1 
 
High score or game over? The financial returns of retro video games on the secondhand market 
By 
Thomas Farrell 
 
 
 
Abstract 
Recent high-profile sales of retro video games, including the $1.5 million sale of Super Mario 
64, have spurred renewed interest in purchasing retro video games as investments, not merely as 
collectibles. While existing literature suggests that collecting anything solely as a financial 
investment is unlikely to result in returns greater than those seen in traditional financial assets, 
the prominence of video games as one of the dominant forms of entertainment and the utility 
associated with owning a retro game may make them a more attractive asset. There is a lack of 
existing research into the secondhand market for retro games or the financial returns of similar 
pop culture items. This thesis explores that gap and determines if there is a significant financial 
interest in secondhand retro video games as an investment. eBay sales data from January 2014 to 
January 2022 were collected for games across ten different consoles and several standardized 
financial indexes were assembled according to the repeat-sales methodology. For an index 
consisting of all games for which sales data was obtained, there is an annualized real rate of 
return of 10.2% and a Sharpe Ratio of 0.73 compared to a 9.8% real return and 0.25 Sharpe Ratio 
for the S&P 500. Consequently, secondhand retro video games are a compelling alternative 
investment. Additionally, their low correlation and high risk premium suggest they may be a 
good investment from a diversification perspective, as risk is lower but returns are similar to 
those of the stock market. These results break with the established literature on collectibles, 
which suggests that collectibles are not an efficient or worthwhile investment.  
Keywords: retrogaming, collectibles, financial returns, alternative investment, videogaming, 
repeat-sales methodology 
 
 
 
 
 
Submitted under the faculty supervision of Professor Stuart Webb, in partial fulfillment of the 
requirements for the Bachelor of Science in Business, summa cum laude, Carlson School of 
Management, University of Minnesota, Spring 2022. 
2 
 
Acknowledgements 
Bryce Villanueva, for developing a program in Python to scrape the dataset I used in this thesis. 
JJ Hendricks of pricecharting.com, for his consent and guidance to scrape his website. 
My readers, Professors Veronica Marotta and Corrinne Fiedler, for their detailed feedback, 
advice, and time. 
My thesis advisor, Professor Stuart Webb, for his support, feedback, and time. 
Professor Ken Reily, Bradshaw Irish, and Jacob Braun, for their data analysis and RStudio 
expertise. 
Ali NasserEddine, for his assistance and guidance regarding the kind of analysis to perform.  
3 
 
1.0 Introduction 
Video games are powerful cultural artifacts that transport their players back to a nostalgic 
time when they were younger. This nostalgia is one of the key factors contributing to the recent 
success of retro games – video games released from the 1970s to the early 2000s (Wulf et al., 
2018). Along with the rise in interest in the retro gaming subculture, there is an interest in 
purchasing games as an alternative investment strategy. The same internet-savvy investors who 
set their sights on GameStop, AMC, and NFTs view retro video games as another potential 
market to exploit (Davison, 2021). A pristine copy of the 1996 game, Super Mario 64, recently 
sold for over $1.5 million, spurring interest in buying and selling retro games for a profit 
(McGreevy, 2021). While research exists for the financial returns of other collectibles, no such 
study exists for retro video games.  
Retro video games serve as unique pop culture artifacts with a utility that cannot be 
perfectly replicated. For example, comic books can be accessed on the web with almost no 
difference to the reading experience, which is not the case for video games. While LEGO sets 
share some of the same traits as games, recent analysis of LEGO does not use the more robust 
and comprehensive repeat-sales methodology to calculate returns. This paper bridges that gap by 
using the repeat-sales methodology to assemble indexes and calculate returns for secondhand 
retro video games. 
While other similar hobbies with collectibles markets are no longer as popular as they 
once were, video games are a new kind of asset that is growing in popularity. For example, the 
comic book industry crashed in 1993 and never recovered (Last, 2011). In comparison, Juniper 
Research values the current global video game market at $155 billion and predicts that it will 
eclipse $268 billion by 2025 (2021). Video games are popular and will remain popular for the 
4 
 
foreseeable future. Perhaps the popularity of video games explains why speculators and investors 
are not deterred by prior research suggesting that “the majority of collectibles yield lower 
financial returns than stocks… and embody more risk than most other financial assets” (Burton 
& Jacobsen, 1999).  
Many investors of games physically own what they invest in, but other options exist. The 
apps Otis and Rally enable customers to fractionally invest in “shares” of many different video 
games without the need to physically own them (Davison, 2021). Those customers will see a 
profit once the game is sold at auction or if they choose to sell their share to another user. As the 
market for physical trading is more mature than that of the digital share market, this physical 
trading is the focus of this paper. To facilitate the discovery of interesting insights, the main 
research question guiding this paper is: How do the financial returns of retro video games on the 
secondhand market compare to those of equities, bonds, gold, and general market inflation in the 
United States? 
To explore this question, this paper utilizes an abductive research philosophy in the same 
vein as prior research performed regarding collectibles (Burton & Jacobsen, 1999; Dimson & 
Spaenjers, 2011). Rather than asserting specific hypotheses, this paper follows the guidance of 
Behfar and Okhuysen and substitutes abductive reasoning in place of a more traditional 
hypothetico-deductive inquiry (2018). Given the format of prior literature and the relative lack of 
research on retro video games or similar classes of collectibles, the hypothetico-deductive style 
of inquiry may limit the scope of potential findings. 
I use the repeat-sales index assembly methodology described by Burton and Jacobsen to 
apply a common scale by which the financial performance of secondhand retro video games can 
be compared to that of traditional financial assets (1999). The calculation of additional financial 
5 
 
metrics, including the Sharpe Ratio and market risk estimates, will provide insight into the over- 
or under-performance of these assets and how suitable they are as an investment. Consequently, 
this thesis is of relevance to investors looking for alternative assets in which to invest. 
The subsequent section looks at available literature on the financial returns of 
collectibles, the motivations behind collecting, and factors influencing the value of collectibles. 
The third section details my methodology, pre-analysis limitations, and data analysis. The fourth 
section contains the results of my research. Section five discusses the implications of the results 
while section six contains a post-hoc analysis of limitations and potential avenues for future 
research. Finally, the last section lists closing thoughts. 
  
6 
 
2.0 Literature Review 
There is a large body of research into the financial returns of collectibles. Some of the 
most examined collectibles include art, wine, and stamps, but such research exists even for 
Beanie Babies and LEGO sets. A broad examination of the returns for collectibles determined 
that they embody higher levels of risk than other financial assets and yield lower returns than 
stocks and bonds (Burton & Jacobsen, 1999). The accepted consensus is that one should only 
collect items if they enjoy doing so, and not as an alternative investment strategy. However, there 
is little research regarding returns on collectibles with utility except for wine. To get the utility 
associated with wine – drinking it – one must physically own the bottle. Similarly, one must own 
the original game and hardware to experience a retro video game as intended. While there exist 
ways to play retro games on modern gaming consoles, phones, or computers, they use different 
control input methods or resort to emulation that introduces technical glitches. Items like comic 
books or art lack utility, as they can be appreciated through other means with relatively little loss 
in the experience.  
Additionally, there is no financial research on retro video games, very little on pop 
culture associated items in general, and none using a repeat-sales methodology. This paper 
attempts to bridge the gap in existing research by providing a detailed analysis of the financial 
returns of such a pop culture item – retro video games. The following section briefly describes 
the motivations for collecting before a detailed discussion of the hedonic factors that influence 
the value of a collectible and the investment feasibility of different collectibles.  
  
