Advancements in Input-Output Models and Indicators for Consumption-Based Accounting
输入输出模型和基于消费会计的指标进展

Global Multiregional Input-Output Tables
全球多区域投入产出表

While the basic I-O methodology applied to industrial ecology (IE) has not changed fundamentally, there have been several innovations around increasing global coverage, resolution, and accuracy of I-O data. Since the reviews by Wiedmann (2009) and Hertwich (2011), computing power has grown considerably. This has enabled the development of MRIO databases containing data for hundreds of countries. Several global MRIO data sets were summarized in a 2013 special issue of Economic Systems Research (Tukker and Dietzenbacher 2013) and elsewhere (Murray and Lenzen 2013). Here, we focus on four, which have been updated over the years: the Global Trade Analysis Project MRIO table (GTAP-MRIOT), the World Input-Output Database (WIOD), EXIOBASE, and Eora.
尽管应用于工业生态学（IE）的基本 I-O 方法没有发生根本变化，但在增加全球覆盖范围、分辨率和 I-O 数据准确性的方面已经出现了一些创新。自从 Wiedmann（2009）和 Hertwich（2011）的评审以来，计算能力有了显著增长。这使得开发包含数百个国家数据的 MRIO 数据库成为可能。2013 年，《经济系统研究》特刊（Tukker 和 Dietzenbacher 2013）以及其他地方（Murray 和 Lenzen 2013）总结了几个全球 MRIO 数据集。在此，我们关注四个在近年来得到更新的数据集：全球贸易分析项目 MRIO 表（GTAP-MRIOT）、世界投入产出数据库（WIOD）、EXIOBASE 和 Eora。

The GTAP-MRIOT is based on the GTAP database, developed by the Center for Global Trade Analysis at Purdue University. The current version, called GTAP 9 Data Base, features 140 countries and 57 sectors, with information on carbon dioxide (CO₂) emissions and five labor skill categories (GTAP 2017). It is a collaboration of researchers and policy makers who contribute to database development and, along with many others, use the database to answer questions, for example, about value added in global production chains or the footprints of products (Andrew and Peters 2013). The WIOD project (Dietzenbacher et al. 2013), updated in November 2016, provides detail on 56 sectors for 28 European Union (EU) countries and 15 other major countries for years 2000–2014 (compared to years 1995–2011, in the 2013 release). The recent update of the WIOD includes detailed data on manufacturing and business services sectors to enable a more comprehensive analysis of global value chains (Timmer et al. 2015). The 2013 data set offers information on socioeconomic and environmental accounts. The socioeconomic accounts are expected to be made available for the 2016 version in early 2018 (WIOD 2018). Recently, the WIOD was extended at the subnational level to include all NUTS (Nomenclature of territorial units for statistics) regions of Europe (Thissen et al. Forthcoming). The first version of EXIOBASE was developed under the EXIOPOL project (Tukker et al. 2013). Since then, the database has been updated under the CREEA3 and DESIRE4 projects to yield EXIOBASE2 and EXIOBASE3, respectively (Wood et al. 2015; Stadler et al. 2018). The database includes information on 15 land-use types, 48 types of raw materials, 172 types of water use, and three employment skill levels for comprehensively quantifying resource footprints of nations (Tukker et al. 2014, 2016). EXIOBASE3 features data on 44 countries plus five ROW regions, 200 products, and 163 industries. A special issue on the DESIRE project provides information about the construction of EXIOBASE3, and the use of this updated database for assessing the environmental impacts embodied in trade (Tukker et al. 2018). Eora was developed at the University of Sydney (Lenzen et al. 2013a). At the time of construction, it covered 187 countries, with a range of 25 to 400 sectors depending on the country, and 35 environmental indicators over the period 1990–2011. Since then, it has been updated to extend the time series to 2014 and to include up to 220 countries and a range of social accounts (e.g., employment, corruption and poverty). It has been applied to questions on global supply-chain GHG emissions (Malik and Lan 2016) as well as biodiversity (Lenzen et al. 2012) and water use (Lenzen et al. 2013b).
GTAP-MRIOT 基于普渡大学全球贸易分析中心开发的 GTAP 数据库。当前版本称为 GTAP 9 数据库，包含 140 个国家和 57 个部门，涵盖二氧化碳（CO ₂ ）排放和五个劳动力技能类别（GTAP 2017）的信息。它是研究人员和政策制定者的合作项目，他们为数据库开发做出贡献，并与其他人一起使用数据库来回答问题，例如关于全球生产链中的增值或产品足迹（Andrew 和 Peters 2013）。WIOD 项目（Dietzenbacher 等人，2013 年），于 2016 年 11 月更新，为 2000-2014 年的 28 个欧盟国家和其他 15 个主要国家提供了 56 个部门的详细信息（与 2013 年发布的 1995-2011 年相比）。WIOD 的最新更新包括制造业和商业服务业的详细数据，以实现全球价值链的更全面分析（Timmer 等人，2015 年）。2013 年的数据集提供了社会经济和环境账户的信息。预计社会经济账户将在 2018 年初的 2016 年版中提供（WIOD 2018）。 最近，WIOD 在次国家层面得到扩展，包括欧洲所有 NUTS（统计地区名称）地区（Thissen 等人，即将出版）。EXIOBASE 的第一个版本是在 EXIOPOL 项目下开发的（Tukker 等人，2013）。从那时起，数据库在 CREEA 3 和 DESIRE 4 项目下进行了更新，分别产生了 EXIOBASE2 和 EXIOBASE3（Wood 等人，2015；Stadler 等人，2018）。该数据库包括关于 15 种土地利用类型、48 种原材料类型、172 种用水类型以及三个就业技能水平的信息，以全面量化国家的资源足迹（Tukker 等人，2014，2016）。EXIOBASE3 包含 44 个国家和五个区域外地区的数据，200 种产品和 163 个行业。关于 DESIRE 项目的一个特刊提供了有关 EXIOBASE3 构建的信息以及使用此更新数据库评估贸易中体现的环境影响（Tukker 等人，2018）。Eora 是在悉尼大学开发的（Lenzen 等人，2013a）。 在建设时期，它覆盖了 187 个国家，根据国家不同，涵盖 25 至 400 个部门，以及 1990 年至 2011 年期间的 35 个环境指标。此后，它已更新，将时间序列扩展到 2014 年，并包括多达 220 个国家以及一系列社会账户（例如，就业、腐败和贫困）。它已被应用于关于全球供应链温室气体排放（Malik 和 Lan 2016）以及生物多样性（Lenzen 等人 2012）和水资源利用（Lenzen 等人 2013b）的问题。

