Professor Thomas Sargent, winner of the 2011 Nobel Prize in Economics, has been working on empirical research on causes and effects in macroeconomics. Professor Sargent, who has taught at Harvard University, New York University and other world-renowned universities, joined Peking University HSBC Business School in 2017 and serves as the director of the Sargent Institute for Quantitative Economics and Finance. In his opinion, "in this way, you can learn about the latest developments in China, and you can constantly correct your misunderstanding of China in the process of communication."
In an exclusive interview with Peking University Financial Review, Professor Sargent said that ChatGPT is a product of the era of big data, and it is very helpful in editing manuscripts, assisting in writing code, and reviewing reports. As for whether artificial intelligence will completely replace some people's work, Professor Sargent believes: "ChatGPT is like an assistant to us, which can help do some basic and repetitive work, but the final creation, revision, review, judgment and other work need to be done by humans." ”
This article was published in the 19th issue of Peking University Financial Review.
Peking University HSBC PFR
, like47
Rather than economist, Thomas Sargent prefers to position himself as an economic scientist, a positivist who likes to speak with numbers. In order to better explain the economic problems he studied, Sargent devoted a great deal of time to his studies in mathematics, sometimes auditing undergraduate and graduate courses in mathematics. "We're just talking on paper, looking at the numbers to try to figure out what's going on," he said. ”
In 2011, the high-profile Nobel laureate in economics was finalized at 7 p.m. on the 10th, and Thomas Sargent, a professor at New York University, and Christopher Sims, a professor at Princeton University, shared the award, and the Royal Swedish Academy of Sciences recognized them for their "empirical research on causes and their effects in macroeconomics". The official announcement of the Nobel Prize Committee in Economics said: "Thomas has made a very significant contribution to the interpretation of data in macroeconomics, and people can discover and learn from his research the impact of national policies, as well as other factors such as sudden price changes and sudden changes in product supply. ”
On the morning of winning the Nobel Prize in Economics, Sargent took the train to Princeton University to give lectures as usual, and it is said that the organizing committee of the Nobel Prize in Economics called Sargent several times that morning, but he did not receive the call because he was an unfamiliar number, and when he was told that he had won the Nobel Prize, he behaved very calmly. In fact, as early as 1995, Thomas Sargent "missed" the Nobel Prize in economics. Thomas Sargent and Robert Lucas, the leading exponents and long-time collaborators of the rational expectations school, were thought to be jointly awarded the Nobel Prize in economics by academics, but the 1995 Nobel Prize was unexpectedly awarded to Robert Lucas. One of the reasons is that Sargent's proposition of "policy ineffectiveness" and related theories have been questioned by many scholars.
When Sargent was a graduate student, he was interested in how to parameterize distributed time delays, an important topic in macroeconomics. Many famous economists of the time, such as Solow, Jorgensen, Griches, etc., tried various methods, but Sargent was not satisfied with these studies. In Carnegie. When he was a student at Mellon, Sargent was recommended to read a paper by John Moose (known as the father of rational expectations). In that paper, Mooss rationalized Friedman's adaptive expectation model by inferring Philip Kagan's expectation equation backwards. Mooss concludes that predictions of stochastic processes should also include distributed time lags and people's expectations of the future. "It set the stage for my later research and made me realize that I should delve in this direction," Sargent said. I think that in the model of rational expectations, people's expectations should not be the input in the theoretical model, but should be the output, part of the outcome. ”
The significant impact of the rational anticipation approach is not only emphasized by its adherents, but also acknowledged by its critics. James Tobin, a Nobel laureate economist, once noted that this kind of thinking is "too unusual and convincing to be a passing visitor on the theoretical stage." No matter what orthodoxy emerges, this idea is destined to develop. This enduring thinking, which goes beyond economic theory itself, is more methodological – it inherently consistently derives rational expectations and rational behaviors contained in the structural equations of the general equilibrium macroeconomic model. This thinking has been mobilized not only to raise the status of the 'invisible hand', but also to explain the causes and consequences of incomplete information, long-term contracts and other agreements, imperfections in capital markets, and many other well-known phenomena. Because of the application of this approach, the questions raised by Keynes must be interpreted from a new perspective".