7 
 
2.1 Factors motivating interest in collecting and investing in collectibles  
Research shows that while there may be financial factors driving the decision to collect, 
the primary driver is the collector’s emotional attachment to his or her purchases. One study 
examining the motivation behind women who collect dresses found causes as varied as wanting 
to mimic a celebrity to wanting a collection to pass on to friends and family (Bishop, 2018). A 
study directed at nonmonetary collectors of art lists several motivations driving their behavior, 
including the desire to surround oneself with beautiful objects and seeking art as a source of 
comfort (Baekeland, 1981). Research from Carey found that additional factors that “enhance the 
ordinary value of a good” can make items more desirable (2008). In the case of retro games, one 
such factor that increases their value and provides motivation for collectors is their nostalgic 
nature. Typically, people will not collect something as a hobby without some kind of personal 
attachment.  
However, there is a group of collectors called “investor collectors” who connect their 
collecting interests to financial motives (Kleine et al., 2020). This group has a different 
psychological and demographic makeup to regular investors and the general population. Kleine 
et al. consider this niche to be about 70% male, with above-average education and income 
(2020). The current set of “investor collectors” is a unique group with their own traits and their 
own motivations. They are interested in collectibles as both a new experience and something that 
can offer a potential financial return. This research suggests that while all collectors are 
motivated by emotional means, some would not collect items if they did not believe they would 
appreciate. 
One study of a niche internet forum for video games found that the desire to be 
recognized as a pure “true gamer” is a significant motivator behind someone choosing to buy a 
8 
 
physical copy of a retro game (Downing, 2011). However, this group views collectors who own 
games out of a desire for ownership instead of a desire to play the game the way it was intended 
as inauthentic (2011). The primary dialectic in the retro video game market is a struggle between 
those who view retro video games as an art form worth playing and those who view retro video 
games simply as something to sell for a financial return. Popular gaming press outlets lament the 
latter group as culpable for a recent perceived increase in the prices of retro games (Devore, 
2021).  
2.2 Hedonic factors influencing the value of collectibles 
There exist several standard characteristics influencing the value of most collectibles. A 
hedonic analysis of American collectible comics determined that rarity and condition are two of 
the primary factors influencing a given comic’s price (Wyburn & Roach, 2012). All else held 
equal, the item in better condition will be worth more. Likewise, the rarer item will usually be 
worth more. Koford and Tschoegl found that rarer coins commanded higher prices but that 
“rarity in the form of original mintage” was the most valuable (Koford & Tschoegl, 1998). 
However, an analysis of vinyl records determined that rarity does not necessarily have a 
premium price (Cameron & Sonnabend, 2020). Just because an item is rare does not mean that 
collectors are interested. 
Cameron and Sonnabend assert that “cultural importance” is the most critical factor 
driving the price of a vinyl record (2020). A Prince record is likely to be worth more than the 
record from a small, local band. Even if the local band’s record is scarcer than a Prince record, it 
lacks the same cultural capital. Other goods have similar types of soft factors that influence 
value. Wyburn and Roach found that the title, content, characters, artists, and media presence of 
a comic all impacted its value (2012). This research suggests that while an item’s purely factual 
9 
 
traits may influence its price, the most crucial factor is whether the collector has an emotional 
attachment to it. These soft factors determine the strength of an individual’s emotional 
attachment.  
Retro video games lack any hedonic analysis determining what factors influence their 
value the most. There is some research on what traits archivists can use to curate a collection of 
retro video games. Leblanc mentions completeness, condition, media type, and instruction 
manual format as some examples (2021). However, this research lacks any analysis on which 
factors have the most influence on the value of a given video game. 
2.3 Investment feasibility for different types of collectibles  
Due to the popularity of art, stamps, and other collectibles as an alternative investment, 
there is a substantial body of research answering the question of whether they are a worthwhile 
investment. A general census regarding the returns of various collectibles, determined that they 
are riskier than traditional financial assets and usually carry inferior returns (Burton & Jacobsen, 
1999). Burton and Jacobsen state that collectibles are more illiquid than other assets and have 
significantly higher storage and transaction costs, which drives the actual returns even lower 
(1999). 
Table 1 contains the real returns for various kinds of collectibles explored in prior 
literature. The research implies that traditional collectibles like fine art and stamps underperform 
the market. A study from Grable and Chen found that U.S. stamps only performed better than 
bonds and that only classic stamps were worthwhile from an investment strategy standpoint 
(2015). A similar study from Dimson and Spaenjers examining British postage stamps came to 
the same conclusion (2011). 
10 
 
Returns for Other Collectibles 
Paper Item 
Annualized  
Real Return (%) 
Grable & Chen (2015) U.S. Stamps 2.90% 
  Classics (1969-2013) 5.50% 
  Modern (1969-2013) 3.29% 
  Airmail (1969-2013) 1.60% 
Dimson & Spaenjers (2011) British Postage 
Stamps 
2.70% 
Baumol (1985) Fine Art 0.55% 
Burton & Jacobsen (2001) Wine 7.90% 
  1982 Vintage 16.20% 
Dobrynskaya & Kishilova (2022) LEGO sets 8.00% 
Table 1: Returns for Other Relevant Collectibles 
Even fine art is not necessarily a good investment, with a real rate of return even lower 
than stamps (Baumol, 1985). Baumol suggests that it is impossible to “beat the game” of art 
investment (1985). Despite this, Mei and Moses state that many investors pursue the accepted 
strategy of buying the most expensive piece of art available (2002). However, their study found 
significant underperformance of these “masterpieces” (2002). Additionally, one 2009 study 
found that the returns on fine art correlate with stock returns (Hiraki et al., 2009). This suggests 
that investing in such collectibles is a waste of time as even employing a specific strategy will 
not reliably beat the market.  
However, stamps and art are unlike retro video games. Most significantly, neither asset 
holds much practical utility beyond its aesthetics. One collectible that does have a utility is wine. 
To drink a glass of wine, one must own that wine. Therefore, wine is consumable, as once it is 
drunk, it loses all value. Consequently, “investment-grade” wine appears to have the most 
significant investment potential of any collectible (Burton & Jacobsen, 2001). Burton and 
Jacobsen assert that without considering storage and shipping costs, wine is more attractive than 
11 
 
