Lomborg-errors: "Cool it!"

Poor, but from global warming or bad policies?
 
    Home                                                                                                                                   Cool it

  Comments to pages 157 - 164 in Cool it.

THE NUMBER OF POOR

FLAW
Page 159 top: ". . in essence have started at almost 50 % poor 30 years later in 1980". Flaw: As explained below concerning the right graph of Figure 48, this is not so. The level was about 40 % poor in 1981.

(COMMENT)

Page 159 and note 850:  ". . current opinion considers the population component of SRES A2 ´out of the range´."  Comment: This is true. Also Stern agrees with that in his review.

ERROR
Page 159 and note 851: " . . but it can only be considered an absolute worst-case scenario."
Error: This is not so. There have been constructed a number of scenarios of global GDP growth other than those of the SRES (IPCC Special Report on Emissions Scenarios). When the four SRES scenarios A1, A2, B1 and B2 are compared to what other scenarios have been published, all four lie within the range covered by the scenario literature. A1 is very high though, and lies for some part of the 21st century at the absolute upper boundary of the range. B1 is close to the average, and A2 and B2 are somewhat below average, but far from the absolute bottom of the range. Source: Nakicenovic et al. 2003.
In addition, the quote from Fischer et al. (2005) is wrong. What they say is that their own projections of crop growth based on the A2 scenario are representative of a worst case scenario, not that A2 itself is so. Furthermore,  Fischer et al. (2005, page 2072) point out that the other three SRES scenarios have been criticized for economic growth rates that are probably too strong, wherefore they have also simulated lower economic growth scenarios. Lomborg´s formulation, using the words "only", "absolute" and "worst-case", is so categorical that it is downright wrong.

FLAW
Page 159, Figure 48, left graph: " The proportion of poor in the world, 1950-2015." Flaw: The data from Berry et al. 1983 and 1991 (the two sources have identical data) give the number of people with less than $200 (presumably per year). This amount is given in 1970 U.S.$. To compare it with later Worldbank estimates, it should be adjusted to 1993 as the base year (defined as lying 8 % above the previous base year 1985), which means multiplying by a factor of 2.77 x 1.08 using the consumer price index, or 2.53 x 1.08, using the GDP deflator. This means that the poverty line of Berry et al. corresponds to approximately $1.5 - 1.6 per day, not $ 1 per day. However, the data from the Worldbank on number of poor below $1 per day, covering later years, do not fit together with Berry´s data, being much higher, not lower as expected. The reason for the discrepancy might be explained by the statement in Berry et al. 1983: "In developing countries, lack of adequate data is a major problem".  Methodologically uniform data exist for the period 1981 onwards (Worldbank working papers by Chen & Ravallion 2004 and Chen & Ravallion 2008).  They imply that the curve that Lomborg attempts to draw should have been steeper in the first part (40 % below $1 per day in  1981). What Lomborg has done, is to splice two data sets that do not fit together. A person teaching statistics should not do that.

FLAW
Page 159, Figure 48, right graph: " The proportion of poor . . . for south Asia and Sub-Saharan Africa, 1981-2100." Flaw: Lomborg´s curve starts at 1981, but the references given by him do not have data from 1981. However, data from 1981 onwards exist elsewhere (Worldbank working paper by Chen & Ravallion 2004 and more recently Chen & Ravallion 2008). The figures from 1990 onwards differ from those used by Lomborg. In the recent data, the number of persons living below $1 per day goes from 42 % in 1981 to 32 % in 2002. Thus, the actual curve has a much flatter slope than the one given by Lomborg. The number of poor in these regions does not drop as fast as he indicates.
These are percentage figures. The absolute number of people living with less than $1 per day has decreased only slightly in south Asia from 1981 to 2002, and in Subsaharan Africa, the number has nearly doubled during this period (from 169 million to 310 million). Only after 2002 is there probably a small decline in the African figures. It is conceivable that the absolute number of poor will decline faster in the future in south Asia and eventually go towards zero. It is less conceivable that the same will happen in Africa. Lomborg´s graph would fit with a situation where the absolute number of poor in south Asia and Subsaharan Africa has dropped to just above one fourth of the number in 2002. This is a rather optimistic scenario, considering that this drop has hardly started yet.

