Personal Reflections

Some personal reflections on my academic career: Life is a series of coincidences, and academic life is no exception.

"What do you think is human life like here or there?

It seems like a swan's traces on mud or on snow.

See the claw and nail prints by chance mud and snow bear.

Will the flying swan care what it has left below?"

(1) Admitted to the Doctoral Program

My undergraduate, master's, and even doctoral degrees were all in economics. As the poet Su Shi once said, 

"What do you think is human life like here or there?

It seems like a swan's traces on mud or on snow.

See the claw and nail prints by chance mud and snow bear.

Will the flying swan care what it has left below?"

It was purely by chance that I entered the field of Information Systems (IS). It has been eight years since I graduated with my Ph.D., and looking back, I can vividly recall the scene when I received the Ph.D. admission offer. During my master's studies, I was enrolled in the China Center for Economic Research (CCER) at Peking University. At that time, pursuing a Ph.D. abroad was a natural choice, and more than half of my classmates planned to go abroad for their doctoral studies. I followed the trend and applied to the top 30 economics departments, including a few doctoral programs in business schools, based on the rankings from US News. 

Many of my CCER classmates who went abroad are now teaching in North America, Asia, and Europe. However, that is a story for another time. During the application process, the most anxious moment for everyone was waiting for Ph.D. admission offers, and the level of anxiety grew exponentially as time passed. Not even visiting the Temple of Sleeping Buddha (the pronunciation is similar to Offer Temple in Chinese) could alleviate it. Due to the time difference, offers often appeared in the email early in the morning. I remember that checking my email was the first thing I did every morning back then. To be honest, my application results were not ideal, and the only offer with financial aid was from UT Austin, so naturally, I ended up going there to pursue a Ph.D. in economics. At that time, I didn't know that UT Austin had a strong business school, particularly in the field of Information Systems. This accidental discovery led me to shift from economics to Information Systems, but again, it was purely by chance and not part of my initial plan. Another unexpected turn was that my wife would be a classmate in my Ph.D. program (both my wife and I graduated from Peking University with master's degrees, but she was from the Guanghua School of Management, while I was from the Center for Economic Research). We met before starting our Ph.D. studies at UT Austin, but we hadn't begun dating at that time. Perhaps it is true that the most wonderful moments in life are often unplanned. Because both life and academic careers are composed of a series of coincidences, the assistance of others becomes even more important. Without their help, the coincidences would have been different, and my academic career and life path would have been entirely different.

Written on May 6, 2022

• （一）被博士项目录取：人生到处如何似？应似飞鸿踏雪泥                      

我的本科，硕士，直到博士读的都是经济学。如同苏轼诗中所说：“人生到处如何似？应似飞鸿踏雪泥。泥上偶然留指爪，鸿飞那复计东西，“ 我踏入信息系统（Information Systems, IS）领域，实属偶然。现在离博士毕业已有八年，蓦然回首，当初拿到 offer 的情景还历历在目。硕士时候，就读的是北京大学中国经济研究中心（CCER）。而在当时，出国读博是自然的选择，班上一半多的同学都会选择出国读博。而我也是随大流，按照 US news 的排名申请了前30的经济系，也包括少量商学院的博士项目。当时和我一起出国的 CCER 的同学现在大多在北美，亚洲，和欧洲执教。当然这都是后话，申请的时候，大家最焦急的就是等待 offer，而且焦虑指数随着时间的流逝，而成指数级增长，就连去卧佛寺也缓解不了。因为时差关系，offer 往往是清早出现在电子邮箱里的，所以记得当时每天早上的第一件事就是查邮件。平心而论，我申请的成绩不太理想，唯一有 financial aid 的 offer 就是 UT Austin，所以自然也就只能去 UT Austin 读经济学的博士了。当时的我还不知道，UT Austin 有很强的商学院，信息系统领域更是出类拔萃。这成为了我由经济学转向信息系统的近水楼台，不过这点又纯属偶然，不在我的计划之内。我没想到的另一点是，我的太太会是我的博士班的同学（我和我太太都是北大硕士毕业，不过她在光华管理学院，我在经济中心。在 UT Austin 读博之前，我们就认识，但是那时还没有开始 dating)。也许无巧不成书正是人生的精彩之处吧。正因为人生和学术生涯都是由一连串偶然构成的，旁人的帮助更显得重要：如果缺少了他们的帮助，偶然便会不同，我的学术生涯和人生路径也会迥然不同。

写于2022年5月6日

(2) My Master's Advisor, Professor Wu

On my academic journey, I was fortunate to encounter several excellent mentors. Professor Xiangrong Jin at Zhejiang University was my mentor in economics, while Professor Qiren Zhou, Professor Dingding  Wang, Professor Chia-Shang Chu, and Professor Yifu (Justin) Lin at Peking University greatly enriched my knowledge (during my time at Zhejiang University, I had the privilege of taking courses by Professor Qiren Zhou and Professor Dingding Wang). Before pursuing my Ph.D., the most influential figure on my academic path was my master's advisor, Professor Ho-Mou Wu. Having earned his Ph.D. from Stanford and studied under Mordecai Kurz and Kenneth Arrow, Professor Wu had profound expertise in industrial organization, game theory, and financial economics. By chance, I had already taken one of Professor Wu's courses before starting my master's at Peking University. During my undergraduate studies at Zhejiang University, Professor Wu taught a course on game theory as a visiting professor. It was this course that sparked my strong interest in the application of game theory and became a crucial coincidence in my academic journey (I remember that Professor Wu only taught this course once at Zhejiang University). 

In my final year of undergraduate studies, I participated in the summer camp at CCER, Peking University, with the hope of securing admission to their master's program. The camp required presenting a paper in front of CCER faculty members, and my chosen topic was the application of game theory in mutual financial organizations. This paper, from its topic to methodology, was inspired by Professor Wu's course on game theory. Additionally, during my Ph.D. studies at UT Austin, game theory became one of the two main fields I focused on.

After successfully securing admission to CCER for my master's, another significant coincidence occurred for me. Professor Wu left his position at National Taiwan University and officially joined CCER at Peking University as a faculty member. Naturally, I hoped to choose him as my advisor. I remember there were many students who wanted him as their advisor, and luckily, I became his student. The weekly group meetings at that time marked the beginning of my systematic academic reading and writing. We would present the latest working papers in industrial organization and financial economics during these meetings, and later, we would present our own papers. Under Professor Wu's guidance, I began researching the impact of higher-order beliefs on financial markets, which was also my first time writing a paper in English. I remember the densely packed revisions by Professor Wu on the printed initial draft, but unfortunately, the whereabouts of that draft have been lost due to several moves. This paper not only served as my master's thesis but also held two other significances for my academic career. First, due to the natural similarity between financial markets and prediction markets, the papers I studied on heterogeneous beliefs in financial markets during my master's studies, as well as the experience of writing my master's thesis, laid the foundation for my subsequent research on prediction markets. Second, this higher-order beliefs paper became an opportunity for me to become a student of Andy (my Ph.D. advisor, Andrew Whinston) at UT Austin. In my academic journey, a series of new coincidences were brewing.

Written on May 7, 2022

• （二）我的硕士导师巫老师

在我的求学路上，幸运的遇到了不少良师。浙大的金祥荣老师是我的经济学启蒙老师，北大的周其仁老师，汪丁丁老师，朱家祥老师，和林毅夫老师的课也让我获益良多（在浙大的时候，就有幸修读了周其仁老师和汪丁丁老师的课）。在读博士之前，对我的学术道路影响最大的是我硕士的导师巫和懋老师。巫老师从 Stanford 博士毕业，师从 Mordecai Kurz 和 Kenneth Arrow，对产业组织，博弈论，和金融经济学，都有很深的造诣。有缘的是在去北大读硕士之前，我已经修读了巫老师的课。在浙大读本科的时候，巫老师作为客座教授开了一门本科生的博弈论。正是这门课，让我对博弈论的应用产生了浓厚的兴趣，是我学术道路上一个非常重要的偶然（记得巫老师好像只在浙大开了一次课）。本科的最后一年，我参加了北大 CCER 的夏令营，希望获得保送读硕士的资格。夏令营的要求是在 CCER 的老师面前 present 一篇自己的习作。而我当时的选题就是博弈论在互助金融组织中的应用。这篇习作从选题到方法，都受到巫老师的博弈论课的启发。另外，我在 UT Austin 读博的时候，博弈论也成了我修读的两个主要 fields 之一。

我成功保送到 CCER 读研以后，对我而言，又出现了一个重要的偶然。巫老师辞去台湾大学的教职，正式来北大 CCER 任教。顺理成章，我非常希望选择他作为我的导师。记得当时想选他做导师的学生很多，我幸运的成为了他的学生。当时每周一次的组会，让我开始了系统的学术阅读和学术写作。我们在组会上 present 最新的产业组织和金融经济学的工作论文，后来也会 present 自己的习作。在巫老师的指导下，我开始研究 higher order beliefs 对金融市场的影响，这也是我第一次用英语写论文。记得打印出来的初稿上，有巫老师密密麻麻的修改，可惜几次搬家，论文的初稿早已不知去向。这不仅仅是我的硕士论文，对于我的学术生涯还有另外两重意义。首先，由于金融市场和 prediction market 的天然相似性，硕士阶段研读的各种关于金融市场异质信念的论文以及写硕士论文的经验是我日后开展一系列 prediction market 研究的起点。其次，这篇 higher order beliefs 的文章也是我在 UT Austin 成为 Andy (我的博士导师 Andrew Whinston) 学生的一个契机。在我的学术道路上，一系列新的偶然又在酝酿。

写于2022年5月7日

(3) First Year of Ph.D. Program

The first year of the Ph.D. program was quite enjoyable, and I quickly adapted to studying and living at UT Austin while making many new friends. Austin is a very livable city, and in 2009, the cost of living was still relatively low, with a good number of Chinese restaurants available. During the first year of the Ph.D. program, my focus was on economics, and I had never imagined that I would later enter the field of information systems (in fact, I didn't even know about this field at the time). Unlike the Ph.D. program in business schools, the first year of the economics Ph.D. program was entirely dedicated to coursework with no research involvement. There were many courses to take, including Advanced Microeconomics I and II, Advanced Macroeconomics I and II, Advanced Econometrics I and II, Mathematical Statistics, and Mathematical Economics. Each course was rigorous, with a significant amount of homework and three exams. Many classmates in the Ph.D. program complained about the frequent exams and assignments. In comparison, I seemed to handle them with ease, primarily due to the rigorous training I received during my master's program at CCER in Peking University. The master's program at CCER was modeled after the U.S. standard for economics Ph.D. programs, with identical course offerings and even similar textbooks (the difficulty and intensity of exams and assignments were also very similar). So, when I arrived in the United States, it felt like I was essentially relearning the courses, which made it relatively easier. Of course, this process itself was very beneficial and deepened my understanding of economics.

One major difference between the economics Ph.D. program and the business school Ph.D. program is the qualifying exams. Business school Ph.D. programs admit very few doctoral students, and the qualifying exams rarely eliminate students. The economics Ph.D. program follows a completely different business model and admits a large number of students, but the end-of-first-year qualifying exams eliminate nearly half of the students. For example, in our Ph.D. cohort, we had over 20 classmates in the first year, but less than 10 graduated. I don't have enough expertise to judge which model is superior or inferior, but personally, I found it difficult to accept the ruthless elimination mechanism of the Chicago economic tradition. Thanks to my rigorous training at CCER, I successfully passed the qualifying exams.

Because I felt comfortable with the coursework and qualifying exams in the first year, I continued to devote my time to the research on higher order beliefs in collaboration with Professor Wu. We submitted the paper to the Econometric Society World Congress, and it was accepted for presentation at the conference. In the summer of my first year in the Ph.D. program (summer 2010), I was scheduled to present this paper at the conference. Apparently, the department attached great importance to this conference and, unusually, provided me with a $1,000 travel fund. This was my first time attending a conference, and it was during this conference that another fortuitous encounter happened: I met Huaxia Rui, Andy's student and my future academic brother. Later, Huaxia introduced me to Andy, and I became Andy's student. Of course, these are stories for later.

Written on May 8, 2022

• （三）博士班第一年

博士班的第一年是相当惬意的，我很快适应了在 UT Austin 的学习和生活，结识了很多新的朋友。Austin 是一个非常适宜居住的城市，2009年的时候生活成本还很低，不错的中餐馆也挺多。在博士班的第一年里，我专注于经济学，从来没有想过以后会进入信息系统领域（甚至当时还不知道这个领域）。 与商学院的博士项目不同，经济学博士项目的第一年完全集中在修读课程上，几乎没有一点研究。上的课程很多，有高级微观一、二，高级宏观一、二，高级计量一、二，数理统计，和数理经济学。每门课都非常扎实，有大量作业和三次考试。博士班里的很多同学都抱怨频繁的考试和作业。比较而言，我显得游刃有余。主要得益于在北大 CCER 硕士阶段的严格训练。CCER 的硕士项目完全仿照美国标准的经济学博士项目，课程设置一模一样，就连所用的教科书也大致相同（考试和作业的难度和强度也非常类似）。所以到了美国后，我的感觉基本就是把课程重新再学一遍，相对轻松。当然这个过程本身还是很有益的，对经济学的理解大大加深。

经济学博士项目与商学院博士项目还有一个重大的不同，就是博士资格考（qualifying exams)。商学院博士项目招的博士生很少，一般博士资格考也很少会淘汰博士生。经济学博士项目的 business model 完全不同，会招很多博士生，而第一年末的博士资格考会淘汰近一半的博士生。比如我们的博士班第一年有20多名同学，毕业的时候不到10人。这两种模式孰优孰劣，我没有足够的能力去评判，不过从个人情感上说，Chicago 经济学传统的残酷的淘汰机制，我是很难接受的。同样是得益于我在 CCER 的基础，我的博士资格考顺利过关。

正因为第一年在课程和资格考试上游刃有余，我继续把时间投入到与巫老师合作的 higher order beliefs 的研究中。我们把文章投到了 Econometric Society World Congress。最后，文章被这个会议接受了，而我也要在博士班第一年暑假（2010年夏天），要去会议上 present 这篇文章。显然系里对这个会议比较重视，破天荒的给了我1000美元的 travel fund。这是我第一次参加会议，也正是这次会议，让另一个偶然发生：我遇到了 Andy 的学生，我后来的师兄芮华夏。之后，华夏把我介绍给 Andy，我也成为了 Andy 的学生。当然，这些都是后话了。

写于2022年5月8日

(4) Huaxia, a Mentor and Friend

Huaxia Rui is my academic brother, graduating two years ahead of me, and currently teaching at the University of Rochester. Meeting Huaxia was also a fortuitous event. Our first encounter was not at UT Austin but in Shanghai. As mentioned earlier, my paper with Professor Wu was accepted by the Econometric Society World Congress (ESWC), which happened to be held in Shanghai that year. So, in the summer of 2010, I planned to return to China to present at the conference. Huaxia also had a paper to present at the same conference. However, based on my habits at the time, it was challenging to get to know Huaxia. I was unaware of the social aspect of attending conferences, so I would typically present my paper, attend a few sessions of my interest, and consider it done without paying much attention to socializing with others. Recognizing my shortcomings at the time, I have always emphasized to my own Ph.D. students the importance of dedicating at least 50% of their time at conferences to socializing. What led to my acquaintance with Huaxia was an email he sent me on July 19, 2010 (yes, I can still find that email in my inbox). In the email, Huaxia mentioned seeing that I would be presenting at ESWC and suggested that since we were both at UT Austin, we should get to know each other at the conference. So, I went to the venue and had a lengthy conversation with Huaxia. He mentioned a series of research projects he and Andy were working on, which I found fascinating and opened up a new world for me. At that time, I didn't know that those were studies in the field of information systems, and Huaxia's presentation at the conference was also on research in financial economics (later published in the Journal of Financial Economics). However, at that moment, I asked Huaxia to help me get in touch with Andy so that when I returned to the United States, I could have a discussion with him about research. Eventually, this conversation led me, by chance, into the field of information systems.