T-Bills and could be an alternative to fixed-income securities (2001). Retro video games have a 
utility like wine but have a more immature market and are not consumable, so their returns will 
likely differ and may be higher. 
More recent research, focusing on pop culture collectibles associated with Millennials, 
suggests that investing in such collectibles may be a good idea. Dobrynskaya and Kishilova 
found that investments in LEGO from 1987 to 2015 outperformed all market benchmarks with 
an annualized real return of 8% and an annualized Sharpe Ratio of 0.4 (2022). Research from 
Shanaev et al. reflects this, stating that LEGO carries diversification benefits to a portfolio, 
making it an attractive investment (2020). Video games are much like LEGO, as both are 
nostalgia-driven, pop culture items with a utility beyond just the aesthetic. As much of the 
research on such pop culture items is very new, it is unclear how applicable it is to collectibles 
overall or why these assets carry greater investment potential. 
Research from Choi and Lee may provide one explanation. They assert that the consumer 
perception of the sneaker resale market is that sneakers are “now recognized as a market with 
new economic benefits,” (2021). This implies that retail investors now view all collectibles as a 
potential investment. Additionally, auction websites, online forums, and price tracking websites 
makes it easy for collectors and investors to buy items and see what those items are worth. A 
positive feedback loop could occur whereby the price of an item goes up, so investors buy more 
of such item, and the price continues to rise.  
Such a positive feedback loop could lead to potentially catastrophic speculative bubbles. 
For example, a positive feedback loop was responsible for the bubble in the late 1970s for Dutch 
collectible postage stamps (Franses and Knecht, 2016). Other research found that sudden 
rightward shifts in demand are just as likely to bring a sudden leftward shift as speculative 
12 
 
expectations change (Stoller, 1984). While most studies agree that collectibles are not a good 
long-term investment, except under particular circumstances, the returns could significantly 
exceed those of other assets during these speculative bubble periods. However, following the 
bursting of the bubble, asset prices will drop significantly and are unlikely to reach the heights 
seen during the bubble period again. 
In summation, existing research asserts that traditional collectibles are generally not a 
good investment and that one should only choose to collect these if they enjoy it. The literature 
on the motivations of collectors suggests that while there is usually a financial component, the 
primary motivator is emotional. That motivator could be nostalgia or self-satisfaction, but it is 
usually just as powerful or even more potent than financial motivation. While factors like rarity 
or condition can drive the price of an item, the emotional interest component seems to be the 
most significant factor. However, more recent research indicates that pop culture collectibles, 
such as sneakers and LEGO, not only may have higher investment returns, but also carry 
diversification benefits. 
While there is some research regarding the investment potential of pop cultural 
collectibles, it is from the last few years and does not cover the same scope of objects as the 
research regarding traditional collectibles. Additionally, research for these classes of items 
analyze fewer items and do not use the repeat sales methodology for pricing index assembly. 
Retro video games represent a unique set of criteria that has yet to be analyzed. Consequently, 
the following paper bridges the gap of research that excludes analysis of secondhand retro video 
games, usage of the repeat sales methodology for pop culture collectibles, and items with a 
practical, non-replicable, non-aesthetic utility. Table 2 shows how research on retro video games 
are differentiated from other collectibles. 
13 
 
  
14 
 
Table 2: Comparing Collectibles with Prior Research to Retro Video Games 
  
  
15 
 
3.0 Methods 
3.1 Datasets 
 To obtain a dataset of the prices of retro video games on the secondhand market, I 
scraped data from the website pricecharting.com after receiving their consent. This website 
provides monthly pricing data from eBay sales dating back to 2008. At the end of each month, 
pricecharting.com calculates the price of a given game based on the average of what it sold for 
on eBay. For example, if Super Mario Bros sold two copies in January 2022, then the price 
recorded on pricecharting.com would be the average of the sale value of the two games. The 
dataset obtained includes a release ID for a given game, the name of the console of release, the 
name, the level of completeness, the genre, the month of sale, and the average price it sold for 
that month. Game genres were inconsistently applied depending on the game and the consoles. 
Consequently, I combined overly specific genres into a smaller set of more general categories. 
Table 3 shows these changes. 
  
16 
 
 
Bundled Genres Resulting Genre 
Categories 
Number of  
Games 
Proportion of  
Total 
Baseball, Basketball, 
Extreme Sports, Football, 
Golf, Soccer, Wrestling, 
Sports 
Sports 1114 20% 
Board & Card, Casino, 
Compilation, Dance, 
Educational, Mini-Games, 
Horror, Light Gun, Music, 
Party, Trivia, Visual Novel, 
Other 
Other 465 8% 
First Person Shooter, Third 
Person Shooter, Action & 
Adventure 
Action & 
Adventure 
1893 34% 
Shoot’em Up, Beat’em Up, 
Pinball, Arcade 
Arcade 397 7% 
Fighting Fighting 280 5% 
Platformer Platformer 336 6% 
Puzzle Puzzle 220 4% 
Racing Racing 496 9% 
RPG RPG 238 4% 
Strategy Strategy 149 3% 
Table 3: Genre Bundling 
Starting in April 2013, this sales data began to distinguish between whether a sale of a 
game was loose, complete-in-box, or brand new. Loose describes a sale that consists only of the 
loose cartridge or disc. Complete-in-box describes a sale including the game, its original box, 
and any manuals or documentation that was originally included. New describes a sale that is 
complete-in-box and unopened. To bypass data coverage and accuracy issues regarding the new 
classification of game sales into different level of completeness, my research covers the period of 
January 2014 to January 2022. Since this market is in its relative infancy compared to other 
collectibles, this time period is representative of the kind of returns one can expect when 
choosing to invest in retro video games.  
17 
 
I only obtained sales data on the games released for popular consoles with high N values 
to get more precise results. I selected a range of systems using both cartridge and disc formats as 
well as home console and handheld formats. Additionally, I focused only on systems released by 
Nintendo, Sega, and Sony. Consequently, all games in the dataset are guaranteed to be licensed 
and commercially available. I selected systems released prior to the year 2000 to fit the definition 
of a retro video game and ensure the secondhand market for each console is relatively mature. 
See Table 4 below for a complete list of systems.  
Console Name American 
Release 
Year 
Media 
Format 
Console Type Number of 
Games 
Proportion of 
Total 
Nintendo Entertainment 
System  
1985 Cartridge Home Console 775 14% 
Nintendo Game Boy 1989 Cartridge Handheld 518 9% 
Super Nintendo 
Entertainment System 
1991 Cartridge Home Console 732 13% 
Nintendo 64 1996 Cartridge Home Console 300 5% 
Nintendo Game Boy 
Color 
1998 Cartridge Handheld 452 8% 
Sony PlayStation 1 1995 Disc Home Console 1,330 24% 
Sega Genesis 1989 Cartridge Home Console 727 13% 
Sega Game Gear 1991 Cartridge Handheld 235 4% 
Sega Saturn 1995 Disc Home Console 253 5% 
Sega Dreamcast 1999 Disc Home Console 266 5% 
Table 4: Video Game Consoles Included in Dataset 
  
18 
 
 Tables 5 and 6 contain a set of summary statistics describing the dataset at the console 
level. Table 5 includes summary statistics after excluding null values, but prior to the exclusion 
of outliers. Table 6 contains the same statistics but after excluding outliers. To exclude extremely 
high-priced games that are not a feasible purchase for a retail investor, outliers were defined as 
those entries with sales prices greater than 1.5 times the inter-quartile range. Since the most 
expensive games are not indicative of what a typical collector might own, they are not relevant to 
the analysis.  This excluded 252,283 transactions from final index assembly. “Transaction” refers 
to a row containing a monthly price.  
Table 5: Summary Statistics by Console Prior to Filtering for Outliers 
Table 6: Summary Statistics by Console After Filtering for Outliers 
19 
 
