Hongbin Li
Subject: How High-Stakes Chinese College Exams Can Change Your Life
Bio: Professor of Economics at Stanford and Author of The Highest Exam: How Gaokao Shapes China
Transcript:
Larry Bernstein:
Welcome to What Happens Next. My name is Larry Bernstein. What Happens Next is a podcast that covers economics, politics, and culture. This episode is How High-Stakes Chinese College Exams Can Change Your Life.
Our speaker is Stanford Economics Professor Hongbin Li who will discuss his new book The Highest Exam: How Gaokao Shapes China.
I want to find out why a high stakes exam is critical for the Chinese in their college admission as well as getting a job. I want to learn why this exam is believed to be the one thing in Chinese society that cannot be corrupted.
Hongbin, can you please begin with six minutes of opening remarks.
Hongbin Li:
The book I wrote with my co-author Ruixue Jia is on China’s Gaokao is The Highest Exam. This is the college entrance exam every kid needs to take to go to college in China, scheduled in early June, between two to four days.
I took it almost 40 years ago. Every student’s score is ranked within a province. Colleges come to each province to admit students according to their score.
We call this the centralized hierarchical tournament. It’s a competition and it’s an education system. The exam followed the imperial exam, or Keju in Chinese, which was started 1400 years ago in the Sui dynasty. The emperor designed this imperial exam for the purpose of selecting officials who can help him govern the country. This was not designed as an education system, it was a governance tool, a selection mechanism for the emperor.
The exam has been going on for 1400 years, There were some brief interruptions in the past century, but after the interruption, the new ruler immediately reinstated it.
In the book we talk about why the ruler, the society and the households are so obsessed with the exam. We also talk about how this exam system can be used as a policy tool for technology advancement. Finally, we talk about how, when people migrated from China to the US, these immigrants affect the local community with their ideas for how education should be structured.
Larry Bernstein:
In 1998 I managed the Tokyo arbitrage department at Salomon Brothers. All my colleagues had attended Tokyo University, the top school in Japan. The Japanese high-stakes exam was the prerequisite for admission, and once there, your job prospects were assured.
According to my buddies, Tokyo University itself was a breeze. What mattered was having attended. What you describe in China sounds like a pan-Asian phenomenon. Korea even keeps planes from flying overhead during its national exam. Why did China, Japan, and Korea all choose a single high-stakes exam?
Hongbin Li:
This all came from China. If you look at who are the government officials in these countries, in Tokyo, a lot of the officials are from this university.
What happened historically in China is society has a hierarchy. In China it’s government officials. Who can become officials? Whoever can pass this test will become an official immediately, no matter who you are, whether you’re a farmer, worker or businessperson. If officials are on top of the hierarchy, then everyone wants to become an official. Then you have to study hard to pass the test.
Larry Bernstein:
Why do these societies front-load the studying before college, rather than rewarding grades and performance in college as we do in the United States? Why is the emphasis before college and not during it?
Hongbin Li:
For economists, education has three functions. What you are describing is human capital, what you learn in school. But for economists, even more important are two other things. One is called a social network. You go to Tokyo University, go to Tsinghua University to make lifetime friends and business partners. This is the social networking or club effect.
Another is the signaling effect. This is contribution by Michael Spence, who got a Nobel Prize for this. Once you graduate from a famous university, say Peking University, people immediately know you’re both smart and work hard to achieve a high score. Employers, when they screen your resumes, look at where you received a degree from.
Larry Bernstein:
When I applied to college 40 years ago, we took standardized tests like the SAT and the ACT, plus subject-matter achievement tests and AP exams for college credit. The Chinese exam appears to combine elements of all of these test.
This single high-stakes exam tests for IQ, grit through rote memorization and covers narrow subjects like language, political history, and poetry. Why combine IQ, grit, and subject knowledge in a single exam?
Hongbin Li:
This is like the combination of SAT, SAT subject and AP tests. But everyone has to take it at the same time. This test needs to be taken by every kid in their senior year, and all the kids can be ranked from number one to the bottom. It’s very efficient as a ranking system.
