LOMBORG-ERRORS                       Home
                        List of chapters
Chapter 9:

Will we have enough food?

                       

 

MAIN ISSUES:

Lester Brown from Worldwatch Institute has repeatedly claimed that the world´s agricultural production cannot keep up with the increase in world population. Lomborg tries to demonstrate that on the contrary, the world´s food situation is slowly, but steadily improving, especially due to the ever increasing crop yields per hectare. The criticism of Lester Brown is justified in some respects, especially concerning the situation in China. However, Lomborg´s presentation is also flawed, with a bias to the optimistic side.

The reader will get a fine overview of the issues treated here by studying a report from the USDA Economic Research Service (ERS). It is called "Linking land quality, agricultural productivity, and food security", and is written by Keith Wiebe (2003). It may be downloaded here. It will be referred to several times in the comments to Lomborg´s text.

Wiebe´s report ends up with model calculations of the food and hunger situation in the near future in the developing world. The model is dynamic and includes changes in crop yields, prices, income etc. Furthermore, most interestingly, the population in each country is divided into five parts (quintiles) according to differences in income. This allows calculation of the distribution gap, i.e. the amount of additional food required to nourish also the poorest people who at the current price and income levels cannot afford to buy what is necessary for them. This, in turn, allows a calculation of the total number of malnourished people in all of the 67 selected developing countries. The results are reported separately for each of five major regions. In the standard run of the model, the total food production is expected to increase by 1.8 % per year, of which 0.5 % per year is due to expansions of the agricultural area, and 1.3 % per year is due to increased yields per area. In this run, the total number of malnourished people in the 67 countries is projected to decline overall from 774 millions in 2000 to 694 millions in 2010. However, this trend is composed of an increase in the number of hungry people in Subsaharan Africa by 91 millions, and a decrease in the rest of the countries by 171 millions. Thus, the projected trend is composite: although the overall trend is positive, the hunger problem will get more severe in the most vulnerable countries and/or among the lower income groups in these countries. That is, pessimists focusing at the black spots will be right, at the same time as optimists looking at the overall situation will also be right. In addition to this baseline run, two other runs are made in which the negative effects of erosion and degradation of agricultural land are left out, either by setting the loss of agricultural land area to zero, or by reducing the productivity loss of unfertilized land. The effects of these changes are felt especially in Subsaharn Africa where increase of agricultural production is mainly based on expansion of the cultivated area rather than on increased input of chemicals and technology. The number of malnourished people may, in these alternative runs, be reduced by a further 68 millions. These results are very sensitive to changes in the extent and severity of soil erosion. In any case, it seems not possible to reach the World Food Summit targets.

 

P. 94-95 and note 619: FLAW

On Fig. 50: "But this selective figure gives the wrong impression and is guided by a faulty logic" (94 left). "In the developing countries, however, production has kept growing . . "(94 right). "Thus, a careful examination of the available evidence shows that food production has not lost its momentum." (95 left) "The last four years have been below the top 1996 production . . but this is normal for such statistics with large, natural variability." (note 619) Flaws: There are many flaws in this text. Firstly, Lomborg says that a curve illustrating the situation for the whole world is selective - in marked opposition to his own principles stated on p. 7 that "we can only elucidate global problems with global figures". This is probably more true here than elsewhere, because a large proportion of the grain consumption in the third world is imported from North America and EU, so the right thing to consider must be the global production. But Lomborg wants to give a more optimistic picture (he has stated explicitly that he wants to contradict Lester Browns pessimism), so he prefers the lower curve in fig. 50 which gives the situation for the developing countries only. In his claim, this figure shows that the growth keeps on. He has a remark, though, in note 619 that the last four years (1997 to 2000) have been lower than the preceding top, but he feels sure that this is just a temporary fluctuation. Concerning this, we may judge Lomborg be the same standards that he applies against Brown: Lomborg uses a lot of space to stress that Browns interpretation of Fig. 48 does not hold true, because Fig. 49 shows that the continued trend was opposite of what Brown believed. Now, in Fig. 50, Lomborg believes that the growth will continue. When we consult the newest data from the FAO database (for 2001 and 2002) we see, however, that the downward trend from 1997 onwards has continued. So the trend is opposite of what Lomborg believed. The "temporary downward fluctuation" has now lasted through six seasons. 2002 marks the lowest downward fluctuation for several years. This is true both for the world in total, and for the developing world alone. So there is less basis than before for being cocksure that growth continues. In any case, the percentage of annual growth has clearly declined since 1984 and been below zero since 1995. This is evidently what is meant by that growth has lost its momentum. So Lomborg´s conclusion is downright wrong.

