By Paul Tomkins.
On Saturday, Channel 4’s Economics Editor Faisal Islam tweeted:
“No one ever inflation-adjusts “record” transfer fees, do they? Sigh. #mata”.
Except, of course, we do.
In 2010 I joined forces with über-accountant and statistician Graeme Riley to create the Transfer Price Index (TPI), which converts football prices to ‘modern money’ (or what we call the Current Transfer Purchase Price) with its own inflation index. After all, standard inflation relates to the price of a loaf of bread, amongst other everyday things placed into a shopping basket. Football transfers occur outside of that economic bubble, existing within their own much bigger bubble (one that gets heartily blown, I might add, but rarely by West Ham).
Now, as we note in our 2010 book, Pay As You Play, it’s not a perfect model; any given transfer fee can be obscured by the parties involved, but we do the best we can, using reputable sources, to give a pretty good idea.
And obviously, wages are another factor which have to be taken into account; a lot of good work has been done on this in numerous studies, but ours was the first to find a connection between transfer fees and success. (Remember, if Liverpool sign no-one this winter they have still committed an additional £25m-30m to Luis Suarez over the remaining four-and-a-half years of his contract. There aren’t many clubs out of the Champion League for four consecutive seasons who can agree to pay a player £200k a week. Every club has to balance its transfer spending with its wage bill.)
With the TPI project we have showed that transfer fees, once inflated, can be used to show the correlation between spending and success. Whilst we hope to fully resurrect the TPI project at some point soon, and hopefully do another book, this is a small update relating to the questions raised this weekend.
There are two aspects on which I’d like to focus: first, the relationship between transfer spending and the Premier League television deals; and second, a rundown of the top 100 transfers in the Premier League era once adjusted for inflation.
(Note: we use the Premier League era for a number of reasons, not least the management logistics of a database that has already millions of data points; and also, the clear way in which money has played a larger part in the game since the change in 1992, ensuring an era of the Haves and the Have Nots. We appreciate that football existed prior to 1992, but that represented a rational cut-off point when working back.)
Currently (and this is based on figures prior to the Mata deal), this season’s average transfer fee is £5.6m; almost ten times the average of £594, 309 in 1992/93, when this whole era got underway. Now, this season’s second (and final) window isn’t yet closed, so things can change – albeit there’s a likelihood of further increase once the Mata deal is taken into account.
But so far, based on these estimates, the average price of a transfer has risen 33.6% on last season, up from £4.2m. Interestingly, there’s a new TV deal in place for 2013/14, which means more money swimming about. The last new deal was introduced in 2010, when fees rose by 22% on the previous season; before that it was 2007, which coincided with a whopping 51.3% increase; the new deal of 2004 saw a 10.2% rise; and in 2001 the average transfer fee increased by 35.8%.
Therefore, the last five TV deals have seen immediate increases in the average cost of a transfer. There have also been increases in the middle of TV deals in that time, but only one of any note (which came in 2008/09, the second season of a new deal), with the addition of two minimal increases and three decreases all within the last eight years.
Still, overall the trend is one of transfer fees rising, as you’d expect.
The lowest increase in the season when a new TV deal commenced was just 10.2%, back in 2004; coincidentally, according to the excellent Swiss Ramble, the only new deal that paid less than the previous one, falling from £1.56bn to £1.45bn. To put that into context, the next rise was to £2.5bn, then £3.3bn, and now £5.2bn, and the rises in transfer fees in the seasons of the increases were 51.3%, 22.0% and currently an estimated 33.6%.
In other words, the new TV deals seem to act as the guaranteed credit which clubs will trade on. (As long as the broadcasters don’t do ‘an ITV Digital’, that is.)
First of all, as this season’s rise hasn’t yet been finalised, I am using 2012/13 figures; i.e. correct as of the end of last season, before the summer transfer window opened. However, for ‘fun’ (okay, it’s not like skydiving naked with Girls Aloud) I have added a column of what the fees could look like based on current projections.
The best way to get a sense of the ‘fairness’ of what a player cost in current money is to think of how much his fee was then, in relation to what else was going on at that time. For instance, in 2003/04, the average price of a Premier League transfer was just £1.97m. This was the final year of a TV, and most clubs weren’t awash with money … bar, of course, Chelsea. So if a club spent £24m on a player around this time, it would be a lot of money. Come forward just three or four years, and £24m suddenly wasn’t such a big deal; by 2008 the average price had more than doubled, to £5m.