In college admission in China, they only look at the total score on the test. In the US, AP, SAT are only two dimensions of college application. It’s called a holistic system — consider forty different indications. That’s the difference.
Larry Bernstein:
During COVID, American colleges turned against the SAT. Critics said that the test favored wealthy families who could afford tutors. Recently, colleges brought SATs back after admitting students who couldn’t do their homework.
The SAT’s test designers claim that it successfully predicts freshman-year grades. Does the Chinese exam predict college success? And since it was originally designed to select good government bureaucrats, does it do a good job for that purpose?
Hongbin Li:
In China, kids care less about college grades. Once they pass the exam, they get into a top college. That’s it. This itself is more important than anything else. It can predict college grades, but because kids don’t work that hard anymore, it’s different.
China has a so-called hard to get in, easy exit policy in college. Almost every college student graduates. Here half do not graduate because college classes are tough. In China it’s the opposite. It’s a waste of resources.
Larry Bernstein:
That sounds like a catastrophe. The reason we have college is to learn. If the learning happens prior to entering rather than at college, we’ve put the cart before the horse.
Hongbin Li:
College is designed to learn. Especially these days, kids don’t even show up to class. They can learn faster by themselves from AI than from their professors.
College is a place where people meet. Another dimension is the label. Once you get a degree from college, the market knows who you are. Learning is important, but most of the learning didn’t happen in class.
Larry Bernstein:
Benjamin Bloom ran a famous study at the University of Chicago comparing classroom teaching, mastery learning, and one-on-one tutoring. The tutored students scored two standard deviations above the typical student, though replications found an effect of just under one standard deviation.
AI can now be that tutor. How should we think about the importance of the Chinese exam in this new AI world, both for learning and for signaling to the marketplace what you’re capable of?
Hongbin Li:
In terms of the signaling, it doesn’t change. You still need to pass the test without the help of AI, so people know you’re smart. What China does is they require all kids to study and work hard from K to 12. What they learn may not matter because whatever you learned, you cannot beat AI. But training when kids are young could be important.
Larry Bernstein:
Childhood is about a quarter of your life. It seems that Chinese kids are using an inordinate amount of their time and resources to prepare for this exam.
There are real results. Chinese students far outscore Americans on international math and science exams, yet later they are less productive than the Americans who scored poorly. Should children sacrifice their youth by memorizing material that may have little relevance?
Hongbin Li:
My kids went to school in Beijing, Hong Kong and Palo Alto. China’s education focused on STEM. They are good at math for a reason.
There are two tracks in China’s school system. One is social science, the other is STEM. About 70% kids in China are STEM track. They don’t do politics or history because they are not tested for those. They have memorization, but it’s more like how fast you can solve a physics question on the test.
In the US, a lot of kids spend a lot of time doing sports. But even to do sports, youngsters who do well in the club or in school need to focus on repetition. If you want to play basketball you have to do the three-point drills 500 times a day.
Larry Bernstein:
In a famous Supreme Court case, Griggs v. Duke Power made it difficult for employers to give job candidates generic IQ tests if minorities underperformed on them. China has no such ruling. Nothing stops Chinese employers from testing candidates before they start work. Why hasn’t China adopted tests that instead of one high-stakes exam taken in high school?
Hongbin Li:
In the labor market, China’s employers also do tests. They copied a lot of these tests from US employers. If you want to work for Google, there’s tests. In China, if you want to work for Alibaba, they gave a similar test asking you to program something.
In the screening process they receive maybe 100,000 resumes. They will keep only a hundred. They will use the diploma to screen.
Larry Bernstein:
Chinese students who were not selected by one of the top schools can earn a graduate degree in the United States. They will argue that even though they attended a second tier Chinese university that employers should instead look at their first-tier US graduate school.
Hongbin Li:
If they got into the top college in China, they don’t want to go to the graduate school of the same college. Those students will come to the U.S. for graduate school. If you go to a second-tier college, in China, you want to go to a first-tier graduate school.
