The Scale and Scope of Online Retail, GFP Ottawa 2018

1. The Scale and Scope of Online Retail
Alex Chernoff
Senior Economist, Bank of Canada
June, 2018
The views in this paper are those of the author and do not necessarily reflect those of the Bank of Canada

2. This Paper
Studies the growth of online retail over the period 1999-2012,
using firm and product level data from Canada
Decomposes online retailers’ revenue into the contributions of
product scope (the number of product categories), and
product scale (average revenue per product category)
Main descriptive result: average product scope dropped from
59 to 5 product categories over the period 1999-2012
Using data on Amazon.ca’s entry in to Canada, I find
evidence that this reduction in product scope may have been
driven by increased competition in the online retail sector
2 / 19

3. Related Literature
Multi-Product Firms
Arkolakis and Muendler (2010), Bernard, Redding and Schott
(2010), Bernard, Redding and Schott (2011), Goldberg,
Khandelwal, Pavcnik and Topalova (2010), Mayer, Melitz and
Ottaviano (2014)
E-Commerce: The Long-Tail
Anderson (2004), Anderson (2006), Bar-Isaac, Caruana and
Cunat (2012), Brynjolfsson, Hu and Smith (2003),
Brynjolfsson, Hu and Smith (2006), Brynjolfsson, Hu and
Smith (2010), Brynjolfsson, Hu and Simester (2011)
3 / 19

4. Data
Statistics Canada’s Annual Retail Non-Store Survey (RTNS),
1999-2012
For purposes of this study, a firm is an online retailer if:
1. It has NAICS classification 4541, and
2. It reports a positive value of online sales
Examples of RTNS Commodity Annex product categories:
Televisions and audio/video equipment
Still cameras and related photographic equipment and supplies
Computers, components and related equipment
Computer software, business and entertainment
Telephones and home office electronics
4 / 19

5. Product Scope and Scale Decomposition of Revenue
¯pqt = ¯Nt
¯pqt
¯Nt
⇐⇒ ln ( ¯pqt)
ln(Revenue)
= ln ¯Nt
ln(Product Scope)
+ ln ¯pqt/¯Nt
ln(Product Scale)
where, ¯x denotes the sample average of x in period t
5 / 19

6. Average Revenue, Scale, and Scope
6 / 19

7. Product Scope Decomposition by Firm Size
¯Nt =
1
2
¯NSmall,t +
1
2
¯NLarge,t
¯NSmall,t: Small firms (TOR <= Median TOR)
¯NLarge,t: Large firms (TOR > Median TOR)
7 / 19

8. Product Scope Decomposition by Firm Size
8 / 19

9. Product Scope Life Cycle Decomposition
¯Nt = θI,t
¯NI,t + θE,t
¯NE,t + θX,t
¯NX,t + θS,t
¯NS,t
¯NI,t: Incumbent (observed in t-1, t, t+1)
¯NE,t: Entrant (observed in t, t+1, not in t-1)
¯NX,t: Exiting firm (observed in t-1, t, not in t+1)
¯NS,t: Singleton (observed in t, not in t-1, t+1)
9 / 19

10. Product Scope Life Cycle Decomposition
Fraction of each type
10 / 19

11. Summary of Descriptive Analysis
Product scope dropped from 59 to 5 product categories over
the period 1999-2012
This decline was apparent for both small and large firms
Entrants, incumbents, and exiting firms all experienced a
similar decline in product scope
Possible explanations:
Technological change
Increased competition in the online retail sector
11 / 19

12. Amazon.ca
Overview of Amazon.ca “store” openings:
2002: Books, Music, Videos, and DVDs
2003: Software, Computer/Video Games, Marketplace
2008: Electronics
2009: Home and Garden, Kindle
2010: Sports and Outdoors
2011: Tools and Building Supplies
2012: Baby
Amazon’s staggered store openings are used to study the effect of
foreign entry on Canadian online retailers’ product scope
12 / 19

