Editing IPCC:AR6/SR15/Chapter-2 (section)

== Cross-Chapter Box 5: Economics of 1.5°C Pathways and the Social Cost of Carbon ==

<span id="section-14"></span>
<span id="lead-authors-1"></span>
====== Lead Authors ======

* Luis Mundaca (Sweden, Chile)
* Mustafa Babiker (Sudan)
* Johannes Emmerling (Italy, Germany)
* Sabine Fuss (Germany)
* Jean-Charles Hourcade (France)
* Elmar Kriegler (Germany)
* Anil Markandya (Spain, United Kingdom)
* Joyashree Roy (Thailand, India)
* Drew Shindell (United States)

<div id="section-2-5-1-block-1"></div>

Two approaches have been commonly used to assess alternative emissions pathways: '''cost-effectiveness analysis (CEA)''' and '''cost–benefit analysis (CBA)''' . '''CEA''' aims at identifying emissions pathways minimising the total mitigation costs of achieving a given warming or GHG limit (Clarke et al., 2014) <sup>[[#fn:r528|528]]</sup> . '''CBA''' has the goal to identify the optimal emissions trajectory minimising the discounted flows of abatement expenditures and monetized climate change damages (Boardman et al., 2006; Stern, 2007) <sup>[[#fn:r529|529]]</sup> . A third concept, the '''Social Cost of Carbon (SCC)''' measures the total net damages of an extra metric ton of CO <sub>2</sub> emissions due to the associated climate change (Nordhaus, 2014; Pizer et al., 2014; Rose et al., 2017a) <sup>[[#fn:r530|530]]</sup> . Negative and positive impacts are monetized, discounted and the net value is expressed as an equivalent loss of consumption today. The SCC can be evaluated for any emissions pathway under policy consideration (Rose, 2012; NASEM, 2016, 2017) <sup>[[#fn:r531|531]]</sup> .

Along the optimal trajectory determined by CBA, the SCC equals the discounted value of the marginal abatement cost of a metric ton of CO <sub>2</sub> emissions. Equating the present value of future damages and marginal abatement costs includes a number of critical value judgements in the formulation of the social welfare function (SWF), particularly in how non-market damages and the distribution of damages across countries and individuals and between current and future generations are valued (Kolstad et al., 2014) <sup>[[#fn:r532|532]]</sup> . For example, since climate damages accrue to a larger extent farther in the future and can persist for many years, assumptions and approaches to determine the social discount rate (normative ‘prescriptive’ vs. positive ‘descriptive’) and social welfare function (e.g., discounted utilitarian SWF vs. undiscounted prioritarian SWF) can heavily influence CBA outcomes and associated estimates of SCC (Kolstad et al., 2014; Pizer et al., 2014; Adler and Treich, 2015; Adler et al., 2017; NASEM, 2017; Nordhaus, 2017; Rose et al., 2017a) <sup>[[#fn:r533|533]]</sup> .

In CEA, the marginal abatement cost of carbon is determined by the climate goal under consideration. It equals the shadow price of carbon associated with the goal which in turn can be interpreted as the willingness to pay for imposing the goal as a political constraint. Emissions prices are usually expressed in carbon (equivalent) prices using the GWP-100 metric as the exchange rate for pricing emissions of non-CO <sub>2</sub> GHGs controlled under internationally climate agreements (like CH <sub>4</sub> , N <sub>2</sub> O and fluorinated gases, see Cross-Chapter Box 2 in Chapter 1). <sup>[[#fn:13|13]]</sup> Since policy goals like the goals of limiting warming to 1.5°C or well below 2°C do not directly result from a money metric trade-off between mitigation and damages, associated shadow prices can differ from the SCC in a CBA. In CEA, value judgments are to a large extent concentrated in the choice of climate goal and related implications, while more explicit assumptions about social values are required to perform CBA. For example, in CEA assumptions about the social discount rate no longer affect the overall abatement levels now set by the climate goal, but the choice and timing of investments in individual measures to reach these levels.

Although CBA-based and CEA-based assessment are both subject to large uncertainty about socio-techno-economic trends, policy developments and climate response, the range of estimates for the SCC along an optimal trajectory determined by CBA is far wider than for estimates of the shadow price of carbon in CEA-based approaches. In CBA, the value judgments about inter- and intra-generational equity combined with uncertainties in the climate damage functions assumed, including their empirical basis, are important (Pindyck, 2013; Stern, 2013; Revesz et al., 2014) <sup>[[#fn:r534|534]]</sup> . In a CEA-based approach, the value judgments about the aggregate welfare function matter less, and uncertainty about climate response and impacts can be tied into various climate targets and related emissions budgets (Clarke et al., 2014) <sup>[[#fn:r535|535]]</sup> .

