FDI’s Imact on Domestic Firms:
spillover through backward linkage
Javorcik (AER, 2004)
Paul Deng
March 22, 2011
1
Big Picture
2
Big Picture
3
The Impact of FDI on Host Countries

MNEs are the most productive firms in their home
countries

MNEs, most of the time, are more productive than firms
in host countries, especially compared to those in
developing countries

Most MNEs are skill intensive, knowledge intensive, and
heavy in R&D investment

Host countries want to attract FDI because they may
benefit from MNE’s presence, through spillover effect
4
The Impact of FDI on Host Countries

The spillover effect could be positive, because

personnel (both workers and executives) trained at MNEs are
more skilled, and later they may open their firms, or work in other
domestic firms

Technology may leak to domestic firms, through domestic firms’
interactions with MNEs

Above two are the most obvious and easiest spillover channels

There are other channels, the mechanism of which economists
are still trying to untangle, and we will discuss them later
5
The Impact of FDI on Host Countries

The spillover effect could also be neutral or even
negative

MNEs’ incentives to protect technology from leaking so to
maintain their lead in innovation put a brake on technology
transfer

MNE’s entry into domestic industry may out-compete
domestic firms, sometimes forcing them to shut down or
switch to other industries

Again, there are more complicated channels, which we will
discuss later
6
Spillover and Its Relation to Type of FDIs

Horizontal spillovers – related to horizontal FDI


Spillover from MNEs to domestic firms within the same industry
Vertical spillovers – related to vertical FDI

Backward linkage
spillover from downstream firms to upstream firms
 e.g., spillover from foreign firms to their domestic suppliers
 The focus of this paper


Forward linkage


spillover from upstream firms to downstream firms
e.g., foreign circuitboard producer and domestic PC maker
7
Javorcik (2004), FDI and Its Spillover Effect

Research question:
Does FDI increase doemstic firms’ productivity?
 Through what channel?


Javorcik investigated spillover effect through the following channels
or linkages:



Horizontal, i.e., within the same industry
Backward, i.e., downstream industry to upstream industry
Forward, i.e., upstream to downward industry

The author argues spillovers from FDI are more likely to be vertical
than horizontal. Why?

Pay special attention to how he defines and measures the vertical
linkages
8
Javorcik (2004), Data Description

Lithuanian firm-level data, with the whole sample covering
85% of total output

This paper only focuses on manufacturing firms, in over
20 industries

Unbalanced panel data from 1996 to 2000, each year
around 2000 to 2700 firms, after data cleaning process
9
A Snapshot of FDI in Lithuania
10
Javorcik (2004), Estimation Strategy
(1)
time effect
regional effect
Industry effect
i: firm j: industry r: region t: year
Note that the first 4 variables are indexed at firm i level, while the rest 3
variables are indexed at industry level
Also note firm-level fixed effect is not controlled in this regression equation
11
How might the linkages work through?

Horizontal linkages



Knowledge spillover thru personnel turnover
Competition effect – negative and positive?
Vertical linkages

Backward linkage



Selection effect
Scale of economy effect
Forward linkage

Competition effect – more efficient production or cheaper
inputs
12
Javorcik (2004), Linkage Measures
13
Javorcik (2004), Linkage Measures
backward linkage

Firms in
industry j

forward linkage
MNEs in
downstream
industry
MNEs in
upstream
industry
MNEs in
downstream
industry
MNEs in
upstream
industry
MNEs in
downstream
industry
MNEs in
upstream
industry



Firms in
industry j

14
Spillover Linkages
15
Javorcik (2004), Estimation Results
16
Fixed Effect with Difference Estimator
Reminder:
17
Estimation Results with Fixed Effects
18
Foreign Ownership and Vertical Spillover
19
Summary of Empirical Findings

Backward linkage is quite robust in various different
estimations

Forward and horizontal linkages are much less robust

Backward linkage seems to work best when a foreign firm has
a local partner, i.e., joint ventures --- important policy
implications for host countries
20
Some Further Thoughts

The specific mechanisms through which backward linkage
operates are still not very clear

Does backward linkage operate through a selection effect by
MNEs?
 Higher quaility control?
 Picking more productive suppliers?
 Competition among suppliers (in winning MNE’s contract)
lead to more efficient production?

Economists are still trying to figure out…
21
Next Time…

Our last class; Afterwards, Niels will take over

Read Harrison (1999), AER, ”Do Domestic Firms Benefit from
FDI.”

Really start to think hard on your term paper, don’t wait until too
late.
22