/
2018-02-13-dangerous-assumptions-revisited.Rmd
516 lines (436 loc) · 24 KB
/
2018-02-13-dangerous-assumptions-revisited.Rmd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
---
title: "Dangerous Assumptions Revisited"
author: "Jonathan Gilligan"
date: '2018-02-13'
keywords:
- climate change
- climate policy
- global warming
- RCPs
- representative concentration pathways
- CO2 emissions
- decarbonization
- Roger Pielke, Jr.
- greenhouse gas emissions
slug: dangerous-assumptions-revisited
tags:
- decarbonization
- RCPs
categories:
- blog posts
---
```{r setup, include=FALSE, cache=FALSE, message=FALSE, warning=FALSE}
knitr::opts_chunk$set(echo = FALSE,
cache = FALSE,
message=FALSE, warning=FALSE, error=FALSE)
if (! require(pacman)) {
stop("Before proceeding, you must install the pacman package.\n",
"In R, run \"install.packages('pacman')\".")
}
p_load(tidyverse)
p_load(magrittr)
p_load(readxl)
p_load(ggrepel)
p_load(janitor)
p_load(knitr)
p_load_gh("gilligan-ees-3310/kayadata")
message("Dep path = ", blogdown::dep_path())
message("fig.path = ", knitr::opts_chunk$get("fig.path"))
```
Comparisons of observed trends of energy and carbon intensity in the global
economy to trends implied by emissions scenarios used in policy analysis
suggested that those scenarios were severely over-optimistic about the rate at
which the world would spontaneously decarbonize its economy.
I update these analysis, using global emissions since 2005, and find
that observed rates of decarbonization are not far behind those implied by
the RCP 4.5 policy scenario. This suggests that the policy challenge may not
be as difficult as previous work has reported.
<!-- more -->
# Introduction
## Background and the Kaya Identity
In 2000, the Intergovernmental Panel on Climate Change published a
[Special Report on Emissions Scenarios](http://www.ipcc.ch/ipccreports/sres/emission/){target="_blank"},
in which it developed many possible
scenarios for the paths that greenhouse gas emissions might follow during the
21st century under "Business as Usual" (BAU) conditions, meaning without any
policy measures to reduce emissions.
Different scenarios made different assumptions about population growth,
economic development, and technological innovation and thus produced a diverse
range of emissions pathways. The hope was that the actual emissions path that
the world would follow in the 21st century would lie somewhere within this
range.
From a policy perspective, it is useful to analyze emissions in terms of the
[Kaya Identity](https://en.wikipedia.org/wiki/Kaya_identity){target="_blank"}:
$$
\text{Emissions} = \text{Population} \times
\frac{\text{GDP}}{\text{Population}} \times
\frac{\text{Energy}}{\text{GDP}} \times
\frac{\text{Emissions}}{\text{Energy}}.
$$
There is nothing magic about this: If you cancel terms that appear on both the
top and bottom of the fractions, it says nothing more than
$\text{Emissions} = \text{Emissions}$.
However, the identity is useful for thinking about policy because each of the
terms multiplied together is something that has direct relevance, so let's
assign each a symbol:
$$
\begin{aligned}
g &= \frac{\text{GDP}}{\text{Population}} = \text{per-capita GDP}\\
e &= \frac{\text{Energy}}{\text{GDP}} = \text{energy intensity of the economy}\\
f &= \frac{\text{Emissions}}{\text{Energy}} = \text{carbon intensity of the energy supply}.
\end{aligned}
$$
Now, we can write the identity as
$$
F = P \times g \times e \times f,
$$
where $F$ stands for total emissions, $P$ stands for population, and
$g$, $e$, and $f$ are defined by the equations above.
Roger Pielke, Jr.'s book, [_The Climate Fix_](https://www.amazon.com/Climate-Fix-Scientists-Politicians-Warming/dp/0465025196/){target="_blank"} has a very clear explanation of this
in Chapter 3.
This way of looking at emissions is useful for policymakers because trying to
directly control population is both morally questionable (I would go much
farther and say that I would find attempts by governments or others to forcibly
control population to be a horrifying violation of people's basic rights, and
especially women's rights to autonomy over their bodies).
It would also be politically unfeasible and morally questionable at best to
try to reduce per-capita GDP, which essentially would mean making people
poorer on average.