These databases vary in the number of regions, sectors, and physical account extensions and follow different approaches for the compilation of global MRIO data. Some of the differences include varying data sources used for the construction of the tables and different approaches used for harmonizing data. For example, the construction of the EXIOBASE relied on a multistage process of harmonization, whereas the Eora database followed a single automated reconciliation step at the time of construction (Geschke et al. 2014). Understanding what questions each MRIO database was developed to address, and why there are differences in results, assists researchers in identifying the best one for the job. To this end, another special issue of Economic Systems Research was devoted to analyzing the differences between the several databases presented above (Inomata and Owen 2014). Unsurprisingly, differences in the compilation of MRIO tables mean that no two databases yield exactly the same analytical outcomes for CBA (Owen 2017). For example Steen-Olsen and colleagues (2014) analyzed the effect of sectoral aggregation on CO₂ multipliers, concluding that those databases with more detailed sectoral resolution showed more accurate results.
这些数据库在区域、部门和物理账户扩展的数量上有所不同，并采用不同的方法编制全球 MRIO 数据。其中一些差异包括构建表格所使用的数据来源不同以及用于数据协调的方法不同。例如，EXIOBASE 的构建依赖于多阶段的协调过程，而 Eora 数据库在构建时遵循了一个单一的自动化协调步骤（Geschke 等人，2014 年）。了解每个 MRIO 数据库旨在解决哪些问题以及为什么结果存在差异，有助于研究人员确定最适合这项工作的数据库。为此，经济系统研究的一个特别问题被用于分析上述几个数据库之间的差异（Inomata 和 Owen，2014 年）。不出所料，MRIO 表格编制的差异意味着没有任何两个数据库为 CBA（Owen，2017 年）产生完全相同的分析结果。 例如，Steen-Olsen 及其同事（2014 年）分析了部门汇总对 CO ₂ 乘数的影响，得出结论：那些具有更详细部门分辨率的数据库显示的结果更准确。

Subnational Multiregional Input-Output Tables
次级多区域投入产出表

Recently, advances have been made in adding subnational, regional detail to IE studies, such as for Australia (Daniels et al. 2011; Lenzen et al. 2014), Spain (Escobedo-Cardeñoso and Oosterhaven 2011; Cazcarro et al. 2013), China (Feng et al. 2013; Wang et al. 2015; Zhao et al. 2015; Wiedenhofer et al. 2017), and Germany (Többen and Kronenberg 2011, 2015). One of the main challenges for constructing subnational MRIO tables is the absence of detailed data at a subnational level. In such a case, nonsurvey methods are often used for constructing subnational input-output tables (IOTs), as demonstrated by Többen and Kronenberg (2015), who used the CHARM nonsurvey method for their study.
近期，在将次国家、地区细节纳入 IE 研究中取得了进展，例如澳大利亚（Daniels 等人，2011；Lenzen 等人，2014）、西班牙（Escobedo-Cardeñoso 和 Oosterhaven，2011；Cazcarro 等人，2013）、中国（Feng 等人，2013；Wang 等人，2015；Zhao 等人，2015；Wiedenhofer 等人，2017）和德国（Többen 和 Kronenberg，2011，2015）。构建次国家 MRIO 表的主要挑战之一是缺乏次国家层面的详细数据。在这种情况下，通常使用非调查方法来构建次国家投入产出表（IOTs），如 Többen 和 Kronenberg（2015）所示，他们在其研究中使用了 CHARM 非调查方法。

Subnational MRIO tables are useful for understanding within-country impacts of inter-regional interactions. Understanding these interactions is crucial for a large country such as China. For example, Feng and colleagues (2013) and Su and Ang (2014) used a subnational MRIO table of China for quantifying emissions embodied in China's inter-regional trade. Likewise, Feng and colleagues (2012) used China's subnational MRIO table for appraising the regional flows of water. Dietzenbacher and colleagues (2012) demonstrated that a conventional IOT cannot distinguish between domestic production and production of both processed and normal exports. The authors therefore used a tripartite IOT to make a distinction between the three classes of production, concluding that for the case of China assessments using an ordinary IOT result in an overestimation of emissions embodied in China's exports (see also Su et al. 2013).
次级区域 MRIO 表有助于理解国内区域间互动的影响。对于像中国这样的大国，理解这些互动至关重要。例如，冯及其同事（2013 年）和苏、安（2014 年）使用中国的次级区域 MRIO 表来量化中国区域间贸易中体现的排放。同样，冯及其同事（2012 年）使用中国的次级区域 MRIO 表来评估区域水流。迪岑巴赫及其同事（2012 年）证明，传统的 IOT 无法区分国内生产和加工及普通出口的生产。因此，作者使用三分法 IOT 来区分三类生产，得出结论：对于中国的情况，使用普通 IOT 进行评估会导致对中国出口中体现的排放高估（参见苏等，2013 年）。