FLAW
Page 160 top: " . . . the World Bank explicitly says that $2 -a-day reflects "poverty lines more commonly used in lower-middle-income countries."" Flaw:  The quoted sentence could not be found in Lomborg´s source. Instead, World Bank experts Chen & Ravallion state in a later publication (2008) that "a $2.00 line is the median poverty line found amongst developing countries as a whole."

FLAW
Page 160 top: " Finally, Stern chose to assume that the world will warm extra much . . " Flaw:  This is a misleading formulation.  Stern operated with two versions, one being the A2 baseline scenario, and the other being the same, with extra heating due to possible feedback effects. Lomborg´s text should have read: "Finally, Stern chose to include an additional version in which the world will warm extra much." Without such a formulation, the text is misleading.

(COMMENT)
Page 160:  "After that it still only slowly starts declining . ." Comment: Lomborg´s graph for the so-called worst-case world does not seem to fit with the parameter values on which it is based. See the comment below to Figure 49.

(COMMENT)
Page 160: " With a high degree of warming, it is likely that . . . ". Comment:  It seems that Lomborg mixes up things here. It is not clear if his graph originating from the A2 scenario includes the negative effects of climate change on the economy. What Anderson (2006) has done, is to project the development first with the A2 scenario, and then with the same scenario but now including the effect of climate on the GDP growth rate. He does that with two versions, one with a 3.9°  C temperature rise by 2100, and one with extra warming, i. a 4.3° C temperature rise by 2100. But Lomborg´s presentation does not tell whether the grey graph labelled "worst-case world" includes any climate effect or not. He only tells that extra warming (from 3.9° to 4.3° C) shifts the graph from the black curve to the white curve.

ERROR
Page 160. "By 2100, global warming will cause the world to have 51 million more people below $2-a-day." Error: At this point  in the text, Lomborg has finally mixed up things completely. The figure of 51 million seems to have been computed by Lomborg (not by Anderson), and the reader will understand that it is the difference between the grey curve and the white curve in 2100. The text on the graph says "with high warming" for the white graph, but the main text says that the difference is due to "global warming". In any case, the figures computed by Anderson (2006), citing the mean values, are that global warming will cause an extra 12 million poor (3.9° warming) og 17 million poor (4.3° warming). Thus the difference between the two extents of warming is just 5 million, and the difference between no climate change and expected climate change is 12 - 17 million poor. None of these mean estimates fit with Lomborg´s graphs, irrespectively of how these graphs are interpreted.

(COMMENT)

Page 160:  "If Stern . . also tweaks the risk of things going very wrong. . ."  Comment: To say that Stern tweaks the risk is an  unkind way of describing his text. The issue is described in Chapter 4 of the Stern review, on page 109-110, and here Stern presents the mean values and the 95th % percentile values for two climate models. He does not present the 5th % percentile values. The graphs that he presents give the mean values. The text is better than what Lomborg usually presents, because Lomborg usually gives just one figure, with no indication whatsoever about the uncertainty range around the figure, even when the uncertainty is extremely large. Presenting the mean and the upper range (95th percentile) is at least somewhat better, because it gives an idea of the size of the uncertainty. However, Stern´s  text is not quite neutral when the 5th percentile is not also given. And Stern does put rather much emphasis in his text on the unlikely 95th percentile values. The reason for this is, of course, that society should be prepared for the worst, not just for the mean, but it brings the text closer to scaremongering.