During my Ph.D. studies, besides Andy, the person who had the greatest influence on my academic path was Huaxia. We collaborated on a total of four papers, from my earliest prediction market series to my job market paper, "WiFi Auction," and we were deeply involved in each of them. It is difficult to find a scholar who excels in both analytical modeling and empirical research as he does. I have learned a lot from him and continue to benefit from his teachings to this day.

Written on May 8, 2022

• （四）亦师亦友的华夏

芮华夏是我的师兄，早我两年毕业，现在 University of Rochester 执教。认识华夏也是一个偶然。我们第一次见面不是在 UT Austin, 而是在上海。正如前面所说，因为我和巫老师的文章被 Econometric Society World Congress (ESWC) 接受了，而那一年的会议是在上海召开，所以2010年的夏天，我准备回国去 present。华夏也有一篇论文要在同一个会议上 present。虽然如此，但是按照我当时开会的习惯，其实是很难认识华夏的。当时我对开会 social 的功能一无所知，所以基本就是讲完自己的 paper，然后听几个自己感兴趣的 session，就算结束了，根本不会注意和别人 social。正是因为认识到当时的不足，现在对我的博士生，我一直强调，开会至少要准备 50% 的时间去 social。促成我和华夏认识的是他在2010年7月19日发给我的一封邮件（是的，我在电子邮箱里还能找到这封邮件）。在邮件里，华夏说看到我会在 ESWC 上 present。因为我们都在 UT Austin, 大家应该在会上互相认识一下。于是，我就去会场与华夏聊了很久。他提到了他和 Andy 的一系列研究，我都觉得很有趣，对我来讲是打开了一个新天地。当然那时，我还不知道这些是信息系统领域的研究，而华夏在会上 present 的也是偏金融经济学的研究（后来发表在 Journal of Financial Economics上）。不过，当时我就请华夏帮我联系 Andy，等我回美国以后可以和 Andy 一起聊聊研究。后来，这一聊就把我偶然的带入了信息系统领域。

在读博阶段，对我的学术道路影响最大的，除了 Andy，就是华夏了。我和华夏一共合作了四篇论文，从我最早的 prediction market 系列一直到我的 job market paper, WiFi Auction，每一篇论文，我们都深度参与。很难找到像他那样 analytical modeling 和 empirical research 都做得非常好的学者。从他身上，我学到很多，一直受益至今。

写于2022年5月8日

(5) First Meeting with Andy

In the Fall of 2010, introduced by Huaxia, I went to the Center to meet Andy (the Center was the base for Andy and his students, and later I also had a small office there). Meeting Andy for the first time naturally made me very nervous, and I have forgotten the specifics of our conversation. The only thing I remember is that he gave me a printed working paper (authored by Yiling Chen from Harvard) and asked me to read it and see if I found it interesting. He also suggested that I present that paper to him the next time we met. The paper was about prediction markets, a concept I was unfamiliar with at the time, but I quickly became fascinated by it and delved into related literature. In fact, prediction markets are closely related to information aggregation in economics and finance, which is also connected to my collaboration with Professor Wu on higher-order beliefs in financial markets. Later, after several discussions with Andy, we realized that prediction market literature assumed participants were making predictions independently, which might not hold true in the prevalence of social media and social networks. Moreover, the structure of social networks could potentially influence participants' predictions and consequently affect the overall predictive performance of the prediction market. Following this line of thought, Andy, Huaxia, and I began conceptualizing our series of research on social media-based prediction markets. At that time, we had a great opportunity to showcase our idea to the industry. Austin hosts a prominent technology event called SXSW every March, and, for example, Twitter was first introduced to the public at SXSW. We were thrilled to have the chance to present our idea at SXSW 2011 and received many interesting suggestions from the industry, which proved to be very helpful.

Since the articles I collaborated with Andy on were progressing smoothly, I wanted to choose Andy as my advisor. At the time, although Andy was a professor in the business school, he also held joint appointments in the economics department and the computer science department. This meant that I didn't need to switch departments to choose Andy as my dissertation committee chair. So, I ultimately selected Andy as my Ph.D. advisor and ventured into the field of information systems. However, for convenience, I continued pursuing my Ph.D. degree in economics and did not transfer to the business school. Andy's style was to go to the Center from Monday to Friday and also on Sundays. He would have brief chats with students every day, ranging from a few minutes to an hour, depending on the progress of the research. I also went to the Center six days a week, which gave me ample time to chat with Andy, and our collaborative papers progressed rapidly. There was a period when every two to three days, Andy would ask me to briefly present the latest articles in economics or information systems. Looking back, that was the time when I made the most progress. Now, I also have my own Ph.D. students, but I feel it's challenging to replicate Andy's experience, as it's difficult to have six interactions with students every week. However, I strive to communicate with my students at least once a week and, if necessary, we can have multiple meetings.

Written on May 9, 2022

• （五）初见 Andy

2010年秋天，我在华夏的引见下，去 Center 见 Andy（Center 是 Andy 和他的学生们的基地，后来我也在 center 有一小间办公室）。第一次见 Andy，自然非常紧张，聊天的具体内容我都忘记了，唯一记得的是他给了我一篇打印好的工作论文（作者是 Harvard 的 Yiling Chen），让我回去看看感不感兴趣，同时下次见面的时候向他 present 这篇文章。这篇文章是关于 prediction market 的，我之前对这个概念相当陌生，但很快被它吸引住了，还翻出了很多相关文献。实际上，prediction market 与经济学和金融学里的 Information Aggregation 密切相关，也与我和巫老师合作的 higher order beliefs in financial markets 大有关系。后来，我和 Andy 又讨论了几次，发现 prediction market 的文献都假设了参与者是独立做出预测的，而这一点在 social media 和 social networks 盛行的情况下未必成立。更有趣的是 social network 的结构可能会影响参与者的预测，进而影响到 prediction market 的整体预测绩效。顺着这条思路，我和 Andy 还有华夏开始构思我们一系列关于 social media based prediction markets 的研究。这时我们得到一个很好的机会向工业界展示我们这个想法。Austin 每年3月有一个科技界的盛会 SXSW，比如 Twitter 最早就是在 SXSW 向公众展示的。当时，我和华夏很开心可以有机会在 2011年 SXSW 上 present 我们的想法，收到了工业界很多有趣的建议，非常有帮助。

由于跟 Andy 合作的文章进展的比较顺利，所以我想选 Andy 为我的导师。当时 Andy 虽然是商学院的教授，但他也同时是经济系和计算机系的兼职教授，这意味着我并不需要转系，就可以选择 Andy 做我的 dissertation committee chair。所以，我最终选择了 Andy 做我的博士论文导师，踏进了信息系统领域。但是为了方便，我继续在经济系修读博士学位，并没有转入商学院。Andy 的风格是周一到周五再加上周日，都会去 center，每天都会跟学生简单的聊一聊，短则十几分钟，长则一个小时，取决于文章的进展情况。我每周也是去 center 六天，这使得我和 Andy 有非常多的时间聊天，合作的文章进展得也比较快。有一段时间，Andy 每过两三天就会让我简单的 present 一篇最新的经济学或信息系统领域的文章，回头来看，我进步最快的也就是这段时间。现在，我自己也带博士生了，总感觉无法重复 Andy 的经验，很难有时间一周与学生交流六次。不过，我也会尽量做到每周至少交流一次，如果有需要，可以多次见面。

写于2022年5月9日

(6) Andy's Anecdotes

Since I had six chats with Andy every week, our conversations covered a wide range of topics. In addition to project progress, we discussed academic history, as well as movies and TV shows (such as Downton Abbey). Even now, every three or four weeks, Andy would call to discuss recent research projects, and we would exchange insights on stock market investments. Similar to me, Andy could be considered as someone who transitioned from economics to the field of information systems. His first job was in the Economics department at Yale (of course, in the 1960s, the field of information systems had not yet taken shape, and Andy was involved in its creation and development). Among his students, there were also notable figures in the field of economics, such as Preston McAfee. Preston McAfee is an expert in auctions and has served as the President of the American Economic Association (AEA) and the Chair of the Economics department at UT Austin (during my Ph.D. study, Preston McAfee came back to UT Austin to give a talk). Andy's early work in economics focused on cooperative games and optimization problems, and he has published in the top five economics journals. The ideas from this early work have been applied in our recent articles. Andy's connection to economics goes beyond that; his cousin, Michael Whinston, is one of the authors of the famous MWG (Microeconomic Theory) textbook.

In Andy's 60-year career, he has had numerous successful students. The Mathematics Genealogy Project alone records 108 Ph.D. students, and there are many more who are not included in this record. In 2015, a celebration of Andy's career was held at WISE/WITS at UT Dallas, where his academic journey was reviewed, including the process of founding the journal DSS and the Center.

Written on May 13, 2022

• （六）Andy 的趣事

因为每周和 Andy 聊六次，所以聊天的内容无所不包，除了项目的进展，还有学界的往事甚至电影和电视剧（比如唐顿庄园）。直到现在，每隔三、四周，Andy 都会打电话来聊聊最近的研究项目，我们也会交换股票市场投资的心得。与我类似，Andy 也可以说是从经济学转向了信息系统领域：他的第一份工作在 Yale 的 经济系 （当然，在60年代，信息系统领域尚未成型，Andy 参与了整个领域的创建和发展），他的学生中也有经济学领域的大家，比如 Preston McAfee。Preston McAfee 是 Auction 方面的巨匠，曾是 AEA 的主席和 UT Austin 经济系的系主任 （我读博士的时候，Preston McAfee 有回 UT Austin 给过 talk）。Andy 早期在经济学中的主要工作是合作博弈和优化问题，经济学期刊的 top five，他也全部发过。这些早期的思想在我和 Andy 最近的文章中都有应用。Andy 和经济学的渊源还不止于此，他的一个 cousin, Michael Whinston，就是著名的 MWG （最流行的微观理论教科书）中的作者 W。

Andy 入行60年来，桃李满天下，单单 Mathematics Genealogy Project 记录的博士生已经有108个，还有很多不在这个记录之内的。2015年在 UT Dallas 召开的 WISE/WITS上举办了一个 Andy 的庆祝会，回顾了他的学术生涯，包括创办期刊 DSS 和 Center 的过程。

写于2022年5月13日

(7) Second Year of Ph.D. Program

In the second year of the economics Ph.D. program, I had to choose two fields and take field courses. Due to my interest in game theory, I chose game theory as my first field. As for the second field, I noticed the similarities between research in information systems and industrial organization (IO) theory in economics. Therefore, I selected industrial organization as my second field. In the area of game theory, I took two different game theory courses taught by Max Stinchcombe and Tom Wiseman. Later, I also took a course on Mechanism Design. Max Stinchcombe is a brilliant scholar, but his lectures were quite challenging to follow due to their leaps in logic. Tom Wiseman's lectures were very clear, and he explained each step of the model derivations thoroughly (which reminded me of the clarity of Professor Hongbin Cai's lectures on Information Economics and Professor Liutang Gong's lectures on Dynamic Optimization when I was pursuing my master's degree). Tom's discussions on Network Games and Observational Learning were particularly inspiring for my later research on social media-based prediction markets and observational learning in social networks. Later on, both Max and Tom became members of my dissertation committee. In the field of industrial organization, I took two courses. The first course covered structural models in industrial organization, focusing on the empirical estimation of single-agent optimization problems. The second course was divided into two parts: the first half covered empirical estimation of dynamic games in structural models, and the second half focused on theoretical models in industrial organization, with an emphasis on two-sided platforms. Soon after completing these courses, I realized that structural estimation had also gained popularity in the field of information systems, which made me feel a strong connection and strengthened my determination to switch to the field of information systems.