Subnational MRIO tables are also being used for enumerating the environmental, social, and economic impacts of a new product or industry in a regional economy. This is evident in a study undertaken by Malik and colleagues (2015), who analyzed the triple-bottom-line impacts of biofuel production in Western Australia. Likewise, Rodríguez-Alloza and colleagues (2015) enumerated the energy and GHG requirements of a new technology for road pavement construction in New South Wales (NSW), Australia. The advent of virtual laboratories (section: Advent of Virtual Laboratories) has supported construction of subnational IOTs for countries such as China (Wang 2017) and Indonesia (Faturay et al. 2017), paving way for regional assessments for enumerating the impacts of cities, in particular GHG emissions (see section: Example Application of Global Consumption-Based Accounting: Cities).
次国家 MRIO 表也被用于统计区域经济中新产品或行业的环境、社会和经济影响。这在 Malik 及其同事（2015 年）进行的研究中表现得尤为明显，他们分析了澳大利亚西澳大利亚州生物燃料生产的综合影响。同样，Rodríguez-Alloza 及其同事（2015 年）统计了澳大利亚新南威尔士州（NSW）道路路面建设新技术的能源和温室气体排放需求。虚拟实验室的出现（章节：虚拟实验室的兴起）支持了中国（王，2017 年）和印度尼西亚（Faturay 等，2017 年）等国家次国家 IOTs 的建设，为评估城市影响，特别是温室气体排放的影响铺平了道路（参见章节：全球消费基础会计的示例应用：城市）。

Hybrid Models  混合模型

Recently, the accuracy of I-O models has been improved by replacing monetary data (which may be affected by price inhomogeneity) with physical data in hybrid-unit models. Examples are: the construction of hybrid tables using the EXIOBASE database (Merciai and Schmidt 2017); calculation of raw material consumption (the material footprint) for the EU (Schoer et al. 2012); inland marine transportation (Ewing et al. 2011); disaggregation of the electricity sector in a Chinese I-O model to evaluate the primary energy embodied in Chinese final consumption (Lindner and Guan 2014); and the construction of a multiregional solid waste account (Tisserant et al. 2017).
最近，通过在混合单位模型中将货币数据（可能受价格非均匀性影响）替换为物理数据，提高了 I-O 模型的准确性。例如：使用 EXIOBASE 数据库（Merciai 和 Schmidt 2017）构建混合表；计算欧盟的原材料消耗（物质足迹）（Schoer 等，2012）；内陆海运运输（Ewing 等，2011）；对中国 I-O 模型中的电力部门进行分解以评估中国最终消费中体现的初级能源（Lindner 和 Guan 2014）；以及构建一个多区域固体废物账户（Tisserant 等，2017）。

Another advance has been the adaptation of global and subnational MRIO frameworks to include process-based, life cycle inventory data to enable hybrid life cycle assessment (LCA) applications. Applications have focused on the assessment of renewable energy technologies based on integrated hybrid LCA by linking process data to I-O matrices (Acquaye et al. 2011, 2012; Wiedmann et al. 2011; Hertwich et al. 2015) or by inserting new sectors derived from process information into the IOTs (Malik et al. 2015; Moran et al. 2015; Teh et al. 2017). MRIO-based hybrid LCA represents a way forward in I-O–assisted LCA because with increasing globalization LCA applications will increasingly deal with functional units that draw on inputs sourced from many countries. Only an MRIO model underpinning a hybrid LCA exercise can ensure that country-specific production recipes as well as international trade are being considered during the enumeration of a functional unit's supply chain. Hybrid I-O/LCA will increasingly be recognized as an important tool as civil society, organizations, and governments seek to report progress toward the Sustainable Development Goals (SDGs), more so as the United Nations (UN) prepares to update the guidelines on social LCA (Ekener 2017).
全球和次国家层面 MRIO 框架的适应，包括基于过程的、生命周期库存数据，以实现混合生命周期评估（LCA）应用。应用主要集中在基于综合混合 LCA 对可再生能源技术的评估，通过将过程数据与 I-O 矩阵相连接（Acquaye 等人，2011，2012；Wiedmann 等人，2011；Hertwich 等人，2015）或通过将来自过程信息的新部门插入到 IOTs 中（Malik 等人，2015；Moran 等人，2015；Teh 等人，2017）。基于 MRIO 的混合 LCA 代表了在 I-O 辅助 LCA 中的前进方向，因为随着全球化的增加，LCA 应用将越来越多地处理从许多国家获取输入的功能单元。只有支撑混合 LCA 活动的 MRIO 模型才能确保在功能单元供应链的列举过程中考虑了特定国家的生产配方以及国际贸易。 混合 I-O/LCA 将越来越被视为一个重要工具，因为民间社会、组织和政府寻求报告实现可持续发展目标（SDGs）的进展，尤其是在联合国（UN）准备更新关于社会 LCA 的指南时（Ekener 2017）。