The CEA- and CBA-based carbon cost estimates are derived with a different set of tools. They are all summarised as integrated assessment models (IAMs) but in fact are of very different nature (Weyant, 2017) <sup>[[#fn:r536|536]]</sup> . Detailed process IAMs such as AIM (Fujimori, 2017) <sup>[[#fn:r537|537]]</sup> , GCAM (Thomson et al., 2011; Calvin et al., 2017) <sup>[[#fn:r538|538]]</sup> , IMAGE (van Vuuren et al., 2011b, 2017b) <sup>[[#fn:r539|539]]</sup> , MESSAGE-GLOBIOM (Riahi et al., 2011; Havlík et al., 2014; Fricko et al., 2017) <sup>[[#fn:r540|540]]</sup> , REMIND-MAgPIE (Popp et al., 2010; Luderer et al., 2013; Kriegler et al., 2017) <sup>[[#fn:r541|541]]</sup> and WITCH (Bosetti et al., 2006, 2008, 2009) <sup>[[#fn:r542|542]]</sup> include a process-based representation of energy and land systems, but in most cases lack a comprehensive representation of climate damages, and are typically used for CEA. Diagnostic analyses across CBA-IAMs indicate important dissimilarities in modelling assembly, implementation issues and behaviour (e.g., parametric uncertainty, damage responses, income sensitivity) that need to be recognized to better understand SCC estimates (Rose et al., 2017a) <sup>[[#fn:r543|543]]</sup> .

CBA-IAMs such as DICE (Nordhaus and Boyer, 2000; Nordhaus, 2013, 2017) <sup>[[#fn:r544|544]]</sup> , PAGE (Hope, 2006) <sup>[[#fn:r545|545]]</sup> and FUND (Tol, 1999; Anthoff and Tol, 2009) <sup>[[#fn:r546|546]]</sup> attempt to capture the full feedback from climate response to socio-economic damages in an aggregated manner, but are usually much more stylised than detailed process IAMs. In a nutshell, the methodological framework for estimating SCC involves projections of population growth, economic activity and resulting emissions; computations of atmospheric composition and global mean temperatures as a result of emissions; estimations of physical impacts of climate changes; monetization of impacts (positive and negative) on human welfare; and the discounting of the future monetary value of impacts to year of emission (Kolstad et al., 2014; Revesz et al., 2014; NASEM, 2017; Rose et al., 2017a) <sup>[[#fn:r547|547]]</sup> . There has been a discussion in the literature to what extent CBA-IAMs underestimate the SCC due to, for example, a limited treatment or difficulties in addressing damages to human well-being, labour productivity, value of capital stock, ecosystem services and the risks of catastrophic climate change for future generations (Ackerman and Stanton, 2012; Revesz et al., 2014; Moore and Diaz, 2015; Stern, 2016) <sup>[[#fn:r548|548]]</sup> . However, there has been progress in ‘bottom-up’ empirical analyses of climate damages (Hsiang et al., 2017) <sup>[[#fn:r549|549]]</sup> , the insights of which could be integrated into these models (Dell et al., 2014) <sup>[[#fn:r550|550]]</sup> . Most of the models used in Chapter 2 on 1.5°C mitigation pathways are detailed process IAMs and thus deal with CEA.

An important question is how results from CEA- and CBA-type approaches can be compared and synthesized. Such synthesis needs to be done with care, since estimates of the shadow price of carbon under the climate goal and SCC estimates from CBA might not be directly comparable due to different tools, approaches and assumptions used to derive them. Acknowledging this caveat, the SCC literature has identified a range of factors, assumptions and value judgements that support SCC values above $100 tCO <sub>2</sub> <sup>−1</sup> that are also found as net present values of the shadow price of carbon in 1.5°C pathways. These factors include accounting for tipping points in the climate system (Lemoine and Traeger, 2014; Cai et al., 2015; Lontzek et al., 2015) <sup>[[#fn:r551|551]]</sup> , a low social discount rate (Nordhaus, 2007a; Stern, 2007) <sup>[[#fn:r552|552]]</sup> and inequality aversion (Schmidt et al., 2013; Dennig et al., 2015; Adler et al., 2017) <sup>[[#fn:r553|553]]</sup> .

The SCC and the shadow price of carbon are not merely theoretical concepts but used in regulation (Pizer et al., 2014; Revesz et al., 2014; Stiglitz et al., 2017) <sup>[[#fn:r554|554]]</sup> . As stated by the report of the High-Level Commission on Carbon Pricing (Stiglitz et al., 2017) <sup>[[#fn:r555|555]]</sup> , in the real world there is a distinction to be made between the implementable and efficient explicit carbon prices and the implicit (notional) carbon prices to be retained for policy appraisal and the evaluation of public investments, as is already done in some jurisdictions such as the USA, UK and France. Since 2008, the U.S. government has used SCC estimates to assess the benefits and costs related to CO <sub>2</sub> emissions resulting from federal policymaking (NASEM, 2017; Rose et al., 2017a) <sup>[[#fn:r556|556]]</sup> .

The use of the SCC for policy appraisals is, however, not straightforward in an SDG context. There are suggestions that a broader range of polluting activities than only CO <sub>2</sub> emissions, for example emissions of air pollutants, and a broader range of impacts than only climate change, such as impacts on air quality, health and sustainable development in general (see Chapter 5 for a detailed discussion), would need to be included in social costs (Sarofim et al., 2017; Shindell et al., 2017a) <sup>[[#fn:r557|557]]</sup> . Most importantly, a consistent valuation of the SCC in a sustainable development framework would require accounting for the SDGs in the social welfare formulation (see Chapter 5).