ERROR
Page 160, Figure 49. "Millions of poor in south Asia and Sub-Saharan Africa, 2002-2100, with Stern´s worst-case scenario. . ." Error: Although the figure is based on Lomborg´s own calculations and the graphs are therefore not easily checked, a sort of check is possible. With the parameters given in Anderson (2006), the income-per-capita doubles in Africa and south Asia from 2002 to 2050. Assuming that the relative income distribution is constant (which is reasonable), one may therefore assume that the proportion (32 %) living with less than $1 per day in 2002 is the same as the proportion living with less than $2 in 2050. As we have also figures for the population increase (+ 86 % from 2002 to 2050), we can calculate the total number of people living with less than $2 in 2050. The result is about 1270 millions. Similar calculations based on the proportion today living below $1.25 per day tell that there will be about 1470 millions living below $2 in 2033. These figures are for a scenario without economic growth reduction caused by climate change. If the mean estimates in Anderson (2006) of the number of poor extra because of climate change are added, then that means adding about 1 million extra in 2033 and 2 million extra in 2050, i.e. the graph is essentially not changed in this part of the time period. Thus the figures are much lower than in Lomborg´s curve for the "worst-case world", and we may therefore assume that Lomborg has dramatized the course of the graph.

FLAW
Page 161: " Perhaps even more important is to look at the other, more likely, policy option, namely going down a different - and more likely - development path, the A1" Flaw:  Lomborg misunderstands or misrepresents what the socio-economic scenarios - A1 and A2 - are. They are not policy options; they are not predictions; and there is no likelihood associated with them. IPCC explicitly states that they are equally valid. They are just scenarios to be used for testing the vulnerability to climate change in different versions of the future world. A1 is constructed to show a world where the gap between rich and poor countries closes, whereas A2 shows a world where the Third World countries do not manage fully to catch up with the rest of the world. Many would consider the latter more likely. Furthermore, A1 presupposes that it will be possible in the future to sustain high economic growth rates continuously to obtain ever larger income levels, not only in the Western World and Russia, but also the Third World countries will obtain such high income levels. It is doubtful if there are resources enough on the globe to make this possible.

(COMMENT)

Page 160:  "The World Bank´s expectation for poverty in south Asia and Sub-Saharan Africa is spot on A1 . . "  Comment: Yes, of course. The World Bank´s projection for 2015 is based on the following premises (World Bank 2004, p. 21): "Developing country economic performance has been strong since 2002, and this is projected to continue over the next two years and beyond." So IF economic performance remains strong, then we will get the same result as in A1, which is built on the assumption that there will be a strong economic growth. No surprise - that is a tautology. But nobody knows IF economic growth will remain strong. When Lomborg wrote his text, nobody knew that a global financial crisis was waiting just around the corner.

FLAW
Page 161: " . . .  again indicating that Stern has picked the wrong scenario. " Flaw:  There is no such thing as a wrong scenario. All scenarios are equally valid, as explained above.
The reason that Stern gives why he chose to run his PAGE model on the basis of the A2 scenario, is that this produces a likely mean estimate of the global temperature increase in the year 2100 - the increase is 3.9° C, which compares well with the mean of the temperature projection in the IPCC Third Assessment Report which is 4.1° C (Stern report chapter 6, page 154).

FLAW
Page 161: " . . .  this future is also much more robust toward climate change - you virtually can´t see the different climate line in Figure 49 . . .  " Flaw:  The "different climate line" (presumably?) gives the difference between 3.9° warming and 4.3° warming. It does not (?) give the difference between climate change and no climate change.

(COMMENT)

Page 161:  "Again, we are at a point where we have to realize that we have several opportunities . . "  Comment: This is a misunderstanding. The scenarios are not opportunities. We do not choose ourselves if we want a global financial crisis or not, or a severe drought that sets back economic growth in Africa, or if we want a per capita growth of income in Africa of 2 % or 4 %. These things just happen, and the scenarios present some possibilities for what might happen. If we indeed had opportunities and could choose, who would choose 2 % growth rather than 4 % growth?

FLAW
Page 161 bottom: " . . .  yet choosing A1 instead will drop that number to 320 million . . .  " Flaw:  As stated repeatedly in the comments above, the scenarios are not something that you can choose.

FLAW
Page 162 top: " If we manage to move from A2 to A1, we will on average reduce the number of poor people by a billion. . . "  Flaw:  As stated above, the graphs (at least the upper graph) in Figure 49 are probably not right, and so the conclusion that the average difference between the graphs is one billion people is dubious.