At the end of the second year of the Ph.D. program, we were required to submit a second-year paper as an assessment. The department's policy was for each Ph.D. student to submit an initial draft of their second-year paper, and then the department would provide an anonymous review report written by faculty members in relevant fields, simulating a journal review process. We would revise our papers based on the review reports and submit the final versions. I always felt that this policy was beneficial to me. As I was particularly interested in Network Games from Tom Wiseman's course, I attempted to apply Network Games to the problems I was interested in. At this time, I had already started researching prediction markets under Andy's guidance, and we were also interested in integrating social media and social networks into prediction markets, but we didn't have a handy tool for it. The emergence of Network Games provided that handy tool, which became another serendipitous event in my academic journey. In my second-year paper, I applied Network Games to prediction markets and constructed a game-theoretic model. The anonymous reviewing faculty showed great interest in this model and provided many constructive comments (although to this day, I don't know who the anonymous reviewers were, but I suspect it was Tom Wiseman). Finally, I submitted the revised second-year paper: "Information Acquisition and Exchange in Social Networks." In this paper, I developed a game-theoretic framework for information exchange in social networks and discussed one particular application: Network-Embedded Prediction Markets, especially the impact of different social network structures on the prediction performance of markets. It also served as the theoretical foundation for my 2014 publication in JMIS titled "Effects of Social Networks on Prediction Markets: Examination in a Controlled Experiment."

I found the anonymous review report for that second-year paper from back then. Here is the first paragraph:

”This paper uses the framework of social networks to examine the incentives to acquire independent information. The social network (i.e. links between agents) is taken as exogenous, and information acquired independently is assumed to be transmitted immediately and truthfully to all direct neighbors. The author derives formal results on the extent of the free-rider problem in this framework. To potentially mitigate the free-rider problem, the author adds a prediction market to the model. Agent-based simulations are employed to show how the prediction market improves efficiency. Overall, the paper is well-written and satisfies the criteria for a second-year paper.“

Written on May 15, 2022

• （七）博士班二年级

在经济系的博士班二年级，我需要选择两个 fields，修 field courses。由于我一直对 game theory 比较感兴趣，所以我选择了 game theory 作为我的第一个 field。至于第二个 field，因为我看到很多信息系统的研究和经济学中的产业组织（IO）理论很类似，所以我选择了产业组织作为我的第二个 field。在 game theory 方面，我先后修读了 Max Stinchcombe 和 Tom Wiseman 的两门不同方面的 game theory 课程，后来又修了一门 Mechanism Design。Max Stinchcombe 是一位天才的学者，不过上课跳跃性很大，我很难完全跟得上他的思路。Tom Wiseman 的课上得非常清晰，模型的每一步推导都解释得很透彻（让我想起了读硕士时在光华上蔡洪滨老师的信息经济学和龚六堂老师的动态优化，他们都讲的非常清楚）。Tom 课上讲的 Network Game 和 Observational Learning 对我后来关于 social media based prediction market 和 observational learning in social networks 的研究都很有启发。后来，Max 和 Tom 也都是我的 dissertation committee members。在 IO 方面，我修了两门课，第一门讲业组织里的 structual model，主要集中在 single agent optimization problems 的实证估计。第二门课分两部分，前半部分讲 structual model 中 dynamic game 的实证估计，后半部分讲产业组织里的理论模型，主要集中在 two-sided platforms 上。刚上完这些课，就发现信息系统领域里也掀起了 structual estimation 的热潮，让我觉得格外亲切，也坚定了我转向信息系统领域的决心。

博士班第二年末需要提交一篇 second year paper 作为考核。系里的制度是每一个博士生提交一篇 second year paper 的初稿，然后系里会找相关领域的老师模仿期刊评审写一篇匿名的审稿报告。我们会根据审稿报告进行修改，提交最终版本。我一直觉得这一制度很好，让我受益良多。因为在 Tom Wiseman 的课上，我对 Network Game 非常感兴趣，所以我尝试着把 Network Game 应用到我感兴趣的问题上。这时，我已经在 Andy 的指导下开始研究 prediction market 了，我们也想尝试着把 social media 和 social network 融入到 prediction market 中，但还没有顺手的工具。而 Network Game 的出现，正好提供了这个顺手的工具，这也成了我学术道路上的另一个偶然。在 second year paper 里，我把 Network Game 应用到 prediction market 中，构建了一个 game-theoretic model。匿名评审的老师对这个模型也很感兴趣，提出了很多建设性的评论（虽然至今我也不知道匿名评审的老师是谁，但我猜测是 Tom Wiseman）。最后，我提交了修改后的 second year paper: Information Acquisition and Exchange in Social Networks。在这篇文章里，我建立了一个 social network 下信息交换的博弈理论框架，并讨论了其中的一个 application: Network-Embedded Prediction Market，尤其是不同 social network structures 对 prediction market 的预测绩效的影响。它也成为了我2014年在 JMIS 上发表的文章 "Effects of Social Networks on Prediction Markets: Examination in a Controlled Experiment“ 的理论部分的雏形。

我找出了当年 second year paper 的匿名审稿报告，这是其中的第一段：

”This paper uses the framework of social networks to examine the incentives to acquire independent information. The social network (i.e. links between agents) is taken as exogenous, and information acquired independently is assumed to be transmitted immediately and truthfully to all direct neighbors. The author derives formal results on the extent of the free-rider problem in this framework. To potentially mitigate the free-rider problem, the author adds a prediction market to the model. Agent-based simulations are employed to show how the prediction market improves efficiency. Overall, the paper is well-written and satisfies the criteria for a second-year paper.“

写于2022年5月15日

(8) First Attendance at IS Conference

In December 2011, I attended the ICIS conference, my first major conference in the field of information systems. Of course, this was not my first academic conference experience. As mentioned before, I attended ESWC in the summer of 2010. Then, in the summer of 2011, I also participated in the Econometric Society North American Summer Meeting held in St. Louis, where I presented my second-year paper. That conference was hosted by the Economics Department of Washington University in St. Louis. At that time, my CCER classmate Feng Dong (now a professor at Tsinghua University School of Economics and Management) was pursuing his Ph.D. at Washington University in St. Louis, and he happened to be back in China, so I stayed in the apartment he rented. I also met another CCER classmate, Wei Wang (now a professor at the University of International Business and Economics).

To get back on track, in December 2011, ICIS was held in Shanghai, at the same venue as the ESWC conference I attended in 2010, which was the Shanghai International Convention Center. It was quite a coincidence. At that time, Huaxia was also on the job market, and we shared a hotel room. The paper I presented at ICIS was a Twitter-based prediction market, co-authored by me, Huaxia, and Andy. It was an application of the theory from my second-year paper. Since I didn't have much experience with presentations, I was very nervous. I remember being in the same session as Lynn Wu from Wharton, and she gave an excellent presentation. I learned a lot from her. As it was my first time attending an IS conference, I focused all my energy on the presentation and didn't have much time for socializing (which is something I now consider a big mistake when it comes to conferences). Fortunately, Andy organized a gathering for his students, and I got to know many academic siblings. After ICIS, I also attended the Chinese Economics Annual Conference. I submitted my second-year paper to that conference, which happened to be held at Shanghai University of Finance and Economics. It was efficient to attend two conferences during one trip to Shanghai.

Written on May 17, 2022

• （八）第一次参见 IS 会议

在2011年12月，我第一次参加信息系统领域的大会 ICIS。当然，这不是我第一次参加学术会议。之前说过，我在2010年夏天参加了 ESWC。之后，在2011年夏天，我还参加了在 St. Louis 召开的 Econometric Society North American Summer Meeting，present 了我的 second year paper。那次会议是由 Washington University in St. Louis 的经济系主办的。当时，我 CCER 的同学董丰（现在清华经管执教）正在 Washington University in St. Louis 读博士，而他又正好回国，所以我就住在他租的公寓里。同时，还见到了我 CCER 的另一名同学王暐（现在对外经贸执教）。

言归正传，2011年12月的 ICIS 在上海举办，和我2010年参加的 ESWC 是同一个会址，都是上海国际会展中心，也算是个巧合。那时正好华夏上 job market，我跟他 share 一个旅馆房间。我在 ICIS 上 present 的文章是我和华夏还有 Andy 合写的 Twitter-based prediction market，是我的 second year paper 里的理论的一个应用。当时没有太多的 presentation 经验，所以非常紧张。记得和我同在一个 session 的是 Wharton 的 Lynn Wu，她 present 得非常好，我学到了很多东西。第一次参加 IS 的会议，我把全部精力都放到了 presentation 上，自然无暇 social （这也是现在我对我的博士生所说的开会的大忌）。好在 Andy 召集门下弟子聚餐，我也认识了不少师兄师姐。ICIS 结束以后，我又转战中国经济学年会。当时，我把我的 second year paper 投到了年会，而年会恰好在上海财经大学举行。我去上海一次，可以参加两个会议，自然感到非常高效。

写于2022年5月17日

(9) Experimental Economics

In my second-year paper, I completed the game theory model for the social network-based prediction market. However, I always wanted to find some empirical evidence to test the theoretical model. During my undergraduate years at Zhejiang University, I took Professor Qiren Zhou's course on New Institutional Economics. In that class, Professor Zhou emphasized the difference between real-world economics and chalkboard economics and recommended Steven Cheung's "Economic Explanations." With great enthusiasm, I finished reading all three volumes of "Economic Explanations" and found it captivating to explore the real world. In the past two years, a new five-volume edition of "Economic Explanations" was released, and I bought it to revisit the feeling of the past. To validate the theoretical model with real-world evidence, I decided to leverage experimental economics.

At the time, I was in the Economics Department at UT Austin, and the department chair, Dale Stahl, was an expert in game theory and experimental economics. He was also a close friend and long-term collaborator of Andy. He had a dedicated experimental laboratory on the fourth floor of the economics department for conducting experiments. So, I visited his office to seek advice on the specific experimental design for the prediction market. He provided detailed suggestions and guidance on every aspect of the experiment (later, Dale Stahl also became a member of my dissertation committee). Based on his advice, Huaxia and I designed the computer system interface for the experiment. Later, Andy helped me apply for $1,000 as funding for the experiment from the university (I still remember the scene of going to the UT Main Tower with Andy to collect the $1,000 check). With everything in place, we just needed the final push. Speaking of the final push, one advantage of the Economics Department at UT Austin was that any Ph.D. student could reserve the economics department's fourth-floor laboratory. So, I booked the lab and posted promotional posters recruiting participants for the experiment (they were plastered all over the entrances of the economics department and the business school). After that, I calmly waited for the day of the experiment.

Overall, the experiment went smoothly, and we quickly reached the target number of participants. It was all hands on deck at that time. I was responsible for introducing the experiment's purpose and specific steps to the participants, Huaxia handled the data entry into the database, and my wife (who was my Ph.D. classmate) calculated the participants' rewards based on their performance. Afterward, I analyzed the experimental data that provided empirical evidence for my JMIS article titled "Effects of Social Networks on Prediction Markets: Examination in a Controlled Experiment." It also formed the basis for the second chapter of my doctoral thesis.

This experience of conducting the experiment was very helpful for my subsequent research. Over the years, experimental methods have been widely applied in the field of information systems, and I have used various experimental methods in several of my later research articles. This complete experimental experience allowed me to understand many details of experimental design clearly. Another benefit is that I introduced experiments into my teaching, which has been well-received by students. I have always believed that research and teaching should complement each other: good research should be applicable to teaching, and likewise, good teaching can inspire research. Speaking of teaching, there is recent news that brings joy. In the Poets & Quants 2022 Best 40-Under-40 MBA Professors selection, three scholars with connections to me won awards: Lizhen Xu, Andy's student and my academic brother (currently teaching at the Georgia Institute of Technology); Yiting Deng, my fellow master's classmate at CCER in Peking University (now teaching at University College London); and Professor Wu's daughter, Houwei Wu (currently teaching at the China Europe International Business School).

Written on May 23, 2022

• （九）实验经济学

在我的 second year paper 里，我写完了 social network based prediction market 的博弈理论模型。然而，我心中一直想找到一些经验证据来检验理论模型。在浙大本科的时候，就上了周其仁老师的新制度经济学。在课上，周老师强调了真实世界的经济学和黑板经济学的区别，并推荐了张五常的《经济解释》。当年一鼓作气读完了《经济解释》的三卷本，觉得酣畅淋漓，真实世界引人入胜。前两年，《经济解释》出了新修的五卷本，我又买了回来，重温一下当年的感觉。为了用真实世界来验证理论模型，我决定借助实验经济学。

当时在 UT Austin 的经济系，系主任 Dale Stahl 是博弈论和实验经济学的专家，也是 Andy 的好朋友和长期合作者。他在经济系四楼有一个专门的实验室，用来做实验。于是，我上他的办公室请教 prediction market 的具体的实验设计。实验的每一个细节，他都给了我详细的建议和指导（之后，Dale Stahl 也成了我 dissertation committee 中的一员）。根据这些建议，我和华夏一起设计了实验的电脑系统界面。后来，Andy 又帮我向学校申请了 1000 美元作为实验的经费（我还记得和 Andy 一起去 UT Main Tower 拿1000 元 check 的情景）。这样一来，就万事具备，只欠东风了。说到东风，UT Austin 的经济系有一个好处， 任何博士生都可以预约经济系四楼的实验室。所以，我预定了实验室并张贴了招募实验参加者的宣传海报（贴满了经济系和商学院的出入口）之后，就气定神闲的等着实验那天的到来。

总的来说，实验进行得相当顺利，实验参加者很快达到了预定的人数。当时，全体总动员，我负责向实验参加者介绍实验目的和具体步骤，华夏负责把实验数据导入数据库，我太太（她是我的博士班同学）按照实验参加者的 performance，给他们报酬。之后，我分析了实验数据，于是我的 JMIS 文章 "Effects of Social Networks on Prediction Markets: Examination in a Controlled Experiment“ 有了实验的证据。我的博士论文的第二章也是由此发展而来的。

这次做实验的经历，对我之后的研究很有帮助。这些年来，实验方法在信息系统领域广泛应用，我之后的好几篇文章也用了不同的实验方法。这次完整的实验经历让我清楚的了解到实验设计中的很多细节。另一个好处是在我的教学中，我也引入了实验，受到了学生的欢迎。我一直觉得研究和教学应该是相辅相成的：好的研究应该可以应用到教学上，同样的，好的教学会给研究带来灵感。说到教学，最近有件令人高兴的事。在 Poets & Quants 2022 Best 40-Under-40 MBA Professors 的评选中，有三位和我有渊源的学者获奖：Andy 的学生，我的师兄 Lizhen Xu（现在 Georgia Institute of Technology 执教）；我在北大 CCER 的硕士同学邓一婷（现在 University College London 执教）；以及巫老师的女儿巫厚玮（现在中欧国际工商学院执教）。

写于2022年5月23日

(10) Friendship beyond Generations: Ira

I was deeply saddened to hear the news of my colleague Ira Horowitz's passing this morning. Ira was not only my colleague and mentor but also a dear friend. He had been teaching at UF ISOM for a long time, while his wife Nancy taught in the UF Economics department. When I joined UF in 2014, Ira was already retired, but he remained very active in the department. He would frequently visit the office, actively participate in seminars, and as a result, we quickly became acquainted.

