"five selected but critical shortcomings"

 A Google search for econ led me to the Economics page at arXiv, a collection of econ papers. In that archive I found the landing page for Macroeconomic and financial management in an uncertain world: What can we learn from complexity science? (PDF, 23 pages) by Thitithep Sitthiyot of the Public Debt Management Office, Ministry of Finance, Thailand.

I know I'm not supposed to judge a book by its cover, but the title of this paper induced me to download the thing. Glad I did. I got as far as the Introduction where I read

Section II discusses five selected but critical shortcomings of existing knowledge in macroeconomics and finance.

That's what I want to present today. This is from the opening paragraph of Section 2:

The main shortcomings in macroeconomics and finance have to do with models and assumptions. For macroeconomics, the model that is being widely used by economists to analyze and forecast the effects of economic policies on the economy is known as the Dynamic Stochastic General Equilibrium or DSGE model. This model has been heavily criticized by many scientists outside the field of economics and by a number of economists that many assumptions imposed in the DSGE model are not consistent with empirical observations.

In a footnote at that point, Sitthiyot says: "Ormerod (2006) refers to Kenneth Arrow, the winner of the Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel in 1972, who regards the DSGE model as being empirically refuted." The paragraph resumes:

This paper chooses to discuss three main assumptions imposed in the DSGE model that do not fit with what happen in reality, namely, the equilibrium state of the economy, the external random shocks as the only factor that could affect the system, and the representative agent with rational expectations.

I like this paper because it is well organized, and because it gathers and presents specific complaints. Below I present a shortened version of the five critical shortcomings given in Section 2 of the paper:

The main shortcomings in macroeconomics and finance have to do with models and assumptions. For macroeconomics, the model that is being widely used by economists to analyze and forecast the effects of economic policies on the economy is known as the Dynamic Stochastic General Equilibrium or DSGE model...

1. Models
This paper chooses to discuss three main assumptions imposed in the DSGE model that do not fit with what happen in reality, namely, the equilibrium state of the economy, the external random shocks as the only factor that could affect the system, and the representative agent with rational expectations.

1a. The Equilibrium State
As its name suggests, the DSGE model assumes an equilibrium state of the economy. Casti (2010) illustrates that, in the real world, an economy where supply and demand are in balance never happens even approximately. According to Helbing (2015), economic system is unlikely to be in equilibrium at any point in time. Rather, it is expected to show a complex non-equilibrium dynamics. Ball (2012) notes that the equilibrium assumption originates from microeconomic theory as an analogue of equilibrium physical systems such as gases which have stable and unchanging states. The physical sciences, however, have long moved on to describe non-equilibrium process such as weather system but economics has not...

1b. The External Random Shocks
Kirman (2010) argues that, by and large, the fluctuations of the economy are the result of interaction among agents who make up the system and not due to some exogenous shocks. Kirman also refers to Sornette (2003) who makes a similar point that a stock market crash is not the result of short-term exogenous events but rather involves a long-term endogenous build-up with exogenous events acting merely as triggers. The idea that endogenous factors could gradually cause the system out of equilibrium is not entirely new, however. It has long been recognized and studied by many disciplines such as physics, biology, ecology, and sociology...

1c. The Representative Agent with Rational Expectations
In addition, the DSGE model assumes that the whole population in the economy can be represented by a representative agent with rational expectations who tries to maximize expected utility at any given period subject to inter-temporal budget constraint. It is as if one person’s thought can be used to represent the way in which everyone else in the entire economy thinks. Clearly, the model ignores the interaction among different agents comprising the economy and positive feedback which could cause emergent phenomena.

There is more. I am trying to present the minimum that shows Sitthiyot's approach and the concept presented in the paper. (The outlining, in bold text, is mine, added to be sure I captured all five points in this short overview.) Section 2 continues:

2. Assumptions
While the drawbacks of macroeconomics are due mainly to the model and assumptions, the problem in finance has to do with assumptions. There are two assumptions this paper chooses to discuss. The first assumption often imposed in finance is that price changes are independent... The second assumption is about the distribution of data.

2a. Price Changes are Independent
The first assumption often imposed in finance is that price changes are independent. One could think of tossing a fair coin as a metaphor. The results coming out of a coin tossing are independent from each other. The coin does not have a memory whether it landed head or tail in the past... In reality, changes in prices are not independent. Financial data have a property of path dependence or long memory. Normally, big changes are followed by big changes, the so-called clustered volatility...

2b. The Distribution of Data
In addition to the assumption that price changes are independent which is not consistent with empirical observations, financial data do not follow normal distribution as assumed in modern finance. Rather, it exhibits power law distribution with fat tails. According to Haldane and Nelson (2012), the power law distribution with fat tails implies that the probability of large events decreases polynomially with their size while, in the normal distribution world, the probability of large events declines exponentially with their size, making large events increasingly rare at a rapid rate. In contrast, under power law distribution, these large events are much more likely.

To provide a numerical example of how risky it might be if one assumes normality of distribution of data, this paper refers to a study conducted by Benoit Mandelbrot using daily index movement of Dow Jones Industrial Average during the period of 1916-2003. Based on Mandelbrot’s empirical findings, normal distribution implies that there should be fifty-eight days when the Dow moves more than 3.4% while in fact there are one thousand and one. In addition, normal distribution predicts six days where the Index swings beyond 4.5% whereas there are three hundred and sixty-six days according to the empirical observations. And lastly, the Index that swings more than 7% should come once every three hundred thousand years as predicted by normal distribution while the twentieth century already observed forty-eight days. It is crystal clear based on this empirical evidence that assuming that data have normal distribution would highly underestimate risk.

That's all five points. Gathering them together makes them easy to understand and makes a strong critique of economic thought. And I have to say the presentation of Mandelbrot's numbers is awesome!