<span id="economic-and-investment-implications-of-1.5c-pathways"></span>
=== 2.5.2 Economic and Investment Implications of 1.5°C Pathways ===

<div id="section-2-5-2-1"></div>

<span id="price-of-carbon-emissions"></span>
==== 2.5.2.1 Price of carbon emissions ====

<div id="section-2-5-2-1-block-1"></div>

The price of carbon assessed here is fundamentally different from the concepts of optimal carbon price in a cost–benefit analysis, or the social cost of carbon (see Cross-Chapter Box 5 in this chapter and Chapter 3, Section 3.5.2). Under a cost-effectiveness analysis (CEA) modelling framework, prices for carbon (mitigation costs) reflect the stringency of mitigation requirements at the margin (i.e., cost of mitigating one extra unit of emission). Explicit carbon pricing is briefly addressed here to the extent it pertains to the scope of Chapter 2. For detailed policy issues about carbon pricing see Section 4.4.5.

Based on data available for this special report, the price of carbon varies substantially across models and scenarios, and their values increase with mitigation efforts (see Figure 2.26) ( ''high confidence'' ). For instance, undiscounted values under a Higher-2°C pathway range from 15–220 USD2010 tCO <sub>2-eq</sub> <sup>−1</sup> in 2030, 45–1050 USD2010 tCO <sub>2-eq</sub> <sup>−1</sup> in 2050, 120–1100 USD2010 tCO <sub>2-eq</sub><br />
<sup>−1</sup> in 2070 and 175–2340 USD2010 tCO <sub>2-eq</sub> <sup>−1</sup> in 2100. On the contrary, estimates for a Below-1.5°C pathway range from 135–6050 USD2010 tCO <sub>2-eq</sub> <sup>−1</sup> in 2030, 245–14300 USD2010 tCO <sub>2-eq</sub> <sup>−1</sup> in 2050, 420–19300 USD2010 tCO <sub>2-eq</sub> <sup>−1</sup> in 2070 and 690–30100 USD2010 tCO <sub>2-eq</sub> <sup>−1</sup> in 2100. Values for 1.5°C-low-OS pathway are relatively higher than 1.5°C-high-OS pathway in 2030, but the difference decreases over time, particularly between 2050 and 2070. This is because in 1.5°C-high-OS pathways there is relatively less mitigation activity in the first half of the century, but more in the second half. The low energy demand (LED, P1 in the Summary for Policymakers) scenario exhibits the lowest values across the illustrative pathway archetypes. As a whole, the global average discounted price of emissions across 1.5°C- and 2°C pathways differs by a factor of four across models (assuming a 5% annual discount rate, comparing to Below-1.5°C and 1.5°C-low-OS pathways). If 1.5°C-high-OS pathways (with peak warming 0.1–0.4°C higher than 1.5°C) or pathways with very large land-use sinks are also considered, the differential value is reduced to a limited degree, from a factor 4 to a factor 3. The increase in mitigation costs between 1.5°C and 2°C pathways is based on a direct comparison of pathway pairs from the same model and the same study in which the 1.5°C pathway assumes a significantly smaller carbon budget compared to the 2°C pathway (e.g., 600 GtCO <sub>2</sub> smaller in the CD-LINKS and ADVANCE studies). This assumption is the main driver behind the increase in the price of carbon (Luderer et al., 2018; McCollum et al., 2018) <sup>[[#fn:r558|558]]</sup> . <sup>[[#fn:14|14]]</sup>

The wide range of values depends on numerous aspects, including methodologies, projected energy service demands, mitigation targets, fuel prices and technology availability ( ''high confidence'' ) (Clarke et al., 2014; Kriegler et al., 2015b; Rogelj et al., 2015c; Riahi et al., 2017; Stiglitz et al., 2017) <sup>[[#fn:r559|559]]</sup> . The characteristics of the technology portfolio, particularly in terms of investment costs and deployment rates, play a key role (Luderer et al., 2013, 2016a; Clarke et al., 2014; Bertram et al., 2015a; Riahi et al., 2015; Rogelj et al., 2015c) <sup>[[#fn:r560|560]]</sup> . Models that encompass a higher degree of technology granularity and that entail more flexibility regarding mitigation response often produce relatively lower mitigation costs than those that show less flexibility from a technology perspective (Bertram et al., 2015a; Kriegler et al., 2015a) <sup>[[#fn:r561|561]]</sup> . Pathways providing high estimates often have limited flexibility of substituting fossil fuels with low-carbon technologies and the associated need to compensate fossil-fuel emissions with CDR. The price of carbon is also sensitive to the non-availability of BECCS (Bauer et al., 2018) <sup>[[#fn:r562|562]]</sup> . Furthermore, and due to the treatment of future price anticipation, recursive-dynamic modelling approaches (with ‘myopic anticipation’) exhibit higher prices in the short term but modest increases in the long term compared to optimization modelling frameworks with ‘perfect foresight’ that show exponential pricing trajectories (Guivarch and Rogelj, 2017) <sup>[[#fn:r563|563]]</sup> . The chosen social discount rate in CEA studies (range of 2–8% per year in the reported data, varying over time and sectors) can also affect the choice and timing of investments in mitigation measures (Clarke et al., 2014; Kriegler et al., 2015b; Weyant, 2017) <sup>[[#fn:r564|564]]</sup> . However, the impacts of varying discount rates on 1.5°C (and 2°C) mitigation strategies can only be assessed to a limited degree. The above highlights the importance of sampling bias in pathway analysis ensembles towards outcomes derived from models which are more flexible, have more mitigation options and cheaper cost assumptions and thus can provide feasible pathways in contrast to other who are unable to do so (Tavoni and Tol, 2010; Clarke et al., 2014; Bertram et al., 2015a; Kriegler et al., 2015a; Guivarch and Rogelj, 2017) <sup>[[#fn:r565|565]]</sup> . All CEA-based IAM studies reveal no unique path for the price of emissions (Bertram et al., 2015a; Kriegler et al., 2015b; Akimoto et al., 2017; Riahi et al., 2017) <sup>[[#fn:r566|566]]</sup> .