13. Product Scope: CND Online Retailers/Amazon.ca
13 / 19

14. Econometric Framework
Difference-in-difference regression specification:
sharej,t = δj + αt + βAmazonj,t + uj,t
sharej,t is the share of firms with positive sales in product
category j in year t
Amazonj,t ≡ 1 if Amazon.ca has entered product category j in
year t, and 0 otherwise
Distribution of shares
14 / 19

15. Difference-in-Difference Results
1 2
Amazonj,t+3 -0.0657***
(0.0115)
Amazonj,t+2 -0.00667
(0.00592)
Amazonj,t+1 -0.00910**
(0.00369)
Amazonj,t -0.0564*** -0.0150***
(0.0113) (0.00503)
Amazonj,t−1 -0.00709
(0.00454)
Amazonj,t−2 -0.0174***
(0.00439)
Amazonj,t−3 -0.0217*
(0.0128)
Number of Lags 0 3
Number of Leads 0 3
Product and Year Yes Yes
Dummy Variables
Observations 1,485 1,485
R-squared 0.915 0.921
Standard errors (in parentheses) are clustered at the product level; *** p<0.01, ** p<0.05, * p<0.1
15 / 19

16. Conclusion
Average product scope of Canadian online retailers dropped
from 59 to 5 product categories over the period 1999-2012
My results suggest that this reduction in product scope may
have been driven by increased competition in the online retail
sector
Next steps: Map Amazon.com product categories into the
RTNS Commodity Annex to address the endogeneity problem
in the current difference-in-difference framework
16 / 19

17. References
Anderson, Chris (2004) ‘The long tail.’ Wired Magazine 12(10), 170–177
Anderson, Chris (2006) The Long Tail: Why the Future of Business is Selling Less of More (New York: Hyperion
Books)
Arkolakis, Costas, and Marc-Andreas Muendler (2010) ‘The extensive margin of exporting products: a firm-level
analysis.’ NBER Working Paper 16641, 1–50
Bar-Isaac, Heski, Guillermo Caruana, and Vicente Cunat (2012) ‘Search, design and market structure.’ American
Economic Review 102(2), 1140–1160
Bernard, Andrew B., Stephen J. Redding, and Peter K. Schott (2010) ‘Multiple-product firms and product
switching.’ American Economic Review 100(1), 70–97
(2011) ‘Multiproduct firms and trade liberalization.’ Quarterly Journal of Economics 126(3), 1271–1318
Brynjolfsson, Erik, Yu (Jeffrey) Hu, and Duncan Simester (2011) ‘Goodbye pareto principle, hello long tail: the
effect of search costs on the concentration of product sales.’ Management Science 57(8), 1373–1386
Brynjolfsson, Erik, Yu (Jeffrey) Hu, and Michael D. Smith (2003) ‘Consumer surplus in the digital economy:
Estimating the value of iincrease product variety at online booksellers.’ Management Science 49(11), 1580–1596
(2006) ‘From niches to riches: anatomy of the long tail.’ MITSloan Management Review 47(4), 67–71
(2010) ‘Research commentary - long tails vs. superstars: the effect of information technology on product
variety and sales concentration patterns.’ Information Systems Research 21(4), 736–747
Goldberg, Pinelopi K., Amit K. Khandelwal, Nina Pavcnik, and Petia Topalova (2010) ‘Multiproduct firms and
product turnover in the developing world: evidence from India.’ The Review of Economics and Statistics
92(4), 1042–1049
Mayer, Thierry, Marc J. Melitz, and Gianmarco I. P. Ottaviano (2014) ‘Market size, competition, and the product
mix of exporters.’ American Economic Review 104(2), 495–536
17 / 19

18. Fraction of Firms by Life Cycle Type
Back
18 / 19

19. Share of Firms in Each Product Category
Econometric Model
19 / 19