I still remember our first meeting in February 2014 when I came to UF for a job talk. I immediately sensed a resemblance between him and Andy. Despite his age, Ira was intellectually vibrant and sharp. Perhaps it was because we both had an economics background that we connected so well. Ira once mentioned that he had the opportunity to become the chair of the Economics department at UT Austin in the 80s or 90s. Unfortunately, for various reasons, it didn't work out. Unlike most retired professors, Ira never truly retired. Before the pandemic, he would visit the office two or three times a week, and naturally, we would often cross paths. Whenever we bumped into each other, we would make plans to have lunch together. He would often take me to unique and hard-to-find places to eat. This may be why Andy referred to him as a foodie. Andy mentioned that he had a pleasant conversation with Ira during his visit to UF in the early 90s when Ira was the department chair.

Ira was approachable and helped me revise my papers multiple times. His contributions were instrumental in the publication of my article titled "Platform Policies and Sellers' Competition in Agency Selling in the Presence of Online Quality Misrepresentation" in JMIS. Each revision would be filled with meticulously tracked changes, accompanied by explanations of why certain modifications were necessary. My current writing style owes a great deal to his guidance. Ira's research mainly focused on decision-making and sports economics. Whether it was theoretical research or empirical analysis, he excelled in both areas. For example, his articles "Some Aspects of the Effects of the Regional Distribution of Scientific Talent on Regional Economic Activity" and "Estimation Uncertainty and Optimal Advertising Decisions," published in Management Science in 1966 and 1989 respectively, delved into deep theoretical exploration while incorporating rich empirical evidence. He once told me that he was most proud of being able to publish single-authored articles even after retiring.

When my wife fell ill, as soon as Ira found out, he and Nancy came to our house to provide us with various forms of support. Because Nancy had experienced a similar illness before, their shared experiences and encouragement greatly helped alleviate our mental burdens. Since then, Ira continued to keep abreast of our situation and showed genuine concern. Due to the pandemic, we rarely saw each other in the past couple of years, but we frequently exchanged emails. Every Chinese New Year, he would send me an email with the subject line "Kung Hei Fat Choi" to extend his wishes. About a year and a half ago, he mentioned that due to his declining health, he moved into a nursing home, but he would still go to the office on weekends. My colleague Shubho mentioned that when he visited Ira earlier this month, Ira asked about me and my wife. Regrettably, unexpected events always occur.

When I received tenure, Ira sent me an email with the following message: 

“Liangfei, this is really a Congratulations message to the tenured faculty of ISOM for having the good sense to confirm your status as the brightest and best young star to grace our department in years. It's a privilege to call you my friend and I look forward to seeing you and shaking your hand in the fall.

More importantly, I hope that all is well and proceeding accordingly with Bei.

Warmest regards, IRA”

I want to say to Ira: I am so lucky to have you as my friend and mentor. 

Written on May 29, 2022

• （十）我的忘年交 Ira 

今早听闻我的同事 Ira Horowitz 去逝，心里非常难受。Ira 是我的同事，师长，也是我的朋友。他在 UF ISOM 执教了很长时间，他的太太 Nancy 在 UF 的经济系执教。我2014年加入 UF 的时候，Ira 已经退休，但他还是在系里非常活跃，经常来办公室，也积极参加各种 seminar，所以我们很快就变熟了。记得第一次和他见面是2014年2月，我来 UF 给 job talk。当时就觉得他跟 Andy 有点类似，年纪虽大，但思维非常活跃。也许是因为我们都有经济学的背景，所以非常投契。Ira 说过他在80，90年代的时候曾有机会去 UT Austin 的经济系当系主任，不过由于种种原因，未能成行。与一般的退休老师不同，Ira 可以说是退而不休。在疫情之前，他基本每周去办公室两三次，大家自然也就常常见面。遇到之后，我们会约定中午一起吃饭。他经常会带我去一些很有特色，但又不容易找到的地方吃饭。这可能是 Andy 称他为美食家（Foodie）的原因。Andy 说他在90年代初去过 UF 一次，而那时 Ira 是系主任，他和 Ira 相谈甚欢。

Ira 平易近人，多次帮我修改 paper。我在 JMIS 上发表的 “Platform Policies and Sellers’ Competition in Agency Selling in the Presence of Online Quality Misrepresentation" 就得益于他的多次修改。每次修改总是密密麻麻的 tracked changes，然后和我解释为什么要这样修改。我现在的写作风格在很大程度也得益于他的指导。Ira 的研究主要集中在 decision making 和 sports economics。无论是理论研究还是实证检验，他都信手拈来。比如他1966年和1989年发表在 Management Science 上的两篇文章 “Some Aspects of the Effects of the Regional Distribution of Scientific Talent on Regional Economic Activity” 和 “Estimation Uncertainty and Optimal Advertising Decisions” 既有深入的理论探讨，也有丰富的经验证据。他跟我说他最自豪的事情就是退休之后，还可以发表 single author 的文章。

在我太太患病的时候，Ira 一知道，就和 Nancy 来我们家，给我们各种支持。因为 Nancy 之前也有过同样的病患，所以他们的经验和鼓励给了我们很大的帮助，减轻了我们心理上的负担。之后，他也一直关切的了解各种情况。由于疫情原因，这两年，我们很少见面，不过我们常常交换电子邮件。每年春节，他总是会发给我恭喜发财为 title（Kung Hei Fat Choi）的邮件。一年半前，他提到由于身体大不如前，搬入了 nursing home，不过周六周日，还有时会去办公室。我的同事 Shubho 说他这个月初去看 Ira 的时候，Ira 还问起了我和我太太。遗憾的事情总是突然发生。

在我拿到 tenure 的时候，Ira 写了一封电子邮件给我：

“Liangfei, this is really a Congratulations message to the tenured faculty of ISOM for having the good sense to confirm your status as the brightest and best young star to grace our department in years. It's a privilege to call you my friend and I look forward to seeing you and shaking your hand in the fall.

More importantly, I hope that all is well and proceeding accordingly with Bei.

Warmest regards, IRA”

我想对 Ira 说的是：I am so lucky to have you as my friend and mentor. 

写于2022年5月29日

(11) Center's Academic Siblings

One prominent characteristic of studying under Andy's guidance is the presence of many academic siblings in the center, and Andy encourages everyone to collaborate. We often visit the center together, have meals, and engage in numerous conversations, naturally leading to more collaborations. I remember Andy particularly enjoying a Chinese restaurant in Austin called Din Ho, so we would often gather there.

Most of our collaborations among academic siblings often arise from casual conversations, while meticulously planned collaborations are rare occurrences. Among the fellow students with whom I interacted more frequently at the center were Huaxia, Zhan Shi (now teaching at Arizona State University), Qian Tang (now teaching at Singapore Management University), Gene Moo Lee (now teaching at the University of British Columbia), He Shu (now teaching at the University of Connecticut and will join us at UF in the fall as a colleague), Shun-Yang Lee (now teaching at Northeastern University), and Ying-Yu Chen. I have had numerous collaborations with them as well.

I am grateful for the inclusive nature of the information systems field. Even though many of my academic siblings hold doctoral degrees in fields other than information systems, we have been able to smoothly enter this field and gradually establish ourselves. For instance, Zhan Shi, Shu He, myself, and our later colleague, Yingda Zhai (now teaching at the National University of Singapore), all hold doctorates in economics, while Gene Moo Lee holds a doctorate in computer science. This is not commonly seen in other fields.

Written on June 4, 2022

• （十一）Center 的师兄弟姐妹

在 Andy 门下学习的一大特点是 center 里有很多师兄弟姐妹，Andy 也鼓励大家一起合作。大家经常去 center，也常常一起吃饭，聊天的机会很多，合作自然也就多了。记得 Andy 非常喜欢 Austin 的一家叫顶好（Din Ho）的中餐馆，所以大家常常去那里聚餐。

很多时候，我们师兄弟姐妹的合作是基于偶然的闲聊，而事先有着周密计划的合作则是极少数的情况。在 Center 和我交流比较多的同门有华夏，施展 (现在 Arizona State University 执教），唐倩（现在 Singapore Management University 执教），Gene Moo Lee （现在 University of British Columbia 执教），何舒（现在 University of Connecticut 执教，秋季会来我们 UF，和我成为同事)， 李舜扬（现在 Northeastern University 执教），和陈盈佑。我和他们之间也都有不少合作。

我很感恩信息系统领域的兼容并包。即使我们师兄弟姐妹中很多人的博士学位并不是信息系统领域的，我们也可以顺利进入这个领域，并逐渐生根。比如施展，何舒，我，还有再后来的师弟翟颖达（现在 National University of Singapore 执教）都是经济学的博士；Gene Moo Lee 是 computer science 的博士。这在其它领域并不多见。

写于2022年6月4日

(12) Collaborating with Scholars Outside of Advisors

Ph.D. students often ask me how to develop academic connections with other scholars (non-advisors) during their doctoral studies. I am happy to share my personal experience: my first collaboration with a non-advisor scholar was more of a fortunate accident than a deliberate plan. During my Ph.D., many scholars in the information systems field would come to UT Austin to give talks, and after their presentations, they would have roundtable discussions with Ph.D. students (in 2018, I had the honor of returning to UT Austin to give a talk myself). For me, this was a valuable opportunity because I wasn't a Ph.D. student in the information systems field. I wanted to take advantage of this chance to get to know scholars in the information systems field and introduce myself to them. I attended almost every roundtable discussion. And it was during one of these roundtables that my first collaboration with a non-advisor scholar began, completely by chance.

On April 6, 2012, Sarah Rice (who was teaching at the University of Connecticut at the time and now teaches at Texas A&M) gave a talk at UT Austin about her research on social ties. Her study employed experimental economics to investigate social ties on Facebook. Coincidentally, I was also using experimental economics to study social network-based prediction markets, so I was very interested in her research. During a subsequent roundtable discussion, I had a conversation with Sarah about possible extensions to her current work. After the conversation, I went home without realizing that it could be a potential collaboration opportunity. The next day, Andy forwarded an email from Sarah to me:

"Hi Andy,

One of your students talked about a relatively clever idea for an experiment yesterday in the round table, and I wanted to follow up with him, but I can't find him on the PhD student website. He is working with you; I think he said he's in his third year, a very skinny Chinese guy with glasses. Not sure if that helps, but if you know who I'm talking to, maybe I could talk with him on Monday too? S."

Afterward, I contacted Sarah, and she introduced me to her collaborator, Ravi Bapna. From there, we began using experimental economics to study social ties. Eventually, this led to my publication in MIS Quarterly (MISQ) in 2017 titled "Repeated Interactions Versus Social Ties: Quantifying the Economic Value of Trust, Forgiveness, and Reputation Using a Field Experiment." It's hard to imagine that an MISQ publication could stem from a chance encounter at a roundtable discussion, but that's how life unfolds.

Written on June 12, 2022

• （十二）与非导师的学者合作

经常有博士生问我，在读博期间，应该如何发展同其他学者（非导师）的学术联系。我也很乐意分享个人的经历：我第一次与非导师的学者的合作是一个偶然多于必然的过程。读博的时候，很多信息系统领域的学者会来 UT Austin 给 talk，讲完文章以后都会和博士生有个 round table（在2018年，我很荣幸的回 UT Austin 给了一个 talk)。对我而言，这是个难得的机会。因为我本身不是信息系统领域的博士生，所以很想借这个机会了解信息系统领域的学者，把自己介绍给他们。几乎所有的 round table，我都参加了。而我第一次与非导师的学者的合作也是起源于一个偶然的 round table。

2012年4月6日，Sarah Rice（当时在 University of Connecticut 执教，现在 Texas A&M 执教）在 UT Austin 讲了她的一篇关于 social tie 的文章。那篇文章是采用实验经济学的办法来研究 Facebook 的 social tie。当时我正好也在用实验经济学的方法来研究 social network based prediction market，所以对她的研究非常感兴趣。在之后的 round table 中，我和 Sarah 聊了一些她现在文章的可能扩展。聊完之后，我就回家了，并没有意识到这是一次可能的合作机会。第二天，Andy 转发给我了一封 Sarah 发给他的电子邮件：

”Hi Andy, 

One of your students talked about a relatively clever idea for an experiment yesterday in the round table and I wanted to follow up with him but I can't find him on the PhD student website.  He is working with you, I think he said he's in his third year, very skinny Chinese guy with glasses.  Not sure if that helps but if you know who I'm talking to maybe I could talk with him on Monday too?  S.“

之后，我就和 Sarah 联系，Sarah 又介绍了她的合作者 Ravi Bapna 给我认识。接着，我们就开始用实验经济学的方法来研究 social tie 了。最后，也就有了我2017年在 MISQ 上发表的文章 ”Repeated Interactions Versus Social Ties: Quantifying the Economic Value of Trust, Forgiveness, and Reputation Using a Field Experiment"。很难想象一篇 MISQ 会起源于一个偶然的 round table，然而世事就是如此。

写于2022年6月12日

(13) What is IS

When I first transitioned from economics to the information systems field, I had a significant question. I noticed that top-tier journals in the information systems field had many discussions about "What is IS." It was hard to imagine such a foundational question being discussed in top-tier journals in economics. Later, when I started submitting articles to information systems journals, I found that reviewers would often ask whether the article addressed an IS-related question. Initially, I struggled to understand why everyone was so concerned about this issue. In economics articles, it is rare to extensively argue whether a research topic is within the realm of economics. However, in the information systems field, such examples are not uncommon. Over time, I gradually understood the logic behind it. Economics has been around for hundreds of years and has developed to the point where almost every university has an economics department. In contrast, the information systems field has only been developing for a few decades, and many prestigious universities do not have an information systems program. In this context, emphasizing "What is IS" is a natural progression.

