Jobs are important. Yet much—if not most—of the literature on macroeconomic-demographic linkages in low-income countries adroitly tap-dances around this simplest of assertions. The reason behind this handwaving is simple—classical macroeconomic theory states that in the long run, supply-side factors—such as the number and education level of workers, physical capital, and land—dominate demand-side factors. This idea was pithily summarized in a common reformulation of Jean-Baptiste Say's 1803 work, known by economists as Say's law—that “supply creates its own demand.” Yet if this is so clearly the case, then why worry about the demand for labor—that is, the provision of jobs by firms—at all?

In Labour Markets in Low-Income Countries, Ahmed Elsayed and David Lam answer this question in a compelling and accessible way. And, ironically, they do so inadvertently. The purpose of this volume is not to engage in higher level debates about macroeconomic theory, but rather to simply report on the decade-long Growth and Labour Markets in Low-Income Countries Programme (commonly known as GLM|LIC), cofunded by the IZA Institute of Labor Economics and the UK Foreign, Commonwealth and Development Office. (Disclosure: I am a Research Fellow of the IZA). It outlines 43 research projects the program has funded since 2012 and presents a “lessons learned” thereof. However, the actual book reads like a primer on labor markets in low-income countries in general. And the main takeaway from this volume is that the rationale behind Say's law is better regulated to philosophical—even utopian—ideals of how markets could work in theory, but in reality, has led to a poor understanding of labor markets in low-income countries.

Yet while economics is the focus of the volume, demography is the inspiration. The first major section of the book is a description of recent demographic changes in low-income countries, which is immediately followed by an assertion that “with the[se] demographic challenges. . . economic growth will need to be extremely robust. . .. This will require labor markets that are dynamic and efficient, and able to absorb millions of additional workers each month.” This broad idea was a major reason behind the rebranding of the potential for demographic change to affect macroeconomic performance from the “demographic gift” to the “demographic dividend”: a gift implies a benefit which would accrue naturally as in classical theory, while a dividend implies that one must actively invest—either through budgets or policies—to harness this demographic potential. Lam in particular is no stranger to these debates. He has written prolifically on the labor market implications—unemployment in particular—of “youth bulges,” a problem which should not matter at all if demand for workers magically created jobs as working-age populations shift.

Since the 1990s, the field of economics has transformed drastically in two ways. First, there has been a change in the notion of human behavior itself. Before this period, economics as an intellectual field had its roots in the premise that—on average—people are both rational and self-interested and make choices accordingly. This view of human behavior culminated in the “rational expectations revolution” of the 1980s, which leaned heavily on economics’ classical, supply-side traditions and culminated in now mostly defunct theoretical models such as real-business cycle theory, as well as disastrous global policies such as structural adjustment programs in sub-Saharan Africa. Fortunately for the field, around this time there was a broad recognition that traditional paradigms surrounding human behavior were perhaps too parochial. Behavioral economics—which sought to incorporate insights regarding human behavior from psychology—began to become more accepted, and this trend has continued to this day. And for a good reason. Humans do not have the infinite cognitive capacity and recognizing such limitations is surely superior to assuming we are hyperrational, optimizing automatons with perfect information and perfect foresight.

The second major shift beginning in the 1990s has been the increasing importance of empirics relative to theory—with an emphasis on causal inference—within economics. Early iterations of causal empirics began with the use of instrumental variables in the mid-1990s, which quickly expanded to difference-in-differences and regression discontinuity designs by the early 2000s. By the mid to late 2000s, randomized control trials (RCTs) began to rise in popularity and now dominate the field of development economics.

Labour Markets in Low-Income Countries perfectly encapsulates this new state of development economics today, both in terms of its methods and conclusions. Indeed, it seems as if it could be put in a time capsule, taken out in the far future, and used as a snapshot of our current moment in time for future students of the history of economic thought. Virtually all of the 43 funded studies outlined in the volume are RCTs, and the evidence discussed therein reflects this reality. More importantly, the research output of the vast majority of these studies reinforces the lessons learned about human behavior within the field of economics over the past 30 years. Specifically, that (1) focusing on supply-side labor market interventions alone is problematic and (2) many well-intended economic programs are destined to fail without robust consideration of their behavioral implications.

