Once again, thank you for allowing us to consider your paper. We regret not being
able to publish this paper in its present form, but we hope that you will revise it
and resubmit it to the Australian Journal of Agricultural and Resource Economics.
Sincerely,
Prof. Bob Lindner
Editor, AJARE
Editorial Office
ajare@blackwellpublishingasia.com
Editor Comments to Author:
Water (or lack of it) is very much in news, so the topic of the paper has some
currency.
However, the referees are not convinced that this is a publishable paper. While they
see some merit in the inclusion of relative endowments as an explanatory variable in
a gravity equation, they think the scope of the paper is rather limited.
Furthermore, they did not find the arguments sufficiently convincing, and the paper
is not well written, with many grammatical errors. Statements and objectives are
often not clear or incorrect. Hence, the comments provided by each referee require
an entire re-write of the paper to address them satisfactorarily. You would need to
pay attention to all comments made by these experienced referees before
re-submitting a revised version of your paper for further consideration.
Reviewer Comments to Author:
Reviewer: 1
Comments to the Author
see attached .pdf file
Reviewer: 2
Comments to the Author
Comment Summary
The paper would benefit from a clearer statement of the research question. The
author seems to be exploring the general question of how water endowments affect
food trade. The model motivating the work is apparently a model with two factors
(water and labour or water and land) and two goods (food and other goods). The
research question appears to be whether such a model is helpful for understanding
the trade pattern. It does not surprise me that this framework is not especially
useful.
A more interesting paper would compare water to other endowments (especially land)
as a source of comparative advantage in food exports. It would also disaggregate
food exports to look at specialization within the food sector. Another possibility
would be to explore the role of water endowments in choosing production techniques.
Major comments
1. Most of the paper reads as if the author has a two-factor world in mind: water
and some other factor (be it labour or land). I have difficulty accepting this as a
realistic framework. Obviously land endowments matter for agricultural production,
and land is a complement to water, not a substitute. The inclusion of the
water-to-land ratio in the regressions and correlations suggests that the implicit
model would predict that an increase in the land endowment (a la Rybczynski) would
decrease food production. I dont think anyone has that as a model of food
production.
When the author uses water per capita as the relevant endowment measure, then one is
left wondering about the role of land. The relatively weak role of water endowment
suggests that land endowments are more important than water in determining
comparative advantage.
2. One possible way to interpret this paper is that the author is trying to ask
whether or not water endowments are sufficient to explain the pattern of trade in
food products.
In that case, it would seem more useful to run a horse race between
water and other endowments (land, population, capital), rather than simply to
evaluate the role of water alone. In that case, the gravity model would include
each endowment separately, rather than as a ratio with water (with other endowments
excluded). In the revealed comparative advantage tests, the comparative advantage
measure would be regressed on the four factors, normalized by real GDP. These
specifications would allow the role of water to be compared to other endowments in
determining the trade pattern.
3. The author is far too quick to aggregate food products. Table 1 shows that
various food products have substantially different water requirements. There may
well be specialization within food products that can be better explained. How do
water requirements correlate with prices. What is the implied price of the water
embodied in food exports?
4. In a related point, Table 1 reveals substantially different water requirements
across countries for the same crops. This suggests that there are differences in
production techniques across countries. These differences are quite plausibly
related to water scarcity or water prices. Is there any evidence that this is the
case? Israel uses drip irrigation to produce some of the same things that California
produced with flooded fields. What does that tell us about the ability of countries
to adjust to water shortages with land- labour- or capital-intensive techniques?
5. The regression tables should include information on the number of observations
in each regression. The author should also be clear about what has been done with
the observations in which bilateral trade takes the value of zero. Presumably,
these have been dropped when the author took logs of the trade flow. It would be
more useful to have included the zero observation somehow (either by running a
tobit, or by adding $1 to every flow). Without some sort of probit or tobit
regression, the author should not be making statements about determining whether
there will be trade, as he does on page 7. A regression on levels of trade (which
only include positive observations) tells us nothing about the question of whether
there is trade.
Minor comments
1. The author interprets the coefficients on the water endowment in the gravity
regression (larger for exporter endowments than for importer endowments) in the
following way: The asymmetry means that water is more important for the exporter
thn the importer. It is not clear what the word important means here. I interpret
this fact as implying that the value of bilateral trade more sensitive to the
exporters endowment than to the importers endowment.
2. It is somewhat surprising to see the Netherlands as the worlds largest net
exporter of food. My guess is that this is due to re-exports of food products
shipped from elsewhere. Am I wrong about this, or are re-exports an important part
of the data? If trade is badly recorded, why should we expect endowments to matter.
At a minimum it would useful to exclude major transshipment points like the
Netherlands, UAE, Hong Kong and Singapore.