Although there are many discussions about "What is IS," it seems that "What is Economics of IS" is mentioned less frequently. When I first entered the information systems field, many friends from economics departments asked me about my research, and my answer was "Economics of IS." Few friends from economics departments understood what "Economics of IS" meant. Even now, my own understanding of this question may still be incomplete. From my perspective, the core topic of information economics (including game theory and mechanism design) is studying how "Who Knows What" influences "Who Gets What." Alvin Roth, the Nobel laureate in 2012, wrote a book called "Who Gets What — And Why: The New Economics of Matchmaking and Market Design," which provides a straightforward introduction to market design and is one of my favorite books. Following this line of thought, my understanding of the Economics of IS is studying how IT changes "Who Knows What," thereby further influencing "Who Gets What." Many of my research studies are conducted within this paradigm. For example, my early research on social network-based prediction markets examined how social media changes information flow, consequently affecting predictive performance. In this case, social media makes information flow between friends more relevant, thereby altering "Who Knows What" and subsequently influencing "Who Gets What." Recently, many scholars have been studying the relationship between social media and political polarization, and the mechanisms involved are quite similar: social media makes information flow between friends more relevant, thereby altering "Who Knows What" and subsequently influencing "Who Gets What." In more general terms, in empirical research within the Economics of IS, the (IT-induced) "Who Knows What" often serves as an independent variable, while "Who Gets What" tends to be the dependent variable. Similarly, in analytical models within the Economics of IS, the (IT-induced) "Who Knows What" is often an exogenous variable, while "Who Gets What" is frequently an endogenous variable.

Written on June 17, 2022

• （十三）What is IS

刚刚从经济学转入信息系统领域的时候，我有一个很大的疑问。我发现在信息系统领域的顶级期刊上有很多关于 What is IS 的讨论。很难想象在经济学的顶级期刊上会讨论这样的原问题。后来向信息系统领域的期刊投稿，也发现审稿人经常会问这篇文章讨论的是不是一个 IS 的问题。一开始，我很难理解为什么大家这么关心这个问题。在经济学的文章里，很少有人会长篇累牍的论证某个研究论题是不是一个经济学问题。而在信息系统领域，这样的例子却并不罕见。后来，我逐渐明白了其中的逻辑。经济学诞生了几百年，发展到现在，几乎每个大学都有经济系。而信息系统领域才发展了几十年，不少名校都没有信息系统专业。在这个背景下，强调 What is IS，也是顺理成章的。

虽然有很多关于 What is IS 的讨论，但似乎 What is Economics of IS 被提到的较少。我刚刚转入信息系统领域的时候，很多经济系的朋友问我在研究什么，我的回答是 Economics of IS。几乎没有经济系的朋友明白什么是 Economics of IS。直到现在，我自己对这个问题的理解可能也是一知半解。在我看来，information economics（包括 game theory 和 mechanism design）的核心话题是研究 Who Knows What 是如何影响 Who Gets What 的。2012年的诺奖得主 Alvin Roth 写过一本书，Who Gets What — And Why: The New Economics of Matchmaking and Market Design，深入浅出的介绍 market design，也是我最喜欢的书之一。沿着这个思路，我理解的 Economics of IS 是研究 IT 是如何改变 Who Knows What，从而进一步影响 Who Gets What 的。我的很多研究也是在这个范式下展开的。比如我早期研究的 social network based prediction market，其实就是看 social media 是如何改变 information flow，从而进一步影响预测绩效的。在这里，social media 使得朋友之间的 information flow 变得更加相关，从而改变了 Who Knows What，然后进一步影响了 Who Gets What。最近不少学者都在研究 social media 和 political polarization 的关系，其中的机制也大致相似：social media 使得朋友之间的 information flow 变得更加相关，从而改变了 Who Knows What，然后进一步影响了 Who Gets What。更一般的，在 Economics of IS 的实证研究中，（因为 IT 而改变的）Who Knows What 常常是自变量，而 Who Gets What 则多是因变量。同样的，在 Economics of IS 的 analytical model 中，（因为 IT 而改变的）Who Knows What 多是外生变量，而 Who Gets What 则常常是内生变量。

写于2022年6月17日

(14) MIS 301

Since deciding to transition to IS (around 2011), I started exploring how to pursue a tenure-track position in a business school. Besides the differences in research, there is a significant distinction between business schools and economics departments: business schools place greater emphasis on teaching, and most doctoral students preparing for the job market would have previous independent teaching experience. With Andy's assistance, I eventually obtained an opportunity for independent teaching. In the spring of 2013, I taught MIS 301: Introduction to Information Technology Management, a foundational undergraduate course in IS, at the McCombs School of Business, UT Austin. It was my first experience teaching a course independently.

Given the rarity of this opportunity, I made extensive preparations. In the semester preceding my teaching assignment, I attended lectures given by all the instructors teaching MIS 301 in the business school. I am particularly grateful to Wen Wen for her excellent teaching when she had just joined UT Austin as an Assistant Professor. I learned many teaching techniques by attending her classes. To this day, I believe the best way to improve teaching is by observing how others teach. Additionally, my junior academic brother, Shun-Yang Lee, and I were teaching simultaneously, both instructing MIS 301 (as it was a foundational course for undergraduate students, there were multiple sessions taking place concurrently). We frequently exchanged teaching experiences and lessons learned. Shun-Yang had a natural talent for teaching, and our interactions were highly beneficial for me. Later, he received a teaching award from the business school. Despite the extensive preparations, I was still incredibly nervous when it came to teaching the class myself and barely remembered what I said. At that time, the class met for three hours each week, divided into three sessions (Monday, Wednesday, and Friday), and almost the entire week was dedicated to preparing for teaching. Looking back, those early teaching experiences were incredibly helpful for me. If there's something you eventually have to do, the sooner you start, the better it might be.

Written on June 18, 2022

• （十四）MIS 301

自从决定转入 IS 以后（大约在2011年），我就开始打听应该如何在商学院寻求教职。除了研究上的区别，商学院与经济系还有一个很大的不同：商学院更加看重教学，大部分要上 job market 的博士生都会有曾经独立教学的经验。最终在 Andy 的帮助下，我也获得了一个独立教学的机会。在 2013 年的春天，我在 UT Austin 的商学院教了一门 IS 本科生的基础课 MIS 301: Introduction to Information Technology Management，这也是我第一次独立教课。

因为机会来之不易，所以我做了很多准备。在教课的前一个学期，我几乎听了商学院里所有教 MIS 301 的老师们的课。尤其要感谢文雯。那时她刚去 UT Austin 担任 Assistant Professor，课教得非常好。通过听她的课，我学到了很多教课的小技巧。时至今日，我仍然觉得提高教学的最好办法，就是去听别人是怎么教课的。另外，我的师弟李舜扬和我同时教课，也是上 MIS 301（MIS 301 是本科生的基础大课，所以有很多 sessions 同时进行）。我也和他经常交流教课的经验和教训。舜扬教课很有天赋，与他的交流也使我得益良多。后来舜扬还获得了商学院的教学奖。虽然做了很多准备，但真正自己上课的时候还是紧张万分，完全不记得讲了什么。当时一周上三小时的课，分三次（周一，周三，和周五）上完，几乎一周全在准备教学。现在想来，这些早期的教学经验对我非常有帮助。一件事如果迟早要做，可能越早做越好吧。

写于2022年6月18日

(15) What is A Profound Theory?

During my Ph.D. studies, apart from asking "What is IS," another question that troubled me was, "What is a profound theory?" Perhaps every doctoral student asks this question. My professor at UT Austin, Max Stinchcombe, liked to classify theories into two categories: theory theory and applied theory. Theory theory is not directly used to explain phenomena but provides a new tool or perspective to generate other theories or implications. On the other hand, applied theory is directly used to explain phenomena.

Based on this classification, in my view, profound theories have two criteria. First, they are conceptually profound and open up a new perspective. This criterion is more applicable to theory theory. For example, Arrow security is a highly profound theory that introduces the concept of contingent contracts, extending general equilibrium theory to uncertain situations and becoming the foundation of asset pricing. The core concept of prediction markets is also contingent contracts. Another conceptually profound example is the Coase Theorem, which states that, under zero transaction costs, the allocation of property rights is not important: property rights will always end up in the hands of the person with the highest value. This theory cannot directly explain phenomena because its premise does not exist in reality. In reality, transaction costs are always greater than zero. However, its profundity lies in providing a new approach. The conventional wisdom suggests that if externalities occur, the government should impose a Pigovian tax to correct the situation and restore optimality. If a company pollutes the environment, it should be taxed to internalize the externality. However, the Coase Theorem suggests that the optimal level of pollution by a firm depends on the value comparison between the firm and the surrounding residents regarding pollution. If transaction costs are low, market mechanisms will naturally come into play. This has had a significant impact on subsequent economics and legal practices, especially in pollution permit trading. Regarding theory theory, Box's statement is highly appropriate (although Box referred to statistical models, I believe it applies equally to theory theory): "Essentially, all models are wrong, but some are useful."

The second criterion is substantive depth, which is more suitable for applied theory. Thomas Schelling once said, "Models tend to be useful when they are simultaneously simple enough to fit a variety of behaviors and complex enough to fit behaviors that need the help of an explanatory model." Daniel Kahneman also made a similar statement: "Richer and more realistic assumptions do not suffice to make a theory successful. Scientists use theories as a bag of working tools, and they will not take on the burden of a heavier bag unless the new tools are very useful." A good applied theory needs to have profound content and be capable of simplifying complexity. In this regard, Kahneman and Tversky's prospect theory is an excellent example. Daniel Kahneman's book, "Thinking, Fast and Slow," summarizes his academic contributions. I remember seeing Andy reading the book when it was first published. Over the years, I have read it multiple times, gaining new insights each time, and many of my ideas are inspired by this book. Kahneman's new book, "Noise: A Flaw in Human Judgment," is a continuation of "Thinking, Fast and Slow," and reading it is also rewarding.

Written on June 30, 2022

• （十五）什么是深刻的理论

在读博阶段，除了 What is IS，另一个困扰我的问题是什么是深刻的理论。可能每一个博士生都会问这个问题。我在 UT Austin 的老师 Max Stinchcombe 很喜欢把理论分成两类，一类叫 theory theory，而另一类是 applied theory。Theory theory 并不是直接用来解释现象的，而是提供一种新的工具或一种新的视角来产生其它的理论或推论，而 applied theory 则是直接拿来解释现象的。

根据这个分类，在我看来，深刻的理论有两个标准。一是概念上的深刻，可以打开一个崭新的视角。这对于 theory theory 更加适用。比如 Arrow security 就是非常深刻的理论，提出了一个崭新的 contingent contract 的概念，把一般均衡理论扩展到了不确定的情况，成为了整个 asset pricing 的基础。同时，prediction market 的核心概念也是 contingent contract。另一概念上深刻的例子是 Coase Theorem。说的是在交易成本为零的情况下，产权的分配并不重要：产权总会通过市场交易到达价值最高的人手里。这个理论无法直接解释现象，因为它的前提条件在现实中并不存在。在现实中交易成本总是大于零。但它的深刻之处在于提供了一个新的思路。传统之见认为如果出现外部性，需要政府施加一个 Pigovian tax，矫正回最优的情况。如果有企业污染环境，需要对其征税，内化其外部性。然而 Coase Theorem 的思路是企业的最优污染程度取决于企业和周边居民对污染的价值比较。如果交易成本很低，那么市场机制自然会起作用。这对之后的经济学和法学实践都产生了重要影响（尤其是污染许可证的交易）。对于 theory theory, Box 的一句话是非常贴切的（尽管 Box 说的是 statistical models, 但我觉得对于 theory theory 也同样适用）：Essentially, all models are wrong, but some are useful.

第二个标准是内容上的深刻性，这对 applied theory 更加合适。Thomas Schelling 有一句话是 Models tend to be useful when they are simultaneously simple enough to fit a variety of behaviors and complex enough to fit behaviors that need the help of an explanatory model。 Daniel Kahneman 也有一句类似的话：Richer and more realistic assumptions do not suffice to make a theory successful. Scientists use theories as a bag of working tools, and they will not take on the burden of a heavier bag unless the new tools are very useful。好的 applied theory 需要有深刻的内容，可以做到化繁为简。在这一点上 Kahneman 和 Tversky 的 prospect theory 就是一个很好的例子。Daniel Kahneman 的著作 Thinking, Fast and Slow，总结了他的学术贡献。记得书刚出版的时候，就看见 Andy 在读。我这些年来也读过几遍，每次读都有新的体会，很多想法也是来源于这本书。Kahneman 的新书 Noise: A Flaw in Human Judgment 是 Thinking, Fast and Slow 的延续，读完也很有收获。

写于2022年6月30日

(16) Florida ICIS

My second participation in ICIS was in December 2012, and it's been almost ten years in the blink of an eye. The conference was held in Orlando that year, and it was also my first time visiting Florida. During the conference, I presented my research on prediction markets using experimental economics methods. At that time, I had no idea that I would later be teaching at the University of Florida. This conference was the most relaxed one during my Ph.D. years, not only because of the palm trees in Orlando that gave a vacation vibe. Shortly before the conference, I had just gotten married, so this conference was a true honeymoon trip. My wife and I had planned to visit Orlando first, then drive to Miami, and finally go to Key West, the southernmost point in the United States. Everything went smoothly along the way, and we had a great time.

Written on July 9, 2022.