P. 97 left and note 641: FLAW

Lomborg gives a summary of Lester Brown´s statements in his "The state of the world 1998". He writes inter alia ". . does it signal a future leveling off in other countries . .? . . Lester Brown answers this latter question in the affirmative." Flaw: Lomborg´s quotations of Lester Brown are very selective and do not in a fair way represent Brown´s text. To use Lomborg´s own phrasing on p. 97: Notice how these quotations have been chosen with great care. Actually, it is not true that Brown simply postulates that the agricultural yeilds will level off soon everywhere, and that this will lead to an age of scarcity with unprecedented political instability in Third World cities and setting the entire monetary system at risk. Brown uses many pages to describe the scope for further improvements in yield. He does list factors which will allow further increases in yield, and after a thorough survey over many pages he finally concludes that overall, furhter increases will be relatively small. He says it is possible that the global grain production will grow by less than 1 % annually after 2000 - i.e. he imagines continued increases, although smaller than now. He says that if wheat prices double for an extended period, then this will lead to political instability in the cities etc., but he does not claim that this will definitely happen. The quote that "eventually the rise in grain yields will level off everywhere" is criticized by Lomborg in note 641 with the remarks that this is not interesting, the interesting is when the leveling off will take place. However, Lomborg has taken this sentence - like many other sentences - out of its context. If Lomborg had quoted the whole context, we would see that Brown says exactly that which Lomborg criticizes him for not saying. The overall impression of Lester Brown´s text is that of growth rates that gradually become smaller and smaller and eventually will reach zero. This is true both in "The state of the world 1998" and in the preceding 1997 book. The impression that production will `hit the wall´ is just a metaphor which Brown uses twice, and which does not characterize his text.

P. 96-98: ERROR

In the section "Limits to yields?" Lomborg claims that the yield of wheat production does not level off. He denies Lester Brown´s postulate that since 1983, the wheat yield in USA has not risen (96 right), and he writes "Actually, the growth rates of top wheat production have in all likelihood not decelerated but remained constant at about 1 percent over the last 30 years." (98 right). Error: To investigate "top wheat production", we may consider the FAO database on wheat yields in EU (12 countries). Up to about 1990, the yields grew by about 3 % per year, but after that, up to 2002, the growth rate has only been about 1.5 % per year. Thus, contrary to Lomborg´s claim, there has indeed been a marked deceleration. Considering wheat production in USA (where yields per area are much lower than in EU), Lomborg is right that there has not been a complete halt of further increase since 1983, and that here has been continued growth. But he fails to acknowledge that there has indeed been a deceleration: The previous rate of 1.6 % growth per year has gone down to 1 % per year from 1984 onwards.

P. 96-98 and Fig. 54: FLAW

Concerning the yields in rice production, Lomborg criticizes Lester Brown´s claim that there is an upper limit. He writes inter alia: "The "wall" Brown has found for Japanese rice yields seems to be caused by a deliberately misleading example." (98 left). Flaw: Firstly, the concept of the "wall" is stressed by Lomborg, not by Brown. More importantly, however, a more careful investigation of the development in rice yields is necessary. For this purpose, I (K.F.) have used data on rice yields for 1961-2001 in Japan, South Korea and China from the FAO database (http://apps.fao.org). These data are generally higher than the USDA data and are only comparable with them after application of a correction factor. I have used 5-year averages and studied the annual rates of increase in yields for the periods 1963-1975, 1975-1987, and 1987-1999. For China, I neglect data from the Mao period (pre 1975). When the growth rates, calculated for each 12 year period, are plotted against yield levels at the midpoint of each period, the combined data from all three countries show a strong and clearly significant correlation telling that the growth rate declines with increasing yield (for more details, see my "example 1" by clicking here). Therefore, it is justified to fit a regression line, and by extrapolation, we predict that the growth rate will be zero at a level of 6.7 tons per hectare. Actually, the time series for all three countries seem to converge at this level. Relative to this level of zero growth, the average yields during the latest 5-year period was 100 % in South Korea, 97 % in Japan, and 94 % in China, and the level of the socalled "wall" (the 1984 yield in Japan), is 96 %. Thus, when we base our conclusions merely on the actually reported yields, and do not consider whether new technology might change the picture, we see that Lester Brown was approximately true when he described the upper limit to production. This upper limit should not be considered a "wall", but a fixed level around which the yields will continue to fluctuate up and down. Thus, the data seem to fit what the ecologists call a logistic growth curve, i.a. a sigmoid curve that approaches an asymptote. In contrast to this, Lomborg claims that data from e.g. South Korea clearly show a steady increase in yields (98 left). As we can see now, this is downright wrong. Also, he cites persons who are confident that further, stable increases are still possible (98 right); but this is just a subjective interpretation. If we stick to what the data show, and leave out all subjective evaluation - and that is allegedly Lomborg´s intention - then the data indicate that the growth will halt, and that China has already nearly reached the upper limit. It should be admitted that later on, on p. 100, Lomborg says something else and now acknowledges that the growth rates are declining, but he states that this is no problem as long as the growth rates in yields are larger than the growth rates in population.