Socio-economic conditions and policy assumptions also influence the price of carbon ( ''very high confidence'' ) (Bauer et al., 2017; Guivarch and Rogelj, 2017; Hof et al., 2017; Riahi et al., 2017; Rogelj et al., 2018) <sup>[[#fn:r567|567]]</sup> . A multimodel study (Riahi et al., 2017) <sup>[[#fn:r568|568]]</sup> estimated the average discounted price of carbon (2010–2100, 5% discount rate) for a 2°C target to be nearly three times higher in the SSP5 marker than in the SSP1 marker. Another multimodel study (Rogelj et al., 2018) <sup>[[#fn:r569|569]]</sup> estimated the average discounted price of carbon (2020–2100, 5%) to be 35–65% lower in SSP1 compared to SSP2 in 1.5°C pathways. Delayed near-term mitigation policies and measures, including the limited extent of international global cooperation, result in increases in total economic mitigation costs and corresponding prices of carbon (Luderer et al., 2013; Clarke et al., 2014) <sup>[[#fn:r570|570]]</sup> . This is because stronger efforts are required in the period after the delay to counterbalance the higher emissions in the near term. Staged accession scenarios also produce higher mitigation costs than immediate action mitigation scenarios under the same stringency level of emissions (Kriegler et al., 2015b) <sup>[[#fn:r571|571]]</sup> .

It has been long argued that an explicit carbon pricing mechanism (whether via a tax or cap-and-trade scheme) can theoretically achieve cost-effective emission reductions (Nordhaus, 2007b; Stern, 2007; Aldy and Stavins, 2012; Goulder and Schein, 2013; Somanthan et al., 2014; Weitzman, 2014; Tol, 2017) <sup>[[#fn:r572|572]]</sup> . Whereas the integrated assessment literature is mostly focused on the role of carbon pricing to reduce emissions (Clarke et al., 2014; Riahi et al., 2017; Weyant, 2017) <sup>[[#fn:r573|573]]</sup> , there is an emerging body of studies (including bottom-up approaches) that focuses on the interaction and performance of various policy mixes (e.g., regulation, subsidies, standards). Assuming global implementation of a mix of regionally existing best-practice policies (mostly regulatory policies in the electricity, industry, buildings, transport and agricultural sectors) and moderate carbon pricing (between 5–20 USD2010 tCO <sub>2</sub> <sup>−1</sup> in 2025 in most world regions and average prices around 25 USD2010 tCO <sub>2</sub> <sup>−1</sup> in 2030), early action mitigation pathways are generated that reduce global CO <sub>2</sub> emissions by an additional 10 GtCO <sub>2</sub> e in 2030 compared to the NDCs (Kriegler et al., 2018a) <sup>[[#fn:r574|574]]</sup> (see Section 2.3.5). Furthermore, a mix of stringent energy efficiency policies (e.g., minimum performance standards, building codes) combined with a carbon tax (rising from 10 USD2010 tCO <sub>2</sub> <sup>−</sup> <sup>1</sup> in 2020 to 27 USD2010 tCO <sub>2</sub> <sup>−</sup> <sup>1</sup> in 2040) is more cost-effective than a carbon tax alone (from 20 to 53 USD2010 tCO <sub>2</sub> <sup>−</sup> <sup>1</sup> ) to generate a 1.5°C pathway for the U.S. electric sector (Brown and Li, 2018) <sup>[[#fn:r575|575]]</sup> . Likewise, a policy mix encompassing a moderate carbon price (7 USD2010 tCO <sub>2</sub> <sup>−</sup> <sup>1</sup> in 2015) combined with a ban on new coal-based power plants and dedicated policies addressing renewable electricity generation capacity and electric vehicles reduces efficiency losses compared with an optimal carbon pricing in 2030 (Bertram et al., 2015b) <sup>[[#fn:r576|576]]</sup> . One study estimates the carbon prices in high energy-intensive pathways to be 25–50% higher than in low energy-intensive pathways that assume ambitious regulatory instruments, economic incentives (in addition to a carbon price) and voluntary initiatives (Méjean et al., 2018) <sup>[[#fn:r577|577]]</sup> . A bottom-up approach shows that stringent minimum performance standards (MEPS) for appliances (e.g., refrigerators) can effectively complement explicit carbon pricing, as tightened MEPS can achieve ambitious efficiency improvements that cannot be assured by carbon prices of 100 USD2010 tCO <sub>2</sub> <sup>−</sup> <sup>1</sup> or higher (Sonnenschein et al., 2018) <sup>[[#fn:r578|578]]</sup> . In addition, the revenue recycling effect of carbon pricing can reduce mitigation costs by displacing distortionary taxes (Baranzini et al., 2017; OECD, 2017; McFarland et al., 2018; Sands, 2018; Siegmeier et al., 2018) <sup>[[#fn:r579|579]]</sup> , and the reduction of capital tax (compared to a labour tax) can yield greater savings in welfare costs (Sands, 2018) <sup>[[#fn:r580|580]]</sup> . The effect on public budgets is particularly important in the near term; however, it can decline in the long term as carbon neutrality is achieved (Sands, 2018) <sup>[[#fn:r581|581]]</sup> . The literature indicates that explicit carbon pricing is relevant but needs to be complemented with other policies to drive the required changes in line with 1.5°C cost-effective pathways ( ''low to'' ''medium evidence'' , ''high agreement'' ) (see Chapter 4, Section 4.4.5) (Stiglitz et al., 2017; Mehling and Tvinnereim, 2018; Méjean et al., 2018; Michaelowa et al., 2018) <sup>[[#fn:r582|582]]</sup> .

