Thursday, March 24, 2016

Wealth Inequality and Baseball Card Collecting: Some Statistical Observations

One of the biggest changes in the baseball card hobby over the past 25 years that doesn`t seem to get talked about too much is the obvious stratification of collectors according to economic class. To illustrate what I am talking about, at the risk of oversimplification, I will divide these into two groups:

1) A small group of extremely wealthy collectors who vie with each other for a select group of the very top end stuff.
2) Everybody else vying with each other for whatever the first group hasn`t driven the price on out of reach.

These two groups seem to approach the hobby, and how they spend their money on the hobby, in different ways.  Collectors in Group 1 only really go after the top stuff - vintage holy grails in top condition and things like that- while the latter, not having the money to go after that stuff, focuses more on whatever is within their means depending on their personal interests.  Ironically, given the volume and variety of lower value stuff out there the average collectors might have much larger and more varied collections than the wealthy, who might just have a few dozen cards (worth thousands or hundreds of thousands of dollars each of course).

This isn`t necessarily a new thing, back in the 80s there were also wealthy collectors driving the price of stuff up, but this division was nowhere near as distinct then as it is now.  A 1952 Topps Mickey Mantle was a very expensive card in the 80s, but the relative price disparity between it and a common was much less than it is today and the premium for cards in top condition also was nowhere near as high as it is now. I remember Beckett Baseball Card Monthly back in the 80s used to assign a premium of up to 50% over Near Mint prices for cards in perfect mint condition, which is completely out of line with what that premium is today.

An obvious explanation for this change in the hobby is the changing wealth structure of American society as a whole.  When the hobby was emerging in the 1970s and 1980s the American middle class was still in relatively good financial health, meaning that the hobby had a large population of people with some disposable income to spend on the hobby.  Today that situation is much different, the middle class has seen its incomes stagnate while its main asset, homes, have fallen in value.  The bottom 2/3 of Americans according to US census data are financially worse off today than they were in the 1990s - they actually own less than they used to.  Americans above that, in the top third of society as measured by net wealth, are the only ones who have seen their position improve over that same time - they collectively have more stuff now than they did in the 1990s. 

When put into a line chart, the distribution of wealth in the US today has a kind of hockey stick shape to it.  It increases at a linear rate when you are going through the lower and middle income groups, then increases rapidly when you hit the top group.  If you divide the US population into fifths by their net wealth, this is what it looks like (data coming from the US Census Bureau):



The difference between the first four groups of 20% more or less increases a little bit at a time, then it skyrockets when you hit the top 20%.  Even that doesn`t really give an accurate picture of how sharp the final curve is, since the distribution within that top 20% is even starker.  If you are at the bottom end of that 20% (ie you own more than 80% of Americans, but less than 20%), you are still a more or less normal middle class person who, though you are doing well, still has to work for a living.  At the top end of that 20% you have billionaires.  So within that top 20% there is a huge difference in wealth.

If we focus just on the top 10%, you see there is another hockey stick curve within that group, with the average person in top 0.01% owning about 150 times more than the average person in the top 10% as a whole (2012 data, from economists Emmanuel Saez and Gabriel Zucman):



If you look at historical data from the 1970s-1980s, while there was still wealth inequality then the sharp upticks you see on the right side of these charts was much less pronounced.

Now consider what this means for the hobby.  Basically the Group 1 collector I described above, being a member of the right side groups in these charts, has a lot more money now than they did 30 years ago.  The Group 2 collector, in contrast, has a lot less money than they used to.  So there is now a much larger amount of money available to be spent on the extremely small number of things that Group 1 collectors are interested in (hobby holy grails and other top end vintage stuff) and a much smaller amount of money available to be spent on things Group 2 collectors are interested in (basically everything else).

 This graph is roughly how the hobby used to value cards based on their condition (I am doing these numbers based on memory, I don`t have an old copy of Beckett with me as I write this so I might be a bit off but basically accurate) when Group 2 collectors had more disposable income and Group 1 collectors less (compared to what they do now).  Basically it was linear, each increase in grade was associated with an incremental increase in value.  There was no hockey stick shaped curve in the line at the right of the graph.  This of course only applied to decent cards.  A 1986 Topps common in anything less than nrmt condition was basically worthless, but if you had say a 1963 Topps Willie Mays or something nice like that, its value would more or less follow this curve.



Contrast that with how the hobby distinguishes card values based on condition today.  To get a rough idea, I went to PSA`s Sports Market Report Online and consulted their values for the iconic T206 set, which they set out by condition.

If you put the values for most of those cards into graph form, they look nothing like the above one.  This is Walter Johnson`s Portrait card goes for:





And fellow Hall of Famer Christy Matthewson`s Dark Cap card:



Those are basically just two hall of famer cards with relatively decent populations (more than 500 PSA graded copies) I chose as examples, but if you look at almost any other card (but not all as I`ll get to below) they more or less follow the same curve.  The value of the card increases increases incrementally at a linear pace until you hit near mint, then all hell breaks loose for the (few) cards that exist above that.  

This of course is where the two groups of collectors draw the line between themselves.  Group 1 collectors chase the few copies of those cards in the Nrmt-mt/ Mint range into the stratosphere while completely ignoring everything below that.  So the cards below Near Mint basically still follow the 1980s logic of condition pricing and Group 2 collectors might actually aspire to own a copy of Johnson or Matthewson in lesser grades.  

What I am saying might seem obvious to most collectors familiar with what these cards go for, but consider how closely these graphs correlate with the graphs on wealth distribution in society above.  If middle class group 2 collectors had more disposable income today they would be spending that on  cards like the Johnson and Matthewson ones (at least, those who like Johnson and Matthewson would).  This would bump up the prices in the middle of the graph and make it a bit more linear like it used to.  If Group 1 collectors had a bit less money this too would lower the right end of the graphs a bit.

There is one really interesting exception to the above trend, which is the value of holy grail type cards.  The T206 set has a few cards - Honus Wagner, the Magie error card and Eddie Plank - which are extremely rare and much more sought after than the others.  The prices for two of these (Wagner and Magie) don`t follow that hockey stick pattern but rather follow the 80s pattern of incremental price increases based on condition.  Here is the Wagner values:


And the Magie:

 The Plank follows a slightly more hockey stick shaped curve, but it too isn`t as pronounced as with the more common Johnson and Matthewson cars above.
 

The obvious explanation for this discrepancy (which as far as I can tell only applies to rare holy grails) is that the cards are so rare that all of them, regardless of condition, are collected by Group 1 collectors who, owing to their scarcity, are willing to spend big time on lower grade copies as well.

The really interesting thing this tells us is actually just how small the number of Group 1 collectors is.  The three T206 holy grails all have less than 100 copies graded by PSA.  The Johnson portrait card, in contrast, has 721 copies and the Matthewson 1,061.  The number of PSA graded copies of a card is not of course an accurate reflection in and of itself of how many copies of that card exist, but it might at least suggest the relative scarcity of each, with Johnson and Matty about 10 to 20 times easier to find than the Holy Grails.
 
This basically means that while there are enough Group 1 collectors to drive the price of all copies of the holy grails up, there are not enough of them to do the same for the only slightly more plentiful cards of Johnson and Matthewson, so there are probably less than a thousand such persons out there. 

Anyway, these are just some thoughts I had been wanting to put in writing.  Personally I prefer the old 80s structure of the card market, for all its faults it wasn`t as stratified as it is today and it at least felt like a level playing field.  The future of the card market, I should note, is also likely to be shaped by these trends.  A more prosperous middle class would probably do wonders for the hobby.