P. 97 right: FLAW

"EU produces more than twice the amount of wheat per hectare than the US. This is primarily because the EU farms much more intensively because it has much less farmland." Flaw: Here, Lomborg leaves the mere inspection of data and advances a subjective interpretation. This interpretation is probably wrong. The primary explanation for lower yields in USA is probably differences in soil types and climate, especially rainfall. There is no indication that USA could rise its yields to the EU level (or to the present level in China) if only it wanted to.

P. 97 right: FLAW

"For neither the EU, the US nor the world at large do data seem to support the hypothesis of a "wall". Flaw: Here, Lomborg criticizes Brown with no right. Brown has not postulated that there exists an overall "wall" that limits wheat production worldwide. Brown has clearly and in detail explained that the maximal attainable wheat yields are widely different in different countries due to differences in soil type, day length, sunshine hours and rain fall. Therefore, Lomborg is not warranted to say that since yields are higher elsewhere than in USA, the postulate of an upper limit to wheat yields in USA is false.

P. 98 right, notes 662 to 668: FLAW IN REFERENCES

Lomborg claims that large productivity increases will be possible in the future, due to technological progress. As a reference supporting this claim, he cites two sources referred to in notes 662 to 668. Flaw: Both the two sources are indicated as "not available" in the reference list. Thus, the reader has no possibility at all to check if the forecasts of improved yields are made on a sound basis, or are just hopeful speculations. Lomborg´s postulate that there are many opportunities for further yield improvements is crucial for his argumentation, and therefore it is a crucial flaw that this postulate cannot be verified.

P. 98 right: (COMMENT)

Here again, Lomborg criticizes Lester Brown, saying that Brown admits that "corn yields do not yet appear to be leveling off" and mentions that medium-term increases of 20 percent in rice yields are within sight. Comment: Firstly, it is wrong to use the word "admits" in relation to Browns text concerning corn (i.e. maize). Brown has made no claim that all crop yields are close to an upper limit, and when he comes to the subject of maize, he simply states that the levelling-off in yields for this particular crop is only very slight up to now. Concerning rice, Brown mentions that a new prototype of rice that is being developed on the Philippines will make it possible to increase yields by up to 20 %, whereby a considerable increase in annual global rice production will be possible; but this protoptype is fit for tropical and subtropical regions, not for e.g. Japan. Thus, the person who omits something is Lomborg, not Brown. Lomborg forgets to mention that Brown does acknowledge the importance of this new rice variety, and he also forgets to mention that this variety is not fit for all rice producing regions. It may be added that rice yields in Bangladesh are stagnant or declining despite rising input use (Pagiola 1995 cited by Wiebe 2003).

P. 100 bottom left: (COMMENT)

"In other words, there are good reasons to presume . . " Comment: At this point, Lomborg takes a leap from describing how the situation is today, to describing what could be expected in the future. But although we know that there is still scope for increasing crop yields e.g. in India, we cannot logically deduce if this will actually take place in the future. As is evident from Lomborg´s pages 62-63, high-yield food production is heavily dependent on fertilizers, and the price of fertilizers depends on the price of energy. During the period referred to by Lomborg - the latest decades - oil prices remained stable, but by now they are rising, and we may therefore presume that the growth rate of per capita agricultural production will become generally lower.

P. 102- 104 and figure 56: (REMARK)

Lomborg criticises Lester Brown´s book, "Who will feed China". Remark: The criticism is justified. Brown´s projections about China´s future food situation are extremely pessimistic and quite unrealistic. The main flaw here is not Brown´s projections regarding population increase or grain yields, but his prediction that total cropland area in China will be halved within 40-50 years. Brown´s main explanation why this should happen is that housing and industrialisation will take up much of the arable land, a very unlikely explanation. The reader may get a much clearer picture of the flaws of Brown´s book, however, by reading Fan and Agcaoili-Sombilla (1997), which Lomborg cites. - It is correct as stated by Lomborg in figure 56 and on p. 104 that China, contrary to all pessimistic projections, turned into a net exporter of grain about 1999 (with 2004 as an exception). It may be added, however, that imports of soybeans have increased greatly, and the combined situation for grain plus soybeans remains a net import.

P. 105 top left and note 714: FLAW

"Pimentel found in 1974 that the US lost 30 tons of topsoil per hectare, whereas we now know . . " Flaw: Lomborg neglects the change in erosion rates with time. At one locality in the US midwest, the rate of erosion is known to have declined by a factor of 4 between the 1930s and the 1970s (Goudie 1993). Nationwide estimates, made by the USDA, are available from 1982 onwards. The average loss of top soil due to erosion in the USA declined from 17.9 t/ha in 1982 to 13.9 t/ha in 1992 and 12.5 t/ha in 1997. Thus, by extrapolation, it is reasonable that the rate could have been 30 t/ha some time around 1960, and that Pimentel, in a publication from 1974, could refer to this point of time. When Pimentel wrote again on the same subject in an article in Science in 1995, he had adjusted the estimate down to 17 t/ha, citing the 1982 data of USDA. Criticising this, Crosson states that this should have been 13 t/ha, citing the 1992 data of USDA. Thus, the differences are due to different years being cited. This is clear from the paper by Crosson (1995), but Lomborg, when referring to Crosson, leaves this aspect out. In his main text, Lomborg cites only Pimentel´s high early estimate, whereas the lower estimate from a paper published 21 years later, is referred to only in note 714, which few readers will read. So Lomborg contrasts what we "now know" - 12 tons/ha - with the very much earlier figure of 30 t/ha, thereby obtaining the effect that the latter figure appears as a wild exaggeration. As Lomborg knows well from the paper of Crosson that this is not a fair comparison, we may conclude that Lomborg deliberately tries to put Pimentel in an unfavourable light.