In summary, new analyses are consistent with AR5 and show that the price of carbon increases significantly if a higher level of stringency is pursued ( ''high confidence'' ). Values vary substantially across models, scenarios and socio-economic, technology and policy assumptions. While an explicit carbon pricing mechanism is central to prompt mitigation scenarios compatible with 1.5°C pathways, a complementary mix of stringent policies is required.

<div id="section-2-5-2-1-block-2"></div>

<span id="figure-2.26"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Figure 2.26'''

<span id="global-price-of-carbon-emissions-consistent-with-mitigation-pathways."></span>
<!-- IMG CAPTION -->
'''Global price of carbon emissions consistent with mitigation pathways.'''
<!-- IMG FILE -->
[[File:23e135d56afd2dc5c4933599ebaceccb Figure-2.26-815x1024.jpg]]

Panels show (a) undiscounted price of carbon (2030–2100) and (b) average price of carbon (2030–2100) discounted at a 5% discount rate to 2020 in USD2010. AC: Annually compounded. NPV: Net present value. Median values in floating black line. The number of pathways included in box plots is indicated in the legend. Number of pathways outside the figure range is noted at the top.

Original Creation for this Report using IAMC 1.5°C Scenario Data hosted by IIASA

<!-- END IMG -->
<div id="section-2-5-2-2"></div>

<span id="investments"></span>
==== 2.5.2.2 Investments ====

<div id="section-2-5-2-2-block-1"></div>

Realizing the transformations towards a 1.5°C world would require a major shift in investment patterns (McCollum et al., 2018) <sup>[[#fn:r583|583]]</sup> . Literature on global climate change mitigation investments is relatively sparse, with most detailed literature having focused on 2°C pathways (McCollum et al., 2013; Bowen et al., 2014; Gupta and Harnisch, 2014; Marangoni and Tavoni, 2014; OECD/IEA and IRENA, 2017) <sup>[[#fn:r584|584]]</sup> .

Global energy-system investments in the year 2016 are estimated at approximately 1.7 trillion USD2010 (approximately 2.2% of global GDP and 10% of gross capital formation), of which 0.23 trillion USD2010 was for incremental end-use energy efficiency and the remainder for supply-side capacity installations (IEA, 2017c) <sup>[[#fn:r585|585]]</sup> . There is some uncertainty surrounding this number because not all entities making investments report them publicly, and model-based estimates show an uncertainty range of about ±15% (McCollum et al., 2018) <sup>[[#fn:r586|586]]</sup> . Notwithstanding, the trend for global energy investments has been generally upward over the last two decades: increasing about threefold between 2000 and 2012, then levelling off for three years before declining in both 2015 and 2016 as a result of the oil price collapse and simultaneous capital cost reductions for renewables (IEA, 2017c) <sup>[[#fn:r587|587]]</sup> .

Estimates of demand-side investments, either in total or for incremental efficiency efforts, are more uncertain, mainly due to a lack of reliable statistics and definitional issues about what exactly is counted towards a demand-side investment and what the reference should be for estimating incremental efficiency (McCollum et al., 2013) <sup>[[#fn:r588|588]]</sup> . Grubler and Wilson (2014) <sup>[[#fn:r589|589]]</sup> use two working definitions (a broader and a narrower one) to provide a first-order estimate of historical end-use technology investments in total. The broad definition defines end-use technologies as the technological systems purchasable by final consumers in order to provide a useful service, for example, heating and air conditioning systems, cars, freezers, or aircraft. The narrow definition sets the boundary at the specific energy-using components or subsystems of the larger end-use technologies (e.g., compressor, car engine, heating element). Based on these two definitions, demand-side energy investments for the year 2005 were estimated about 1–3.5 trillion USD2010 (central estimate 1.7 trillion USD2010) using the broad definition and 0.1–0.6 trillion USD2010 (central estimate 0.3 trillion USD2010) using the narrower definition. Due to these definitional issues, demand-side investment projections are uncertain, often underreported, and difficult to compare. Global IAMs often do not fully and explicitly represent all the various measures that could improve end-use efficiency.

Research carried out by six global IAM teams found that 1.5°C-consistent climate policies would require a marked upscaling of energy system supply-side investments (resource extraction, power generation, fuel conversion, pipelines/transmission, and energy storage) between now and mid-century, reaching levels of between 1.6–3.8 trillion USD2010 yr <sup>−</sup> <sup>1</sup> globally on average over the 2016–2050 timeframe (McCollum et al., 2018) <sup>[[#fn:r590|590]]</sup> (Figure 2.27). How these investment needs compare to those in a policy baseline scenario is uncertain: they could be higher, much higher, or lower. Investments in the policy baselines from these same models are 1.6–2.7 trillion USD2010 yr <sup>−1</sup> . Much hinges on the reductions in energy demand growth embodied in the 1.5°C pathways, which require investing in energy efficiency. Studies suggest that annual supply-side investments by mid-century could be lowered by around 10% (McCollum et al., 2018) <sup>[[#fn:r591|591]]</sup> and in some cases up to 50% (Grubler et al., 2018) <sup>[[#fn:r592|592]]</sup> if strong policies to limit energy demand growth are successfully implemented. However, the degree to which these supply-side reductions would be partially offset by an increase in demand-side investments is unclear.