From a policy perspective, it is not feasible nor desirable to try to
reduce people's wealth or coerce their reproductive choices.
Thus, climate policy must focus on the last two terms:
Improving energy efficiency can reduce the energy intensity of the economy
and switching from carbon-intensive energy sources, such as coal, to less
carbon-intensive sources, such as nuclear or renewable energy, can reduce the
carbon-intensity of the energy supply.
Reducing $e$ (energy intensity) and $f$ (carbon intensity) can reduce emissions
without reducing either population or per-capita GDP.
Thus, feasible climate policy focuses on ways to reduce energy and carbon
intensity.
## Comparing Observed and Implied Rates of Decarbonization
In 2008, Roger Pielke, Jr., Tom Wigley, and Christopher Green
("[Dangerous Assumptions](https://pdfs.semanticscholar.org/afdc/e9bfa0fb51768394407d086cdeb5e9f77207.pdf){target="_blank"},"
_Nature_ **452**, 531 (2008)) compared the observed rate of change of energy
and carbon intensity in the years 2000--2005 to the rates of change implied
by the SRES scenarios. They observed that most of the scenarios assumed that
even without any specific climate policies, the world's economy would steadily
become more energy efficient and less carbon-intensive. However, they pointed
out, between 2000 and 2005 the opposite happened: the world's economy became
both _more_ energy intensive and _more_ carbon intensive.
![Assumed versus observed decarbonization rates, from Pielke _et al._, "[Dangerous Assumptions](https://pdfs.semanticscholar.org/afdc/e9bfa0fb51768394407d086cdeb5e9f77207.pdf){target="_blank"}"](/img/post/gilligan-dangerous-assumptions/pielke_assumed_decarbonization.jpg)
If there is a fairly rapid rate of "spontaneous decarbonization" in the absence
of climate policy, then making climate policy is relatively easy, because it
only has to make modest changes to something that's happening already:
![Under the IPCC's SRES scenarios, spontaneous decarbonization (blue bars) does most of the hard work and climate policy only needs to add a small amount of additional decarbonization (red bars). However, if spontaneous decarbonization falls short of the predicted blue bars, then the policy task is harder. From Pielke _et al._, "[Dangerous Assumptions](https://pdfs.semanticscholar.org/afdc/e9bfa0fb51768394407d086cdeb5e9f77207.pdf){target="_blank"}"](/img/post/gilligan-dangerous-assumptions/pielke_cumulative_emissions.jpg)
However, Pielke, Wigley, and Green found that the spontaneous decarbonization
rate was much smaller than the IPCC expected, and emphasized that this implied
that the policy challenge was much more difficult than expected.
More recently, Samantha Stevenson and Roger Pielke, Jr.
("[Assumptions of Spontaneous Decarbonization in the IPCC AR5 Baseline Scenarios](https://pdfs.semanticscholar.org/0aef/69f0117b6f4e1d2926e1acaa6db7af963924.pdf){target="_blank"}")
conducted a similar analysis of the
"[Representative Concentration Pathways](https://skepticalscience.com/rcp.php?t=1){target="_blank"}"
(RCPs),
which were published in 2011.
Unlike the SRES scenarios, the RCPs are policy-active pathways in which
government action, such as imposing a price on greenhouse gas emissions,
is used to reduce emissions.
In their analysis, Stevenson and Pielke compared the observed global emissions
between 2000 and 2008 to the baseline pathways corresponding to the RCPs
(the baseline pathways are the trajectories of greenhouse gas emissions the
RCP authors assumed would occur if there were no climate policies in effect).
Stevenson and Pielke found that, just as with the SRES scenarios,
the baseline emissions trajectories assumed far greater spontaneous decarbonization
than actually occurred.
## Problems with the Previous Analyses
An important problem that the analyses by Pielke, Wigley, & Green and
Stevenson & Pielke face is that they compared the scenarios to observed trends
during the years 2000--2005 and 2000--2008, respectively.
This may not have been the best choice.
World energy and carbon intensity steadily declined from 1980--2000 and then
suddenly jumped for several years starting around 2002, which corresponded to
a period of tremendous economic growth in China.