P. 105 top left and note 714: ERROR

". . . whereas we now know that the true figure was 12 tons per hectare. " Error: As a source of the estimate of 12 tons per hectare, Lomborg cites Goudie (1993). However, there is no mention of this figure in Goudie pp. 161-162. There is only a mention of Pimentel´s figure of 30 tons per hectare. On p. 163, Goudie cites another figure, viz. a loss of 1.8 mm topsoil per year in Colorado during the last 100 years up to 1979; this corresponds to much more than 18 tons per hectare, and it is stated that this rate is about sixfold the rate during the preceding 300 years. On p. 161, Goudie says: "That soil erosion is a major and serious aspect of the human role in environmental change is not to be doubted."

P. 105 top left and note 714: FLAW OF OMISSION

Lomborg completely leaves out the information that erosion rates in the USA have declined markedly since the 1930s. Flaw: The context is the question: "Should we worry about erosion?". Now, the reason that erosion rates in the USA have declined so much, is that americans have worried about erosion. As we may read in a report published by the USDA (Wiebe 2003), p. 16: "Much of the decline in soil erosion can be attributed to Federal programs, especially conservation compliance provisions . . ". Also, A. Goudie (The human impact on the natural Environment, 5th ed., 2000, p. 200) writes that the 40 % reduction in erosion rates from 1982 to 1997 is a result of conservation efforts. From this, it would be natural to conclude that it has been good to worry about erosion. Instead, Lomborg distorts the whole issue by postulating that early, high estimates of erosion rates were exaggerated, and that we should therefore not worry about erosion.

P. 105 left and notes 717 and 735: (BIAS)

"IFPRI concludes that "the early, high estimates of soil degradation have not been substantiated." Comment: The cited paragraph in the IFPRI publication contains a mixture of optimistic and pessimistic sentences. The optimistic sentence cited by Lomborg in his main text comes first. It is immediately followed by a less optimistic sentence, containing the phrase ". . prices and malnutrition may rise", which Lomborg has not included in his main text, but deferred to a note (note 735) that few will read. And a few lines further down, the IFPRI text says: "The area of degraded soils is extensive, however, and the effects of soil degradation . . are significant." This latter sentence is not cited at all. So, the more pessimistic the sentence, the less is Lomborg willing to cite it.

P. 105 left : (COMMENT)

". . the topsoil layer probably did not grow significantly thinner between the 1930s and the 1980s. .". Comment: This conclusion is dubious. It is based on the fact the depth of the top soil layer was reported in more rounded figures in 1930s than in the 1980s. On this basis, some intractable corrections are carried out, which make the difference between the depth of the topsoil layer in the two periods disappear.

P. 105 top right : FLAW

"The FAO adds that much of the disappearing soil is simply deposited further down . .". Flaw: This is not exactly what is said in the cited FAO source. The text here (p. 357) reads: "It is also possible that yield loss in one area may be compensated by gains further down the slope, valley or plain, where the soil is eventually deposited", and the text proceeds to state that deposition may also have negative external consequences. To say "it is possible" is very different from stating that much of the soil is deposited further down. Furthermore, what remains in the landscape is the least valuable part of the soil. It is chiefly the nutrient-rich organic soil and clay particles that tend to be carried away. See also the comment below.

P. 105 top right : (COMMENT)

" . . has moved only 5 percent of the eroded soil all the way into a river.". Comment: The cited reference (Crosson 1997d) gives two such examples from the USA concerning water erosion, and states that comparable studies of wind erosion have not been done. The same reference also says: "Many studies show that most of the soil eroded from a place on the landscape in any year is stored somewhere else on the landscape for years, decades or even centuries before it is being delivered to the oceans, where it becomes truly "lost" to agriculture". On a global scale, there are conflicting points of view as to how much of the eroded soil ends up in rivers. Concerning the Yellow River in China, for instance, man has been blamed for much of the silt that gives a yellow colour to the river water. This is contradicted by the reference FAO 1995b, which suggests that over 60 % of the erosion here is due to natural processes, i.e. less than 40 % is due to man´s activities. An opposite view is expressed by Goudie (5th ed., 2000, p. 189), namely that the rate of deposition of eroded soil in the sea outside the mouth of the Yellow River during the period of human agriculture has been 10 times the rate in prehistoric times, i.e. 90 % of the erosion must be due to man´s activities. Similar increases in sedimentation have been observed elsewhere on the globe. The impression given by Lomborg, that very little eroded soil ends up in rivers, is unlikely to be universally true.

