The Diffusion of Wal-Mart and Economies of Density by 670c2f55bd90e297


									The Diffusion of Wal-Mart
and Economies of Density
               Paper by: John Holmes
             Discussant: Brett Wendling

  Disclaimer: The views expressed in this presentation are those
  of the discussant and do not necessarily represent the views of
  the Federal Trade Commission or any individual Commissioner.
               Paper Summary
   Economic Question:
    – How do the competing effects of “cannibalization” and
      “density economies” affect store location choice?
    – Provides a lower bound for size of density effects.

   Data: 1962-2005 Wal-Mart store location,
    distributor location, and store profits.

   Estimation: Structural choice model with
    endogenous location.
    – Profits key explained variable
    – Store location (distance) and population density key
      explanatory variables
    Paper Summary (cont.)
 Results:
  – Store Density:
     Density   is Important
       – Store-specific cost savings of $220K with a
         regional distribution center 100 miles closer
   Tie this paper to urban literature
    – Many analogies in city formation literature:
                choice explains variable of interest
        Location

        “Agglomeration Economies” vs. “Rents”/”Cannibal”

    – Contextual Justification:
        Why   location choice is important in other literatures

    – Model Assumptions:
        Similarmathematical techniques
        Previous experimentation with specific functions
Location Choice not Exhaustive
   Only observe 2 of 4 Wal-Mart types
    – Do not observe “Sam’s Club”
    – Do not observe “Neighborhood Stores”

   May believe location choice will depend on
    the location of these stores
    – Unobserved store types may affect both
      “cannibalization” and “density economies”

   Very little discussion in paper about how
    the omission of these decisions affect
           Product Heterogeneity
   Wal-Mart offers numerous products
    – Product availability is endogenous
          Related to location choice
             – Minnesota vs. Florida
          Could potentially affect profits
          May lead to biased estimates

    – Products may depend on unobserved competition

    – Product mix could have changed over time
          “Chow” test of results before 1980 and after 1980

    – Product prices vary dramatically
          CD: $7.88 ; Power Generator: $547.47
               Better data choices

   Wal-Mart unnecessary for question
    – Not asking why Wal-Mart is a profitable firm
          Inter-firm distribution could be important
          Price negotiations could be important
    – Question only needs retail firms making location choices
      that affect profits

   Many retail firms face similar problems
    – Distribution is important part of cost structure for nearly
      every retail firm
    – Population density affects sales
    – Proximate locations cannibalize sales
          Better Data (cont.)
 Otherfirms can address identical
 problem and offer relatively
 homogenous products
  – Fast food firms
     RaphaelThomadsen has paper using fast
     food data
  – Starbucks
  – Bookstores
   Size and effect of agglomeration
    economies are important in many contexts

   Model provides a reasonable exposition for
    how firm-specific agglomeration
    economies are formed

   Choice of Wal-Mart data unnecessarily
    introduces heterogeneity bias into analysis

To top