• （十六）Florida ICIS

我第二次参加 ICIS 是在2012年的12月，转眼已经快十年了。那一年的举办地点是 Orlando，也是我第一次去佛罗里达。在会议上我 present 了用实验经济学方法做的 prediction market 的研究。当时完全没有想到日后会在佛罗里达大学执教。这次会议是我在博士期间最轻松的会议，倒不完全是因为 Orlando 的棕榈树让人有一种度假的感觉。在会议前不久，我刚刚结婚，所以这次会议是名副其实的蜜月之旅。我和太太一早计划好了，先去 Orlando，然后再开往迈阿密，最后去美国的最南端 Key West。一路上非常顺利，玩得也很开心。

写于2022年7月9日

(17) Job Market Paper

During the last three years of my Ph.D. program, my main focus was on my Job Market Paper (JMP). And my JMP was another stroke of luck. At that time, Huaxia had just had a paper accepted by JFE (Journal of Financial Economics). We discussed many details about mechanism design. It was around the same time that Wen-Ling Hsu, one of Andy's former students working at AT&T Research Lab, visited and talked to me about various applications of Mobile Data Offloading in the industry. That's when I realized that the details of mechanism design I had discussed with Huaxia could be applied to designing a Wi-Fi Procurement Auction. This became the prototype for my JMP titled "Hotspot Economics: Procuring Third-Party WiFi Capacity for Mobile Data Offloading." Later, Huaxia and I spent a lot of time refining the mechanism design model and computational algorithm, and it ultimately became one of my most satisfactory papers. Although the publication of this paper had its ups and downs, it was eventually accepted by ISR (Information Systems Research) as "Optimal Auction Design for Wi-Fi Procurement." In May 2014, thanks to the recommendation from my senior colleague Wen-Ling Hsu, I gave a talk at AT&T Research Lab, and everyone liked this model.

Choosing this paper as my JMP was a gamble at the time. In 2013, I had already felt that the research in the Economics of IS was leaning more towards empirical aspects, and purely analytical modeling work might have limited acceptance in job interviews. For me, since I had already published articles in both empirical and analytical modeling aspects, a safer approach would have been to choose an empirical paper as my JMP. Additionally, the most popular topics in IS at the time were social media and social networks. Since I had also been working on research in this area (social network-based prediction market), I had a strong urge to choose a paper in that domain as my JMP. However, in the end, I chose the paper on Mobile Data Offloading that I was most satisfied with at the time (which was also an analytical modeling paper) as my JMP. Whether it was a good or bad decision, even today I'm not entirely sure. It's hard to imagine how things would have turned out if I had chosen another paper as my JMP in a counterfactual scenario.

Written on July 31, 2022.

• （十七）Job Market Paper

博士班的最后三年，我的主要精力投放在 Job Market Paper (JMP) 上。而我的 JMP 称得上是另一个偶然。当时，华夏刚刚有一篇文章被 JFE 接受。我跟他也讨论了很多关于 mechanism design 的细节。恰恰在这个时候，Andy 以前的学生，在 AT&T Research Lab 工作的 Wen-Ling Hsu 来访，和她聊了工业界里 Mobile Data Offloading 的许多应用。这时，我意识到与华夏讨论的关于 mechanism design 的细节可以应用到设计 Wi-Fi Procurement Auction 上面。这也成了我 JMP, "Hotspot Economics: Procuring Third-Party WiFi Capacity for Mobile Data Offloading" 的原型。后来我和华夏又花了大量时间打磨 mechanism design model 和 computational algorithm，最终它成为了我比较满意的一篇文章。这篇文章的发表虽然一波三折，不过最后还是被 ISR 接受（"Optimal Auction Design for Wi-Fi Procurement"）。2014年5月，在师姐 Wen-Ling Hsu 的引荐下，我去  AT&T Research Lab 做了一个 talk，大家也很喜欢这个 model。

采用这篇文章作为我的 JMP, 在当时来说是一个冒险。在2013年，我已经感觉到 Economics of IS 的研究越来越偏向 empirical 方面，纯粹的 analytical modeling work 很有可能接受面会窄一些。对于我而言，由于之前在 empirical 和 analytical modeling 方面，都有文章，所以稳妥一点的方法是选择一篇 emprical 的文章作为我的 JMP。另外，当时在 IS 里最流行的主题是 social media 和 social networks。因为我之前也在做这方面的研究（social network-based prediction market），所以我也有很大的冲动去选择这方面的文章作为我的 JMP。不过，最终我还是选择了当时我个人最满意的关于 Mobile Data Offloading 的文章（也是一篇 analytical modeling 的文章）作为我的 JMP。是好是坏直到今天我也不太确定，似乎很难想象如果选择另一篇文章作为 JMP 的 counterfactual 会是如何。

写于2022年7月31日

(18) Job Market Preparation

The second half of 2013 to early 2014 was when I went on the Job Market. Going on the Job Market is indeed a comprehensive process, with many things to prepare. First, because the main conference that year, ICIS, was held in Milan, Italy, I needed to return to my home country in advance and ensure that my U.S. visa was valid (at that time, U.S. visas were renewed on an annual basis). Then, I also needed to go to the Italian embassy in Houston to obtain an Italian visa. Of course, these were just minor hassles. The real preparation was about how to handle job interviews and job talks.

In the Economics Department at UT Austin, there is a great tradition. Every Ph.D. student going on the job market has two faculty members who conduct practice job interviews for them. I personally found it extremely helpful. Both job interviews and job talks are processes that improve with practice. The more you experience them, the more composed and calm you become. Even today, I often conduct practice job interviews for Ph.D. students who are about to go on the market, and I encourage them to seek out friends and professors they know for practice sessions. Additionally, the department also organizes practice job talks for every Ph.D. student on the market, with the scale and format identical to regular seminars. I still remember vividly that my practice job talk took place on the afternoon of September 27, 2013. The placement coordinator at the time, Eugenio Miravete, recorded my entire talk (watching my own talk again afterward was a very beneficial experience). I distinctly remember feeling extremely nervous during the practice job talk but exceptionally relieved after it ended. On my way home, I even saw a legendary double rainbow. Years later, I saw a similar double rainbow at my home in Gainesville.

Written on August 4, 2022.

• （十八）Job Market 的准备

2013年下半年到2014年初是我上 Job Market 的时候。上 Job Market 的确是一个系统工程，千头万绪，很多事情需要准备。首先因为那年的主要会议 ICIS 在意大利的米兰召开，我需要提前回国，把美国的签证弄好（那时美国的签证还是一年一签）。然后，我还需要去休斯顿的意大利使馆，办好去意大利的签证。当然这些只是小折腾，真正需要准备的是如何应对 job interview 和 job talk。

在 UT Austin 的经济系，有一个很好的传统。每一个要上 market 的博士生，都会有两位老师为其进行 practice job interview。我个人觉得非常有用。无论是 job interview 还是 job talk，都是一个熟能生巧的过程。见得多了，自然就可以处乱不惊，镇定自若。时至今日，我也经常为即将上 market 的博士生开展 practice job interview，也鼓励他们去找认识的朋友和老师去做练习。另外，系里也会为每个上 market 的博士生举办 practice job talk，规模和形式和常规的 seminar 一模一样。至今还记得我的 practice job talk 是在2013年9月27日的下午。那时的 placement coordinator, Eugenio Miravete 还给我的 talk 全程录了相（事后再看一遍自己的 talk 是非常有益的经历）。我清楚的记得在做 practice job talk 的时候，非常紧张，而结束之后，又格外轻松。在回家的路上，还看到了传说中的双彩虹。多年后在 Gainesville 的家中也看到了类似的双彩虹。

写于2022年8月4日

(19) Another Graduation Season

Yesterday was UF's doctoral graduation ceremony. Due to the pandemic, I haven't attended it for three years. From my own doctoral graduation until now, witnessing many students graduate, it has already been eight years in the blink of an eye. Over these eight years, I have collaborated with numerous doctoral students (together with UF's doctoral students, we have published six papers in ISR, one in MISQ, four in POM, and four in JMIS), serving as their dissertation committee Chair, Co-Chair, or member. They have gradually grown, and many of them have become rising stars in academia, taking on prominent roles. This kind of relationship may only exist in academia. In the industry, the supervisor-supervisee relationship often becomes a competitive one: the success of newcomers often poses a threat to the old guard. However, academia is completely different; the more successful the students, the higher the academic reputation of the professors. Perhaps this is one of the appealing aspects of academia.

Written on August 6, 2022.

• （十九）又是一年毕业季

昨天是 UF 的博士生毕业典礼。因为疫情的缘故，我已经三年没有参加了。从自己博士毕业到现在看着许多学生毕业，一晃就已经八年了。在这八年里，我跟不少博士生一起合作 （我和 UF 的博士生在这八年里合作发表了六篇 ISR，一篇 MISQ，四篇 POM，四篇 JMIS），担任他们博士论文委员会的 Chair, Co-Chair, 或者 member。 他们逐渐成长，不少人已经在学术界独当一面，成为冉冉升起的新星。这种关系可能只有在学术界才会存在。在工业界，supervisor-supervisee relationship 常常是一种竞争关系：新人的成功往往会对旧人造成威胁。而学术界则完全不同，学生越成功，老师的学术声誉就越高。这或许是学术界吸引人的一个地方吧。

写于2022年8月6日

 2022年和我的学生 Mariia (Chair)                                2017年和我的学生梅小伟 (member)                    2016年和 UF 的博士毕业生 Gulver

(20) IS Job Market

I officially entered the IS job market primarily through two conferences: the first was the INFORMS/CIST held in Minneapolis in the fall of 2013, and the second was the ICIS/WISE/WITS held in Milan, Italy in December 2013. At the INFORMS/CIST in Minneapolis, I had only a few interviews. That time, I flew with Andy from Austin to Minneapolis. Andy drove me from the school to the Austin airport, and then we went to Minneapolis together. As a newcomer to the conference, one observation I made was that the hallway could be considered a second venue. Regarding this, Michael Lewis captures it well in his intriguing new book (well, all of his books are intriguing), "The Premonition: A Pandemic Story": "You go to a formal meeting. The important conversation is not in the meeting. It's in the Halls during the breaks. And usually what's important is taboo. And you can't say it in the formal meeting."

I stayed in Milan for a week, and I had several interviews every day, so apart from the conference venue and the hotel for interviews, I didn't get to see much of Milan. I attended ICIS and WISE consecutively, and I had a paper presentation and a poster presentation at WISE. That poster presentation was the one I mentioned earlier with Sarah and Ravi's paper. At the time, Sarah printed the poster, sent it to Austin, and I brought the poster with me to Milan.

Interviewing is a skill that improves with practice. When you have four or five interviews in a day, day after day, you become more adept at handling them. I still vividly remember the UF interview at ICIS. It took place in a classroom at Bocconi University, and I was interviewed by Haldun, Janice, and Shubho. A few weeks later, I received an email from Haldun, and my job market entered the next stage: Campus Visit.

"Liangfei,

We would like to invite you for a campus interview. Can you send me two or three available dates (Mondays and Fridays) starting with Jan 24th?"

Written on August 20, 2022.

• （二十）IS Job Market

我正式上 IS 的 Job Market 主要是在两次会议上：一是在2013年秋天 Minneapolis 举办的 INFORMS/CIST，另一个是在2013年12月意大利米兰举办的 ICIS/WISE/WITS。在 Minneapolis 的 INFORMS/CIST，我只有少量面试。那次是和 Andy 一起从 Austin 飞往 Minneapolis。我坐着 Andy 的车从学校到了 Austin 机场，然后一起去 Minneapolis。那时，作为会议的新丁，我的一个观察是 Hallway 可以算成第二个会场。关于这点，Michael Lewis 引人入胜的新书（当然，他每本书都引人入胜），The Premonition: A Pandemic Story，有个很好的总结：“ You go to a formal meeting. The important conversation is not in the meeting. It's in the Halls during the breaks. And usually what's important is taboo. And you can't say it in the formal meeting." 

在米兰住了一周，每天都有不少面试，所以除了会场和面试的旅馆，基本没有去过米兰的其它地方。我连续参加了 ICIS 和 WISE，并在 WISE 上有一个 paper presentation 和 一个 poster presentation。而那个 poster presentation 就是之前我提到的和 Sarah 以及 Ravi 的文章。当时，Sarah 把 poster 打印好，寄到 Austin，我又把 poster 带到了米兰。

面试是一个熟能生巧的事。当你一天有四，五个面试，连续几天，你就可以驾驭自如了。我还清楚的记得 ICIS 上 UF 的面试。那是在 Bocconi University 的一间教室，面试我的是 Haldun, Janice, 和 Shubho。过了几周，我收到了 Haldun 的电子邮件，而我的 Job Market 也进入了下一阶段：Campus Visit。

”Liangfei,

We would like to invite you for a campus interview. Can you send me two or three available dates (Mondays and Fridays) starting with Jan 24th? “

写于2022年8月20日

(21) Econ Job Market

In addition to the IS job market, I also simultaneously entered the job market for economics. At that time, my main focus was IS, but I thought it would be a good idea to participate in the economics market as well to broaden my horizons. The main event for the economics job market that year was the AEA annual meeting held in January 2014 in Philadelphia. I remember arriving in Philadelphia a little after 3 p.m. Soon after reaching the hotel, it started snowing heavily, and I heard that many job market candidates' flights were delayed. This heavy snowfall made the interviews particularly challenging. Interviewers from different schools stayed in different hotels, not far from each other, with walking distances ranging from 5 to 20 minutes. Under normal weather conditions, this wouldn't be an issue, but in freezing temperatures and knee-deep snow, a 20-minute walk became a significant test. Additionally, interviews were often scheduled with less than half an hour between them, so in the hotel area in Philadelphia, you could see job market candidates rushing from one interview to another.

This was my first time attending the AEA annual meeting. I was delighted to meet my fellow classmates and friends from CCER, Ben Zou and Haojun Yu. After the interviews, we also explored Philadelphia together. Additionally, I had a small gathering with Professor Wu and other fellow students, and I attended the CCER alumni gathering organized by Professor Yang Yao, where I met many classmates I hadn't seen in a long time. The next time I could have such a gathering would be when I attended the AEA annual meeting again.

At the AEA annual meeting, I presented my prediction market paper and participated in more than a dozen interviews. I was honored to have an interview with Harvard Business School (HBS), although I realized it was more of an eye-opening experience. At the AEA annual meeting, I also had interviews with some Chinese universities. At that time, Chinese universities were offering positions very quickly, often within a day or two after the interview. During the AEA annual meeting, I received an offer from Wuhan University, and a few days later, I received an offer from Shanghai Jiao Tong University, Antai. At that time, with over a dozen interviews at ICIS and many more at AEA, I thought I would have several campus visits in North America. However, in the end, I only had one campus visit in North America, which was UF. While waiting for my flight at the airport in Philadelphia, I received an email from Haldun inviting me for a campus visit at UF. This was my only opportunity in the North American job market.