For equities, I gathered a dataset from DataPlanet tracking the price performance of the 
S&P 500. The S&P 500 is a standard proxy to measure the performance an individual investor 
can expect to see. I obtained a dataset tracking the price of gold from the S&P Capital IQ 
database and an index of U.S. Treasury Bills from S&P Dow Jones Indices.  To determine the 
level of inflation, I sourced Consumer Price Index (CPI) data from the Bureau of Labor 
Statistics. I converted all data to a monthly basis if required and set the base month for all data to 
January 2014 by dividing all subsequent entries by the value of the given index in April 2014. 
 All datasets examine the same period of April 2014 to January 2022. The retro video 
game price dataset solely includes North American released games as data for other markets is 
more limited and significantly increases the complexity of the analysis required. Additionally, 
my analysis excludes variants, non-commercially released games, and unlicensed video games. 
These types of releases are unrepresentative of the collector’s market and are excluded from the 
analysis. In summation, the dataset’s controls ensure that each game only has one ID and is 
limited to the North American geographical market. Table 7 provides a detailed list of variables 
available in the dataset. 
  
20 
 
Variable Description 
Date Month and Year of sale 
Game ID The database ID associated with a given 
release 
Name Name of video game 
Price The average price a game sold for in a given 
month 
Completeness Loose disc/cartridge, complete-in-box, or new 
Genre The genre of a game. See Table 1 for a 
detailed list 
Console Game system the video game released on 
Console Year The release year of a given console 
Console Type Home or handheld 
Console Format Disc or cartridge 
Table 3: Variables Available in the Dataset 
  
21 
 
3.2 Retro Video Game Price Index Justification 
In a detailed analysis of previous studies, Burton and Jacobsen list three methods of 
calculating returns for assets: market basket analysis, hedonic regression, and repeat sale 
regression (1999). According to Burton and Jacobsen, none of these approaches “will necessarily 
understate or overstate returns as compared to the other approaches” (1999). Consequently, we 
are free to choose the method that best fits the dataset. Market basket is the least sophisticated 
method of analysis and involves selecting a basket of items and comparing their prices over time. 
The basket can be chosen at random or hand-picked and can be varying or fixed. A varying 
basket changes for each year of analysis, while a fixed basket includes the same items over the 
entire period of analysis. Unlike wine and fine art, which both have well-credentialed experts 
who could hand-select an adequate portfolio, retro video games lack these experts or research 
that could inform a typical portfolio strategy. While I could have selected a basket of retro games 
randomly, it may exclude commonly purchased games that add significant value to a collection. 
A hedonic regression involves regressing the price of an item based on certain 
characteristics. Consequently, one can track the price performance over time while allowing for 
different levels of quality for different items. However, such a regression is linear, resulting in a 
constant rate of return that doesn’t allow for detailed analysis or comparison with different 
assets.  
The final method, running a repeat-sales regression, is the most sophisticated method of 
analysis. This method involves tracking the performance of repeat sales of a given item. Repeat 
sales refer only to subsequent sales for a given retro video game, not the repeat sale of one 
individual copy of that game. For each pair of sales, the log-price is calculated and regressed 
based on a set of dummy variables which are equal to +1 at the time of the sale and -1 at the time 
22 
 
of the prior sale. This method enables the creation of an index for retro video games that 
accurately captures the price movement of the entire market without limiting the analysis to a 
constant rate of return. The repeat-sales method is most apt when studying heterogeneous assets 
and comparing their performance to other asset classes. Consequently, this paper uses the 
standard repeat-sales methodology put forward by Bailey et al. to create an index for housing 
prices (1963).  
3.3 Repeat-Sale Price Index Assembly 
 For this paper, I calculated a variety of repeat-sale indexes based on the characteristics 
available in the dataset. This included one index containing all game sales, three completeness 
indexes, three console manufacturing company indexes, ten genre indexes, and ten console 
indexes. The base methodology followed to assemble each index was the same. While they all 
provide interesting insights, I will primarily focus on the overall index and the three 
completeness indexes for the subsequent analysis. Doing this enables the development of more 
useful and specific insights.  
 To assemble the various indexes, I imported the video game price index into the data 
analysis tool RStudio. I cleaned the data by filtering out all rows containing null entries and any 
outliers. Following the exclusion of outliers, I filtered for additional traits depending on the kind 
of index to be assembled.  
To assist with the analysis, I used the R Package “rsmatrix” based on the “rsi” package 
detailed by Kirby-McGregor and Martin (2019).  This statistical package was designed for non-
economists to calculate repeat-sales matrix and convert data into a sales pair format. First, I 
converted the filtered dataset into the sales pair format using the function rs_pairs. This function 
creates columns for the previous date of sale, what it sold for on that previous date, the current 
23 
 
date of sale, and what it currently sold for. Secondly, I used the function rs_matrix to assemble 
the dummy variable matrix using January 2014 as the base year. Each sale was given a dummy 
variable of +1 for the date of sale and -1 for the date of the prior sale. All other dates were 
assigned a value of 0. 
After assembling of the dummy variable matrix, I calculated the geometric repeat sales 
(GRS) index based on the methodology in Bailey et al. (1963). This consisted of solving a 
system of equations where the cross products of the dummy variable matrix are on the left-hand 
side and the cross products of the dummy variable matrix and the log prices are on the right-hand 
side. Then the exponentials of the calculated values were multiplied by 100 to generate the final 
pricing index. The resulting price index starts at a value of 100 in the base month of January 
2014. 
  
24 
 
3.4 Additional Data Analysis Methods 
 To provide useful insights for investors, I calculated several other financial metrics for 
each index to compare to those of more traditional assets. These included the annualized rate of 
return, real rate of return, monthly return rates for each asset, and the standard deviation of 
returns. The methodology for these are described in Appendix 1.  Additionally, I calculated the 
Sharpe Ratio, alpha and beta estimates, and the correlations of each index against the others. 
Sharpe Ratio 
 The Sharpe Ratio is a measure that represents how well an asset performs compared to a 
risk-free rate given the asset’s level of risk. This measure was calculated on a monthly basis 
unless otherwise specified. The risk-free rate provided by the data library of Kenneth R. French 
is derived from the one-month Treasury bill rate. The following formula was used to calculate 
the Sharpe Ratio for each index: 
𝑅𝑎−𝑟𝑓
𝜎𝑎
, where: 
 Ra-rf = Average of monthly excess returns, calculated by subtracting the risk-free rate 
from the monthly rate of return and 
 σa = Standard deviation of the asset. 
  
25 
 
Alpha and Beta Estimates 
 Alpha and beta estimates measure the relative over- or under-performance of a given 
index based on its level of risk. Estimates were calculated for the all games, loose, complete-in-
box, and new indexes based on the CAPM and Fama and French Three-factor and Five-factor 
Models. These models attempt to find the factors that influence the returns of a given portfolio. 
Therefore, using these models enables us to determine if there are any existing factors that 
explain the over- or under-performance of retro video games. All values for the CAPM and Fama 
and French Models were obtained from Kenneth R. French’s data library. For each index a 
regression was performed on each of the following formulas to calculate the Alpha and Beta 
estimates:  
CAPM: 𝑅𝑎 − 𝑅𝑟𝑓 = 𝛼 + 𝛽1 ∗ (𝑀𝑘𝑡 − 𝑅𝐹), 
Three-factor: 𝑅𝑎 − 𝑅𝑟𝑓 = 𝛼 + 𝛽1 ∗ (𝑀𝑘𝑡 − 𝑅𝐹) + 𝛽2 ∗ (𝑆𝑀𝐵) + 𝛽3 ∗ (𝐻𝑀𝐿), and 
Five-factor: 𝑅𝑎 − 𝑅𝑟𝑓 = 𝛼 + 𝛽1 ∗ (𝑀𝑘𝑡 − 𝑅𝐹) + 𝛽2 ∗ (𝑆𝑀𝐵) + 𝛽3 ∗ (𝐻𝑀𝐿) + 𝛽4 ∗ (𝑅𝑀𝑊) +
𝛽5 ∗ (𝐶𝑀𝐴), where: 
 Ra = Rate of return for a given month, 
 Rrf = Risk-free rate of return for a given month, 
 α = Alpha value, and 
 βx = Beta value for a given factor.  
Correlation 
 A correlation matrix was calculated based on Pearson coefficients. All assets and indexes 
were included in the assembly of the matrix.  
  
26 
 
3.5 Assumptions, Limitations, and Strengths 
 The primary assumption required for this analysis is that the dataset obtained is 
representative of the secondhand market for games. Since the data is from a commercial platform 
that sells access to stores, we can assume it is representative of the value of these games. 
Additionally, pricecharting.com has been gathering such sales data since 2008, a much larger 
time window than any other competing service. Consequently, the dataset contains the price of 
every eBay sale for every official, North American-released, commercially available video game 
sold for all game consoles analyzed over that time period. As eBay is the dominant internet 
auction site in the world and the small physical dimensions of most games makes the online 
listing and shipping of them simple, it is safe to assume that sales on eBay capture a 
representative sample of the market. 
However, eBay sales do not capture the entirety of the retro video game market and 
exclude the most high-value sales seen at auctions. For example, the copy of Super Mario 64 that 
sold for over $1 million sold at auction, not on eBay (McGreevy, 2021). While the following 
analysis is consciously aimed at the average individual collectible investor and not wealthy 
collectors or institutional investors, it does mean the analysis excludes the retro games likely to 
have the greatest return. Also excluded are those sales at specialist stores or video game 
conventions. As these transactions are conducted in-person, there is no way to gather a 
representative sample of historical sales. However, given that many vendors base their pricing on 
prices listed by platforms like pricecharting.com, the exclusion of these sales is inconsequential. 
 Another limitation is the relative immaturity of the retro video game secondhand market. 
Current conclusions based on this data could be inaccurate in only a few years. The Dutch stamp 
bubble of the 1970s was indicative of such a relationship. This bubble spanned several years and 
27 
 
gave a false impression as to the value and financial return potential of the asset (Franses and 
Knecht, 2016).  Therefore, it is possible that retro video games are in a bubble period or that they 
have yet to reach their market potential. In the event the retro video game market is in a bubble 
period, the long-term investment potential is likely very low. 
  
  
28 
 
4.0 Results 
 This section outlines the results from the data analysis, including all relevant graphs and 
tables. Some graphs created are excluded from the results section, such as for the index values 
for the genre indexes. When this occurred, it was because the graph was deemed to not show any 
additional, new, worthwhile insights. Section 4.1 shows the financial returns for each index and a 
graphic representation of the monthly index value over the period. Section 4.2 looks at the Alpha 
and Beta values estimated by the CAPM and Fama and French 3-factor and 5-factor models. 
Section 4.3 contains a correlation matrix to determine how the video game indexes compare to 
traditional financial assets. 
4.1 Pricing Index Returns and Graphs 
 Table 9 shows the nominal returns, real returns, standard deviation of returns, and Sharpe 
Ratio for the traditional financial assets and the video game price indexes representing all games 
and each level of completeness. Of the four video game indexes, the new games index saw the 
highest real return of 12.2%, while the index containing all games carried the most favorable 
Sharpe Ratio of 0.73. Table 10 shows the same information but only for the additional indexes 
that are not the focus of the primary analysis. The highest values for these indexes are 
summarized in Table 8 below. Figures 1-3 show the index values for each month for the various 
indexes from January 2014 to January 2022. These graphs show a steady rise in index value 
along with fairly low variation. This is especially true for Figure 1, which shows similar growth 
alongside market benchmarks, with less significant drops in value. 
  
29 
 
Highest Values for Additional Indexes 
 Highest Real Return Highest Monthly 
Sharpe Ratio 
Company Nintendo (10.77%) Sega (0.70) 
Console Nintendo 64 
(14.10%) 
Nintendo 64 (0.64) 
Genre Strategy (11.46%) Platformer (0.78) 
Table 8: Highest Values for Additional Indexes 
 
Game and Traditional Investment Asset Returns and Measures 
  
Index 
Name  
 Annualized 
Nominal  
Return (%) 
Annualized 
Real  
Return (%) 
Annualized 
SD (%) 
Monthly 
Sharpe  
Ratio1 
T
ra
d
it
io
n
al
  
A
ss
et
s/
M
ea
su
re
s 
CPI 2.3%   1.16% 0.41 
T-Bills 0.8% -1.5% 0.08%  
S&P 500 12.3% 9.8% 13.93% 0.25 
Gold 4.5% 2.1% 13.28% 0.10 
C
o
m
p
le
ti
o
n
  
an
d
 O
v
er
al
l 
All 
Games 12.7% 10.2% 4.52% 0.73 
Loose 11.5% 9.0% 5.21% 0.58 
CIB 11.5% 8.9% 5.03% 0.59 
New 14.9% 12.2% 6.78% 0.57 
Table 9: Game and Traditional Investment Asset Returns and Financial Measures 
  
30 
 
Additional Game Index Returns and Measures 
  
Index  
Name 
Nominal  
Return (%) 
Real 
Return (%) 
Annualized 
SD (%) 
Monthly 
Sharpe  
Ratio 
Number 
of 
Games 
C
o
m
p
an
y
 
Sega 13.09% 10.52% 4.86% 0.70 1,481 
Nintendo 13.35% 10.77% 5.18% 0.67 2,777 
Sony 10.77% 8.25% 4.54% 0.62 1,330 
C
o
n
so
le
 
Genesis 12.40% 9.84% 5.50% 0.59 727 
Game Gear 15.55% 12.92% 8.66% 0.47 235 
Saturn 12.86% 10.29% 6.17% 0.55 253 
Dreamcast 12.55% 10.00% 6.13% 0.54 266 
NES 12.73% 10.17% 5.38% 0.62 775 
Game Boy 11.29% 8.76% 7.06% 0.42 518 
SNES 14.64% 12.03% 6.28% 0.61 732 
Nintendo 64 16.75% 14.10% 6.78% 0.64 300 
Game Boy Color 12.07% 9.52% 6.15% 0.51 452 
PlayStation 10.77% 8.25% 4.54% 0.62 1,330 
G
en
re
 