From Keynesian believers to rational anticipatory revolutionaries
In his early studies of economics, Sargent was a believer in Keynesian economics. At this stage, he fully agreed with the Keynesian view of the use of monetary and fiscal policy to solve the problems of the economic cycle, and was convinced that fiscal and monetary policies have the power to solve macroeconomic problems. However, Sargent's belief in Keynesianism was shaken by the view of rational expectations, and he gradually realized that the introduction of rational expectations into the study of macroeconomics might change his previous understanding and perception of economics. Macroeconomic Theory is a complete account of Sargent's transformation from a devout Keynesian to a rational expectations economist.
The rational expectations hypothesis has come a long way since John Moth's seminal paper "Rational Expectations and the Theory of Price Movement" in 1961. In the 70s of the 20th century, Sargent, along with Lucas, Barrow and Wallace, reconstructed macroeconomic theory on the basis of previous research, including basic assumptions, micro foundations, model forecasts and policy implications. This came to be known as the "revolution of rational expectations." Thomas Sargent argues that this revolution is based on Keynesian economics and is an important result of the continuous development of Keynesianism, a continuous and inevitable evolution.
Rational anticipation refers to the fact that when people are rational in anticipating an economic phenomenon (e.g., market price), they will make the most of the information they receive to make the most of it. This means that any measures taken by policymakers will be counteracted by rational expectations and become ineffective. This was a blow to the idea of government intervention in the economy advocated by the West at the time, and was criticized by the older generation of Keynesian economists. While it is not the only possible, reasonable, or even "rational" choice, it is a method that has been proven in the intrinsic development of economics and has been, and can become, a common ground for independent research in many fields.
For example, the rational expectation hypothesis has been widely used in financial market research under the premise of efficient markets. The efficient market model asserts that security prices are free and flexible and reflect all possible information. That is, the price is related to the condition expectations. The theory is that price movements always follow a stochastic pattern. If past prices or volumes play a role in predicting future price movements, it won't be long before Wall Street technologists spot these patterns. When technologists start acting on their findings, prices adjust so that these patterns disappear. For example, if people predict that the stock price will increase by 5% over the weekend, investors will swoop to buy the security until its price rises by 5%. Prices will be raised then, not until the weekend. As a result, there are no ready-made models estimated from past data that can be used to predict future price movements. Therefore, the price movement is necessarily random.
In practice, the rational expectations model played an important political role in promoting Reagan's economic plan. Republicans have developed a policy platform for massive tax cuts for individuals and businesses. They predict that their policies will reduce inflation and stimulate economic growth. Traditional macroeconomic models hold that higher economic growth means higher inflation, and that lower inflation can only be achieved in conjunction with lower economic growth. The answer to this conundrum is that by announcing policies that lower the rate of money growth, we can reduce inflation without sacrificing real growth. In this way, the simple rational expectation model makes the economic propositions in the Republican platform seem viable.
However, this new economic program did not convince the media circles and most sectors of the economy. The creators, owners and operators of existing econometric models believe that the ideas on which the forecasts are based are just untested speculations. As Thomas Sargent has pointed out, "Economists do not now have reliable, empirically testable, rigorous models that allow us to accurately predict how long it will take and how much it will cost in terms of output and employment for such a change in order for institutional change to be effective."
Sargent recognized that further evidence must be relied upon for examples of an immediate reversal in the direction of inflation after a change in policy regime. In his view, several hyperinflations during the twenties of the twentieth century (Austria, Hungary, Poland and Germany) provided a rare experience for the study of changes in the management system. One of the most striking features of these examples is the rapid and abrupt stabilization of price levels and exchange rates in the wake of drastic monetary and fiscal reforms. These inflationary processes do not end gradually, but come to a halt all at once; The obvious impetus for the inflationary process is completely gone.
The important historical role of Sargent's economic thought lies in the fact that he systematically perfected the more precise formulation of the theory of rational expectations from the perspective of positivism. For example, Sargent constructed a series of mathematical models, taking the rational expectation assumption as a major economic factor in the mathematical model, enriching the development of structural macroeconometrics. At the same time, he carried out the stability analysis of empirical tests, put forward the robust control theory, and proposed the adaptive expectation theory, which constructed a more systematic rational expectation analysis framework and triggered the rational expectation revolution of macroeconomic theory.
Are government policies ineffective?