Written on August 22, 2022.

• （二十一）Econ Job Market

除了 IS Job Market，我还同时上了经济学的 Job Market。当时，我的主要集中点是 IS，但还是觉得可以顺便上一下经济学的市场，开开眼界。我那一年的经济学的 Job Market 主要是2014年1月在费城召开的 AEA 年会。还记得是下午3点多到费城的。刚到旅馆，就开始下大雪，听说很多 job market candidates 的飞机都延误了。这场大雪也使得面试变得格外辛苦。不同的学校的面试老师住在不同的旅馆，相隔不远，走路基本 5 - 20 分钟就可以了。正常天气下，这本不是问题，但是在天寒地冻，大雪埋靴的时候，走20分钟路却是极大的考验。再加上面试之间往往只隔着不到半小时的时间，所以在费城的旅馆区，到处都能看到行路匆匆的 job market candidates。

这是我第一次参加 AEA 年会。令人开心的是见到了 CCER 的同学老友邹奔和俞浩君。面试结束以后，还和他们游玩了一下费城。另外，我和巫老师以及其他同门也小聚了一下，还参加了姚洋老师组织的 CCER 校友聚会，见到了很多许久不见的同学。下一次这样的聚会只有等我什么时候再参加 AEA 年会了。

在 AEA 年会上我 present 了我的 prediction market paper，又参加了十几个面试。非常荣幸的是可以参加 Harvard Business School (HBS) 的面试，不过我也意识到这就是开阔一下眼界，并不会真正拿到 HBS 的 campus visit。在 AEA 年会上我还参加了一些中国学校的面试。那时中国学校给 offer 的速度非常快，常常是面试之后一两天就会给 offer。我还在 AEA 年会的时候，就收获了武汉大学的 offer，过了几天又收到上海交大安泰的 offer。当时觉得在 ICIS 上有十几个面试，在AEA 上也有不少，应该在北美会有好几个 campus visits。没想到，最后在北美只有一个 campus visit，那就是 UF。在离开费城等飞机的时候，我收到了 Haldun 的邮件，邀请我去 UF campus visit，这是我在北美 Job Market 上的唯一机会。

写于2022年8月22日

(22) Florida Job Talk

My job talk at UF took place on January 27, 2014. Since this was my only opportunity in North America, I felt extremely anxious before the trip. I arrived in Gainesville on the evening of the 26th and left on the afternoon of the 28th, and the entire schedule was packed. Although I had rehearsed my job talk countless times, I still felt incredibly nervous on the actual day. I had no idea what I was saying during the hour and a half of the talk. On the morning of the 28th, I participated in a real estate tour, which was a fascinating experience. Later, when I bought a house in Gainesville, I used the same real estate agent.

After the campus visit, it was a period of anxious waiting. Since I didn't receive any positive news from other places in North America, I placed all my hopes on UF. When I returned to Austin and met Andy, I told him I was extremely anxious. Andy said, "Although you only had one job talk in North America, you only need one job." While his words were comforting, facing such uncertainty inevitably made me restless. The department chair at UF, Kenny, said that if there was any news, he would call me. So I kept a close eye on my phone, ensuring it had enough battery power and a stable signal. On the evening of February 3rd, just after I finished taking a shower, I heard my wife calling me to answer the phone. For me, it was an exciting moment. Kenny informed me that UF was preparing to offer me a position. I quickly accepted UF's offer, ready to embark on a new chapter in my academic journey.

Written on September 10, 2022.

• （二十二）Florida Job Talk

我在 UF 的 job talk 是在2014年1月27日。因为这是在北美的唯一机会，所以临行前心中异常忐忑。26日晚上到达 Gainesville, 28日中午离开，整个行程被安排得满满的。虽然 job talk 已经事先排练了无数次，但是亲身临场还是格外紧张。一个半小时的 talk 完全不知道自己讲的什么。28日早上参加了 real estate tour，这是一个很有趣的经历。后来在 Gainesville 买房，我用的还是同一个地产经纪。

Campus visit 以后就是焦急的等待过程了。由于在北美其它地方都没有好消息，我把全部希望都放在 UF 上了。回到 Austin，见Andy 的时候，我说自己非常的焦虑。Andy 说虽然你在北美只有一个 job talk，但你只需要一个工作。话虽如此，但面临这样的不确定性，难免坐立不安。UF 的系主任 Kenny 说如果有消息，会打电话给我，所以我时刻注意着手机，确保它电力稳定，信号通畅。2月3日的晚上，我刚洗完澡，就听到太太叫我接电话。对我来说，这是激动人心的时刻，Kenny 说 UF 准备给我 offer 了。我很快接受了 UF 的 offer，也即将开启学术之旅的新篇章。

写于2022年9月10日

(23) The New and the Old in Phenomena

Compared to economics and other social sciences, business school research has a notable characteristic: the pursuit of new phenomena and new business models. A glimpse into the past decade of IS research reveals this trend: from social media to mobile, and then to big data, IoT, and emerging trends like Fintech, blockchain, and AI. Once a new pattern emerges, most scholars promptly follow suit. This is understandable as everyone is drawn to novelty, and I myself am no exception. However, researching new models does not imply excluding the study of old phenomena and old models. Researching old phenomena can also bring forth new knowledge.

Avner Greif from Stanford studied the group punishment practices of medieval merchants. By combining detailed historical records with game theory models, he greatly expanded our understanding of markets and institutions. I first heard about his research during Professor Qiren Zhou's New Institutional Economics course back in 2003. However, it wasn't until today, a decade later, during my son's chess tournament break, that I finally had the opportunity to read Avner Greif's work. Unlike historians, economists often study history to derive new theories and test their validity. Recently, my senior academic sister at CCER, Ruixue Jia, published an article in the top economics journal QJE about Zeng Guofan's social relationships. It is another excellent example of how studying old phenomena can yield new knowledge. Unexpectedly, this article has sparked controversy among historians. The perspectives of economists and historians are ultimately different.

Written on January 25, 2023.

• （二十三）现象的新与旧

比起经济学和其它社会科学，商学院的研究有一个显著特点：求新，追求研究新的现象和新的商业模式。从近十年的 IS 研究，就可以管中窥豹：从 social media，到 mobile，再到 big data，IOT，和方兴未艾的 Fintech, blockchain, 和 AI。一旦一种新的模式和热点兴起，大部分学者都会及时紧跟。这本来无可厚非，大家都喜欢新鲜事物，我自己也是这么做的。不过，研究新模式并不代表着要排斥研究旧现象和旧模式。研究旧的现象也可以带来新的知识。

Stanford 的 Avner Greif 研究了中世纪商人的群体惩罚模式。他把详细的史料和博弈模型相结合，大大扩展了我们对市场和制度的理解。早在 2003 年周其仁老师的新制度经济学课上，就听说过他的研究。不过直到十年后的今天，趁着儿子 chess tournament 的间隙，我才终于拜读了 Avner Greif 的著作。与历史学家不同，经济学家研究历史，常常是想从历史中得出新的理论，并验证理论的力量。最近，我在 CCER 的学姐贾瑞雪在经济学顶刊 QJE 上发表了一篇关于曾国藩的社会关系的文章。这也是研究旧现象可以带来新知识的好例子。当然，始料未及的是这篇文章引起了史学家的争议。经济学家和史学家的角度终究是不同的。

写于2023年1月25日

(24) Ten Thousand Whys

When I was a child, there was a book I really liked called “Ten Thousand Whys.” After becoming a professor, I often receive various questions from students. Here, I will briefly summarize a few of them and provide my own thoughts. It is important to note that my responses can only be considered as personal opinions and not definitive answers, as many academic questions do not have a single correct answer.

Question 1: Is it always preferable to use randomized controlled trials instead of observational studies to establish causation? 

My answer: No. In general, randomized controlled trials (RCTs) are often considered superior in terms of internal validity for establishing causality, while observational studies excel in external validity. However, it is important to recognize the issue of compliance within RCTs. Although RCTs have traditionally been regarded as the gold standard, many trials encounter challenges with low compliance rates in the treatment group. For instance, the effectiveness of mask-wearing has seen conflicting results between observational studies and some previous RCTs. While numerous observational studies consistently demonstrate a significant reduction in respiratory virus transmission through mask-wearing, certain RCTs have reported mixed or limited findings in preventing influenza-like illnesses. 

Although RCTs are often perceived to address unobserved confounders and yield more convincing results, in this specific case, I would place more trust in the findings of the numerous observational studies. Many RCTs on mask-wearing face difficulties with low compliance rates, particularly among children in the treatment group. Ensuring proper adherence to mask usage is crucial for maintaining the validity of RCTs, but in reality, a relatively low number of participants in the treatment group adhere to mask-wearing guidelines. As John List highlights in his new book, "The Voltage Effect: How to Make Good Ideas Great and Great Ideas Scale,": "Compliance, another way of simply saying use, is a non-negotiable ingredient for just about every policy, program, or enterprise at scale."

Written on May 10, 2023

• （二十四）十万个为什么

小时候很喜欢的一本书叫《十万个为什么》。当了老师以后，经常会有学生问各种各样的问题。我简单总结几个，并提供自己的想法。需要注意的是我的回答只能称作想法，并不是标准答案，而很多学术问题也没有标准答案。

问题1：使用随机对照试验而不是观察性研究来建立因果关系，是否总是更可取？

我的回答：不是。通常情况下，随机对照试验（RCT）在内部效度（因果关系）方面被认为更为优越，而观察性研究在外部效度方面表现出色。然而，需要意识到随机对照试验中的遵从性问题。尽管随机对照试验传统上被视为金标准，但许多试验在治疗组中遇到了低遵从率的挑战。例如，关于佩戴口罩的有效性，观察性研究和某些以往的随机对照试验得出了矛盾的结果。尽管许多观察性研究一致显示佩戴口罩能显著减少呼吸道病毒传播，但某些随机对照试验却报告了不一致或有限的预防类流感疾病的结果。

尽管随机对照试验通常被认为能解决未观察到的混淆因素并产生更具说服力的结果，但在这个特定案例中，我更倾向于相信众多观察性研究的结果。许多有关佩戴口罩的随机对照试验在治疗组中面临低遵从率的困难，尤其是在儿童中。确保正确遵守佩戴口罩的规定对于维护随机对照试验的有效性至关重要，但事实上，治疗组中遵守佩戴口罩指南的参与者相对较少。正如 John List 在他的新书，The Voltage Effect: How to Make Good Ideas Great and Great Ideas Scale，中强调的那样：“遵从性，也就是使用，对于几乎所有的政策、项目或大规模企业来说都是一种不可妥协的要素。”

写于2023年5月10日

(25) Ph.D. Graduation 

Facebook reminded me that it has been 9 years since I graduated with a Ph.D. Time is indeed relative; sometimes it feels like it's passing slowly, and other times it's accelerated. I remember 9 years ago, the graduation ceremony for the Economics Department was held in the gym. After finding a job, I felt particularly relaxed. The Ph.D. regalia at UT is quite distinctive, but it comes with a hefty price tag. I vaguely remember that even after the discount, it was still over $700. However, I thought that in the future, if I have my own Ph.D. students graduating, this regalia will be quite useful. And now, it has proven to be true.

Written on May 22, 2023

• （二十五）博士毕业

Facebook 提醒我已经博士毕业9年了。时间的确是相对的，有时觉得过得慢，有时又被加速了。记得9年前经济系的毕业典礼是在 Gym 的礼堂里举行的。找到工作之后，我的心情格外放松。UT 的博士服很有特色，不过买一件价格不菲，依稀记得打完折之后还要700多。不过想着以后如果自己的博士生毕业，这件博士服会挺有用的，现在果然如此。

写于2023年5月22日

(26) Making a Comeback After a Long Break 

Due to the pandemic, I haven't been able to attend in-person conferences and give talks for a long time. Recently, I have resumed participating in conferences and giving talks. A few weeks ago, I went to Florida State University to present a paper on social trading. Last week, I attended a blockchain conference held at the University of Miami. Apart from presenting a paper, I also took part in a panel discussion, where we discussed the future of blockchain research.

In my view, blockchain research encompasses two aspects: finance and technology. First, it is natural to question the differences between the new blockchain and traditional financial markets. For example, Initial Coin Offerings (ICOs) bear many similarities to traditional Initial Public Offerings (IPOs) in several aspects. We have gained considerable knowledge about IPOs through finance literature. My colleague, Jay Ritter, at the University of Florida's finance department, is a leading authority in IPO research and is known as Mr. IPO. He has conducted extensive research on IPO market efficiency. Overall, the IPO market is quite efficient. However, we know very little about the new ICOs. On the one hand, ICOs lack underlying firms like IPOs, they don't have quarterly financial statements, and they lack government regulation. They only have simple white papers, so it is quite possible that the market efficiency of ICOs is not as good as that of IPOs. However, on the other hand, the ICO market is not like the stock market. It doesn't have trading holidays and operates 24/7, which might make it more efficient. One of my recent working papers aims to compare the market efficiency of ICOs and IPOs.

Second, because blockchain encompasses both finance and technology, the interaction between the two becomes very interesting. For example, another paper of mine investigates how the technology aspect influences the financial attribute. Specifically, most cryptocurrencies have their source codes publicly available on GitHub, and understanding how these source codes affect cryptocurrency prices becomes an intriguing question.

Written on May 23, 2023

• （二十六）久休复出

由于疫情的原因，我很久没有参加 in person 的 conference 和 talk 了。最近重新开始了参加会议和给 talk。几周之前去了 Florida State University 讲 social trading 的 paper。上周又去了迈阿密大学召开的 blockchain conference。除了讲 paper 以外，还参加了 panel discussion，谈了谈 blockchain research 的未来。

在我看来，blockchain research 包含两个方面，一是 finance，二是 technology。首先，很自然的，人们都会想问新的 blockchain 跟传统的金融市场有什么区别。比如说 initial coin offering (ICO) 在很多方面都很像传统金融的 IPO。关于 IPO，我们通过金融学的文献已经了解了很多。我的同事，在UF金融系的 Jay Ritter 是研究 IPO 的权威，被称为 Mr. IPO。他有大量关于 IPO market efficency 的研究。总的说来，IPO market 是相当有效的。可是我们对新型的 ICO 知之甚少。一方面，ICO 不像 IPO 那样有 underlying firm，没有财务报表，也缺少政府监管，只有简单的 white paper，所以很有可能市场有效性不如 IPO。然而，另一方面，ICO 市场不像股票市场，它没有休息日，全天全年都可以交易，所以也有可能更有效率。我最近的一篇论文就是想比较 ICO 和 IPO 的 market efficiecny。

其次，正因为 blockchain 包含了 finance 和 technology 的两个方面，所以两者的互动变得很有意思。比如我的另一篇文章研究了techonology 是怎么影响 finance 属性的。具体来说，大部分加密货币在 GitHub 上都公布了源代码，而这些源代码如何影响加密货币的价格就成为了有趣的问题。

写于2023年5月23日

(27) Is AI a Rising Tide that Lifts All Boats? 