Act. & Ad. 12.44% 9.88% 5.11% 0.63 1,893 
Arcade 13.00% 10.43% 6.40% 0.53 397 
Fighting 11.42% 8.88% 5.32% 0.56 280 
Other 13.27% 10.69% 5.39% 0.64 465 
Platform. 13.09% 10.52% 4.39% 0.78 336 
Puzzle 13.90% 11.31% 5.31% 0.68 220 
Racing 11.20% 8.67% 6.07% 0.48 496 
RPG 8.17% 5.71% 5.13% 0.41 238 
Sports 10.61% 8.09% 5.43% 0.51 1,114 
Strategy 14.05% 11.46% 5.42% 0.68 149 
Table 10: All Additional Index Returns and Measures 
  
31 
 
Figure 1: Primary Game Indexes and Traditional Investment Asset Values 
Figure 2: Console Index Values 
32 
 
Figure 3: Company Index Values 
4.2 Financial Metrics – CAPM and Fama and French Models  
Table 11 shows all Alpha and Beta estimates for the index including all games and the 
completeness indexes. These values were derived from the linear regression methodology 
previously mentioned. In the event that a value for a given model is statistically significant, that 
model can explain the excess returns of the asset and to what degree. The relatively low R-
squared values suggest that none of these models adequately explain the returns seen for the 
game indexes. All alpha measures are statistically significant at a level of p <0.01. The only 
factor with a P-value lower than 0.01 is the RMW (Robust Minus Weak) factor. Also known as 
the profitability factor, this takes the average return of two robust operating profitability 
portfolios and subtracts the average return of two weak operating profitability portfolios. The 
RMW Beta value for the index containing all games is 0.32, meaning such a portfolio would 
have lower exposure than the American Stock Market, which has a Beta value of 1. Table 12 
33 
 
shows only the R-squared, alpha value, standard error, t-stat, and P-value for each Alpha 
estimated. 
 
Table 11: Alpha and Beta Estimates Based on CAPM and Fama and French 3-Factor and 5-
Factor Models 
Table 12: Alpha Estimates and Hypothesis Tests Based on CAPM and Fama and French 3-
Factor and 5-Factor Models 
4.3 Correlation Matrix 
 Figure 4 shows a correlation matrix for all traditional assets, the comprehensive index 
including all games, and each completeness index. Each game index has a high correlation with 
all other game indexes, with the exception of the new index, which has a lower correlation with 
the loose and complete-in-box indexes. All game indexes have low positive correlation with 
inflation, although this correlation is higher than that seen from the S&P 500. They also have 
34 
 
negative correlation with T-Bills and low positive correlation with both the S&P 500 and gold. 
For convenience of viewing, asset returns are included alongside the matrix. 
Figure 4: Correlation Matrix and Asset Returns for Video Game Indexes and Traditional Assets 
  
35 
 
5.0 Discussion 
5.1 The Performance of Secondhand Retro Video Games as an Investment Asset 
 Based on the results of the data analysis, secondhand retro video games are a reasonable 
investment and perform better than more traditional investment assets. In particular, an 
investment strategy based on having an equal mix of games of all different completeness levels 
and from different systems would represent the greatest opportunity. The index containing all 
games reflects this strategy. This is evidenced by both the real return of 12.7%, which is slightly 
higher than the S&P 500’s value of 12.3%, and significantly lower variation. Additionally, the 
Sharpe Ratio of 0.73 is almost three times greater than that of the S&P 500. This finding breaks 
with prior literature examining the returns of collectibles. Secondhand retro video games could 
not be more different from collectibles analyzed in prior research, as they are less risky than 
stocks and provide comparable returns. Even if storage and transaction costs for the games were 
factored into this analysis, they would still prove an attractive asset for potential investors. This 
relationship holds for the other three indexes as well. However, they are not as attractive as 
simply choosing to hold a general portfolio of games as they have a lower Sharpe Ratio and only 
new games result in a greater real rate of return. 
 The index including all games is moderately positively correlated with inflation, only 
slightly positively correlated with the S&P 500 and gold, and negatively correlated with T-Bills. 
These results hold for the three completeness indexes. These findings suggest that while video 
games may not be much of an effective hedge against anything but T-Bills, they could be a better 
store of value than the both the American Stock Market and gold. Additionally, the low 
correlation and high risk premium associated with them suggests that retro games are a good 
investment from a diversification perspective since they could reduce the risk of a portfolio 
36 
 
without reducing the returns. Due to the low standard deviation of 4.5% for the index containing 
all games, secondhand retro video games are an increasingly attractive investment for retail 
investors who want something a bit more interesting than stocks or commodities. In fact, video 
games compare favorably to more traditional collectible investments such as wine as well. 
Burton and Jacobsen’s study of the rate of return on wine found that wine typically sees about an 
8% annual return, with some vintages seeing returns as high as over 16% (2001). However, wine 
requires high storage and shipping costs, making their investment potential debatable. Such high 
costs are not found with video games. 
 Additionally, the comprehensive game index and the three individual completeness 
indexes all have a statistically significant alpha of about 0.009-0.012 on a per-month basis 
according to the CAPM and Fama and French 3-factor and 5-factor models. This suggests that, 
in all cases, a typical basket of retro video games will outperform the market with significant 
excess returns over the risk-free rate and the excess returns of the market. However, these 
models all fail to adequately explain the price performance of secondhand retro video games as 
an investment asset. Therefore, we can assume there are additional factors explaining the price 
performance. The RMW beta value of 0.32 for the comprehensive index including all games 
suggests that this index has lower exposure compared to the market average, which has an RMW 
beta of 1. Thereby, diversifying one’s portfolio with retro games presents benefits even if there 
may be other risks. This reflects positively on the investment potential of secondhand retro video 
games. 
 While not the focus of this analysis, the additional indexes created for the console 
development companies, each individual console, and each genre do present some interesting 
insights. Firstly, Sega appears to be the console manufacturer with the highest investment 
37 
 
potential at a Sharpe Ratio of 0.70 and only marginally lower returns than games released for 
Nintendo consoles. However, this Sharpe Ratio is still lower than that of an index containing all 
games. Therefore, it is not a good idea for an investor to choose to invest solely in games for one 
company’s consoles.   
Secondly, while Sega is the best company overall, games for the Nintendo 64 have the 
highest investment potential for a single console. The high real return rate of 14.1% combined 
with a Sharpe Ratio of 0.64 sets it above all other systems analyzed. However, much like for the 
index of games released for Sega consoles, the Sharpe Ratio is lower than that of an index 
including all games. So, while the role of set completion may play a factor influencing the prices 
of games, it is not a good idea for an investor to attempt to contrive this by obtaining all games 
for a one system (Carey, 2008). 
Finally, platformer genre video games represent the highest valued opportunity for any 
genre. While the real rate of return of 13.09% is not quite as high as strategy or puzzle games, 
the Sharpe Ratio of 0.78 is significantly greater. This relationship is potentially explained by the 
genre including some of the most iconic retro games ever, like Super Mario World and Sonic the 
Hedgehog 2. However, answering this question would require an additional analysis and is 
outside the scope of this paper. 
5.2 Graphic Trend Analysis 
 When examining Figure 1, one sees the stability of the video game indexes compared to 
the stock market. While the S&P 500 does occasionally outperform the index including all 
games, it has significantly higher variability. For example, in early 2020 there was a significant 
drop in value on the S&P 500 in the wake of the Coronavirus. However, the prices of 
38 
 