Advent of Virtual Laboratories
虚拟实验室的兴起

While the creation of a number of MRIO databases has advanced the field of input-output analysis (IOA), it has been a time-consuming process with possible replication of people power and resources. To automate and streamline the process of MRIO table compilation and update, researchers from a consortium of Australian academic institutions constructed the Industrial Ecology Virtual Laboratory (IELab) infrastructure for compiling the most detailed subnational IOT to date, for Australia (Lenzen et al. 2014). Based on a unique root-base-branch relationship, the Australian IELab allows for data to be stored in the most detailed regional and sectoral classification at the root level. While impossible to construct a table at the root classification because of a lack of computing power, the regional and sectoral resolutions can be aggregated to construct tables within the processing power of current computing technology, thus giving users the flexibility of defining the regions and sectors to be included in an MRIO table depending on their research question. This IELab flexibility has resulted in a range of case studies, for example, triple-bottom-line assessment of biofuel production (Malik et al. 2015, 2016), IOA of waste flows (Reynolds et al. 2014, 2015a; Fry et al. 2016), small island assessments (Malik 2016), environmental impacts of food production (Reynolds et al. 2015b), carbon footprinting of cities (Chen et al. 2016a; Wiedmann et al. 2016), renewable electricity generation (Wolfram et al. 2016), and the assessment of construction materials and the built environment (Teh et al. 2015).
尽管创建多个 MRIO 数据库推动了投入产出分析（IOA）领域的发展，但这是一个耗时且可能重复人力和资源的过程。为了自动化和简化 MRIO 表格编制和更新的过程，来自澳大利亚学术机构联盟的研究人员构建了工业生态学虚拟实验室（IELab）基础设施，用于编制迄今为止最详细的澳大利亚次国家级 IOT（Lenzen 等人，2014 年）。基于独特的根-分支关系，澳大利亚 IELab 允许在根级别存储最详细的区域和行业分类数据。由于计算能力不足，无法在根分类级别构建表格，但可以在当前计算技术的处理能力内，将区域和行业分辨率汇总以构建表格，从而使用户能够根据其研究问题定义包含在 MRIO 表格中的区域和行业。 这一 IELab 灵活性导致了多种案例研究，例如，生物燃料生产的“三重底线”评估（Malik 等，2015，2016），废物流量的 IOA（Reynolds 等，2014，2015a；Fry 等，2016），小岛评估（Malik，2016），食品生产的环境影响（Reynolds 等，2015b），城市碳足迹（Chen 等，2016a；Wiedmann 等，2016），可再生能源发电（Wolfram 等，2016），以及建筑材料和建成环境的评估（Teh 等，2015）。

In addition to providing the flexibility to construct customized IOTs (Geschke and Hadjikakou 2017), the IELab provides features for using the IOTs for footprint assessments or for quantifying uncertainty in MRIO data. Wiedmann (2017) reviewed 30 case studies provided by experts using the IELab platform. He surveyed these users to determine the potential of the lab in enabling I-O research. The author concluded that two thirds of the studies were made possible by the IELab and an additional six would have required substantial input of time and resources without it. While the current make-up of the IELab has restricted its update by nonexpert users (Wiedmann 2017), it is nevertheless a powerful tool for undertaking timely and policy-relevant applications, as demonstrated by Lenzen and colleagues (2017a). In addition to uses in IE, sustainability assessment, and economic modeling (Wiedmann 2017), Lenzen and colleagues (submitted) used the IELab for disaster modeling. Within 2 months of Cyclone Debbie hitting Queensland, Australia, in March 2017, Lenzen and colleagues (submitted) were able to assess the employment and value-added supply-chain effects of the disaster; the analysis was made possible by timely data provision and representation of regional detail in cyclone-hit areas.
除了提供构建定制化 IOTs（Geschke 和 Hadjikakou 2017）的灵活性外，IELab 还提供了使用 IOTs 进行足迹评估或量化 MRIO 数据不确定性的功能。Wiedmann（2017）回顾了由专家使用 IELab 平台提供的 30 个案例研究。他调查了这些用户，以确定实验室在促进 I-O 研究方面的潜力。作者得出结论，三分之二的研究得益于 IELab，另外六个研究如果没有它将需要大量时间和资源。尽管 IELab 的当前构成限制了非专家用户的更新（Wiedmann 2017），但它仍然是一个强大的工具，可以用于及时和与政策相关的研究应用，正如 Lenzen 及其同事（2017a）所展示的。除了在 IE、可持续性评估和经济建模（Wiedmann 2017）中的应用外，Lenzen 及其同事（提交）还使用 IELab 进行灾害建模。 在 2017 年 3 月 Cyclone Debbie 袭击澳大利亚昆士兰州两个月内，Lenzen 及其同事（提交）能够评估灾害的就业和增值供应链效应；分析得以实现得益于及时的数据提供和对受飓风影响地区的区域细节的呈现。

The IELab has recently been extended to bring together some of the aforementioned global MRIO databases, such as WIOD and EXIOBASE, on one platform (see Rahman et al. 2017; Reyes et al. 2017). This has helped establish the means for regularly updating global MRIO frameworks, fostering collaboration among researchers around the world, and facilitating global cross-disciplinary research (Lenzen et al. 2017b).
IELab 最近得到了扩展，将一些上述全球 MRIO 数据库，如 WIOD 和 EXIOBASE，整合到一个平台上（参见 Rahman 等人 2017 年；Reyes 等人 2017 年）。这有助于建立定期更新全球 MRIO 框架的手段，促进全球研究人员之间的合作，并促进全球跨学科研究（Lenzen 等人 2017b）。