The number of students who come to US reached a peak before the pandemic. It now started to decline, because of US-China relations. But today I still believe that the brightest students still come here for graduate school, especially for the PhD programs. The majority, if they could, want to stay here.
I did graduate school at Stanford thirty years ago. At my time, almost no one returned to China. Today some will stay, some will go back. There are opportunities in China, big companies and very good colleges.
Larry Bernstein:
In the book you say that many studies put the economic return for a student for a year of college is 10%. What is it in China?
Hongbin Li:
This research I conducted myself. Using a 10% return to education that means comparing a college graduate to a high school graduate, you have four more years of education, so your income is 40% higher. But if you get into an elite college, your income is another 40% higher. This doesn’t count other benefits associated with going to a good college, like housing benefits, medicine, and your children can go to a good city, go to good school. The real return is much larger than 10% per year of education.
Larry Bernstein:
One study looked at marginal students who were waitlisted and one got into Penn and the other ended up at Penn State. It showed little to no economic difference in the students’ outcomes. Does that analysis hold in China?
Hongbin Li:
That’s a friend of mine who did it, Alan Krueger, from Princeton. He compared kids who got into Princeton and chose not to go there to those who went to Princeton. The return to elite education is almost zero in his study.
But in China, if you score 600 points, you got into an elite college. If you score 599 you’re not eligible. We compare those kids who score 600 to those who score 599. The return of going to elite college is 40%.
Larry Bernstein:
Why is there no impact in the US to elite colleges but huge differences in China? Both countries have a social network and a signaling effect. Why is Chinese society different?
Hongbin Li:
What didn’t matter is the average SAT score of the college. But what they found is that the tuition of the school matters. If you get into a college in the US which has high tuition, it’ll help you make more money later.
In China there’s nothing called tuition. Tuition is almost free, The only thing is the ranking. The labor market is also ranked. Here you have free labor markets. It’s different.
Larry Bernstein:
You write that Chinese students must choose either a STEM or a social science path, but not both, and that the social science path has a much lower payoff. Tell us about that.
Hongbin Li:
A lot of jobs in China are engineer jobs because China is big in manufacturing. Even government officials in China have an engineering background. That’s different from here, because in China there’s no political campaign. You don’t need to be a good speaker.
For social science track, there are not many good jobs, even in the US.
Larry Bernstein:
During the Cultural Revolution, Mao eliminated the high-stakes exam and closed the universities. Tell us about the Cultural Revolution and its relationship with the exam and the university system.
Hongbin Li:
He not only shut down the whole education system, but he also turned the social hierarchy upside down. Mao said the poor farmers, workers should be on top of the social hierarchy. A lot of urban kids were sent to the countryside to do farm work.
When Mao passed away in 1976, Deng Xiaoping immediately reinstated the college entrance exam. The 1977 cohort was the hardest to get into college because there are people taking the tests at the same time.
Larry Bernstein:
Because colleges were closed for a decade, ten years of high school graduates all applied at once, so there were ten times as many applicants as normal.
Hongbin Li:
High schools were also shut down. All the students self-learned by themselves. They never went to school.
Larry Bernstein:
When the Cultural Revolution ended and colleges and high-stakes entrance exams were back, what was the response among Chinese society?
Hongbin Li:
They love it. Many believe that’s the most important reform after the Cultural Revolution.
Larry Bernstein:
One conclusion of your book is that the high-stakes exam is here to stay. There’s nothing like taking something away to make people recognize its importance.
Hongbin Li:
The Cultural Revolution was gone. The other time was at the end of the Qing dynasty, around 1911. China’s elites at that time believed that China was backward because of the imperial exams, memorizing classics, but European countries were studying science and technology.
What the Qing Emperor abolished the test then suddenly all the elites in China lost the hope to move up the hierarchy. They went up to protest and overthrow the government. That’s the end of Qing dynasty. In 1911, the Republic was established. This is the other time that China abolished this system and didn’t end well.
Larry Bernstein:
I took AP exams at New Trier High School in Winnetka, Illinois. Because the AP test was well defined, teachers could teach to the test and students could master the material.