Some trends are robust across scenarios (Figure 2.27). First, pursuing 1.5°C mitigation efforts requires a major reallocation of the investment portfolio, implying a financial system aligned to mitigation challenges. The path laid out by countries’ current NDCs until 2030 will not drive these structural changes; and despite increasing low-carbon investments in recent years (IEA, 2016b; Frankfurt School-UNEP Centre/BNEF, 2017) <sup>[[#fn:r593|593]]</sup> , these are not yet aligned with 1.5°C. Second, additional annual average energy-related investments for the period 2016 to 2050 in pathways limiting warming to 1.5°C compared to the baseline (i.e., pathways without new climate policies beyond those in place today) are estimated by the models employed in McCollum et al. (2018) to be around 830 billion USD2010 (range of 150 billion to 1700 billion USD2010 across six models). This compares to total annual average energy ''supply'' investments in 1.5°C pathways of 1460 to 3510 billion USD2010 and total annual average energy ''demand'' investments of 640 to 910 billion USD2010 for the period 2016 to 2050. Total energy-related investments increase by about 12% (range of 3% to 24%) in 1.5°C pathways relative to 2°C pathways. Average annual investment in low-carbon energy technologies and energy efficiency are upscaled by roughly a factor of six (range of factor of 4 to 10) by 2050 compared to 2015. Specifically, annual investments in low-carbon energy are projected to average 0.8–2.9 trillion USD2010 yr <sup>−1</sup> globally to 2050 in 1.5°C pathways, overtaking fossil investments globally already by around 2025 (McCollum et al., 2018) <sup>[[#fn:r594|594]]</sup> . The bulk of these investments are projected to be for clean electricity generation, particularly solar and wind power (0.09–1.0 trillion USD2010 yr <sup>−1</sup> and 0.1–0.35 trillion USD2010 yr <sup>−1</sup> , respectively) as well as nuclear power (0.1–0.25 trillion USD2010 yr <sup>−1</sup> ). Third, the precise apportioning of these investments depends on model assumptions and societal preferences related to mitigation strategies and policy choices (see Sections 2.1 and 2.3). Investments for electricity transmission and distribution and storage are also scaled up in 1.5°C pathways (0.3–1.3 trillion USD2010 yr <sup>−1</sup> ), given their widespread electrification of the end-use sectors (see Section 2.4). Meanwhile, 1.5°C pathways see a reduction in annual investments for fossil-fuel extraction and unabated fossil electricity generation (to 0.3–0.85 trillion USD2010 yr <sup>−1</sup> on average over the 2016–2050 period). Investments in unabated coal are halted by 2030 in most 1.5°C projections, while the literature is less conclusive for investments in unabated gas (McCollum et al., 2018) <sup>[[#fn:r595|595]]</sup> . This illustrates how mitigation strategies vary between models, but in the real world should be considered in terms of their societal desirability (see Section 2.5.3). Furthermore, some fossil investments made over the next few years – or those made in the last few – will ''likely'' need to be retired prior to fully recovering their capital investment or before the end of their operational lifetime (Bertram et al., 2015a; Johnson et al., 2015; OECD/IEA and IRENA, 2017) <sup>[[#fn:r596|596]]</sup> . How the pace of the energy transition will be affected by such dynamics, namely with respect to politics and society, is not well captured by global IAMs at present. Modelling studies have, however, shown how the reliability of institutions influences investment risks and hence climate mitigation investment decisions (Iyer et al., 2015) <sup>[[#fn:r597|597]]</sup> , finding that a lack of regulatory credibility or policy commitment fails to stimulate low-carbon investments (Bosetti and Victor, 2011; Faehn and Isaksen, 2016) <sup>[[#fn:r598|598]]</sup> .

Low-carbon supply-side investment needs are projected to be largest in OECD countries and those of developing Asia. The regional distribution of investments in 1.5°C pathways estimated by the multiple models in (McCollum et al., 2018) <sup>[[#fn:r599|599]]</sup> are the following (average over 2016–2050 timeframe): 0.30–1.3 trillion USD2010 yr <sup>−1</sup> (ASIA), 0.35–0.85 trillion USD2010 yr <sup>−1</sup> (OECD), 0.08–0.55 trillion USD2010 yr <sup>−1</sup> (MAF), 0.07–0.25 trillion USD2010 yr <sup>−1</sup> (LAM), and 0.05–0.15 trillion USD2010 yr <sup>−</sup> <sup>1</sup> (REF) (regions are defined consistent with their use in AR5 WGIII, see Table A.II.8 in Krey et al., 2014b) <sup>[[#fn:r600|600]]</sup> .