```{r world-carbon-plot, include = TRUE, fig.height=7, fig.width=7, fig.cap = "Trend of global carbon intensity, with 2000--2008 highlighted."}
get_kaya_data("World") %>% filter(year >= 1980) %>%
plot_kaya("f", y_lab = expression(paste(MMT * CO[2] / quad)), log_scale = TRUE,
start_year = 2000, stop_year = 2008) +
labs(title = "Carbon intensity of world energy supply") +
scale_y_log10(breaks = seq(62, 78, 4)) +
theme_bw(base_size = 15)
```
```{r world-energy-plot, include = TRUE, fig.height=7, fig.width=7, fig.cap = "Trend of global energy intensity, with 2000--2008 highlighted."}
get_kaya_data("World") %>% filter(year >= 1980) %>%
plot_kaya("e", y_lab = "quad / $ trillion", log_scale = TRUE,
start_year = 2000, stop_year = 2008) +
labs(title = "Energy intensity of world economy") +
scale_y_log10(breaks = seq(6, 11, 2)) +
theme_bw(base_size = 15)
```
China's growth was fueled by a rapid expansion of energy consumption, and both
factories and generating plants were very inefficient, and were largely fueled
by coal. This drove both the energy intensity of China's economy and the carbon
intensity of its energy supply through the roof, and this appears to have played
a large role in reversing for several years the global trend of declining
energy and carbon intensity.
Around 2007 or so, China's energy and carbon intensity began to fall again and
the world energy and carbon intensity fell along with it.
```{r china-carbon-plot, include = TRUE, fig.height=7, fig.width=7, fig.cap = "Trend of China's carbon intensity, with 2000--2008 highlighted."}
get_kaya_data("China") %>% filter(year >= 1980) %>%
plot_kaya("f", y_lab = expression(paste(MMT * CO[2] / quad)), log_scale = TRUE,
start_year = 2000, stop_year = 2008) +
labs(title = "Carbon intensity of China's energy supply") +
scale_y_log10(breaks = seq(75, 95, 5)) +
theme_bw(base_size = 15)
```
Note the very rapid decline in carbon intensity beginning in 2010.
```{r china-energy-plot, include = TRUE, fig.height=7, fig.width=7, fig.cap = "Trend of China's energy intensity, with 2000--2008 highlighted."}
get_kaya_data("China") %>% filter(year >= 1980) %>%
plot_kaya("e", y_lab = "quad / $ trillion", log_scale = TRUE,
start_year = 2000, stop_year = 2008) +
labs(title = "Energy intensity of China's economy") +
scale_y_log10(breaks = seq(10, 60, 10)) +
theme_bw(base_size = 15)
```
And here are the trends for the whole world _except_ China:
```{r not-china-carbon-plot, include = TRUE, fig.height=7, fig.width=7, fig.cap = "Trend of carbon intensity for the world except China, with 2000--2008 highlighted."}
china <- get_kaya_data("China") %>% select(year, G, E, F)
world <- get_kaya_data("World") %>% select(year, G, E, F)
mix <- world %>% full_join(china, by = "year", suffix = c(".world", ".china")) %>%
mutate(G = G.world - G.china, E = E.world - E.china, F = F.world - F.china) %>%
mutate(e = E / G, f = F / E) %>% select(year, e, f) %>%
filter(year >= 1980)
plot_kaya(mix, "f", y_lab = expression(paste(MMT * CO[2] / quad)), log_scale = TRUE,
start_year = 2000, stop_year = 2008) +
labs(title = "Carbon intensity of rest of world") +
scale_y_log10(breaks = seq(50, 70, 5)) +
theme_bw(base_size = 15)
```
```{r not-china-energy-plot, include = TRUE, fig.height=7, fig.width=7, fig.cap = "Trend of energy intensity for the world except China, with 2000--2008 highlighted."}
plot_kaya(mix, "e", y_lab = "quad / $ trillion", log_scale = TRUE,
start_year = 2000, stop_year = 2008) +
labs(title = "Energy intensity of rest of world") +
scale_y_log10(breaks = seq(5, 10, 1)) +
theme_bw(base_size = 15)
```
An alternative analysis of the RCP scenarios, which takes the observed trend
from 2005--2016, seems more sensible, both
because it better matches the period in which we are calculating
the implied decarbonization trends
(the first 15 years of the RCP scenarios, from 2005--2020),
and also because it avoids the anomalous temporary reversal of the historical
trend from approximately 2002--2005.
