Statistics and Inflation

Statistics and Inflation

Statistics and Inflation

The word «inflation» is part of our everyday language; we hear it on the street, in the media, on public transportation, and even at family/friends’ meetings. With dubious background and at the same time pretentious airs we comment on its causes and, even more so, on how to solve it. The information provided about it and thus the one we feed of, is often incorrect or comes from questionable sources. Even worse, statistical data on inflation can be manipulated by modifying values or omitting unfavorable data for several reasons; political, to present a more favorable economic scenario, attract investors, win elections, etc.

At other times, results may stem from methodological errors, inappropriate selection of goods and services for price indices, data collection issues, or changes in calculation methods.

Inflation is measured using price indices that reflect changes in the cost of a basket of goods and services over time. The methods for calculating these indices must follow rigorous procedures to ensure accuracy and objectivity. Even with reliable data, we can build a strategy depending on what we intend to prove.

Consider an example from the book «How to Lie with Statistics» by Darrell Huff and Irving Geis:

«Let’s assume that last year milk cost $100/L and bread $50/kg. This year, the price of milk dropped by $50 and the price of bread increased by $50. Now, what do you want to prove? That the cost of living has increased, decreased, or remained unchanged?

Consider last year as the base period, making the prices of that time represent 100%. Since the price of milk decreased to half (50%) and the price of bread doubled (200%), the average of 50 and 200 is 125, meaning prices have increased by 25%. Let’s try again, taking the current year as the base. Milk cost 200% of what it costs now, and bread was at 50%, averaging 125%. Prices were 25% higher than they are now.

To show that the cost of living hasn’t changed, we use the geometric mean and take any period as the baseline. This mean is somewhat different from the arithmetic mean but is a legitimate figure and sometimes more informative. To obtain the geometric mean of 3 numbers, you multiply them together and take the cube root; for 4 quantities, the fourth root; for 2, the square root, and so on. Let’s take the previous year as base 100; now multiply 10 by each product and take the root of the product, which is 100. For this year, with the price of milk at 50% of the previous year and bread at 200%, multiply 50 by 200 to get 10,000. The square root, which is the geometric mean, is 100. Prices haven’t risen or fallen. The fact is, despite its mathematical base, statistics are as much an art as a science. Many manipulations and even distortions are possible within its jurisdiction.»