Until now, IAM investment analyses of 1.5°C pathways have focused on middle-of-the-road socio-economic and technological development futures (SSP2) (Fricko et al., 2017) <sup>[[#fn:r601|601]]</sup> . Consideration of a broader range of development futures would yield different outcomes in terms of the magnitudes of the projected investment levels. Sensitivity analyses indicate that the magnitude of supply-side investments as well as the investment portfolio do not change strongly across the SSPs for a given level of climate policy stringency (McCollum et al., 2018) <sup>[[#fn:r602|602]]</sup> . With only one dedicated multimodel comparison study published, there is ''limited to medium evidence'' available. For some features, there is ''high agreement'' across modelling frameworks leading, for example, to ''medium to high'' ''confidence'' that limiting global temperature increase to 1.5°C would require a major reallocation of the investment portfolio. Given the limited amount of sensitivity cases available compared to the default SSP2 assumptions, ''medium confidence'' can be assigned to the specific energy and climate mitigation investment estimates reported here.

Assumptions in modelling studies indicate a number of challenges. For instance, access to finance and mobilization of funds are critical (Fankhauser et al., 2016; OECD, 2017) <sup>[[#fn:r603|603]]</sup> . In turn, policy efforts need to be effective in redirecting financial resources (UNEP, 2015; OECD, 2017) <sup>[[#fn:r604|604]]</sup> and reducing transaction costs for bankable mitigation projects (i.e. projects that have adequate future cash flow, collateral, etc. so lenders are willing to finance it), particularly on the demand side (Mundaca et al., 2013; Brunner and Enting, 2014; Grubler et al., 2018) <sup>[[#fn:r605|605]]</sup> . Assumptions also imply that policy certainty, regulatory oversight mechanisms and fiduciary duty need to be robust and effective to safeguard credible and stable financial markets and de-risk mitigation investments in the long term (Clarke et al., 2014; Mundaca et al., 2016; EC, 2017; OECD, 2017) <sup>[[#fn:r606|606]]</sup> . Importantly, the different time horizons that actors have in the competitive finance industry are typically not explicitly captured by modelling assumptions (Harmes, 2011) <sup>[[#fn:r607|607]]</sup> . See Chapter 4, Section 4.4.5 for details of climate finance in practice.

In summary and despite inherent uncertainties, the emerging literature indicates a gap between current investment patterns and those compatible with 1.5°C (or 2°C) pathways ( ''limited to medium evidence, high agreement'' ). Estimates and assumptions from modelling frameworks suggest a major shift in investment patterns and entail a financial system effectively aligned with mitigation challenges ( ''high confidence'' ).

<div id="section-2-5-2-2-block-2"></div>

<span id="figure-2.27"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Figure 2.27'''

<span id="section-15"></span>
<!-- IMG CAPTION -->
Historical and projected global energy investments.
<!-- IMG FILE -->
[[File:40fc06a7417da9e38a79f2144cac80a8 Figure-2.27-1024x853.jpg]]

(a) Historical investment estimates across six global models from (McCollum et al., 2018) <sup>[[#fn:r608|608]]</sup> (bars = model means, whiskers full model range) compared to historical estimates from IEA (International Energy Agency (IEA) 2016) (triangles). (b) Average annual investments over the 2016–2050 period in the “baselines” (i.e., pathways without new climate policies beyond those in place today), scenarios which implement the NDCs (‘NDC’, including conditional NDCs), scenarios consistent with the Lower-2°C pathway class (‘2°C’), and scenarios in line with the 1.5°C-low-OS pathway class (‘1.5°C’). Whiskers show the range of models; wide bars show the multimodel means; narrow bars represent analogous values from individual IEA scenarios (OECD/IEA and IRENA, 2017) <sup>[[#fn:r609|609]]</sup> . (c) Average annual mitigation investments and disinvestments for the 2016–2030 periods relative to the baseline. The solid bars show the values for ‘2°C’ pathways, while the hatched areas show the additional investments for the pathways labelled with ‘1.5°C’. Whiskers show the full range around the multimodel means. T&amp;D stands for transmission and distribution, and CCS stands for carbon capture and storage. Global cumulative carbon dioxide emissions, from fossil fuels and industrial processes (FF&amp;I) but excluding land use, over the 2016-2100 timeframe range from 880 to 1074 GtCO <sub>2</sub> (multimodel mean: 952 GtCO <sub>2</sub> ) in the ‘2°C’ pathway and from 206 to 525 GtCO <sub>2</sub> (mean: 390 GtCO <sub>2</sub> ) in the ‘1.5°C’ pathway.

Original Creation for this Report. The data comes from: – McCollum, Zhou et al. (2018). Nature Energy (https://www.nature.com/articles/s41560-018-0179-z) – IEA (2016) (“World Energy Investment 2016”) – IEA / IRENA (2017) (“Perspectives for the Energy Transition—Investment Needs for a Low-Carbon Energy System”)

<!-- END IMG -->
<span id="sustainable-development-features-of-1.5c-pathways"></span>
=== 2.5.3 Sustainable Development Features of 1.5°C Pathways ===

<div id="section-2-5-3-block-1"></div>

Potential synergies and trade-offs between 1.5°C mitigation pathways and different sustainable development (SD) dimensions (see Cross-Chapter Box 4 in Chapter 1) are an emerging field of research. Chapter 5, Section 5.4 assesses interactions between individual mitigation measures with other societal objectives, as well as the Sustainable Development Goals (SDGs) (Table 5.1). This section synthesized the Chapter 5 insights to assess how these interactions play out in integrated 1.5°C pathways, and the four illustrative pathway archetypes of this chapter in particular (see Section 2.1). Information from integrated pathways is combined with the interactions assessed in Chapter 5 and aggregated for each SDG, with a level of confidence attributed to each interaction based on the amount and agreement of the scientific evidence (see Chapter 5).

Figure 2.28 shows how the scale and combination of individual mitigation measures (i.e., their mitigation portfolios) influence the extent of synergies and trade-offs with other societal objectives. All pathways generate multiple synergies with sustainable development dimensions and can advance several other SDGs simultaneously. Some, however, show higher risks for trade-offs. An example is increased biomass production and its potential to increase pressure on land and water resources, food production, and biodiversity and to reduce air quality when combusted inefficiently. At the same time, mitigation actions in energy-demand sectors and behavioural response options with appropriate management of rebound effects can advance multiple SDGs simultaneously, more so than energy supply-side mitigation actions (see Chapter 5, Section 5.4, Table 5.1 and Figure 5.3 for more examples). Of the four pathway archetypes used in this chapter (LED, S1, S2, and S5, referred to as P1, P2, P3, and P4 in the Summary for Policymakers), the S1 and LED pathways show the largest number of synergies and least number of potential trade-offs, while for the S5 pathway more potential trade-offs are identified. In general, pathways with emphasis on demand reductions and policies that incentivize behavioural change, sustainable consumption patterns, healthy diets and relatively low use of CDR (or only afforestation) show relatively more synergies with individual SDGs than other pathways.

There is ''robust evidence'' and ''high agreement'' in the pathway literature that multiple strategies can be considered to limit warming to 1.5°C (see Sections 2.1.3, 2.3 and 2.4). Together with the extensive evidence on the existence of interactions of mitigation measures with other societal objectives (Chapter 5, Section 5.4), this results in ''high confidence'' that the choice of mitigation portfolio or strategy can markedly affect the achievement of other societal objectives. For instance, action on SLCFs has been suggested to facilitate the achievement of SDGs (Shindell et al., 2017b) <sup>[[#fn:r610|610]]</sup> and to reduce regional impacts, for example, from black carbon sources on snow and ice loss in the Arctic and alpine regions (Painter et al., 2013) <sup>[[#fn:r611|611]]</sup> , with particular focus on the warming sub-set of SLCFs. Reductions in both surface aerosols and ozone through methane reductions provide health and ecosystem co-benefits (Jacobson, 2002, 2010; Anenberg et al., 2012; Shindell et al., 2012; Stohl et al., 2015; Collins et al., 2018) <sup>[[#fn:r612|612]]</sup> . Public health benefits of stringent mitigation pathways in line with 1.5°C pathways can be sizeable. For instance, a study examining a more rapid reduction of fossil-fuel usage to achieve 1.5°C relative to 2°C, similar to that of other recent studies (Grubler et al., 2018; van Vuuren et al., 2018) <sup>[[#fn:r613|613]]</sup> , found that improved air quality would lead to more than 100 million avoided premature deaths over the 21st century (Shindell et al., 2018) <sup>[[#fn:r614|614]]</sup> . These benefits are assumed to be in addition to those occurring under 2°C pathways (e.g., Silva et al., 2016) <sup>[[#fn:r615|615]]</sup> , and could in monetary terms offset either a large portion or all of the initial mitigation costs (West et al., 2013; Shindell et al., 2018) <sup>[[#fn:r616|616]]</sup> . However, some sources of SLCFs with important impacts for public health (e.g., traditional biomass burning) are only mildly affected by climate policy in the available integrated pathways and are more strongly impacted by baseline assumptions about future societal development and preferences, and technologies instead (Rao et al., 2016, 2017) <sup>[[#fn:r617|617]]</sup> .

At the same time, the literature on climate–SDG interactions is still an emergent field of research and hence there is ''low to medium confidence'' in the precise magnitude of the majority of these interactions. Very limited literature suggests that achieving co-benefits is not automatically assured but results from conscious and carefully coordinated policies and implementation strategies (Shukla and Chaturvedi, 2012; Clarke et al., 2014; McCollum et al., 2018) <sup>[[#fn:r618|618]]</sup> . Understanding these mitigation–SDG interactions is key for selecting mitigation options that maximize synergies and minimize trade-offs towards the 1.5°C and sustainable development objectives (van Vuuren et al., 2015; Hildingsson and Johansson, 2016; Jakob and Steckel, 2016; von Stechow et al., 2016; Delponte et al., 2017) <sup>[[#fn:r619|619]]</sup> .

In summary, the combined evidence indicates that the chosen mitigation portfolio can have a distinct impact on the achievement of other societal policy objectives ( ''high confidence'' ); however, there is uncertainty regarding the specific extent of climate–SDG interactions.

<div id="section-2-5-3-block-2"></div>

<span id="figure-2.28"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Figure 2.28'''

<span id="interactions-of-individual-mitigation-measures-and-alternative-mitigations-portfolios-for-1.5c-with-sustainable-development-goals-sdgs."></span>
<!-- IMG CAPTION -->
'''Interactions of individual mitigation measures and alternative mitigations portfolios for 1.5°C with Sustainable Development Goals (SDGs).'''
<!-- IMG FILE -->
[[File:0d13fee60c5f3aa776b3a1bad545915b Figure-2.28-979x1024.jpg]]

The assessment of interactions between mitigation measures and individual SDGs is based on the assessment of Chapter 5, Section 5.4. Proxy indicators and synthesis method are described in Supplementary Material 2.SM.1.5.

Original Creation for this Report.

<!-- END IMG -->
<span id="knowledge-gaps"></span>