# Methods
To calculate carbon intensity and energy intensity, we need economic data
for the scenarios, so to make things simple, we use the
[GCAM simulations of the RCPs](http://www.globalchange.umd.edu/gcamrcp){target="_blank"}.
These are not identical to the official RCP data, but present reasonable
scenarios that produce emissions trajectories very close to the RCPs.
The GCAM versions of the pathways are calculated within a consistent
macroeconomic model of regional economies and energy use, and thus facilitate
comparisons of energy and carbon intensities.
<!--
## Read GCAM Scenarios
First, we read in the GCAM scenarios from the file `GCAMRCPData.xls`:
-->
```{r read-gcam-data, include=TRUE}
data_dir <- 'data'
data_file <- file.path(data_dir, 'GCAMRCPData.xls')
rcp_45 <- read_excel(data_file, 'RCP4.5') %>% clean_names()
gcam_60 <- read_excel(data_file, 'GCAM6.0') %>% clean_names()
gcam_85 <- read_excel(data_file, 'GCAM8.5') %>% clean_names()
gcam_26 <- read_excel(data_file, 'GCAM2.6') %>% clean_names()
gcam_ref <- read_excel(data_file, 'GCAMReference') %>% clean_names()
gcam_by_sector <- read_excel(data_file, 'CO2EmissionsbySector') %>% clean_names()
#
# Clean up a list of variables by converting them into legal names
#
clean_vars <- function(vv) {
new_vv <- vv %>%
str_replace_all(c("['\"]" = "", "%" = "percent", "^[ ]+" = "_")) %>%
make.names() %>%
str_replace_all(c("[._]+" = "_", "_+$" = "")) %>% str_to_lower()
}
#
# The excel sheet runs years in columns and variables in rows.
# Flip this to put years in columns (and we need to fix "x2008" type col names
# and convert years to integers), and clean up variable names to make them
# legal column names.
#
process_scenario <- function(df) {
df %>% gather(key = "year", value = "value", -model, -scenario, -region, -variable, -unit) %>%
mutate(year = str_replace_all(year, "^[^0-9]+", "") %>% as.integer()) %>%
mutate(variable = clean_vars(variable))
}
for(s in c('gcam_26', 'rcp_45', 'gcam_60', 'gcam_85', 'gcam_ref', 'gcam_by_sector')) {
get(s) %>% process_scenario() %>% assign(x = s, value = ., pos = globalenv())
}
all_scenarios <- rbind(gcam_26, rcp_45, gcam_60, gcam_85, gcam_ref)
#
# There are lots of variables. We're interested in the Kaya ones, so filter
# those out, and change the names to the ones we want.
# Change scenario names to what we want.
# Then add up values for each region to get whole world numbers.
#
gcam_scenarios <- all_scenarios %>%
filter(variable %in% c("population_total", "gdp_total_mer", "primary_energy_total",
"emissions_co2_fossil_fuels_and_industry")) %>%
select(scenario, region, variable, year, value) %>%
mutate(variable = str_replace_all(variable, c("_.*$" = "", "^population" = "P",
"^gdp" = "G", "^primary" = "E",
"^emissions" = "F")),
scenario = str_replace_all(scenario, "(GCAM|RCP)([0-9]|Reference)", "\\1 \\2")) %>%
group_by(scenario, variable, year) %>% summarize(value = sum(value)) %>% ungroup()
```
## Observed Historical Data
We get historical emissions data from the 2017 BP Statistical Review of World Energy
and historical economic data from the World Bank.
<!--
For convenience, we have already calculated these in the `kayadata` package,
so we use that.
-->
```{r load-historical-data, include=TRUE}
historical <- get_kaya_data("World")
```
## Calculate Decarbonization Rates
From the Kaya identity, the energy intensity of the economy is
given by $e = E/G$, and the carbon intensity of the energy supply is
given by $f = F/E$, where $E$ is the total primary energy consumption,
$F$ is the total carbon dioxide emissions, and $G$ is the gross world product.
