Chinese / English
Home >> Great Minds China Forum >>  3rd Great Minds China Forum >>  Research Frontiers of Macroeconomics: Theories, Methodologies and ToolsResearch Frontiers of Macroeconomics: Theories, Methodologies and Tools
Xie Danyang:frontier theories of macroeconomics, methodology and tools


Xie Danyang: Haizhou added three topics according to his experience. I will also add three topics that closely linked to the theme of this afternoon, namely, frontier theories of macroeconomics, methodology and tools.


Let’s start with the first one. If you ask me which macroeconomic article has the greatest influence on policies over the past 50 years, my answer Kydland and Prescott's research on time inconsistency. Their groundbreaking research and subsequent discussions on the government's compliance or discretion have far-reaching impact on national monetary and fiscal policies. The governments are aware of the importance of compliance, and more than 20 countries, including major developed and emerging markets have adopted inflation targeting systems. Before the 2008 crisis, the global economy was characterized by high growth and low inflation.


When macroeconomists have argued that macroeconomics has entered the golden age and an era of science, global financial crisis occurred. After the crisis, people started to questioned macroeconomists, Krugman also criticized us. In terms of policies after the crisis, countries can no longer follow the old mindset. In fact, national policies have completely changed from old ones. The adoption of QE and the current negative interest rates are impossible to imagine before. In previous studies, why governments tended to compliance instead of discretion? I think it was because when we design the model, necessarity of discretion and the advantages of discretion under certainty conditions were not given fully consideration. I think it is necessary to add unforeseen risks, or even unaware risks to the model. For example, before the collapse of Lehman, on one thought Lehman will go bankrupt. But how to add such risks into the existing monetary and fiscal policy models? This is a difficult problem and may overturn the conclusions of the existing literatures. This is the topic we should think about.


Let's talk about the second topic. Over the past 30 years, there have been considerable advances in decision-making theory, behavioral economics, and behavioral finance. This is not fully reflected in the macro literature. Professor Miao Jianjun and I have made some attempts to introduce the concept of monetary illusion into the endogenous economic growth model to study how money illusions affect consumption-saving decisions and then affect economic growth. I think that the concept of behavioral economics such as monetary illusion, Prospect Theory and Loss Aversion can be transplanted into the popular DSGE model to discuss how these factors affect the transmission mechanism of monetary policy. I will guide my students to work on this direction.


Before talking about the third issue, I would like to briefly review the history of macroeconomics. Macroeconomic can be divided into two branches, economic growth and the study of the economic cycle. The most influential work in the study of the economic cycle was the "General Theory" of Keynes in 1936, which initiated the so-called Keynesian Revolution. The theory of economic growth can be traced back to the first chapter of Adam Smith's "Wealth of Nations", which is called the Smith theorem: “The division of labor is limited by the extent of the market”. Smith believes that the division and specialization of labor are the source of economic growth. The modern theories on economic growth include Solow's theory of exogenous growth and the theory of endogenous economic growth proposed by Paul Romer, Lucas, Grossman and Helpman, Aghion and Howitt. From the expression of these theories, Adam Smith's age is literal while Solow uses the Reduced Form, and the endogenous growth theory uses dynamic optimization. Our understanding on economic growth is further deepened when the forms of expression and analysis framework became more rigorous. The same trend also takes place in the study of the economic cycle.


Keynes's book also adds Reduced Form Analysis (such as consumer function, backward looking) in the texts. Later, Friedman introduced the Permanent Income concept into a consumer function (forward looking). And then Lucas, Sargent, Prescott, Bernanke/Gertler/Gilchrist, Mankiw, Jordi Gali, adopted the dynamic optimization. Prescott and others concerned about how technology shocks affect the business cycle. Bernanke et al. Studied the role of financial accelerators in the economic cycle. Mankiw, Gali and others introduced a variety of information sticky and price sticky into the dynamic optimization model and formed the new Keynes model. Did these dynamic optimization models reach to the peak of our macroeconomics? I do not think so. No matter it is economic growth research, studies on economic cycle, dynamic optimization models are relatively simple and abstract. Why it is simple and abstract? Because they basically profiled behaviors of representatives’ behavior: consumer representatives, vendor representatives, investor representatives, and even government representatives. In such models, we Herding Behavior cannot be described. There are some models that allow a certain degree of heterogeneity, but these heterogeneity is still very simple. How will macroeconomics develop in the future? What can young people do in the future? I think that with the improvement of computing power and with the progress of artificial intelligence, we can achieve the ideal mentioned by Lucas in 1988: “I prefer to use the term 'theory' in a very narrow sense, to refer to an explicit dynamic system, something that can be put on a computer and run. The construction of a mechanical, artificial world, populated by the interacting robots that economics typically studies, that is capable of exhibiting behavior the gross features of which resemble those of the actual world that I have just described”

So far, we have seen only Robot 1.0 (representative) or 2.0 (heterogeneous representative). What we study at is how these representatives respond to government policies or reforms. The idea is right, but far from sufficient since the problem solving process of Robot 1.0 and 2.0 is not as sophisticated as the thinking and decision-making process of human being. What we need is Robot Alpha. I've seen some hope when AlphaGo beat Kejie: It is the robot with human thought intuition. We can assign these Robot Alpha with differentiated types of thinking, computing power, risk appetite, or even various degree of efforts. Their decision-making processes and behaviors can be incorporate into the model to understand how a group responds to macroeconomic policies. Thus the potential effects of the policy can be studied. It is not impossible. I suggest that young students here focusing on technological progress. If you know anyone who are studying artificial intelligence, become their good friends for future cooperation. Interdisciplinary cooperation can often produce unexpected breakthroughs. Thank you all!


◆please indicate the source if authorized: National Economics Foundation

◆photo:National Economics Foundation