Sargent once said in an interview: "There are two Chinese proverbs that illustrate the essence of rational expectations very well, the first is 'there is a policy above, there is a countermeasure below', and the second is 'people have no long-term concerns, there must be near-term worries.'" Before engaging with rational expectations, my research and that of some scholars was based on the assumption that governments are better able to predict the future than people. When applied to economic policy, such theoretical models provide the government with the basis to change policy by systematically and consistently 'deceiving the people'. Arguably more democratic, the rational expectations hypothesis holds that governments and individuals make predictions about future interests on the same basis, thus eliminating the means by which governments can influence policy. ”
In the late 60s and early 70s of the 20th century, stagflation occurred in many countries around the world, and Keynesian economic theory began to be questioned by the economic community, and its position began to waver. In Keynesian macroeconomics, the government is the main body and focuses on the analysis of a series of economic aggregates, but this analytical framework lacks the behavior analysis of micro units, resulting in the lack of necessary microeconomic foundation in Keynesian macroeconomic theory. When Keynesian macroeconomics fell into a theoretical and practical dilemma, the theory of rational expectations was gradually developed through reflection on Keynesianism and absorbing some ideas of monetarism.
After introducing the rational expectations hypothesis into macroeconomic policy analysis, Sargent first put forward the proposition of "policy ineffectiveness" in 1975, which set off a revolution in rational expectations. Before Sargent put forward the proposition of "policy ineffectiveness", Keynesians insisted that the government can intervene in the operation of macroeconomic policy through fiscal policy and monetary policy, while monetarist economists advocated the role of monetary policy and denied that the government's use of fiscal policy can bring positive effects. Sargent's theory of rational expectations is based on a long sequence of practices, and he argues that over a longer period of time, the effects of macroeconomic causality caused by rational expectations will invalidate government policies. For example, economic agents' expectations for the future of the macroeconomy will affect their current decisions on investment, consumption and savings, while government policymakers' expectations of economic agents' activities will affect current policy making. When the government of a certain country lowers the reserve requirement ratio in order to implement expansionary policy regulation, people predict that the supply of money will increase and the price of commodities will rise, and raise the buying and selling prices of commodities in the current economic activities, so that the increase in commodity prices is equivalent to the level of money supply.
At the same time, Sargent pointed out that if the monetary authorities secretly increase the money supply, then although people will not make any expected judgments, the price level will continue to grow in the short term due to this artificially created information asymmetry, resulting in short-term invisible inflation. This inflation can lead manufacturers and consumers to mistakenly believe that the demand for products and labor increases, which in turn increases the supply of corresponding products and labor, and eventually leads to the level of output deviating from the output corresponding to the level of natural rate, which Sargent believes can only be maintained for a short time. Manufacturers and workers will quickly realize that this increase in demand is only an "illusion", that actual demand has not changed, that price growth is mainly caused by short-term inflation, and thus adjust their expectations again, reducing the supply of products and labor, so that the level of output will return to the natural rate, and monetary policy will only bring inflation in the end, and will not help the actual sustained level of output.
Although this proposition has been widely controversial in economic research, the discussion of this proposition has had a profound impact on the development of macroeconomics, promoted economists' thinking about the role of macroeconomic policies, and also revolutionized the research methods of macroeconomics.
New research in the age of artificial intelligence
Sargent received his B.A. from Berkeley in 1964 and his Ph.D. from Harvard University in 1970. He has taught at the University of Minnesota, the University of Pennsylvania, the University of Chicago, Harvard University, New York University, and many other world-renowned universities. In 2017, Professor Sargent joined Peking University HSBC Business School and served as the Director of the Sargent Institute for Quantitative Economics and Finance. He believes that "in this way, you can learn about the latest developments in China, and you can constantly correct your misunderstanding of China in the process of communication."
Since the late 80s of the 20th century, Thomas Sargent has been involved in research at the Santa Fe Institute and has begun to explore new themes such as complexity, unpredictability, adaptability, etc. During this period, widespread concern was raised about how economic agents consider specific issues for the future when making decisions. Thomas Sargent's research on rational expectation theory gradually shifted to focus on the study of "learning theory". At this stage, Thomas Sargent began to study bounded rational applications in the field of artificial intelligence. Thomas Sargent himself admits that he still encountered frustration at this stage of his research, and that neither adaptive anticipatory learning nor learning through artificial intelligence could establish the symmetry he sought.