Fractional growth rates are given by
$$
\begin{aligned}
r_e &= \frac{\text{d} \ln e}{\text{d}t} \\
r_f &= \frac{\text{d} \ln f}{\text{d}t}
\end{aligned}
$$
We implied rates under RCP and GCAM scenarios from 2005--2020 and
we calculate rates of observed (historical) change from 2005--2016.
```{r decarbonization-rates, include = TRUE}
rates <- gcam_scenarios %>% spread(key = variable, value = value) %>%
mutate(e = E / G, f = F / E) %>% select(year, scenario, e, f) %>%
filter(year >= 2005, year <= 2020) %>% gather(key = variable, value = value, e, f) %>%
nest(-scenario, -variable) %>% group_by(scenario, variable) %>%
mutate(fit = map(data, ~lm(log(value) ~ year, data = .x))) %>%
mutate(coefs = map(fit, coef)) %>% mutate(rate = map_dbl(coefs, 2)) %>%
ungroup() %>%
select(scenario, variable, rate) %>% mutate(variable = str_c("r.", variable)) %>%
spread(key = variable, value = rate)
hist_rates <- historical %>% select(year, e, f) %>% filter(year >= 2005) %>%
gather(key = variable, value = value, -year) %>%
nest(-variable) %>% group_by(variable) %>%
mutate(fit = map(data, ~lm(log(value) ~ year, data = .x))) %>%
mutate(coefs = map(fit, coef)) %>% mutate(rate = map_dbl(coefs, 2)) %>%
ungroup() %>%
select(variable, rate) %>% mutate(variable = str_c("r.", variable)) %>%
spread(key = variable, value = rate) %>% mutate(scenario = "Historical")
rates <- rates %>% bind_rows(hist_rates)
```
# Analysis
Now we have the data, so we plot the implied rates of decarbonization in the
RCP and GCAM scenarios and compare them to the observed historical rates
over the same time frame.
```{r plot-data, include=TRUE, fig.height=7, fig.width=7, fig.cap="Observed vs. implied rates of decarbonization. RCP and GCAM rates are for 2005--2020 and Historical rate is for 2005--2016."}
rcp_scales <- tribble(
~scenario, ~order, ~break_order, ~shape, ~size, ~nx, ~ny,
"Historical", 3, 1, 18, 5, 0, 0.05,
"GCAM Reference", 1, 2, 16, 3, 0, 0.05,
"GCAM 8.5", 2, 3, 16, 3, 0, 0.05,
"GCAM 6.0", 4, 4, 16, 3, -0.02, 0.02,
"RCP 4.5", 5, 5, 16, 3, 0, 0.05,
"GCAM 2.6", 6, 6, 16, 3, 0, 0.05
) %>% arrange(order) %>% mutate(scenario = ordered(scenario, levels = scenario))
rcp_levels <- levels(rcp_scales$scenario)
rcp_breaks <- rcp_scales %>% arrange(break_order) %>% select(scenario) %>% simplify() %>% unname()
rcp_shapes <- rcp_scales %>% arrange(order) %>% select(scenario, shape) %>%
spread(key = scenario, value = shape) %>% simplify()
rcp_sizes <- rcp_scales %>% arrange(order) %>% select(scenario, size) %>%
spread(key = scenario, value = size) %>% simplify()
r <- rates %>%
mutate(scenario = ordered(scenario, levels = rcp_levels)) %>%
left_join(rcp_scales %>% select(scenario, nx, ny), by = "scenario")
rate_plot <- ggplot(r, aes(x = 100 * r.f, y = 100 * r.e, color = scenario,
shape = scenario, size = scenario)) +
geom_hline(size = 1, color = "gray70", yintercept = 0.0) +
geom_vline(size = 1, color = "gray70", xintercept = 0.0) +
geom_point() +
geom_abline(aes(slope = -1, intercept = 100 * (r.e + r.f), color = scenario),
data = filter(r, scenario == "Historical"), size = 1) +
geom_text_repel(aes(label = scenario, nudge_x = nx, nudge_y = ny),
color = "black", size = 4,
box.padding = 0.4) +
labs(y = expression(paste(Delta, " Energy Intensity of GDP (%/year)")),
x = expression(paste(Delta, "Carbon Intensity of Energy (%/year)"))) +
scale_x_continuous(expand = c(0, 0.125)) +
scale_color_brewer(palette = "Dark2", breaks = rcp_breaks, direction = -1,
name = "Scenario", guide = "none") +
scale_shape_manual(values = rcp_shapes, breaks = rcp_breaks, name = "Scenario",
guide = "none") +
scale_size_manual(values = rcp_sizes, breaks = rcp_breaks, name = "Scenario",
guide = "none") +
coord_fixed(ratio = 1) +
theme_bw(base_size = 15)
rate_plot
```
```{r rate-table}
rates %>% mutate(scenario = ordered(scenario, levels = rcp_breaks),
r.ef = r.e + r.f) %>%
arrange(scenario) %>%
select(scenario, r.ef) %>% mutate(`Decarbonization rate`=sprintf("%.1f%%", 100 * r.ef)) %>%