Take the evidence presented on vocational training programs. These programs are a bread-and-butter supply-side intervention and have formed a major pillar of domestic poverty alleviation programs across the world for decades. The rationale is simple—if people are poor because they have few skills, giving them more skills will lift them from poverty. However, pleas to a 40 year-old Ph.D. researcher on her seventh low wage, precarious postdoc to improve her financial situation by investing in more education are unlikely to persuade. Similarly, teaching a subsistence farmer how to diagram a sentence, or that the mitochondria are the powerhouse of the cell, may not lead to better wages if—upon graduating from primary school—her only employment option is to go and work on the very farm from whence she came. This volume argues that the productive capacity of labor is not necessarily enhanced through training unless there are jobs fitting the training—increasing the supply of educated workers alone is not enough. It is then perhaps unsurprising that Elsayed and Lam—after surveying the most recent evidence from the GLM|LIC program—conclude that “vocational programs … have a disappointing record when subjected to rigorous evaluation.” Indeed, one study even showed that “better trained workers may [actually] face higher unemployment risk” —as was the case for our hypothetical postdoc.

However, all is not bleak. While supply-side interventions alone are less effective, pairing such interventions with insights from behavior psychology does seem to work well. Elsayed and Lam note that “the evidence shows that training programs which apply techniques from behavioral psychology tend to have stronger positive impacts, not only on labor market outcomes. . . but also on long-term innovation and capital investment.” Indeed, some of the most successful programs in the book detailed the use of cognitive behavioral therapy as a mechanism to improve labor market outcomes for participants. Other studies emphasized the importance of incorporating realistic expectations about job market outcomes, perceived fairness, and networking into such programs.

Another important conclusion of the book is the central role market frictions play in the failure of purely supply-side interventions. Elsayed and Lam particularly emphasize the role of mobility, information, and gender norms in reducing the effectiveness of these programs. For example, one study showed that vocational training in Bangladesh had no effect on job market outcomes for its participants. However, when the same training was paired with an additional intervention to reduce mobility and opportunity costs—free transit to the training site and reimbursement of the foregone income from attendance—there were significant and lasting labor market improvements.

It is fitting that we review this book in a PDR issue focused on gender. While founded in 2011, the main focus of the GLM|LIC program shifted in 2019 to gender. In fact, so central has this shift been than they rebranded it as G²LM|LIC, adding the second “G” for gender. In this volume, persistent traditional gender norms were found to be an important predictor of program failure or success for women. For example, interventions that increased education or led to lower fertility did not necessarily improve female labor force participation. In fact, the authors argue that if low education leads to poverty, this may force uneducated women into the labor market, and therefore staying home to perform the gendered role of housewife may actually be perceived as a middle-class luxury. This is particularly pronounced in India, where the U-shaped relationship between education and labor force participation is well documented.

Labour Markets in Low-Income Countries is a highly accessible read. In fact, the presentation—with bolding of important points and bullet points for important takeaways at the end of chapters—reads somewhat like an advanced undergraduate textbook. This is a great strength since it succeeds in becoming something larger than a simple report. It could even be used as a primer for nontechnical audiences in government, policy organizations, or advanced students. In addition, the fact that the volume contains research summaries of such a broad cross section of issues in labor markets in low-income countries is a major validation of the success of the G²LM|LIC program; through its funded projects, it spans enough of the important topics in labor markets in low-income countries to cover much of what an incoming student needs to know.

Yet in spite of the volume's focus on labor markets, the implications of their findings are important to the broader theme of demography and the macroeconomy. It is a reaffirmation that the “demographic dividend” is far from automatic. By painting a portrait of labor markets that easily fail, we learn that to meet the challenges of demographic change in low-income countries, perhaps the state-of-the-art should become less abstract and more realistic.