We recognize that the virtual laboratory infrastructure is relatively new and its use as yet is mostly based in Australia. However, we believe that virtual laboratories have the potential to facilitate timely research of topical issues of political importance (such as disasters). It would therefore be interesting to apply this capability to a wide range of applications undertaken by research groups around the world.
我们认识到虚拟实验室基础设施相对较新，其使用目前主要在澳大利亚。然而，我们认为虚拟实验室有潜力促进对政治重要议题（如灾害）的及时研究。因此，将这一能力应用于世界各地研究小组的广泛应用将是有趣的。

Environmentally Extended Multiregional Input-Output Analysis
环境扩展的多区域投入产出分析

When considering environmental impacts, Hertwich (2011) found that there had been few applications of IOA to consumption impact studies beyond energy use and GHG emissions. In his analysis of the use of MRIO analysis for CBA, Wiedmann (2009) also found a strong focus on GHG emissions, in particular CO₂. While energy and GHG emissions are still common consumption impacts studied using IOA, the field has broadened considerably. Environmental footprinting has gone through several cycles, from providing a single number of integrated environmental impacts to a more detailed accounting of a single type of impact (Hoekstra and Wiedmann 2014; Lifset 2014). In general, an environmental footprint summarizes the total pressure exerted on the environment by the use of resources or the generation of wastes or emissions by a consumption activity. Increasingly, there are attempts to express these pressures as actual environmental impacts as defined in LCA, for example, global warming for GHG emissions (ISO 2013), water scarcity or pollution for water use (ISO 2014), or resource depletion for material extraction (Fang and Heijungs 2014). Undoubtedly, footprinting remains a popular application of IOA to evaluate the impacts of consumption. This includes carbon footprints,5 water footprints,6 material footprints,7 biodiversity footprints,8 and other environmental pressures.9 Note that the references listed in footnotes 5 through 9 are examples of a range of footprinting studies.
在考虑环境影响时，Hertwich（2011）发现，除了能源使用和温室气体排放外，投入产出分析（IOA）在消费影响研究中的应用很少。在他的关于 MRIO 分析在成本效益分析（CBA）中应用的分析中，Wiedmann（2009）也发现对温室气体排放，特别是 CO ₂ 的关注很强。虽然能源和温室气体排放仍然是使用 IOA 研究的常见消费影响，但该领域已经大大拓宽。环境足迹分析已经经历了几个周期，从提供单一的综合环境影响数值到对单一类型影响的更详细核算（Hoekstra 和 Wiedmann 2014；Lifset 2014）。总的来说，环境足迹总结了由资源使用或消费活动产生的废物或排放对环境施加的总压力。越来越多人试图将这些压力表达为生命周期评估（LCA）中定义的实际环境影响，例如，对于温室气体排放是全球变暖（ISO 2013），对于水资源使用是水资源短缺或污染（ISO 2014），对于材料提取是资源耗竭（Fang 和 Heijungs 2014）。 毫无疑问，足迹分析仍然是 IOA 在评估消费影响方面的一种流行应用。这包括碳足迹、5 水足迹、6 物质足迹、7 生物多样性足迹、8 以及其他环境压力。9 请注意，脚注 5 至 9 中列出的参考文献是足迹分析研究范围的一个例子。

When calculating footprints using MRIO analysis, it is crucial to understand the influence of aggregation or disaggregation of I-O sectors or physical accounts data on calculating impacts. This is particularly true for environmentally extended assessments. Several authors have analyzed the effect of sector aggregation in yielding uncertain results or what is called aggregation bias for material (de Koning et al. 2015; Piñero et al. 2015; Majeau-Bettez et al. 2016) and carbon flows (Su et al. 2010; Steen-Olsen et al. 2014). Aggregation bias is particularly true for assessments where the IOTs feature an aggregation of all agricultural commodities into one sector called Agriculture. Lenzen (2011) pointed out that this can lead to significant errors and concludes that disaggregation of I-O data or environmental data yields more superior results than aggregation of these data.
在利用投入产出分析计算足迹时，理解 I-O 部门或实物账户数据的汇总或细化对计算影响的影响至关重要。这对于环境扩展评估尤其如此。几位作者分析了部门汇总对产生不确定结果或称为材料汇总偏差的影响（de Koning 等，2015；Piñero 等，2015；Majeau-Bettez 等，2016）以及碳流（Su 等，2010；Steen-Olsen 等，2014）。汇总偏差对于 IOTs 将所有农产品汇总为一个称为农业的部门进行评估的评估尤其如此。Lenzen（2011）指出，这可能导致重大错误，并得出结论，I-O 数据或环境数据的细化比汇总这些数据产生更优越的结果。