Recently, some American high schools revolted and stopped offering AP classes. How do you feel about that loss of accountability, versus the freedom to teach for a broader education? The Chinese high-stakes exam is a perfect example of needing to teach for a test.
Hongbin Li:
What is education before college, from K to 12? And what is the best way to select students into college? These are two separate questions.
One is what you should learn. We don’t know. Schools were created 200 years ago. We haven’t changed it much, how we teach, how we test students. Maybe it’s time to think and change.
The second question is tougher. How should a college select students? You can randomly place students to a college. At least it’s fair. Another extreme is legacy.
Here in US, we have a holistic system, considering both legacy and credential of the kids. In China they choose something in between, but it’s more towards the student’s own credentials. Family background also contributes to how well you are prepared for the test, like tutoring, but less so compared to the US system.
Maybe in the age of AI, learning itself in schools is becoming less important. Then we must design a smarter way to select students into college to consider both efficiency and equity.
Larry Bernstein:
College tuition costs have exceeded inflation for decades. Yet access to information is becoming nearly free. That will put a lot of pressure on schools, because when you lose your monopoly, the moat is gone.
There are two consumers here, the student who needs to learn and the employer who wants an educated hire. If students can learn outside of college and show their quality without a degree, won’t the signaling effect diminish?
Hongbin Li:
If in the future AI can help you to do the recruiting, maybe the search cost becomes zero. Then you don’t need a college degree to show you are able. Then the signaling effect is gone. But college still has the club effect. You still go there to meet people.
Having done all this work on the Chinese exam, do you find it a force for good in the world?
Hongbin Li:
It’s very difficult to say. I just went to Brazil a month ago. They asked what Brazil can learn from China. I said every country has a different history and culture.
It’s so hard to change. The best model would be something between US and China. Neither is perfect.
Larry Bernstein:
The Netherlands has a high-stakes exam at age twelve that splits children into a college track or a trade track.
In the US we don’t have high-stakes exams. Students get many chances. If you attend a second-rate college but you’re sharp, an employer will notice. In China, as you tell it, there aren’t many second or third chances.
Hongbin Li:
It’s not very forgiving for sure.
Larry Bernstein:
That seems crazy to me that a two-day exam in high school would determine the course of your life. The test isn’t that good. Your ability to memorize Chinese poetry may not be that important.
Hongbin Li:
That’s the big cost of this test. It has a lot of noise. When I was in China, we did some experiments, like adopting the AP system into China. You can take the AP test at any time. You don’t have to take everything on one day.
I was in charge of designing China’s AP economics test. I spent two years training high school teachers in how to teach economics. After two or three years, the reform stopped.
They realized that this new economics AP test, can only increase inequality in China. All the high school teachers are from elite high schools in big cities. That means the AP economics can only be offered in the urban areas. Rural areas don’t have resources. They are going to lose more opportunities to go to college. That’s why it was stopped.
Larry Bernstein:
That outcome doesn’t seem right to me. You have a chance to educate a group of children, but because it isn’t available to every child, no child gets it.
Hongbin Li:
It’s not efficient. We talked about the debate in Palo Alto. One idea is to have no levels in math education. Every kid should take the same math every year, no matter how talented you are. That’s called equity. Some people think this is right. Others think that kids should be leveled, allowed to skip grades in math.
Larry Bernstein:
I graduated college a year early and got a job working at Salomon Brothers. After two years I was accepted at Harvard Business School, but because I was only 22, they said I had to work another year before matriculating.
I lived across the street from Brooklyn Law School, so I called admissions and asked if I could take some night classes. The woman said, “sure, just apply and take the LSATs.”
I asked for an exception, since I had been accepted to Harvard. She flipped through her best practices guidebook and said, no, nothing here says that if you get into Harvard, you’re automatically into Brooklyn Law School.
I decided that I would learn the material on my own. I went to the Brooklyn Law School bookstore and picked out the textbook and case materials for corporations and tax, and the woman behind the counter asked if I wanted to buy a nameplate.
