, telecom companies, etc.) and essential services companies. Utilities use this data to justify increasing rates. They say that because of inflation, they be allowed to raise rates. Since their own products form a part of the basket of things used to measure inflation, they basically go into a feedback loop-the basic necessities companies raise rates because of inflation data, and since they raise rates, next year inflation data shows even more inflation, and they raise rates even more. Imagine if every product or service in the inflation basket started raising rates, the next year's inflation would go through the roof! Here in Chile where I live, this is why gas and electricity prices are the highest in the Americas (North and South)..and the Government lets these utility companies increase rates , who use the excuse of higher inflation.

Here's what Smith said about increase of paper money possibly causing inflation (which seemed to be a theory in 1760 as much as it now!) in Book 2 of the Wealth of Nations, chapter on "Money Considered as a particular Branch of the General Stock of the Society"

"The increase of paper money, it has been said, by augmenting the quantity, and consequently diminishing the value of the whole currency, necessarily augments the money price of commodities. But as the quantity of gold and silver, which is taken from the currency, is always equal to the quantity of paper which is added to it, paper money does not necessarily increase the quantity of the whole currency. From the beginning of the last century to the present time, provisions never were cheaper in Scotland than in 1759, though, from the circulation of ten and five shilling bank notes, there was then more paper money in the country than at present. The proportion between the price of provisions in Scotland and that in England is the same now as before the great multiplication of banking companies in Scotland. Corn is, upon most occasions, fully as cheap in England as in France; though there is a great deal of paper money in England, and scarce any in France. In 1751 and in 1752, when Mr. Hume published his Political Discourses,*36 and soon after the great multiplication of paper money in Scotland, there was a very sensible rise in the price of provisions, owing, probably, to the badness of the seasons, and not to the multiplication of paper money."

Related post: Central banks have no influence in jobs creation, unemployment or real interest rates. Here's the post explaining this.

-Sanjay