While carbon and energy footprints have been the focus of many studies, water footprinting has seen an increase in recent years. A number of special issues have been published focusing on water-related research, such as on “Input-output and Water” (Duarte and Yang 2011), “Water Footprints and Sustainable Water Allocation” (Hoekstra et al. 2015), and “Water Footprint Assessment” (Hoekstra et al. 2017). Similar to the developments in carbon footprint accounting, there is also debate about the respective strengths and weaknesses of bottom-up and top-down approaches to water footprinting. Daniels and colleagues (2011) argue that environmentally extended MRIO (EE-MRIO) is well suited to complement process-based approaches to water footprinting by expanding the supply-chain coverage and by establishing the geography of embodied water. Another innovation in water footprinting is the inclusion of scarcity. In calculating physical flow, it is not appropriate simply to add together supply-chain contributions of water from Ireland and water from Uzbekistan, the latter being much scarcer (Lenzen et al. 2013b)—also see Sachs and colleagues (2017) for discussion of this issue in relation to SDG 6 that aims to achieve universal access to clean water and sanitation for all by year 2030. Similar assessments were undertaken for the EU (Serrano et al. 2016). Recent focus of research into water footprinting is to measure progress toward SDG 6 (Hoekstra et al. 2017).
碳足迹和能源足迹一直是许多研究的焦点，但近年来水足迹研究有所增加。已出版了一些特别问题，重点关注与水相关的研究，如“投入产出与水”（杜阿尔特和杨，2011 年）、“水足迹与可持续水资源分配”（霍克斯特拉等人，2015 年）和“水足迹评估”（霍克斯特拉等人，2017 年）。与碳足迹核算的发展相似，关于自下而上和自上而下方法在水足迹研究中的相对优缺点也存在争议。丹尼尔斯及其同事（2011 年）认为，环境扩展多区域投入产出模型（EE-MRIO）非常适合通过扩大供应链覆盖范围和建立体现水地理分布来补充基于过程的水足迹研究方法。水足迹研究中的另一项创新是纳入稀缺性。在计算物理流量时，仅仅将爱尔兰的水供应链贡献和乌兹别克斯坦的水供应链贡献相加是不合适的，因为后者更为稀缺（伦岑等人）。 2013b)—亦参见 Sachs 及其同事（2017）关于此问题与 SDG 6 的讨论，SDG 6 旨在到 2030 年实现所有人都能获得清洁水和卫生设施。对欧盟也进行了类似的评估（Serrano 等人，2016）。近年来，关于水足迹的研究重点在于衡量实现 SDG 6 的进展（Hoekstra 等人，2017）。

In addition to calculating a range of footprints separately (see footnotes 5 through 8), scholars now acknowledge the overlapping nature of footprint indicators (Simas et al. 2017). The first comprehensive and consistent inclusion of carbon, water, and ecological footprint indicators in an EE-MRIO framework was described by Galli and colleagues (2012). The authors concluded that combining these overlapping, interacting, and complementing indicators in a Footprint Family and one modeling framework is of benefit for decision making. More specifically, footprint indicators have been combined with an EE-MRIO model (Weinzettel et al. 2011) in the project One Planet Economy Network Europe (OPEN:EU) funded by the European Commission (EC). A user-friendly analysis and scenario tool was developed from the model. The EUREAPA tool10 allows the user to quantitatively unravel global supply chains using a carbon, ecological, and water footprint indicator (Roelich et al. 2014). The links between the consumption of a product type in one country and its production impacts elsewhere are identified and the top ten sources of greatest impact are displayed. It is worth mentioning that researchers have explored the distinction between pressure- and impact-type indicators. Traditional footprints report on environmental pressures; Verones and colleagues (2017) followed the DPSIR (drivers, pressure, state, impact, and response) framework to link pressures to consequences of consumption. The authors highlight the need for assessing impact footprints for effective policy making.
除了分别计算一系列足迹（见脚注 5 至 8）之外，学者们现在承认足迹指标的重叠性质（Simas 等人，2017 年）。Galli 及其同事（2012 年）描述了在 EE-MRIO 框架中首次全面且一致地纳入碳、水和生态足迹指标。作者得出结论，将这些重叠、相互作用和互补的指标结合在足迹家族和一个建模框架中，对决策有益。更具体地说，足迹指标已被与 EE-MRIO 模型（Weinzettel 等人，2011 年）结合，在由欧洲委员会（EC）资助的项目“一个地球经济网络欧洲”（OPEN:EU）中。从该模型开发了一个用户友好的分析和情景工具。EUREAPA 工具 10 允许用户使用碳、生态和水足迹指标定量解开全球供应链。确定了某一国家产品类型的消费与其在别处的生产影响之间的联系，并显示了影响最大的前十大来源。 值得注意的是，研究人员探讨了压力型和冲击型指标之间的区别。传统的足迹报告关注环境压力；Verones 及其同事（2017 年）遵循 DPSIR（驱动力、压力、状态、影响和响应）框架，将压力与消费后果联系起来。作者强调了评估影响足迹对于有效政策制定的重要性。