Remember, £24m in 2003/04 was fairly close to the British transfer record. In 2013, it’s less than half of the British transfer record. Therefore, £24m from 2004 (just like £1m from 1979) doesn’t give you the correct sense of magnitude. After all, the younger generation won’t get a sense of how much £1m felt back at the time I was a young lad watching The Six Million Dollar Man. (Remember: $6m equates to c.£3m to rebuild an entire man with bionic powers. Now that amount wouldn’t even get you Djimi Traore.)
Converting it to 2013 money does just that; Didier Drogba, signed in the summer of 2004, works out at £47m in 2013 money. That sounds about right, doesn’t it? – a little bit below the transfer record, so what you’d classify as a big buy.
What I’ve tended to find when studying the fees is that the TPI record fee runs a little ahead of the current actual transfer record, but that the true record soon catches up. In 2013 money, Andriy Shevchenko remains the most expensive Premier League player (in terms of purchases, not sales), at £57.7m. This ‘feels’ like the kind of record fee you’d expect a Premier League club to be willing to spend at this moment in time.
However, the way this season is going, the inflated record in 2014 money looks set to rise to £76.8m. This tells me that before too long a Premier League team will pay in that kind of money for a player; perhaps such offers have already been made, as with United’s aim to get Gareth Bale and bring back Ronaldo, both of whom left English football for over £80m.
Shevchenko was about as nailed-on as a transfer got: genuinely world-class, brimming with big-club experience, at the peak age (in theory), and yet it turned out to be a dreadful mistake. In 6th place, Fernando Torres (£50m actual, £44m in 2013 money, likely to be £63m in 2014 money) is another Chelsea flop who, at the time, seemed like a sure-fire hit. The same is true of Juan Sebastian Veron, at 10th, and Jose Reyes in 12th (although this, as with all deals, was based on the maximum fee Arsenal could have paid, and needs recalibrating because he was sold before reaching the peak payment.)
Liverpool’s most expensive player remains Andy Carroll (ranked 20th), although his fee falls to £30.9m in 2013 money; however, by the summer it could rise to £41.1m in ‘real’ terms. Not far behind (30th and 31st) are Emile Heskey and Djibril Cissé – £28m in 2013 money, but likely to rise to a whopping £37m in 2014 money. Whilst on the subject of Liverpool, Rafa Benítez’s most expensive signing, Fernando Torres, ranks 50th, at £24m, or £32.9m in 2014 money, and Brendan Rodgers’ biggest buy is Sakho, just outside the top 100. (FSG have signed three players in the list, with a bizarre quirk of fate meaning that inflation pairs Luis Suarez and Stewart Downing at an identical price.)
Ranking 53rd is Stan Collymore, whose 1995 fee of £8.5m equates to £25.6m in 2013 money (£32.3m 2014 money). This is interesting, as even though the Reds sold him to Aston Villa in 1997 for £7m, he ranks at 46th, at £25.6m (£34.1m) for that deal, seven places higher than his ‘more expensive’ Liverpool move. This is because £7m, in 1997, was “more” than £8.5m in 1995, once inflation is taken into account. In other words, players were generally less expensive the year he moved to Aston Villa than they had been two years earlier.
Right now, Juan Mata’s move to Manchester United ranks as the 14th most expensive in the Premier League era. (Note: Carlos Tevez was listed at £25m at the time City bought him, but the actual fee has since been quoted at around £47m, which, once inflation is added, would place him second overall, at c.£71m. For now we’ve left him where he was, but we will look to further verify such transfer fees and update the database accordingly.)
The value of inflation is seen when noting that whilst announced as United’s record buy, Mata ranks way back in 4th when the fees for Rooney, Ferdinand and Veron are adjusted. (Purely for the purposes of this piece, the fees I’ve used for this season are all taken from transfermarkt.com.) It stands to reason that the £30m spent on Rio Ferdinand over a decade ago would be a lot more money than £37.1m now, and indeed, this is the heart of the Twitter conversation that led to this article.
Mata becomes Moyes’ second signing inside the top 30, after the (hitherto) unsuccessful near-£30m spent on Fellaini.
In total, United have 18 of the 100 most expensive signings of the past 22 years, but that doesn’t make them the kings of this list. That honour belongs to Chelsea, with over one quarter of the top 100 signed, and most of those in the second half of the Premier League era. City are catching up fast, mind; a staggering 15 of their 16 entrants have been signed since 2008, although depending on your view on the Tevez deal, yet to break into the top 10.