P. 105 top right: FLAW

"A comprehensive study on China shows . . no net soil degradation." Flaw: Lomborg has found one study that says so. However, most other evidence points to the opposite. For instance, he has read the Chapter on erosion in Goudie (1993), where it is stated that the rate of sediment removal down the Yellow River in China is ten times higher than before man, because of accelerated erosion. Also, there is at present much concern about the excessive wind erosion in north China, leading to large amounts of sand being deposited in the streets of Beijing. Hardly any inhabitant of that city would agree with Lomborg´s misleading statement.

P. 105 right and note 725: (COMMENT)

" The annual drop in productivity is estimated at 0.3 percent per year." Comment: The reader may easily misunderstand this sentence. As explained in Crosson (1997d), the drop in productivity is 12 percent. If this drop has occurred gradually over 40 years, this means that the decline in productivity has been augmented by 0.3 percent points per year. It seems that Lomborg prefers to state the result in such a way that the decline looks very small.

P. 105 right and note 727: ERROR

" . . while only 0.07 percent is strongly degraded.". Error: Lomborg´s figures are from Oldeman et al. (1990, 1991), although note 727 refers to a secondary source (UNEP). The UNEP source quotes Oldeman et al. correctly. The data are that out of all land utilised by man, 749 million hectares are lightly degraded, 910 million hectares moderately degraded, 296 million hectares strongly degraded, and 9 million hectares extremely degraded. Strong + extreme degradation makes out 2.3 % of all land (13 billion hectares). Extreme degradation makes out 0.08 %. Strong degradation is defined as cases where terrain restoration is not possible on a farm basis, but only through major engineering works, whereas extreme degradation means that the land is beyond restoration. All this is clearly stated in the UNEP source. In the Danish version of Lomborg´s book, he says that 0.07 % are "destroyed", which is correct. But by translating this into "strongly degraded" in the English version, an error has arisen. In his next sentence, on the other hand, dealing with agricultural land only, Lomborg uses a correct figure for "strongly" affected areas.

P. 105 right, note 728: FLAW

Note 728 says: ". . other more careful studies have indicated that at least in South and Southeast Asia the actual extent of erosion is less serious . . ". Flaw: As described in Scherr (1999), the South and Southeast Asian study indicated that the percentage of all agricultural land that is affected by erosion is somewhat less than previously estimated, but on the other hand, relative to the earlier study, there was a 20 times greater decline in soil fertility and organic matter, triple the extent of salinisation, and nearly 100 times the extent of waterlogging. This is not what one may understand by "the extent is less serious".

P. 105 right and note 729: FLAW

" In total, this erosion has cost a cumulative loss of 5 percent of agricultural production . . ". Flaw: First, the statement is not quite correct. What is lost is not agricultural production, but overall soil productivity, including that of grazed areas and forests. If the same calculation were made for agricultural land alone, the cumulative loss would be 7.2 %, i.e. agricultural production is affected more strongly than the average of all production. It may be added that the reader may not quite understand what is meant here. The meaning is that the yearly agricultural production now (in 1990) has dropped to a level c. 5 % below the level found 45 years ago (for a given input of farm chemicals etc.). This means that the decline in productivity has been augmented by c. 0.1 percent points per year. Note 729 gives the impression that we have two independent studies which yield similar results. However, all the studies are calculations made from the same data set, that of Oldeman et al. (1990, 1991). Furthermore, the loss of agricultural production seems to have been underestimated in the calculations above, due to a methodological problem. Local experts had been asked to estimate the severity of degradation of that land which is currently (1990) used by man. Therefore, land that was formerly cropped, but has now been abandoned, does not enter into the calculation. According to Scherr (1999), which Lomborg has read, 5-12 million hectares of agricultural land (most likely 5-6 million hectares) is lost per year due to degradation, turned into wasteland (deserts etc.) and abandoned. This means that c. 0.4 % of all agricultural land is abandoned annually, in addition to the 0.1 % annual productivity loss on land that is still cultivated. Therefore the total loss may be several times the rate indicated by Lomborg.

P. 105 bottom right: FLAW

" Compared to this productivity increase, the effect of soil erosion is so small . . ". Flaw: When dealing with the subject of erosion, Lomborg has dealt with the loss of topsoil, and the loss of soil productivity, but he has completely omitted the subject of land that is so severely degraded that agricultural utilisation is totally abandoned. This subject is treated especially by Scherr (1999), which Lomborg has read, and from which most of the following figures have been taken. The world´s current agricultural area is estimated to be c. 1,500 million hectares. This may be compared with the estimate according to UNEP that 2,000 million hectares of land that was once biologically productive has been irreversibly degraded in the past 1,000 years. So it seems that more than half of all land that has been cultivated some time within the latest 1,000 years is now lost to agriculture. Just during the latest 45 years up to 1990, an estimated 250 million hectares have been permanently lost to agriculture through human-induced soil degradation. The total irrigated agricultural area was in 1990 about 250 million hectares. This may be compared with the estimates that during the latest 300 years, 100 million hectares of irrigated land has been destroyed, and another 110 million hectares have been reduced in quality due to salinisation. Just within the latest 45 years, 77 million hectares have become depreciated due to salinisation. Currently, the annual loss of irrigated crop land is estimated to be 0.2 - 1.5 million hectares. Altogether, the total loss of irrigated cropland approaches the total area that is currently irrigated. On the other hand, during the period 1980-1999, agricutural yield per hectare in developing countries has on average increased by 1.8 % per year (Wiebe 2003, se "Main issue"), which, if it continues, will mean a doubling of yields in c. 40 years. Now, Lomborg´s argumentation is essentially that we should not worry about losing agricultural land, as long as the percentage of yearly loss is less than the percentage of yearly increase in yields. In this view, it does not matter if we lose half of all agricultural land in several hundred years, if productivity on the remaining land is doubled within just 40 years. So, even if half of all agricultural land is lost, that will be "no problem". To many of us, this way of thinking is absurd. Furthermore, it presupposes that the yearly increases in crop yields can go on practically for ever. Clearly, Lomborg believes this. It must be pointed out, however, that the reported increases in yield are net increases. From gross increases due to increased input of chemicals and labour has been subtracted the effect of lost soil productivity due to erosion, salinisation etc. If the yearly loss due to erosion is constant, then a slightly smaller increase in gross yields in the future will mean a considerably smaller increase in net yields. Yield increases due to improved technology etc. are believed to be somewhat smaller in the near future (Wiebe 2003), and the paragraphs in the first half of this chapter indicate that the increases may level off completely for some crops, e.g. rice. The more we approach this situation, the more important will it be to lose no further agricultural land. A final point is that an increased dependence on large inputs in agriculture means increased risks of negative effects on adjoining nature areas.

P. 106 left: (COMMENT)

"Moreover, the FAO gives several examples . . ". Comment: It gives two examples.

P. 106 left: FLAW

". . . waiting for higher food prices before rehabilitating land often makes good sense." Flaw: This is not quite the impression given in the cited text. Here, a table lists 24 types of soil degradation. Of these, only 3 can be rapidly and relatively cheaply reversed.

P. 106 left, note 733: (FLAW OF REFERENCE)

The reference Alexandratos 1995 is correct, except that it is not listed as such in the reference list. It is found under the heading "FAO 1995b".

P. 106 left and note 734: (COMMENT)

". . the 3 percent erosion-induced loss is trivial." Comment: The reference is correctly cited. However, it deals with the expected future rate of erosion in USA, after conservation efforts have reduced the erosion rate to possibly 6-7 times less than what it was in the 1930s. If erosion had continued unabated, it might have induced losses of about 20 %, which are not so trivial. So this should not lead to the conclusion that it is not worthwhile to combat erosion.

P. 106 left: WRONG STATEMENT

". . . the present evidence does not not seem to indicate that soil erosion will to any significant degree affect our global food production . . " Error: Lomborg is not right in drawing this conclusion. Let us assume that yields per hectare increase by 1.5 % per year, and that the loss of cultivated area due to erosion amounts to 0.4 % per year . In 50 years, these figures will cause yields per hectare to increase by an extra 110 % , but the number of hectares will have decreased by 18 % relative to a no-erosion scenario, which means that the world total net increase will be only c. 70 %. This is certainly a "significant" effect of erosion. Furthermore, Lomborg´s conclusion is contrary to what is concluded in the more thorough treatment of the subject in Wiebe 2003 (see "Main issue"). Wiebe reports the net increase in agricultural land area for all of the developing world, and compares this to an estimate of the land area that has been lost to agriculture during the same time period. His conclusion is that out of the gross increase in agricultural area, at least one third is counterbalanced by losses elsewhere. This is hardly compatible with Lomborg´s notion that erosion does not to any significant degree affect food production. Based on historic data from 67 developing countries, and projections concerning population growth, prices, crop yields, expansions of agricultural areas etc., Wiebe runs a "baseline model" that calculates how the number of hungry people (people who cannot meet their nutritional requirements) will change from 2000 to 2010. Next, he runs a "no-erosion" model, in which the gross increases in agricultural area are not reduced by erosion losses. The result is that the number of hungry people in the world by 2010 will then not be 694 millions, as in the baseline model, but "only" 626 millions, i.e. a reduction of 68 millions. That is, without erosion the number of hungry people will be reduced by 10 % in just 10 years. This is certainly a "siginificant" effect.

P. 106 bottom left and note 736: FLAW OF REFERENCE

" . . . for the first time in history, they can no longer count on fishing fleets to help them expand the food supply. ". Flaw: The sentence cited here, starting with "for the first time", is not to be found on p. 34 of the cited reference, but rather on the first page of the reference´s chapter 2. The rest of the quoted text is not to be found. This flaw dates back to the Danish version of the book and seems to be due to haste in writing.