Social Footprints and Socially Extended Input-Output Analysis
社会足迹与社会扩展投入产出分析

One of the applications of IOA not well considered prior to 2010 is the use of socially extended I-O matrices to study social ecology and social impacts. Although post–World War II there had been a focus on using I-O to assess social progress, the field stagnated in the following decades at the expense of the development of environmentally extended IOA (McBain and Alsamawi 2014). A recent special issue on social LCA (Gloria et al. 2017) clearly demonstrates advancements in this growing field. The United Nations Environment Program (UNEP)/Society of Environmental Toxicology and Chemistry (SETAC) guidelines on social LCA, currently being updated (Ekener 2017), and the accompanying methodological sheets (Benoit-Norris et al. 2011) have contributed to the understanding of consumption through the use of IOA-assisted LCA. The use of socially extended MRIO is particularly useful for considering the human impacts of consumption from global supply chains. Examples include consideration of the human toll of supplying tantalum to the global marketplace from a conflict zone (Moran et al. 2015) and the impact of commodities produced for U.S. domestic consumption on global inequalities (Prell et al. 2014). MRIO analysis is now being used for enumerating social footprints, for example, employment (Alsamawi et al. 2014a), labor (Simas et al. 2014, 2015), inequality (Alsamawi et al. 2014b), and other social indicators (Hardadi and Pizzol 2017).
2010 年之前，IOA 的一个未充分考虑的应用是使用社会扩展的 I-O 矩阵来研究社会生态和社会影响。尽管二战后人们曾关注使用 I-O 来评估社会进步，但在此后的几十年里，该领域因环境扩展的 IOA（McBain 和 Alsamawi 2014）的发展而停滞。最近的一期关于社会 LCA 的特刊（Gloria 等人，2017 年）清楚地展示了这一增长领域的进展。联合国环境规划署（UNEP）/环境毒理学与化学学会（SETAC）关于社会 LCA 的指南，目前正在更新中（Ekener 2017），以及相关的方法论表格（Benoit-Norris 等人，2011 年），通过使用 IOA 辅助的 LCA，有助于理解消费。社会扩展的 MRIO 的使用特别有助于考虑全球供应链对人类消费的影响。例如，包括从冲突区向全球市场供应钽的代价（Moran 等人，2015 年）以及为美国生产的商品的影响。 国内消费与全球不平等（Prell 等人，2014 年）。现在，MRIO 分析被用于列举社会足迹，例如就业（Alsamawi 等人，2014a），劳动力（Simas 等人，2014，2015），不平等（Alsamawi 等人，2014b），以及其他社会指标（Hardadi 和 Pizzol，2017 年）。

One of the challenges to conceptualization of social footprint work is that of causality. Environmental footprints establish a causal link between trade and, say, emissions or water use. We could say that: “production of this good caused these emissions.” In the case of social footprints, there is no such clear-cut relationship between product, consumer, and social indicator. For example, we cannot say that production of this good caused this amount of corruption (or inequality or ill-health, etc.). Social footprint researchers have addressed this issue by saying that consumers who knowingly purchase goods embodying high levels of, for example, hazardous employment, are implicated in the continuation of such employment. By purchasing contaminated goods, the consumer could be said to tacitly endorse dangerous employment conditions. Purchase of such goods can be characterized as a missed opportunity to pressure governments and global brands to improve conditions (Xiao et al. 2017). At the same time, researchers have been at pains to emphasize that simply not purchasing a contaminated good is no solution to improving worker conditions; any temporary suspension of purchase needs to be accompanied by pressure on suppliers for improvements.
社会足迹概念化所面临的挑战之一是因果关系。环境足迹在贸易与排放或用水等方面建立了因果关系。我们可以这样说：“这种商品的生产导致了这些排放。”在社会足迹的情况下，产品、消费者和社会指标之间没有如此明确的关系。例如，我们无法说这种商品的生产导致了这么多的腐败（或不平等或健康问题等）。社会足迹研究人员通过指出，明知商品具有高风险就业等特征而购买这些商品的消费者，参与了这种就业的持续。通过购买受污染的商品，消费者可能被视为默许危险的工作条件。购买此类商品可以被视为错失了向政府和全球品牌施压以改善条件的良机（Xiao 等人，2017）。 同时，研究人员一直努力强调，仅仅不购买受污染的商品并不能改善工人条件；任何购买暂停都需要伴随对供应商进行改进的压力。

Future of Multiregional Input-Output Development
多区域投入产出发展的未来

We have made considerable progress since the conception of IOA by Wassily Leontief; however, there is still potential for enhancing its capability as a tool for informing policy making at a local, national, and global level. In the area of MRIO development, the future could include construction of nested IOTs. While this has been done for China (Wang and colleagues 2015), a country with a vast geographical area and complexity, nested tables for other countries of the world would pave way for assessing linkages between cities in two different countries. The use of such tables is particularly important for undertaking CBA of cities that rely on imports. Chen and colleagues (2017) carried out CBA of two large cities of Australia—Melbourne and Sydney—and found that the source of a large chunk of imported emissions for these cities lies outside of Australia: 55% for Melbourne and 71% for Sydney. In a separate study, by linking subnational MRIO tables of China and Australia, Chen and colleagues (2016a) were able to identify trade links between different Australian and Chinese cities. The roadmap from national IOTs to global MRIO tables to nested MRIO tables does not come without challenges. At the time of construction of nested MRIO tables linking the subnational MRIO table of China with the global MRIO table, Wang and colleagues (2015) were faced with computational challenges, requiring the authors to adopt a two-step procedure, focusing first on the construction of an MRIO table of China and later integrating that into the global table in a second step. It is worth mentioning that while we have made significant advancements in computational power for constructing and processing MRIO data sets, the construction of a large inter-regional and inter-country data set harboring subnational data for every country of the world is far from being realized.
自瓦西里·列昂季耶夫提出 IOA 概念以来，我们在 IOA 方面取得了相当大的进展；然而，仍有潜力提高其在地方、国家和全球层面为政策制定提供信息的能力。在 MRIO 发展领域，未来可能包括构建嵌套 IOTs。虽然这已经在拥有广阔地理面积和复杂性的中国（王及其同事 2015 年）实现了，但为世界其他国家的其他嵌套表格将有助于评估两个不同国家城市之间的联系。使用此类表格对于进行依赖进口的城市 CBA 尤为重要。陈及其同事（2017 年）对澳大利亚的两个大城市——墨尔本和悉尼进行了 CBA，发现这些城市大量进口排放的来源不在澳大利亚：墨尔本为 55%，悉尼为 71%。在另一项研究中，通过连接中国和澳大利亚的次国家 MRIO 表格，陈及其同事（2016a）能够识别不同澳大利亚和中国城市之间的贸易联系。 国家 IOTs 到全球 MRIO 表再到嵌套 MRIO 表的路线图并非没有挑战。在构建将中国地方 MRIO 表与全球 MRIO 表连接的嵌套 MRIO 表时，王及其同事（2015 年）面临了计算挑战，需要作者采用两步程序，首先构建中国的 MRIO 表，然后在第二步将其整合到全球表中。值得一提的是，尽管我们在构建和处理 MRIO 数据集的计算能力方面取得了重大进展，但构建包含世界上每个国家地方数据的庞大跨区域和跨国数据集仍远未实现。