[Edit: Man City don’t have many players near the very top of the list, because they have often spent £20m-£35m on single deals, at a time when four other clubs – Chelsea, United, Arsenal and Liverpool – have also done so. Go back to 2004, and only United and Chelsea were paying that kind of fee; two clubs then, but five clubs now. Chelsea and United are also the two most recent clubs to hold the British transfer record. Where City are different is in just how many fairly expensive players they’ve bought in a reasonably short space of time.]
In terms of the number of the top 100 signed by clubs, the list is as follows:
Man Utd 18
Man City 16
Aston Villa & Blackburn 2
Plus entrants for Leeds, Wimbledon and Fulham.
Key: Players highlighted in green are from 2013/14.
[Update: Mascherano’s move to Liverpool was mistakenly included twice, the error in part because he arrived on loan in one season and ‘joined’ in another. Also, a couple of players outside the top 20 had vanished, and that has now been amended. Finally, signings attributed to Roberto Mancini in the season he took charge should have been listed under Mark Hughes, and that’s been corrected.]
In terms of the number of the top 100 signed by individual managers, the list is as follows:
Dalglish 5 (across three different clubs)
Plus others with 4 or fewer.
With just two qualifying purchases in two full years, Carlo Ancelotti averaged one top 100 purchase per season, the same as Ranieri around a decade earlier, and, at Liverpool, Rafa Benítez.
By contrast, Jose Mourinho averages over three times as many mega-purchases, in terms of those acquired per season, and already has two qualifying buys this campaign. In the six years Mourinho was away, Chelsea averaged just fewer than one entrant per season; a mere five big signings in all that time. Whilst clearly a top-class manager, Mourinho has received the backing that other Chelsea managers have lacked, albeit they inherited some of the gems from his first tenure.
One thing I will look to update in 2014 is TPIC, the coefficient I created with Graeme’s data, which looks into the success and failure of around 3,000 transfers in the past 20-odd years.
It takes into account the inflation-adjusted fee – so all transfer fees are judged at ‘current’ prices, whether signed in 1993 or 2013 – and looks at how many games each transferred player started, and if applicable, any profit or loss (when adjusted for inflation) from a sale. It can’t discern ‘greatness’ on the pitch, other than by looking at how frequently players were selected (if they rarely played they can’t be called good signings) and the total number of appearances, as well as how much they were sold for (looking at the profit margin, as both a total in sterling, and as a percentage of the original fee).
The last time I updated the figures, in terms of value for money, Arsenal’s purchase of Kolo Toure – who cost peanuts, started 203 Premier League games, and was sold for c.£20m CTPP – just edged out United’s Ronaldo for top spot.
Ronaldo was the better player, clearly, but they were virtually equal in terms of deals. Players like Jaaskelainen, Given, Lundekvam, Shorey and Kevin Davies rank in the top 20 best Premier League signings using TPIC, as well as more illustrious names like Hyypia, Lampard and Fabregas. Presumably when I update the list, Gareth Bale will rocket towards the top.
(Davies also ranks in the worst, with his earlier purchase by Blackburn a clear disaster. Torres and Adebayor are two other players who feature at both ends, although Davies is the only one to be present in both the best 20 and worst 20 transfers of the past 22 years.)
How many of the 100 most expensive deals worked out well? (I’m curious to know what percentage you think succeeded, although some haven’t even played a game yet!)
It’s logical to presume that the more you pay the greater your odds of success. Overall, looking at the 3,000 transfers in the TPIC and apply the coefficient, only c.40% of the deals looked worth the money, and roughly 10% are supremely successful. Therefore, around 60% don’t work out to any meaningful degree, although with squads in excess of 25 men, 60% is pretty close to the amount of players who don’t make the starting line-up each week. Of course, some of those squad players were bought merely as back-ups, while in other cases the bench is warmed by expensive signings who have failed to deliver or whose talent has been curtailed by injury.
The Season So Far
Over the past few years we have demonstrated a clear correlation between how much a club spends on transfers, when adjusted for inflation, and the position it will finish in the table. This correlation has increased since the early days of the Premier League, before the money really started to roll in to the elite clubs.
In terms of squad cost, and the average cost of the XI over 38 games (the ‘£XI’), there is a clearly defined descending scale from 1st to 10th, based on the seasons since the turn of the millennium; i.e. the most expensive squads and XIs average out finishing 1st, the 2nd-most expensive finish 2nd, and so on. In any given season any given club can over- or under-achieve based on its budget, but over a 13-year period it has evened out to a startling degree.
Only once you hit 11th spot does the pattern break, with teams that finish in the lower half of the table providing a jumbled picture.
This, and more, is covered in this recent piece: