TPI©: The Complete Guide to Football Inflation

TPI©: The Complete Guide to Football Inflation
February 9, 2010 Paul Tomkins

Over the coming weeks and months, I will be publishing a number of articles based on some exclusive, highly-detailed transfer research that covers the entire Premier League era. These articles will be free to read, but will contain extra Liverpool FC-based analysis for Subscribers only, as well as the option for members of the site to discuss the findings with myself and those who generated the original data. Some of the results are as to be expected, but there are a few surprises and plenty of interesting patterns.


One of the enduring problems with comparing managers (and teams) across the eras is how inflation distorts their record in the transfer market. After all, when Nottingham Forest’s Brian Clough paid £1m for Trevor Francis in 1979, he was doing much the same as Manchester City 29 years later when purchasing Robinho: paying more money for one player than anyone else had ever done. Only now, that figure had risen to £32.5m.

There is almost infinitely more money in the game now, so that, aside from some natural sense of inflation, where the price of everything tends to rise, there is also ‘football inflation’, which reflects the increased TV revenue and, more recently, the influx of billionaire owners who don’t have to balance the books.

This project came about after Graeme Riley contacted me upon reading my own method of comparing players: the Relative Transfer System (RTS) ©. RTS was devised for my book Dynasty, as a way of comparing the signings of Bill Shankly with those of his seven successors as Liverpool manager. For this, I set every record-breaking transfer at 100%; Francis and Robinho therefore both had that value. Anyone who then cost half of the record, would have a 50% value, and so on. So, in 1980, £500,000 was 50%, whereas in 2009, £16.25m was 50%.

But I made clear that it wasn’t without its flaws; not least how a big anomalous transfer can occur, and yet the mid-range transfers that season might be of a lower average value.

What was the average fee in any given season? And did the average go down as well as up?

It was with this in mind that Graeme, Head of Tax, Treasury and Risk at a large corporation by day and author of several mind-bendingly replete statistics books on European football by night (not to mention being the unofficial Tomkins Times statistician), suggested looking at other ways to solve the problem.

And with this he came up with the Transfer Price Index ©. He ran his idea by me, I made a few suggestions, and with the help of Gordon Fawcett, we researched and double-checked every single transfer fee in the new satellite TV era; no mean feat, as you can imagine.

(Of course, it’s important to stress that deciding on a definitive transfer fee for any given player is not straightforward. Some clubs do their best to cloud the information, and ‘undisclosed’ can start a thousand internet arguments raging; that’s before getting onto part-exchanges and additional bonus clauses. But even though we looked only at trusted sources, it is impossible to be 100% accurate – and as such we had to use their educated guesses. However, we feel that the majority will be between 95-100%.)

And thus was born an immense database encompassing every transfer between 1992-2010 (January), and various calculations as to how that fee relates to current day spending.

And once we had the TPI, all sorts of interesting comparisons could be made, such as how costly each club’s squad was during any given season, and how much it had spent compared with its rivals. It would show us all the expensive players, but also, how much each transfer would equate to had it taken place in 2010.

We could also look at how much of that investment made it onto the pitch during a campaign; it’s one thing paying lots of money, but what if all your big buys were out injured? We could look at which teams spent the largest percentages of the overall Premiership outlay over the course of each 12 month period, in both Gross and Net terms, and how that affected their league position and their silverware collection.

Transfer Price Index ©

“A new method of comparing players’ values”

In everyday life, people are familiar with the concept of the Retail Price Index (RPI) as a measure of inflation. A basket of goods is identified and every month the same items are checked to see what the value would be if these were to be purchased. The difference between the value currently and the previous month’s value is calculated and termed the RPI. By comparing the value this month with the corresponding value for the same month last year, we obtain the annual RPI.

Over the course of time, transfer values of players become distorted. What a club pays for the registration of a player in one year cannot be directly compared to the cost of his replacement a decade later. The Transfer Price Index seeks to enable such comparisons by smoothing out any extremes. In effect, the basket of goods used in calculating the Retail Price Index is replaced with a basket of players i.e. those transferred during one season. (Human Rights note: no players were placed into actual baskets during the course of this study.)


There has been a gradual decline in the number of transfers for cash over the last three years, and for the first time in several seasons there was a reduction in the average value of these transfers. The rate of decline in the average is actually the highest for any season since the start of the Premier League: a fall from £4.8m in 2008-09 to £4m in 2009-10 – a drop of 16.3%.

This is presumably caused by not just the credit crunch and the effect of exchange rates, particularly against the Euro, but also the teams involved. Big-spending Newcastle have been replaced by Burnley and Wolves, who would expect to shell out less combined than the Magpies on their own.

There has also been a large increase in the number of loans since the start of the Premier League; in the first years the figure was usually in single figures, however it is now regularly in the 30s. This is noticeably higher than the number of trainees coming through to play in the current season.

And obviously, the advent of the ‘Bosman’ transfer in the mid-‘90s removed payment of a fee when signing out-of-contract players. While prior to 1995 these fees would have been set by a tribunal, and even though they never matched the ‘full contract’ worth, the ruling has still served to wipe off tens of millions from the league’s transfer expenditure.


Transfers and Trainees

The first time a player appears for a club, his status is assessed. Generally this can be transfer (for a value or “on a Bosman”), loan, trialist or a youth player from the academy (here deemed to be a trainee).

Transfers can be for cash, or they can also be for (part-) exchange or undisclosed. In this latter case the existence of the transfer is acknowledged, it is not however used in calculating TPI © as this would give a false value. We are effectively removing them from the numerator (value) and the denominator (number of transfers) for the calculation.

In cases where several players were acquired as part of the same transfer, the fee is divided equally amongst the players.

Transfer values are taken from a number of sources – these are listed in the database. For players acquired when the team was playing in a lower division, the transfer value is noted with the letter d preceding the fee, but is not used in calculating TPI. If the exact date of a transfer is not known but the month is known, the transfer is deemed to have taken place on the 15th of the month.

Having established all of the transfers for a given season and placed a cash value (or undisclosed) against each of them, it is now possible to calculate the average transfer value across the entire Premier League. By performing this calculation for each season, an index can be produced showing how transfer values have been inflated over time.

Application of the TPI

Given an indexation of the values, it is now easy to produce a Current Transfer Purchase Price (CTPP ©) for any given player. As an example, Stan Collymore cost £8,500,000 in July 1995. The average transfer price in 1995-96 was £1,594,214 and in 2008-09 it was £5,352,080, giving a price inflation since 1995-96 of 236%. This implies that the equivalent transfer in 2008-09 would have cost £8,500,000 x 3.36 = £28,536,114. A year later, however, his ‘price’ dropped to £22,217,140, due to a decrease in the average value of transfers in 2009/10.

Players purchased when the club was in a lower division still have their value inflated according to TPI, even though they have not been included in the calculation of the TPI itself. This effectively benefits managers who have identified a player’s potential and, presumably, paid a lower price for him, similar to trainees of Premier League clubs.

Final Word

None of the articles are intended to definitely prove anything, especially when facts are taken in isolation. It is simply an indication of what has been spent, and an attempt to analyse the data.