Future of Indicator Development
指标发展未来

The future for the development of new, unique, and harmonized indicators for inclusion into an MRIO database needs to be informed by UN goals and targets on sustainable development. The UN SDGs are a prime example of where MRIO data sets could be used for assessing a country's progress, at the same time benchmarking the performance in comparison with other world nations. Xiao and colleagues (2017) demonstrate that MRIO analysis can be used for informing progress toward SDGs; however, there is a pressing need to develop indicators that can inform all 17 goals set by the UN and the accompanying 169 targets. As an example, Goal 2 aims to end hunger by 2030 (UN 2015a). At the time of writing, there was no published research investigating the contribution of international trade in causing hunger. The contribution of international trade in promoting or eradicating hunger is unclear. It has been suggested that international trade opens avenues for developing countries’ access to large global markets allowing them to specialize in production and exploit economies of scale. There is, however, another school of thought that challenges this argument on the basis of unfair trading rules (FAO 2017a; OXFAM 2017). A potential integration of a data set such as the Global Hunger Index (GHI) (IFPRI 2017) with a global trade database, coupled with additional data for harmonizing the GHI data set with the trade model, could yield useful insights into the implications of international trade on hunger in developing and underdeveloped nations. It is important to note that while for environmental indicators such as carbon emissions and energy use, we can enumerate the amount of emissions embodied in the consumption of a particular good or service, such a link is not clear cut for social issues such as hunger. These intrinsically complex issues require exploration of potential indicators that could be coupled with the global database for undertaking a supply-chain assessment. The future of indicator development relies on coupling big data with small data at a local and regional level. The report on the success of the Millennium Development Goals calls for “…a data revolution to improve the availability, quality, timeliness and disaggregation of data” to track progress toward SDGs (UN 2015b, 10). Integration of big and small data is crucial for understanding and quantifying the environmental issues faced by world nations, particularly low (-middle)-income countries, that often lack the resources and expertise in this area. This is definitely an area of research that is open for development: production of comprehensive, detailed, and complete small data for integration into big data sets for informing policy making. The World Bank is leading the way in initiatives for collecting data for low (-middle)-income countries, as indicated by Kaushik Basu, Chief Economist of the World Bank: “Data gives representation to people who may otherwise be marginalized and forgotten, hence our decision to greatly step up efforts to collect more and better quality data in developing countries” (The World Bank 2015). The I-O research community has a crucial role to play in informing such efforts and for developing metrics and methods for integrating small data with big data.
未来开发新的、独特的、协调一致的指标并将其纳入 MRIO 数据库的发展，需要参考联合国关于可持续发展的目标和目标。联合国可持续发展目标（SDGs）是 MRIO 数据集可用于评估一个国家进步、同时与世界其他国家绩效进行基准比较的典型案例。Xiao 及其同事（2017 年）证明，MRIO 分析可用于了解 SDGs 的进展；然而，迫切需要开发能够了解联合国设定的 17 个目标和 169 个相关目标的指标。例如，目标 2 旨在到 2030 年结束饥饿（联合国 2015a）。在撰写本文时，没有发表的研究探讨国际贸易在导致饥饿方面的贡献。国际贸易在促进或消除饥饿方面的作用尚不清楚。有人提出，国际贸易为发展中国家进入大型全球市场开辟了途径，使他们能够专注于生产和利用规模经济。 然而，存在另一种观点，该观点基于不公平的贸易规则（FAO 2017a；OXFAM 2017）对此论点提出质疑。将全球饥饿指数（GHI）（IFPRI 2017）等数据集与全球贸易数据库的潜在整合，以及与贸易模型协调的额外数据，可能有助于深入了解国际贸易对发展中国家和欠发达国家的饥饿问题的影响。值得注意的是，对于环境指标，如碳排放和能源消耗，我们可以列举出特定商品或服务的消费中所包含的排放量，但对于饥饿等社会问题，这种联系并不明确。这些本质上复杂的问题需要探索可能被结合到全球数据库中以进行供应链评估的潜在指标。指标发展的未来依赖于在地方和区域层面将大数据与小数据相结合。 报告呼吁“……一场数据革命，以提高数据的可获得性、质量、及时性和细化程度”，以跟踪实现可持续发展目标（SDGs）的进展（联合国 2015b，10）。整合大数据和小数据对于理解和量化世界各国，尤其是资源和技术往往不足的低（中）收入国家面临的环境问题至关重要。这无疑是一个有待发展的研究领域：生产全面、详细和完整的小数据，以便将其整合到大数据集中，为政策制定提供信息。世界银行在为低（中）收入国家收集数据方面走在前列，如世界银行首席经济学家考希克·巴苏所言：“数据为那些可能被边缘化和遗忘的人们提供了代表，因此我们决定在发展中国家加大收集更多和更高质量数据的努力”（世界银行 2015）。I-O 研究界在提供此类努力的信息和开发将小数据与大数据整合的指标和方法方面发挥着关键作用。