At present, Sargent is working hard to study the latest achievements in the field of science and technology such as artificial intelligence. He said that the economy, like science and technology, is closely related to human development, and the distinguishing feature of intelligence is the ability to identify (recognizing patterns) and (generalizing). In his view, from an economic point of view, the application of artificial intelligence is conducive to reducing global trade barriers and building trust, so that more value will be created, as he pointed out in an interview at Xiamen University, "some outstanding entrepreneurs have also known how to combine these tools and make some new attempts, and at the same time integrate existing things in innovative ways." As more computers are connected to the same network, we can use competition to make monopolists disappear, we want to gradually decentralize and share information, everyone will have a ledger, and we will use economic incentives to build trust."
《北大金融评论》：人们普遍认为ChatGPT这样的生成式人工智能将重塑经济，您在《The Artificial Intelligence Frontier of Economic Theory》一文中到“With Big Data,Faster Computers and Better Algorithms,We Might See Patterns where once We Heard only Noise”，ChatGPT爆火至今，其发展是否符合您这一观测？您认为ChatGPT应用的终极形态会是怎样的？
Thomas Sargent: Historically, scientists like Galileo, Kepler, etc., collected or observed data in a variety of ways, and then simplified it into several equations to generalize it, and then formed a model. In the early stages, humans had to adapt small models because of the limited amount of data they collected. But as the amount of data available grows exponentially, so do the models that can be adapted, and ChatGPT is a product of the big data era.
ChatGPT is very helpful when it comes to editing manuscripts, assisting with writing code, reviewing reports, and more. For example, where I work in the United States, I am required to file various reports. At one point, I asked ChatGPT to write a report on how to make research more diverse and inclusive, and it was very well written. In addition, it can do some interesting things. For example, I used it to write a Shakespearean love poem for my wife. However, if I ask it to summarize relevant information and materials in a certain field that can be used to help me with in-depth analysis and understanding, then ChatGPT will disappoint me. Because ChatGPT can't tell whether information is true or false, the information it can provide you isn't necessarily true either.
Some people say that ChatGPT may replace some people's jobs, but I don't think so. ChatGPT is like an assistant to us, which can help do some basic and repetitive work, but the final creation, revision, review, judgment and other work need to be done by humans.
Peking University Financial Review: According to Schumpeter's theory of creative destruction, under the market model, innovation, as a force that creatively destroys the equilibrium of the market, will constantly innovate the economic structure from within, and dynamic imbalance will become the "normal" of economic development. So, how do you see the "disruptive" nature of future innovation? What kind of challenges will it bring to our social development? How do we deal with it?
Thomas Sargent: Creative destruction happens all the time. For example, the birth of smartphones is one of a kind, which will make young people addicted to mobile games and social media, but you can't stop it. In the United States, startups abound, and of course, there are many people who fail, or almost say, the vast majority of startups fail, but the wind of entrepreneurship is still unstoppable. Years ago, the U.S. government tried to save a company that was about to fail, but it turned out to be a no-brainer.
If it's creative destruction, something new has to be produced. In terms of the current ChatGPT's capabilities, I don't think it's up to the standard of "creative destruction." Because it only replaces some of the basic work we do, and there is no new invention output.
There's a team of geeks at the moment and I think what they're doing falls into the category of "creative destruction". They use machine learning tools to overlay a large number of complex mathematical models to create "geometric deep learning", and do things on this structure, such as advancing the frontiers of biology, which may help us live a little longer.
My father had just passed away, and the job he used to do was now replaced by a computer. I feel that one day, my job, as an economist and a professor, will be replaced by machines. Therefore, our work should not only pass on existing knowledge and skills, but also constantly create new knowledge and skills.
My understanding of creativity is a creative mindset, or rather, it can make predictions more accurate.
Written by: Du Wenxin, Liu Baixiao, Yang Jingwen
Bibliography:
1. Steve M. Schaefflin, trans. Li Zhenning, Rational Expectations, The Commercial Press, 2016.
2. Zheng Xin, Thomas Sargent's Research on Economic Thought, Ph.D. dissertation, 2021.6.
3. Li Dawei and Wu Si, China Economic Report, 2016.
4. Firmament, what can we learn from the doctrine of "rational expectations"? Shanghai Economic Research, 1993.