secondhand retro video games were unaffected by this and appreciated over the same period. The 
only game index that saw a significant decrease in value over any period was the steady decline 
for loose games over 2017. However, this index rebounded and continued to grow alongside 
complete-in-box and new games. 
 Unsurprisingly, new games are worth the most, with steady returns consistently 
outpacing those of all other financial assets examined. There was a significant increase in index 
value at the end of both 2017 and 2021. The other indexes saw a similar leap over the same 
period in 2021, suggesting the sales at auction of rare games may have had a trickle-down impact 
on the prices of games on eBay. The outperformance of the stock market in the same year could 
be indicative of a potential bubble period for games. However, for the time being, prices for 
games continue to rise and it appears as if secondhand retro video games both offer short-term 
and long-term returns and relative stability for investors interested in alternative investments. 
 In summation, secondhand retro video games are an attractive asset for any investor 
interested in assets other than stocks, bonds, and commodities. They provide comparable, often 
greater, returns to the S&P 500 and are much less risky, with a lower level of volatility. This 
contradicts almost all prior research regarding the investment potential of collectibles. Returns 
are greater than those calculated by Dimson and Spaenjers and Baumol for British postage 
stamps and fine art, respectively. However, given the fact that fine art and postage stamps are 
more mature assets, it is possible that the remarkable returns and low volatility seen for 
secondhand retro video games will not continue in the long-term. 
  
39 
 
6.0 Post-Hoc Analysis 
 The results of the data analysis are interesting for retail investors, large investment firms, 
and alternative investment platforms. Retail investors looking for a unique alternative to the 
stock market or inflation hedges like gold and silver are likely to be attracted by the lower 
volatility, high returns, and personal interest. Large investment firms and alternative investment 
platforms may want to offer a fund composed of retro video games. Even retro gaming specialist 
stores may find some benefit to offering curated packs of games that offer good returns at an 
affordable price. 
 The analysis performed is by no means comprehensive. It only focuses on a limited 
number of consoles and ignores home computers with large gaming libraries, such as the 
Commodore 64, and any systems released prior to the Nintendo Entertainment System, such as 
the Atari 2600. Additionally, focusing solely on retro games means we exclude games for any 
console released after the year 2000. Games released for newer systems are still collectible and 
have appreciated in value over the same time period. Since many of these consoles will be 
considered retro in a few years, they potentially have the most interesting games from a growth 
perspective. As the future of gaming seems to be focused on digital downloads on home 
consoles, computers, and mobile phones, it is unclear how large a time period physical retro 
games will eventually encompass. It could be strictly limited to the period analyzed in this paper, 
as more modern releases forgo physical releases altogether. Finally, the dataset utilized lacks 
many interesting factors that could influence the value of a game. For example, knowing the 
publisher and developer of a given game would generate more interesting insights. However, 
while these limitations are present, it provides several interesting directions for future research to 
take. 
40 
 
 One simple path for future research is to provide a more comprehensive analysis 
including more consoles and more geographic markets. Additionally, choosing to focus on 
another geographic market altogether could be worthwhile. One potential market could be Japan, 
which has a strong retro gaming specialty store culture and prefers Yahoo! Auctions to eBay. 
Additionally, they have different cultural tastes when compared to America. These differences 
may make data gathering and analysis difficult but could provide an interesting insight into 
foreign collectors’ markets now that there is an American benchmark. 
 Another potential area of research are game auctions selling high-value games. While this 
selling channel has recently been mired in controversy over accusations of fraud, it is still a 
worthwhile target of research due to the substantial difference in sale prices and quality (Kamen, 
2021). In many ways, they are almost a completely different category of item in the way original 
fine art is distinct from a print. Future research could simply examine the returns on these games 
and how they differ. 
 However, the most interesting direction for future research would be to determine the 
best investment strategy for retro games. A hedonic analysis will show what factors have the 
greatest impact on the price of a game and, therefore, what the best strategy for investment is. 
The research of Cameron and Sonnabend and Wyburn and Roach provide satisfactory models for 
performing such an analysis (2020; 2012). Due to the limited scope of this paper and dataset, 
questions like these are best left for future research papers to analyze. 
  
41 
 
7.0 Conclusion 
 The goal of this thesis was to assemble a set of secondhand retro video game indexes, 
estimate their returns, and compare their performance to more traditional financial assets. To 
accomplish this, I gathered a robust dataset and performed a data analysis based on methods 
explored in the existing literature. The results showed that a typical portfolio of secondhand retro 
video games offers significant, superior returns and Sharpe Ratios to those of traditional financial 
assets. Particularly when compared to the S&P 500, video games offer comparable returns and 
lower volatility with an annualized real rate of return of 10.2% and a Sharpe Ratio of 0.73. These 
findings break with established literature on collectibles, which suggests that collectibles are not 
an efficient or worthwhile investment. The insights derived from this research are relevant for 
retail investors, investment firms, and alternative investment platforms looking for potential 
competitive advantage in a sea of similar offerings.  
However, the data I gathered and the analysis I performed are not comprehensive nor do 
they identify an ideal strategy for potential investors to follow. If future research were to 
determine what factors most influence the value of secondhand games and create a tailored index 
to reflect their findings, it may offer a foundation upon which firms could create funds and retail 
investors could build portfolios. There are still plenty of questions left unanswered that provide 
exciting avenues for future literature to explore. The results of my research imply that retro video 
games are not just entertainment anymore – there is serious money to be earned.  
42 
 
Appendix 1: Methodology for Rates of Return and Standard Deviation 
 
Annualized Nominal Rate of Return  
 The following formula was used to calculate the annualized nominal rate of return for 
each index: (
𝑋𝑖−(𝑋𝑓−𝑋𝑖)
𝑋𝑖
)
1
𝑇
, where: 
 Xi = Initial index value, 
 Xf = Final index value, and 
 T = Number of years. 
Real Rate of Return 
 The real rate of return was calculated based on the annualized rate of inflation from the 
CPI Inflation Index. The following formula was used to calculate the annual real rate of return of 
each index:  (
(1+𝑅𝑎)
(1+𝑅𝑖)
) − 1, where: 
 Ra = Annualized nominal rate of return for a given asset/index and 
 Ri = Annualized rate of inflation.
43 
 
Monthly Return Rates 
 Monthly return rates for each index were calculated beginning in the February 2014. The 
following formula was used each month to calculate the rate of return: 
(𝑋𝑡−𝑋𝑡−1)
(𝑋𝑡−1)
 , where: 
 Xt = The index value for a given month and 
 Xt-1 = The index value for the month prior. 
Standard Deviation of Returns 
 The standard deviation of returns was calculated the sample standard deviation of the 
returns for a given index. The following formula was used: √
∑(𝑥−?̅?)2
(𝑛−1)
 , where: 
 x = Sample mean and 
 n = Sample size. 
  