FLAW
Page 162, Figure 50:  Flaw:  The source for the postulate that Kyoto will give just 1 million fewer people is missing, and it is not stated if "Kyoto" just means the present Kyoto agreement and nothing more. It is wrong to say that there is a choice between A2 and A1 - scenarios are not opportunities to choose or not choose. It is also wrong to say that the difference between A2 and A1 is in essence a difference of economic policy; a main difference is that whereas in A1 the world succeeds to reduce population growth quite fast, in A2 population growth in The Third World remains high for a long time. It is no wonder that if the world population is higher (A2), then the number of poor people is also higher. The figure of 984 million is much too precise, given that the curves in Figure 49 are very uncertain and probably wrong; indicating the figure with three digits implies a very false degree of precision.

THE SRES SCENARIOS

GENERAL COMMENTS:
The IPCC produced a Special Report on Emission Scenarios (Nakicenovic et al. 2000), abbreviated SRES; it may be downloaded here. The SRES report formulates 40 socio-economic scenarios for the world up to the year 2100. They are divided into four main types, called A1, A2, B1 and B2. Lomborg presents these four in his Figure 51. However, he greatly misunderstands or misrepresents what the scenarios signify. This is discussed in detail on this page.
 
(COMMENT)
Page 163, Figure 51: The size of the A1 pie is given as $3005T.
Comment:  It is stated in the SRES report that the size of the world econonomy in the A1 world is $529 trillion per year in 2100. It may be calculated that the total worth of the income over the century would be around $12,000 trillion. Why then does Lomborg tell that the total worth of income over the century is estimated at $3,005 trillion? This is not evident to the reader, nor clearly understandable from note 860. The explanation must be that the future sums have been discounted back to present values. The rate of discount used to do this is not given, but according to note 860, Lomborg has used the same dynamic discounting as Nordhaus (2006), i.e. presumably a discount rate that declines with time.
The sizes of the slices marked with black in the other pies depend on the rate of discount.

ERROR
Page 164 top: " Again, this should be seen in the light of a total cost of global warming   . . "  Error:  No it should not. The black slices in the pies have nothing to do with climate mitigation.

FLAW
Page 164 top: " Again, this should be seen in the light of a total cost of global warming at about $14.5 trillion  . . "  Flaw:  There is no source for this figure, nor any indication of the uncertainty around the figure. Of course, considering how difficult, if not impossible, it is to know what will be the total costs all over the globe up to the year 2100 of global warming, it is ridiculuous to  postulate that one knows what the cost will be with an accuracy of three significant digits and no specification of what scenario it refers to. If the figure is meant to be of relevance when considering the SRES scenarios, then it must have been obtained by calculations that are relevant in relation to these scenarios. This seems not to be the case.

FLAW
Page 164 top: " . . . or a cost of all the other environmental policies . . . of about $60 trillion. . . "  Flaw:  The black slices in the pies in Figure 51 might signify in part the costs of following environmentally-friendly paths, and thus these black slices are what should be compared to Lomborg´s alleged cost of $60 trillion. Obviously the figures do not fit together, which proves that the black slices actually do not represent the cost of any environmental policy.

ERROR
Page 164 bottom: " . . . we can increase wealth by just $0.6 trillion . . "  Error:  Lomborg tries to tell us that the damages from climate change are a thousand times smaller than the costs of fighting climate change. This is very wrong. The figure of $0.6 trillion is not obtained in relation to the SRES scenarios and is thus not comparable to costs calculated from these. Actually the cost of mitigation by 2050 is in the size order of 0.1 - 1.8 % of global GDP (see above), which will probably keep the temperature increase at 2 - 3° C and prevent the most serious damages from climate change in the years after that. The prevented damages might very well be in the range of  1 % of GDP, as an average of several estimates, wherefore it seems that cost and benefits approximately are at a ratio of 1:1. This is very different from Lomborg´s postulate of 1:1000, which illustrates how far out Lomborg´s postulates in this chapter are.