APPENDIX
Decomposing nominal export growth
We examine the extent to which the increase in nominal exports between 1993 and 1996
can be attributed to an increase in exports by incumbents, the net entry of new firms into the
export market, and the size differences between entering and exiting firms.
(A1)
Qf- Qff-ι
Qt-1
(
- Σ
iεm^
Qff )
Qt-1 )
( Qft-⅛ι )+√ Qft I
—Q— +Σ f
V Qit-1 ) iɛb V Qf-1 )
—
ς(f
i≡d V Qt-ι
)
)
|
Σ Sft-1 iεm |
Qft - Qft-1 '' |
+ |
(nf nf ) nbt ndt-1 |
( Qf + Qf ) |
+ |
Qbt - Qdt -1 |
(nf + nf ) |
|
V Qft-1 ) |
V nt 1 ) |
V 2Qtf-1 ) |
Qf V Qt-1 ) |
2nf „ V nt-1 ) |
Here Sift-1 denotes the share of total exports attributable to the fh plant in year -1, nf
refers fo fhe number of planfs fhaf are exporfing, Qff is oufpuf sold in foreign markefs during year
t, and overbars denote temporal averages. In general, i subscripts refer to firms, m subscripts to
incumbent firms, b subscripts refer to plants in their first year of exporting, and d subscripts refer
to firms that exported at time t-1, but ceased doing so at time t. Aggregates without these
subscripts refer to the entire set of exporting plants.
We are concerned mainly with the second line of equation (A1). The first term measures
the contribution of incumbent exporters to sample-wide export growth. It is a weighted average of
the growth in exports among producers who continue to sell abroad (i ∈ m), the weights being
their share in total period t-1 exports. The second term measures the effect of net changes in the
number of exporters on growth, i.e. the difference between the number of firms that begin
exporting between periods t-1 and t (i ∈ b), and the number of firms that cease exporting during
the same period (i ∈ d). The third term describes the effect on export growth of replacing firms
ceasing to export with firms beginning to export. If both groups export the same amount per firm,
the turnover effect is nil. On the other hand, if large exporters drop out of foreign markets and are
replaced by small exporters, turnover may have a contraction effect.
25