From estimating probit models determining which firms exported at any time during
1996-2004, using a weighted FAME dataset, our results for 16 separate UK industry
groups (covering all the main marketed output sectors of the economy) confirm what
most other similar studies have reported in the literature on self-selectivity. We find
that firms with higher (labour) productivity in the previous year are significantly more
likely to sell overseas in the current period. Also firms that are older or that possess
intangible assets (e.g. R&D stock, brand recognition, goodwill, etc.) are generally
much more likely to export.
In terms of the ‘learning-by-exporting’ effect associated with post-entry sales to
overseas markets, we test the relationship between exporting and productivity using
three approaches to controlling for selectivity effects: an IV model (with the age of
the firm and whether it had intangible assets as the additional instruments used to
control for selectivity); a control function approach (with the selectivity terms
obtained from a first stage probability of exporting model included in the production
function to control for bias); and a matching approach (based on the propensity scores
obtained from the probability of exporting model). We have estimated production
function models that incorporate the determinants of TFP, including exporting, and
results show that generally all three approaches produce broadly similar results:
‘learning-by-exporting’ is present but by no means universal, and even within
industry groups there are differences amongst entrants, exitors, and those that
experience both entry and exit into overseas markets. However, in terms of the overall
estimate for the UK economy, our findings suggest a fairly substantial post-entry
productivity effect for firms new to exporting; a negative effect for firms exiting
overseas markets in the year they quit and thereafter; and large productivity gains
whilst exporting, of firms that both enter and exit.
33