“What is that?”
“It is a metal plate with your name on it, so the professor knows your name to call on you in class.”
“Yes, I will take a nameplate.”
I snuck in on the first day of night classes, and I put out my new nameplate. The professor asked, “can anyone define income?”
I raised my hand, and he said, yes, Larry Bernstein, why don’t you start us off?
One day I ran into a college fraternity brother in the hallway who asked what I was doing at Brooklyn Law. I said, “I’m sneaking into law school.”
He said, “Larry, that’s outrageous, I’m paying $50,000 a year and you’re not paying a dime.”
I said, “yes, but you get a degree. I only get an education.”
Education is free, but the credential is expensive. The credential matters today because the employer can’t properly evaluate the individual. But that can change.
Hongbin Li:
That’s going to change with AI for sure.
Larry Bernstein:
If that changes, you don’t need the high-stakes exam or the college education. Ultimately, all the employer wants to know is whether the candidate can do the task and has the chops for the job.
Hongbin Li:
A lot of Silicon Valley AI firms want to have Stanford students without a degree. They want them to drop out.
Larry Bernstein:
They just want the signal that the students passed the selection process.
Another thing I found interesting in your book is that what the Chinese people like about the exam is that it can’t be corrupted. Grades are corruptible, because you can bribe the teacher. Everything is for sale in China except that score. Talk about corruption and the importance of the high-stakes exam.
Hongbin Li:
A lot of sectors in China are corrupt. If you want to see a doctor, you need to pay the doctor some fees that we call red bags. The exam is free from corruption. This is the only place that is really clean.
That’s why it’s so hard to remove it unless you can have a substitute that people trust. With AI people can design other tests with little cost and can get the trust of people.
Larry Bernstein:
What are you optimistic about as regards high-stakes exams?
Hongbin Li:
Kids do learn through a process and they also learn how to remain competitive. If you look at the US-China AI competition, in the US, a lot of these AI scientists are from China. Kids are good at STEM, good at math, and they work hard.
Larry Bernstein:
American educators don’t like kids memorizing for a high-stakes test. They find the idea offensive, wrongheaded, and ridiculous. You’re saying that in a society where children give up their youth to do well on this test, it benefits society. That runs counter to the American ideal.
Hongbin Li:
There’s certainly a lot of memorization in the process. But to be good at physics and math, good memory is not enough. They must have trained the kids to think cleverly to solve a difficult problem. That skill may be important.
The Chinese testing, nobody knows what’s on the test. The test is written in a way that nothing is repeated. They have the top professors for each subject spend half a year to come up with questions for each year’s exam.
There are ten different tests to prevent cheating. They don’t know which one they’re going to use. That’s why you need to be flexible to solve these questions really fast. Those tests are much harder than AP tests.
Larry Bernstein:
There’s a reason the Chinese have a high stakes and the United States doesn’t. It reflects something deep in each society. What is that?
Hongbin Li:
The most important thing is opportunity. In the US, there are so many chances in your life. In China, this is almost the only opportunity. That’s why people work so hard on it.
Larry Bernstein:
Thanks to Hongbin for joining us.
If you missed our previous podcast, it was AI and Human Connections.
Our first speaker was Allison Pugh a Professor of Sociology at John Hopkins and the author of a new book entitled The Last Human Job: Seeing Each Other in an Age of Automation.
Our second speaker was Eddie Monroe who I met on my first day of college and is still a very close friend. He is working on teaching AI to love. Eddie is the founder of Ganesh Guides a startup building AI for inner growth and transformation.
We discussed the role of AI with human connections. Allison was worried that AI will undermine human relationships and risk greater loneliness. Eddie is optimistic that AI can teach us to love.
Please join us next week with Ashley Rubin who will discuss her recent report on the politicization of sociology and its impact on scholarly standards.
Previous episodes and transcripts are available on our website as well as on Apple Podcasts and Spotify.
I am Larry Bernstein with the podcast What Happens Next.
Check out our previous episode, AI and Human Connections, here.