P. 106 right: (COMMENT)

". . oceanic fisheries have collapsed not just off the coast of developing countries . . . but also off those of industrial countries." Comment: Lomborg quotes Brown in a way which could make the reader believe that Brown exaggerates wildly and thus is not to be trusted. Actually, however, Brown´s text covers reality quite well. The world´s largest fishery off the coast of Peru collapsed in 1972 and plummeted to a mere tenth, with further decreases a few years later. Spring herring off the coast of Norway was the largest herring stock worldwide. After 1967, the introduction of the purse seine caused the stock to collapse. Stocks of the North Sea herring had collapsed from 1965 onwards, and there was a total ban on herring fisheries here during 1975-1978. The stock of North Sea mackerel plummeted in 1970, and hs never recovered. And there are many more examples of collapses: Capelin in the Barents Sea: collapse in 1985. Cod off the coast of Canada: total stop to fishing in 1992. Haddock at Georges bank off the US east coast; 1965 and later again. Herring, same location: total collapse in 1977. In 1994 parts of Georges Bank were closed permanently to fisheries. Black Sea: All fisheries collapsed around 1989. Also fisheries in the East China Sea and the Yellow Sea have been reduced to one-tenth of what they used to be. So, Brown´s statement is not at all exaggerated.

P. 106 right: FLAW

". . and it can be expected that these problems will lead to `mounting social disruption, economic pressures, and the threat of violence." Flaw: This refers to a chapter by Anne Platt McGinn from the Worldwatch Institute. However, it is not correct that she expects these problems. Instead, she reports that these things actually happened in Canada in 1997, at an episode that lasted for three days. She then states that as a symptom of overfishing, such incidences are becoming more widespread - which seems to be true.

P. 106 right and note 741: ERROR

" . . is down by 7.5 percent since 1988 . . ". Error: The reference is erroneous. There is no such information on p. 32 in the cited book. There is, however, a text on p. 24, stating that the catch per capita has started to decline and has been reduced by about 9 % since 1988. The error dates back to the Danish version of the book and seems to be partially due to haste in writing.

P. 106 right: FLAW OF OMISSION

"But as is often the case, this does not tell the whole story". Flaw: Actually, Worldwatch Institute does tell the whole story, or at least more than Lomborg does. In their 1995 yearbook (which Lomborg has studied), they have a chapter on fish production, which has essentially all the information that Lomborg also has, including data on fish farming. But relative to Lomborg´s text, their text is both more informative and more balanced, and includes information on the negative effects of fish and shrimp farming, an aspect which Lomborg completely leaves out. Again in their 1998 yearbook, they have a whole chapter on sustainable fishery. The quote about "mounting social disruption" etc. is taken from the introduction to that chapter, obviously with the purpose of derogating the author. The rest of that chapter includes the production from fish farming, and besides that, has much more information e.g. on regulation than Lomborg has. Thus, Lomborg is here very selective, quotes a few sentences which might serve to derogate Worldwatch Institute, and otherwise avoids telling that WI has brought much more information, in a more balanced way, than Lomborg does. In a telephone conversation with me (KF) in 1999, Lomborg admitted that he knew nothing about fisheries, and it seems that it is mainly the WI 1995 yearbook that allowed him to write about the subject. All in all, it seems evident that his text is deliberately manipulated in order to derogate Worldwatch Institute.

P. 106 right: ERROR

"The global fish catch has in the 1990s not increased as much as earlier, as is evident in Figure 57." Error: What Lomborg does not know, is that China, by over-reporting its marine fisheries landings to FAO, generated increasing figures for world landings during the 1990s, when in reality global landings declined, and global fish production including fish farming stagnated (Watson & Pauly (2001): Nature 414: 534-536).

P. 107 left: FLAW

"It is possible to show that there is an optimal level of fishing in between these two extremes." Flaw: Lomborg fails to distinguish between the biological optimum and the economic optimum, which is remarkable, because in other contexts, he relies much more on economic arguments than on biological contexts. In worldwide marine fisheries, in terms of yields, we have not surpassed the biological optimum very much - we are catching about 90 mio. tons annually, where the optimum is about 100 mio. tons. The economic optimum, on the other hand, is widely surpassed. As early as 1989, the global spending on fisheries, including interest and repayments, was about 124,000 mio. US dollars, which widely surpassed the economic gross returns of c. 70,000 mio. US dollars. Thus, fisheries in 1989 caused an annual loss of 54,000 US dollars, and still, this activity continued. If Lomborg had included the economic aspects - which he usually does in other fields - the economic system´s lack of ability to regulate the resource would have stood out much more clearly.