工作很重要。然而，关于低收入国家宏观经济与人口联系的大部分（如果不是大多数）文献都巧妙地围绕着这个最简单的断言进行了踢踏舞。这种挥手的原因很简单——古典宏观经济理论指出，从长远来看，供给侧因素——如工人的数量和教育水平、物质资本和土地——主导着需求侧因素。让-巴蒂斯特·萨伊 (Jean-Baptiste Say) 1803 年的著作被经济学家称为萨伊定律 (Say's law) 的一个共同的重新表述中简洁地总结了这个想法——即“供应创造自己的需求”。然而，如果情况如此明显，那么为什么还要担心对劳动力的需求——即企业提供的工作岗位——呢？

在低收入国家的劳动力市场, Ahmed Elsayed 和 David Lam 以令人信服且易于理解的方式回答了这个问题。而且，具有讽刺意味的是，他们这样做是无意的。本卷的目的不是参与有关宏观经济理论的更高层次的辩论，而是简单地报告由 IZA 共同资助的长达十年的低收入国家增长和劳动力市场计划（通常称为 GLM|LIC）劳动经济研究所和英国外交、联邦和发展办公室。（披露：我是 IZA 的研究员）。它概述了该计划自 2012 年以来资助的 43 个研究项目，并介绍了其中的“经验教训”。然而，这本书读起来就像一般低收入国家劳动力市场的入门读物。这本书的主要内容是说背后的基本原理理论上可行，但在现实中却导致人们对低收入国家的劳动力市场缺乏了解。

然而，虽然经济学是本书的重点，但人口学是灵感。本书的第一个主要部分是对低收入国家最近的人口变化的描述，紧随其后的是一个断言，即“伴随着[这些]人口挑战。. . 经济增长需要极其强劲。. .. 这将需要充满活力和高效的劳动力市场，并且每个月能够吸收数百万额外的工人。” 这一广泛的想法是将人口变化影响宏观经济表现的潜力从“人口礼物”重新命名为“人口红利”的主要原因：礼物意味着按照古典理论自然产生的利益，而红利这意味着人们必须积极投资——无论是通过预算还是政策——来利用这一人口潜力。尤其是林对这些辩论并不陌生。他写了大量关于“青年膨胀”对劳动力市场的影响——尤其是失业——的文章，如果随着工作年龄人口的转变，对工人的需求神奇地创造了就业机会，这个问题根本不重要。

自 1990 年代以来，经济学领域在两个方面发生了巨大变化。首先，人类行为本身的概念发生了变化。在此之前，经济学作为一个知识领域的根源在于——平均而言——人们既理性又自利，并做出相应的选择。这种对人类行为的看法在 1980 年代的“理性预期革命”中达到顶峰，该革命严重依赖于经济学的古典供给侧传统，并最终形成了现在大多已不复存在的理论模型，例如真实商业周期理论，以及灾难性的全球撒哈拉以南非洲的结构调整计划等政策。对于该领域来说幸运的是，大约在这个时候，人们普遍认识到围绕人类行为的传统范式可能过于狭隘。行为经济学——它试图从心理学中吸收关于人类行为的见解——开始变得越来越被接受，并且这种趋势一直持续到今天。并且有充分的理由。人类没有无限的认知能力，认识到这些限制肯定优于假设我们是超理性的，优化具有完美信息和完美远见的自动机。

从 1990 年代开始的第二个重大转变是，在经济学中，经验相对于理论的重要性日益增加——强调因果推理。因果经验的早期迭代始于 1990 年代中期使用工具变量，到 2000 年代初期迅速扩展到差异中的差异和回归不连续性设计。到 2000 年代中后期，随机对照试验 (RCT) 开始流行起来，现在主导了发展经济学领域。

低收入国家的劳动力市场就其方法和结论而言，它完美地概括了当今发展经济学的这种新状态。事实上，它似乎可以放入一个时间囊中，在遥远的未来取出，并用作我们当前时刻的快照，供未来的经济思想史学生使用。该卷中概述的 43​​ 项资助研究几乎都是 RCT，其中讨论的证据反映了这一现实。更重要的是，这些研究中绝大多数的研究成果强化了过去 30 年经济学领域内人类行为的经验教训。具体来说，(1) 仅关注供给侧劳动力市场干预是有问题的，(2) 如果不认真考虑其行为影响，许多用心良苦的经济计划注定会失败。