select(Scenario = scenario, `Decarbonization rate`) %>%
kable(align = 'cc', caption = "Observed and implied total global decarbonization rates")
```
# Conclusions
In contrast to the previous work, using the last 8 years of observed data on
energy and carbon intensity, the world appears to currently be headed along a
trajectory is decarbonizing slightly faster than both the RCP 4.5 reference
scenario and the RCP 6.0 policy scenario, and which
not too far behind the early years of RCP 4.5 policy scenario
(see Table \@ref(tab:rate-table)).
Thus, the challenge of getting from the current trajectory to the RCP 4.5
policy trajectory seems actually to be a bit smaller than anticipated.
This is very different from the conclusions of Pielke, Wigley, & Green
and Stevenson & Pielke, and presents a much more optimistic picture than those
analyses do.
Of course, the big cuts in the RCP 4.5 pathway occur in the future, so we should
not take this as license for complacency.
Much of the acceleration of decarbonization has been driven by the unexpected
fracking revolution, which dramatically reduced the price of natural gas and
thus stimulated a massive shift from coal to gas in electrical generation.
Sustaining a long-term decarbonization trend will require a similar revolution
in clean energy sources (nuclear and renewables) to replace gas.
Moreover, the anomalous reversal of
spontaneous global decarbonization in the early 2000s may be repeated if
other developing countries experience rapid industrialization and economic
growth. This could happen in India and among African nations.
Nonetheless, if we consider that this analysis compares
the observed _spontaneous_ decarbonization rate (with no major climate policies)
over the last eleven years
to the decarbonization rates implied by the policy-regimes of the
different RCP policy regimes, and if we also recognize that RCP 4.5 gives
[roughly 80% probability of keeping warming below 3.0°C](http://www.nature.com/articles/nclimate1385){target="_blank"},
then this makes it much more plausible that we have time to avoid catastrophic
damage.
This analysis also highlights challenges with respect to energy efficiency:
the carbon intensity of the world's energy supply has dropped much faster than
either the baseline or any of the policy scenarios except GCAM 2.6.
However, the energy intensity of the world's economy has lagged far behind
the reference scenario and all of the serious policy scenarios
(the exception, GCAM 8.5 may be seen as reflecting perverse policies that
encourage energy consumption and the use of coal).
This is true at the global scale and also apparent for China, as shown in
figures \@ref(fig:china-carbon-plot) and \@ref(fig:china-energy-plot).
This is consistent with many experts' views that policy measures will be
more effective at reducing carbon intensity by promoting inexpensive
clean energy sources than at slowing or reversing the growth of energy demand.
However, one could also interpret this as a failure by both governments and
markets to provide
[behaviorally effective incentives](/files/pubs/pdfs/stern_2010_design_principles.pdf){target="_blank"}
and opportunities for
businesses and households to implement available energy-efficiency measures.
Michael Vandenbergh and I argue for the latter interpretation in our recent
book, [_Beyond Politics_](https://beyondpoliticsbook.com){target="_blank"}, where we present
evidence that there is a large efficiency gap both in households and in
businesses, and that there are significant opportunities to reduce emissions
while saving money by systematically pursuing inefficiencies.
# Code
The R code for the analysis in this post is available at <https://github.com/jonathan-g/jonathangilligan.hugo/blob/master/content/post/2018-02-13-dangerous-assumptions-revisited.Rmd>
and an alternate version is available at <https://github.com/jonathan-g/dangerous-assumptions-revisited>