Putting Acemoglu's new book, 'Power and Progress,' together with Piketty's magnum opus, 'Capital in the Twenty-First Century,' is not only complementary but also intriguing. Both books' core ideas are closely aligned, suggesting that development and technological progress do not necessarily lead to an increase in income for ordinary people. Piketty's major work argues that growth is not a rising tide that lifts all boats, focusing more on data collection and measurement. On the other hand, Acemoglu's new work emphasizes that technology is not a rising tide that lifts all boats, concentrating more on theoretical propositions. Both books underscore the fact that despite significant technological advancements during the post-industrial revolution period, the income of ordinary workers remained stagnant. Most of the benefits from technological progress were captured by the elite class controlling technology and machines. Substantial wage increases for ordinary workers only occurred in the following decades. This led Acemoglu to his core viewpoint that technological progress, including AI, does not inherently result in increased income for ordinary people. The distribution of technological dividends depends on the underlying political processes, which resonates with some aspects of Marx's classic theory.

Acemoglu holds a similar perspective regarding AI. Of course, AI can either substitute or complement existing human labor, depending on the specific AI technology. However, even if AI complements human labor and enhances labor's marginal productivity, it does not necessarily mean that AI development will increase workers' income. This contrasts with classical economic theory, which suggests that if AI and human labor are complementary and increase labor's marginal productivity, businesses should hire more workers, leading to higher wages for workers. However, Acemoglu points out that labor markets are often not characterized by perfect competition. In historical periods like the Middle Ages or the Industrial Revolution, forced labor was prevalent. Therefore, even if technological advancements made labor more valuable, they did not guarantee a rise in workers' wages. The elite class could simply force workers to work longer hours through political mechanisms. Card's latest review article published at AER also points out that labor markets are not necessarily perfectly competitive, and monopsony power exists in many labor markets.

In my opinion, many labor markets in the sharing economy are indeed characterized by monopsony, where platforms hold significant bargaining power (e.g., in the ride-sharing market). Recently, during a trip back to my home country China, I noticed that using ride-hailing services like DiDi was very convenient, with effective matching technology ensuring a quick arrival of the vehicle after booking. However, did the drivers' income increase due to the improvements in matching technology? This could be an empirical question.

Written on July 30, 2023

• （二十七）Is AI a Rising Tide that Lifts All Boats? 

把 Acemoglu 的新作 Power and Progress 与 Piketty 的大作 Capital in the Twenty-First Century 放在一起看是相得益彰，饶有趣味的。两本书的核心思想非常接近，都是说发展与技术进步未必可以带来普通人收入的增长。Piketty 的大作是说 growth is not a rising tide that lifts all boats，更多聚焦在数据的收集与测度；而 Acemoglu 的新作是说 technology is not a rising tide that lifts all boats，更集中与理论的提出。两本书都强调了一个事实，工业革命后的五、六十年，虽然技术大发展，但普通工人的收入却停滞不前。技术进步的大部分收益都被掌握技术和机器的精英阶层分走了。普通工人工资的显著提高是再之后几十年的事了。Acemoglu 由此引发了他的核心观点，技术（包括 AI 在内）进步并不会自然而然使得普通人的收入得到提高。技术红利的分配取决于背后的政治过程。这与马克思的经典理论有异曲同工的地方。

对于 AI，Acemoglu 的想法是类似的。当然，AI 与现有的人类劳动力可能是替代，也可能是互补关系；这取决于具体的 AI 技术。但即使 AI 与人类劳动力是互补的，可以提升 labor 的 marginal productivity，也并不一定意味着 AI 的发展可以提高工人的收入。这与经典的经济学理论是不同的。经典的经济学分析告诉我们如果AI 与人类劳动力是互补的，可以提升 labor 的 marginal productivity, 那么企业就应该聘用更多的工人；那么工人的收入也会相应的上涨。但这个经典的逻辑链条有一个关键假设：劳动力市场是充分竞争的。因为劳动力市场充分竞争，那么劳动力提升了的 marginal productivity，就会使得劳动力更值钱，同时工人的工资也会上涨。然而 Acemoglu 指出很多时候，劳动力市场都不是充分竞争的。比如在中世纪或是工业革命发生的时候，强制劳动还是非常普遍的。即使技术进步使得劳动力更加值钱，也并不意味着劳动力的工资会上升。精英阶层通过政治机器强迫工人更长时间工作就可以了。Card 在他最新的综述文章（AER）里就指出劳动力市场未必是充分竞争的，monopsony 在很多劳动力市场上都存在。

在我看来，不少 sharing economy 的 labor markets 都属于 monopsony，也就是平台拥有太大的话语权（比如在 ride sharing market 上）。前段时间刚刚回国了一趟，最大的感觉就是滴滴打车非常方便，配对技术行之有效，叫车之后两三分钟车就到了。然而司机的收入是不是因为配对技术的改进而提高了呢？这可能是一个实证的问题了。

写于2023年7月30日

• (28) Remembering Professor Wu

I learned this morning that Professor Wu passed away, and I feel deeply saddened. Over the past few weeks, we, the academic brothers and sisters of the Wu Group, have been offering our final prayers. Unfortunately, the news we dreaded has arrived.

My journey into academia was greatly influenced by Professor Wu. Twenty years ago, while studying for my undergraduate degree at Zhejiang University, I had the privilege of attending Professor Wu's game theory class. It was my introduction to the realm of game theory, where I experienced its wonders and wisdom. To this day, I have several research projects related to applied game theory. At that time, Professor Wu resigned from his tenured position at Tulane and returned to teach at National Taiwan University, while also serving as a visiting professor at Zhejiang University. It was in Professor Wu's class that I first experienced the charm of game theory.

After graduating from Zhejiang University, I went to Peking University's China Center for Economic Research (CCER) for further studies, and it was during this time that Professor Wu joined Peking University as a full-time faculty member. Not only did I take Professor Wu's Advanced Microeconomics II course, but I also became one of his first batch of master's students at Peking University. I remember when Professor Wu first arrived at Peking University, many CCER students wanted to choose him as their advisor. I was very fortunate to become his student. The weekly group meetings, where we discussed the latest academic papers in finance and game theory and presented our own ideas, were immensely beneficial. Because Professor Wu often asked probing questions during these meetings, we all prepared diligently, carefully reading the academic papers to avoid being caught off guard. Those pure academic moments are unforgettable. I remember one time when I had to introduce a paper co-authored by Snehal Banerjee, Ron Kaniel, and Ilan Kremer titled "Price Drift as an Outcome of Differences in Higher-Order Beliefs" (which was later published in Review of Financial Studies, but at that time, it was still a working paper). To fully understand the paper, I visited Banerjee's personal webpage and read three other related working papers. This deepened my understanding of higher-order beliefs in financial markets. Under Professor Wu's guidance, a few months later, I even developed a small model on the coexistence of higher-order beliefs and first-order beliefs in financial markets. This small model later became my first English working paper: Higher-Order Beliefs and Asset Market Dynamics. I still remember Professor Wu helping me revise the paper on higher-order beliefs: from grammar to structure, and from citations to insights, his revisions were meticulous and insightful. Even today, when I supervise doctoral dissertations, I try to be as meticulous as possible. However, I am ashamed that I still cannot match Professor Wu's level of attention to detail. This further emphasizes the value of the education I received under Professor Wu. The experiences from those group meetings greatly helped my later research on prediction markets and recent studies in FinTech. For example, my recent paper titled "Informational Efficiency of Cryptocurrency Markets" published in Journal of Financial and Quantitative Analysis (JFQA) benefited greatly from Professor Wu's group meetings. I remember our first group meeting was on October 9, 2007, at 3 p.m. (in the small room of Zhifu Building). Looking back at Professor Wu's emails from that time, it feels like it all happened yesterday:

"Dear all, Our first meeting will be on Oct 9 at 3pm in the small room of Zhifu Building. Please talk about what you have read or done in the summer etc.

Ho-Mou"

When I graduated from CCER, Professor Wu gave me a book on Fu Peirong's lectures on traditional Chinese culture. Now, when my students graduate with their doctorates, I also give them a book I like. Since graduating from CCER, I have stayed in touch with Professor Wu. After my first year of doctoral studies at UT Austin, I even returned to Beijing to attend the group meeting of Professor Wu's research group during the summer break. At that time, my junior colleagues Yiqing Xing, Yao Zeng, and Yizhou Xiao defended their master's theses during the group meeting. Later, Professor Wu encouraged me to submit our coauthored paper to the 2010 Econometric Society World Congress. After the paper was accepted, Professor Wu helped me revise the presentation slides at the conference venue in Shanghai. Due to the pandemic, prior to last year, I hadn't returned to China for many years. The last time I saw Professor Wu in person was at the AEA meeting in 2014. However, we have always kept in touch on WeChat. The last message Professor Wu sent me on WeChat was on January 10th this year:

"Yesterday, a CCER student forwarded me your personal reflection. I was deeply moved after reading it! You mentioned your time from Zhejiang University to CCER, and the period when you wrote your master's thesis, immersed purely in academic research. Those were also the happiest memories of my academic career at Peking University! Your journey has not been easy, and your achievements are remarkable and sweet. Congratulations on your academic achievements and having a happy and fulfilling family life!"

Written on May 3, 2024

• （二十八）怀念巫老师

今天早上得知巫老师去世了，心里非常难过。前几周已经收到了一些消息，我们巫组的师兄弟姐妹们在做最后的祈祷。不幸的消息还是来了。

我走上学术的道路，很大程度上是受到巫老师的影响。二十年前在浙大读本科的时候，有幸上了巫老师的博弈论课，被引进了博弈论的殿堂，感受了博弈论的奇妙与智慧。至今我还有好几个关于应用博弈论的研究项目。当时巫老师从 Tulane 辞去终身教职，回台大任教，同时也在浙大当客席教授。而我在巫老师的课堂里也第一次领略了博弈论的魅力。

从浙大毕业后，我去北大中国经济研究中心 (CCER) 继续读研，而这时巫老师也全职来到北大任教。我不仅修读了巫老师的高级微观经济学II，还成为了巫老师在北大的第一批硕士研究生。记得巫老师刚来北大时，很多 CCER 的同学想选他当导师，我非常有幸成为了他的学生。之后每周一次的组会，让我受益匪浅。当时组会的主要内容是讨论最新的金融学和博弈论方面的学术论文以及 present 自己的想法。因为巫老师常常会在组会上问出很深入的问题，所以我们都会精心准备，仔细阅读那些学术论文，生怕被问倒。这段纯粹的学术时光令人怀念。记得有一次，我要介绍 Snehal Banerjee, Ron Kaniel, and Ilan Kremer 合写的一篇文章 Price Drift as an Outcome of Differences in Higher-Order Beliefs （这篇文章后来发表在 Review of Financial Studies 上，不过当时还是一篇 working paper)。为了能全面的理解这篇文章，我找到 Banerjee 的个人主页，把其它相关的三篇 working papers 都看了一遍。这样对金融市场上的高阶信念的来龙去脉就有了更深的认识。在巫老师的指导下，几个月后，我还写出了一个 higher order belief 和 first order beleif 在金融市场上并存的小模型。这个小模型后来成为了我的第一篇英文的 working paper: higher order beliefs 对金融市场的影响。至今还记得巫老师帮我修改 higher order beliefs 的文章：从文法到布局谋篇，再到引用，细致无遗，充满洞见。到今天，我自已指导博士生论文时，也尽量认真。可是很惭愧，还是没有办法像巫老师那么细致严谨。这也更使我感受到在巫老师门下受教的可贵。组会里的这些经验对我后来的 prediction market 的研究，和最近 FinTech 方面的研究都有很大的帮助。比如，我最近发在 Journal of Financial and Quantitative Analysis (JFQA) 上的文章 Informational Efficiency of Cryptocurrency Markets 在很大程度上得益于巫老师的组会。记得我们第一次组会是在2007年10月9日下午3点（朗润园的致福轩小教室）。翻出巫老师当年的电子邮件，一切似乎还发生在昨天：

"Dear all, Our first meeting will be on Oct 9 at 3pm in the small room of Zhifu Building. Please talk about what you have read or done in the summer etc.

Ho-Mou"

在CCER 硕士毕业的时候，巫老师送了我一本傅佩荣讲中国传统文化的书。现在我的学生博士毕业，我也会送他们一本我自己喜欢的书。从 CCER 毕业以后，我和巫老师也一直保持着联系。我在 UT Austin 博士班一年级暑假的时候，还回北京参加了巫组的组会。当时师弟邢亦青，曾垚，和肖弋舟在组会上硕士论文答辩。后来，巫老师鼓励我把我们合作的文章投到 了2010年的 Econometric Society World Congress。文章被接受以后，巫老师还在上海的会场里帮我修改了 presentation slides。因为疫情，在去年之前，我多年没有回国。最后一次 in person 见巫老师是在 2014 年的 AEA 上。不过，我们在微信上一直保持着联系。巫老师最后一次发给我微信是在今年的1月10日：

“昨天一位CCER的同學轉來你寫的 personal reflection ，我看了很是感動！你提到從浙大到CCER，還有寫碩士論文的那段時光，純粹的沈浸於學術研究，那也是我在北大與學術生涯中最快樂的回憶！你一路走來也真不容易，成果也相當豐盛甜美，祝賀你取得的學術成就與擁有幸福美滿的家庭！”

写于2024年5月3日