P. 107 right: FLAW

". . the state can (through permits, for example) make sure only the optimal amount of fish is caught." Flaw: It is a widespread conception that a combination of private property rights, e.g. ownership of a territory, and liberal economy make for the best possible use of available resources. In fisheries, such a view is a blatant misconception. Firstly, a state cannot just "make sure". The regulation of fisheries is an extremely difficult exercise, which in the first step requires that very extensive knowledge is provided, and in the second step, great political skill. Lomborg´s nearly complete neglection of these very demanding processes reveals that he understands very little of the field that he writes about here. For instance, he does not discuss the question of how large permits should be. Secondly, a liberal economic system is hardly able to regulate this kind of activity, as stated above for p. 107 left. Necessary tools are either the explicit and exclusive transfer of responsibilities to small local communities, or quota, control, fines etc. A main cause for the failure to regulate fisheries in an economically sensible way is precisely the reluctance of many politicians to intervene effectively by sufficient official regulation.

P. 108 left: FLAW

""Of course, we would love to be able to get hold of that extra 10 million tons, but this in no way decides the outcome of food availability. . . " Flaw: Lomborg, and with him many others that are not fisheries biologists, has not understood why the overfishing makes biologists worry. The point is that as overfishing proceeds, the landings decline only slightly at the same time as the stocks decline drastically. For a long time, this will go on without serious problems, because the remaining stock becomes more and more productive. But at a sudden point, the biological limit of the fish stock is exceeded, it cannot increase its productivity any more, and it suddenly collapses. It is like walking in the dark in flat country in the direction of a steep cliff, and you never know if the next step will bring you out over the edge. So, the concern is not with the size of the landings, as Lomborg supposes, but with the safety of the stocks. Once the collapse has occurred, the loss is typically not just 10 % relative to maximum yield, but may be 90 % or 100 %. As an example explaining the risk, cf. e.g. Cook et al. (1997): Nature 385: 521-522.

P. 108 left: FLAW

"When Lester Brown generally finds that we get less fish per capita it is because he does not include fish farm production. This decision seems somewhat odd . . " Flaw: Should Brown have included firsh farm production ? No, because the issue, as stated on p. 106 bottom left, is whether the fishing fleets can help expand the food supplied by agriculture. The capture of wild marine fish is a way in which proteins in the oceans are directed into human food production, and this harvest of marine proteins cannot expand. Fish in fish farms, on the other hand, are fed with protein from some source. When it comes to farms of marine fish,e.g. salmon, these are fed with fish meal and fish oils, which originate from fish caught in the sea. If you add the amount of farmed salmon to the amount of captured marine fish, then you count the same proteins twice. Thus, Lomborg´s total sum is too high because some items are counted twice. When it comes to fresh water carp farmed in China, then these are fed with protein produced in agriculture. The same protein might as well have been fed e.g. to chicken or ducks, but in the expanding Chinese fish production, the choice has been to feed it to fish. So, this procuction should not be included in the sum of what we can harvest from seas and lakes.

P. 108 bottom left: (COMMENT)

" . . Fish farm production will exceed the catch from tradtional fisheries." Comment: This statement seems mainly based on an extrapolation of present trends into the future. Such an extrapolation disregards the underlying causes. The production of carp fish in China (18 mio. tons out of the 23 mio. tons from global fresh water aquaculture in 1996) has grown in proportion to the growth in Chinese agriculture, and is a byproduct of that agriculture. As the scope for further growth in Chinese agriculture is modest, so is the scope for increase in Chinese aquaculture. There is, however, a growing aquaculture in other tropical counntries, e.g. India; but up to now, nobody knows how much this will expand.

P. 108 right AND NOTE 756: ERROR

" . . the United States, France, and China. All three countries went on to experience annual yield growth rates between 2.3 percent and 5 percent." Error: The reference in note 756 is a web page which seems not to exist anymore and thus cannot be checked. However, the postulate of 2.3 - 5 % growth does not agree with the data on yields in wheat and rice, as explained.in the notes to pp. 96-98. Also, a quick check on the increase in a wider range of food products disproves the postulate. We may add together e.g. the production of cereals, vegetables, roots and tubers in the FAO database (http://apps.fao.org/), and divide the combined production by the area of arable land. For the three countries, viz. USA, France and China, we find that from 1982 to 2001, production per area has grown annually by 1.07 %, 1.22 %, and 0.86 %, respectively. All three percentages have dropped considerably from the previous twenty-year period 1962 to 1982. This is corroborated when we inspect trends in yields for cereals, vegetables or roots seperately. Thus, although the growth has not stopped, growth rates are closer to zero than to the values postulated by Lomborg.

P. 108 bottom right: (COMMENT)

"There is no "wall" for maximum yields in sight". Comment: As stated in the comments to pp. 96-98, this is not true.

P. 109 left: (COMMENT)

"It is expected that there will be fewer malnourished people". Comment: According to the recent FAO report: "The state of the food insecurity in the world 2003", the decrease in the total number of starving people in the developing world seems to have stopped or even been reversed to a slight increase. It has gone from about 780 mio. around 1996, up to about 798 mio. around 2000. See comments to fig. 7 in chapter 1.