以职业培训计划中的证据为例。这些项目是一项基本的供给侧干预，几十年来已成为全球国内扶贫项目的主要支柱。理由很简单——如果人们因为缺乏技能而贫穷，那么给予他们更多技能将使他们摆脱贫困。但是，请一位 40 岁的博士。她的第七个低工资，不稳定的博士后研究员，通过投资更多来改善她的财务状况教育不太可能说服。同样，教一个自给自足的农民如何绘制一个句子，或者线粒体是细胞的动力源，如果——从小学毕业后——她唯一的就业选择就是去农场工作的话，可能不会带来更高的工资。她是从哪里来的。本卷认为，除非有适合培训的工作，否则劳动力的生产能力不一定能通过培训得到提高——仅增加受过教育的工人的供给是不够的。因此，Elsayed 和 Lam 在调查了 GLM|LIC 计划的最新证据后得出结论，“职业计划……在接受严格评估时的记录令人失望”，这也许并不令人惊讶。事实上，一项研究甚至表明，“训练有素的工人可能 [实际上] 面临更高的失业风险”——就像我们假设的博士后一样。

然而，一切都不是黯淡的。虽然单独的供给侧干预效果较差，但将此类干预与行为心理学的见解结合起来似乎效果很好。Elsayed 和 Lam 指出，“证据表明，应用行为心理学技术的培训计划往往会产生更强的积极影响，而不仅仅是对劳动力市场的结果。. . 还包括长期创新和资本投资。” 事实上，书中一些最成功的项目详细介绍了使用认知行为疗法作为改善参与者劳动力市场结果的机制。其他研究强调了将关于就业市场结果、感知公平和网络的现实期望纳入此类计划的重要性。

这本书的另一个重要结论是市场摩擦在纯供给侧干预失败中所起的核心作用。Elsayed 和 Lam 特别强调了流动性、信息和性别规范在降低这些项目的有效性方面的作用。例如，一项研究表明，孟加拉国的职业培训对其参与者的就业市场结果没有影响。然而，当同样的培训与额外的干预措施相结合以降低流动性和机会成本时——免费前往培训地点并报销先前的出勤收入——劳动力市场出现了显着而持久的改善。

我们在一个关注性别的PDR问题上回顾这本书是合适的。GLM|LIC 计划成立于 2011 年，但其主要关注点在 2019 年转移到了性别上。事实上，这种转变比他们将其重新命名为 G ² LM|LIC 更为重要，并为性别添加了第二个“G”。在这本书中，持久的传统性别规范被发现是女性项目失败或成功的重要预测因素。例如，增加教育或降低生育率的干预措施并不一定会提高女性劳动力参与率。事实上，作者认为，如果低教育导致贫困，这可能会迫使未受过教育的女性劳动力市场，因此呆在家里扮演家庭主妇的性别角色实际上可能被视为中产阶级的奢侈品。这在印度尤为明显，那里的教育和劳动力参与之间的 U 型关系有据可查。

低收入国家的劳动力市场是一本易于阅读的读物。事实上，该演示文稿——在章节末尾以粗体显示要点和要点——读起来有点像一本高级本科教科书。这是一个很大的优势，因为它成功地成为比简单报告更大的东西。它甚至可以用作政府、政策组织或高级学生的非技术受众的入门读物。此外，该卷包含对低收入国家劳动力市场如此广泛的问题的研究摘要这一事实是对 G ^{2成功的重要验证}LM|LIC 计划；通过其资助的项目，它涵盖了低收入国家劳动力市场的重要主题，足以涵盖新生需要了解的大部分内容。

然而，尽管本书重点关注劳动力市场，但他们的研究结果对更广泛的人口学和宏观经济主题具有重要意义。这再次确认了“人口红利”远非自动。通过描绘容易失败的劳动力市场，我们了解到，为了应对低收入国家人口变化的挑战，也许最先进的技术应该变得不那么抽象，更现实。