Online Prices for Computing Standards of Living Across Countries (OPSLAC)

对各国经济发展的比较需要以共同货币（即美元）计算各国的实际国内生产总值（GDP）。世界银行（World Bank）对实际人均GDP的计算与达到每天1.90美元的极端贫困线所需的计算基本相同：这两种计算都必须将以当地货币计算的支出转换为以美元计算的支出。名义汇率不能用于这一目的，因为它们没有考虑到每个国家现有的价格水平。特别是，将发展中国家的支出按名义汇率换算成美元，穷国显得更穷，富国显得更富。相反，重要的是要纠正穷国与富国相比价格较低的问题。本项目研究使用“大数据”方法和在线价格将各国消费转换为共同货币（美元），并以共同货币（例如每天1.90美元）确定极端贫困线的可行性。互联网价格数据的来源是“十亿价格项目”（BPP），这是一项由主要研究人员之一阿尔贝托·卡瓦洛（Alberto Cavallo）发起的学术倡议。BPP可以获得非常详细的全球价格信息，这对于比较各国相同商品（例如阿拉比卡咖啡豆）的价格至关重要。与传统的价格数据收集方法相比，这些在线价格有可能提高价格观察的频率和准确性。例如，世界银行目前在所谓的国际比较项目（ICP）中从事一项劳动密集型的各国价格收集工作。因为它涉及到对许多国家商店的实际访问，ICP只能偶尔收集价格（最近几年是2005年和2011年，计划在2017年进行另一次收集）。此外，传统的数据收集方法需要当地统计机构的合作，并需要它们遵守严格的质量和方法标准，而这些标准往往无法控制。研究结果对于世界银行（World Bank）和其他从事衡量实际GDP和极端贫困的国际机构将是有价值的；对于依靠这些测量的准确性来设计和实施解决方案的决策者；从事宾夕法尼亚大学世界排名和相关国际比较数据库的学者的工作；以及经济学、政治学和其他学科的学者，他们依赖于对各国实际GDP的测量。该项目的主要目标是：a)利用来自“十亿价格项目”的“大数据”互联网资源，获得一个独立于世界银行收集价格方法的各国可比产品平均在线价格数据库；b)使用这些在线价格信息来构建各国生活水平的衡量标准，这些标准可以与世界银行的标准进行比较，但由于新的价格数据来源，这些标准可能会有所不同；c)每月进行一次，而不是每6年一次，以更准确地跟踪和查明价格和生活水平的变化。近年来，研究人员已经可以获得来自各种来源（政府调查、扫描数据集或网络在线资源）的大规模微观价格数据，但它们还不适合用于比较各国之间的价格，因为价格不是根据相同的分类或识别系统收集的，使用不同的价格收集方法可能会使国际比较无效。方法将是利用BPP提供的大规模多国价格数据集，建立一种标准化的方法来匹配各国产品，并制定一种适当的方法，从单个产品价格扩大到总体消费者价格比较。该项目将侧重于关键的实践和方法问题，并为在研究应用中使用在线价格数据奠定基础，以便对各国的生活成本进行比较。这项研究将发表在学术论文中，研究人员还将每月发布一个数据库，其中包含在项目期间计算的各国平均价格水平。研究人员将通过哈佛/麻省理工Dataverse等数据存储库与其他研究人员公开共享该数据库。宾夕法尼亚大学世界价格表网站（www.ggdc.net/pwt）还将提供一个平台，传播将在线价格信息与其他宏观经济数据系列相结合的结果。该奖项是“挖掘数据挑战”（Digging Into Data Challenge）第4轮的一部分，这是一个国际资助机会，旨在促进各国之间的研究合作，并鼓励采用创新方法分析社会科学和人文科学领域的大型数据集。美国的研究人员将与加拿大和荷兰的学者合作，以实现该项目的目标。

Comparison of economic progress across countries requires that calculation across countries of real gross domestic product (GDP) in a common currency, i.e. U.S. dollars. The calculation of real GDP per capita, which is made by the World Bank, is much the same as the calculation needed to obtain an extreme poverty line of $1.90 per day: both calculations must convert spending in a local currency into spending in U.S. dollars. Nominal exchange rates cannot be used for this purpose because they do not take into account the price levels existing in each country. In particular, converting the spending in developing countries into U.S. dollars with nominal exchange rates makes poor counties appear even poorer and rich countries even richer. Instead, it is essential to correct for the lower prices found in poor countries as compared to rich countries. This project studies the feasibility of using "Big Data" methods and online prices to convert spending across countries into a common currency (U.S. dollars), and to establish extreme poverty lines in a common currency (e.g. $1.90 per day). The source of internet price data is the Billion Prices Project (BPP), an academic initiative associated with Alberto Cavallo, one of the principal investigators. The BPP has access to global price information at a very fine level of detail, as is essential in order to compare price for the same goods (e.g. Arabica coffee beans) across countries. These online prices have the potential to improve both the frequency and accuracy of price observations in comparison to traditional price data collection methods, For example, the World Bank currently engages in a labor-intensive collection of prices across countries in what is called the International Comparisons Project (ICP). Because it involves actual visits to stores in so many countries, the ICP can collect prices only at infrequent intervals (the most recent years were 2005 and 2011, with another collection planned for 2017). Furthermore, traditional data collection methods require the cooperation of local statistical agencies and their adherence to strict quality and methodological standards that are often impossible to control. The results will be valuable for the World Bank and other international agencies engaged in measuring real GDP and extreme poverty; for policy-makers that rely on accuracy of these measurements to design and implement solutions; for the work of academics engaged in the Penn World Tables and related international comparison databases; and for scholars from Economics, Political Science, and other disciplines who rely on the measurement of real GDP in various countries. The main goals of this project are to: a) utilize "Big Data" internet sources from the Billion Prices Project to obtain a database of average online prices for comparable products across countries that is independent of the methods used to collect prices by the World Bank; b) to use this online price information to construct measures of the standard of living across countries that can be compared to those from the World Bank, but which will potentially differ due to the new sources of price data; and c) do this at a monthly frequency, rather than once every 6 years, to more accurately track and pinpoint changes in prices and living standards. Large-scale micro-price data from various sources (government surveys, scanner datasets, or web-scraped online sources) have become available for researchers in recent years, but they are not yet suited for comparing prices across countries because prices are not collected according to the same classification or identification system and use differing price collection methods can invalidate international comparisons. The approach will be to access the large-scale, multi-country dataset of prices available at the BPP, build a standardized method for matching products across countries, and develop an appropriate methodology to scale up from individual product prices to an overall consumer price comparison. The project will focus on key practical and methodological issues and lay the groundwork for the use of online price data in research applications that make cost-of-living comparisons across countries. The research will be published in academic papers and the researchers will also publish a monthly database of average price levels in various countries computed during the duration of the project. The investigators will share this database publicly with other researchers via data depositories such as the Harvard/MIT Dataverse. The website of the Penn World Table (www.ggdc.net/pwt) will also provide a platform for disseminating the results from integrating online price information with other macroeconomic data series.This award was made as part of Round 4 of the Digging Into Data Challenge, an international funding opportunity designed to foster research collaboration across countries and to encourage innovative approaches to analyzing large data sets in the social sciences and humanities. The U.S based researchers will collaborate with scholars in Canada and the Netherlands to achieve the goals of this project.