44 
 
(1) Kenneth R. French - Data Library; data accessed at 
http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html  
  
45 
 
References 
Baekeland, F. (1981). Psychological Aspects of Art Collecting. Psychiatry, 44(1), 45–59. 
https://doi.org/10.1080/00332747.1981.11024091 
Bailey, M. J., Muth, R. F., & Nourse, H. O. (1963). A Regression Method for Real Estate Price Index 
Construction. Journal of the American Statistical Association, 58(304), 933–942. 
Baumol, W. J. (1985). Unnatural Value: Or Art Investment as Floating Crap Game. Journal of Arts 
Management and Law, 15(3), 47–60. https://doi.org/10.1080/07335113.1985.9942162 
Behfar, K., & Okhuysen, G. A. (2018). Perspective—Discovery Within Validation Logic: 
Deliberately Surfacing, Complementing, and Substituting Abductive Reasoning in Hypothetico-
Deductive Inquiry. Organization Science, 29(2), 323–340. 
https://doi.org/10.1287/orsc.2017.1193 
Bishop, S. J. (2018). Motivations for Private Collecting. Fashion Theory, 22(4–5), 515–521. 
https://doi.org/10.1080/1362704X.2018.1425515 
Burton, B. J., & Jacobsen, J. P. (1999). Measuring Returns on Investments in Collectibles. Journal of 
Economic Perspectives, 13(4), 193–212. https://doi.org/10.1257/jep.13.4.193 
Burton, B. J., & Jacobsen, J. P. (2001). The Rate of Return on Investment in Wine. Economic Inquiry, 
39(3), 337–350. https://doi.org/10.1093/ei/39.3.337 
Cameron, S., & Sonnabend, H. (2020). Pricing the Groove: Hedonic equation estimates for rare vinyl 
records. Applied Economics, 52(50), 5516–5530. 
https://doi.org/10.1080/00036846.2020.1765963 
Carey, C. (2008). Modeling collecting behavior: The role of set completion. Journal of Economic 
Psychology, 29(3), 336–347. https://doi.org/10.1016/j.joep.2007.08.002 
 
46 
 
Choi, Y., & Lee, K. (2021). Recent Changes in Consumer Perception in Sneaker Resale Market. 
International Journal of Costume and Fashion, 21(1), 67–79. 
https://doi.org/10.7233/ijcf.2021.21.1.067 
Davison, E. (2021, April 1). Oh Great, Now Investors Are Buying Shares of Video Games. Wired. 
https://www.wired.com/story/video-game-stocks-vintage-games-mario-pokemon/ 
Devore, J. (2021, September 3). Retro game prices are high enough as is, thanks. Destructoid. 
https://www.destructoid.com/retro-games-expensive-enough-without-price-fixing-scheme/ 
Dimson, E., & Spaenjers, C. (2011). Ex post: The investment performance of collectible stamps. 
Journal of Financial Economics, 100(2), 443–458. https://doi.org/10.1016/j.jfineco.2010.12.005 
Dobrynskaya, V., & Kishilova, J. (2022). LEGO: The Toy of Smart Investors. Research in 
International Business and Finance, 59, 101539. https://doi.org/10.1016/j.ribaf.2021.101539 
Downing, S. (2011). Retro Gaming Subculture and the Social Construction of a Piracy Ethic. 
International Journal of Cyber Criminology, 5(1), 750–772. Criminal Justice Database; Publicly 
Available Content Database. 
Franses, P. H., & Knecht, W. (2016). The late 1970s bubble in Dutch collectible postage stamps. 
Empirical Economics, 50(4), 1215–1228. https://doi.org/10.1007/s00181-015-0974-3 
Grable, J. E., & Chen, X. (2015). Collectible, Investment, or Both: Evaluating the Attractiveness of 
Collectible Stamps. Journal of Financial Service Professionals, 69(5), 78–87. 
Hiraki, T., Ito, A., Spieth, D. A., & Takezawa, N. (2009). How Did Japanese Investments Influence 
International Art Prices? Journal of Financial and Quantitative Analysis, 44(6), 1489–1514. 
Cambridge Core. https://doi.org/10.1017/S0022109009990366 
47 
 
Juniper Research. (May 31, 2021). Global video game market value from 2020 to 2025 (in billion 
U.S. dollars) [Graph]. In Statista. Retrieved December 02, 2021, from https://www-statista-
com.ezp2.lib.umn.edu/statistics/292056/video-game-market-value-worldwide/  
Kamen, M. (2021, August 24). Record video game auction prices may be a result of “fraud.” NME. 
https://www.nme.com/news/gaming-news/record-video-game-auction-prices-may-be-a-result-of-
fraud-3027627 
Kirby-McGregor, M., & Martin, S. (2019). An R package for calculating repeat-sale price 
indices. Revista Română de Statistică, 67(3), p. 17.  
Kleine, J., Peschke, T., & Wagner, N. (2020). Rich men’s hobby or question of personality: Who 
considers collectibles as alternative investment? Finance Research Letters, 35, 101307. 
https://doi.org/10.1016/j.frl.2019.101307 
Koford, K., & Tschoegl, A. E. (1998). The market value of rarity. Journal of Economic Behavior & 
Organization, 34(3), 445–457. https://doi.org/10.1016/S0167-2681(97)00084-X 
Last, J. V. (2011). The Crash of 1993. The Weekly Standard (New York, N.Y.), 16(37). 
Leblanc, K. (2021, April 1). Curating Retro Video Game Collections. Information Today, 38(3), 17. 
Access World News. 
Mei, J., & Moses, M. (2002). Art as an Investment and the Underperformance of Masterpieces. The 
American Economic Review, 92(5), 1656–1668. https://doi.org/10.1257/000282802762024719 
Oosterlinck, K. (2017). Art as a Wartime Investment: Conspicuous Consumption and Discretion. The 
Economic Journal, 127(607), 2665–2701. https://doi.org/10.1111/ecoj.12391 
Shanaev, S., Shimkus, N., Ghimire, B., & Sharma, S. (2020). Children’s toy or grown-ups’ gamble? 
LEGO sets as an alternative investment. The Journal of Risk Finance, 21(5), 577–620. 
https://doi.org/10.1108/JRF-02-2020-0021 
48 
 
Stoller, M. A. (1984). The Economics of Collectible Goods. Journal of Cultural Economics, 8(1), 91–
104. JSTOR. 
WATA-Certified Vintage Video Games. (n.d.). Heritage Auctions. Retrieved December 5, 2021, from 
https://www.ha.com/information/certified-video-games.s 
Wulf, T., Bowman, N. D., Rieger, D., Velez, J. A., & Breuer, J. (2018). Running Head: Video Game 
Nostalgia and Retro Gaming. Media and Communication, 6(2), 60–68. 
https://doi.org/10.17645/mac.v6i2.1317 
Wyburn, J., & Roach, P. A. (2012). An hedonic analysis of American collectable comic-book prices. 
Journal of Cultural Economics, 36(4), 309–326. https://doi.org/10.1007/s10824-012-9166-6