Read an Excerpt

CHAPTER 1

Modeling Financial and Economic Systems

There are no known laws of nature that govern human behavior in the same sense as physical laws govern the motion of particles. This makes it difficult to model socioeconomic and financial systems mathematically. We can always find equations that explain past observations but it is increasingly clear that such models may not have predictive power. If we did have a model that could predict the effects of policy decisions on the economy, would the Board of the Federal Reserve need to vote?

How can a model that explains the past be unable to predict the future? Consider a tribe for which rain is so important that it is a god. Why does it rain? The rain god is benevolent. Why did it not rain this time? The rain god must be angry. Ceremonies must be performed to propitiate him. What if it does not start to rain after the rituals? Perhaps something was missed. More intense ceremonies are now needed. Rain at this stage would only be a lesson to the tribe to treat their traditions with more respect. What if it still does not rain? The council of priests pronounces that the tribe has made the sun god unhappy and ceremonies are needed to correct that. No rain? The sun god and the rain god are fighting. Explanations and ceremonies keep changing till it begins to rain. When it does, as far as the tribe is concerned they not only explained why it had not rained, they even managed to fix the problem. Over time the pantheon expands and becomes a fundamental paradigm for all aspects of life. Life, after all, depends on rains.

A model that starts with an assumption that is not true does not get closer to reality as more complexities are added to explain increasing quantities of past data. It is over-fitting. This is not to argue that such models do not have any practical utility in giving us a sense of control over matters related to our survival. When all else fails, ceremonies to propitiate rain gods are still carried out in drought-hit areas across the world.

Our mathematical models for financial and other social systems may actually be one step worse. The tribe does not have to poll its members as to whether they believed it had rained in some imaginary place to decide if their model worked satisfactorily.

1.1 Fundamental Issues

The quantities we model are not physical values, and belief in models influences realized results. If everyone believes a model that predicts the stock price of a company to be $100, the stock price becomes $100. As the company releases new results it may turn out the stock price should have been different. We then refine parameter values based on the new data and may even add new explanatory variables. Calculating the simplest of parameters, like the β (beta) of a company which describes how its returns change with market returns, is not straightforward. If we use too short a period to measure the relationship the results would not be statistically significant. If we use too long a period the company itself could have changed over time as also the markets, which means the relationship would no longer be valid. At times we even introduce an explicit factor, how many historical values to use, and estimate this dynamically for estimating parameters to come up with a value that would give the best possible fit. Models with parameters whose value can change depending on external conditions are especially difficult to validate because if they are complex enough we can always find parameter values that will fit a given set of observations. In physical sciences we can run experiments under controlled conditions to guard against over-fitting. In economic and financial systems we do not have this luxury and depend on historically realized results. This brings us back to belief in models influencing realized outcomes. If we wish to look at more complex models, derivatives were meant to better manage risk. Borrowing from Keynes in a different context (Kuehn, 2013), we would suggest such models are "an extraordinary example of how, starting with a mistake, a remorseless logician can end up in Bedlam." Computers are the ultimate remorseless logicians. We believe their predictions at our own risk.

Agent-based simulations may provide an alternative but such simulations require that agents and the rules of the system be modeled properly. The basic idea is very simple. Instead of developing complex equations we simply assign individuals some properties, have them follow a set of rules, and run some simulations to see what kind of behaviors and results emerge. Swarm behavior became a popular venue for research when it was realized that extremely simple rules can lead to complex observed behavior, such as the v-shaped patterns migratory birds take up.

The problem, again, is that humans are not particles with simple equations that can describe their behavior. Modeling human agents involves the types of beliefs individuals may hold and how such beliefs form and evolve (Gargiulo and Huet, 2012; Kitto and Boschetti, 2013), their decision-making processes (Yukalov and Sornette, 2014) and their actions and their interactions with other agents in a manner that could be cooperative or competitive in varying measures (Brede, 2012; Schweitzer et al., 2013; Janssen et al., 2014). What constitutes successful behavior may change over time (Kampouridis et al., 2012). Individuals may have significant differences in their power and ability to effect change, and may even have different proclivities as to the extent to which they follow the rules of the game. The enforcement of rules is carried out by individuals and may not always be fair or uniform. The rules themselves are ultimately endogenous to the system. It is the emergent behavior of individuals that leads variously to meta-stable tribes, gangs, kingdoms, empires, nation-states and their alliances, each with its own rules. Such systems can evolve over centuries, or even collapse within days, based on internal and external impetuses. Agent behavior in a specific context, such as in financial markets, could be influenced by broader social, cultural and political systems and by rapidly changing technological regimes, along with individual psychological factors. In a sense the complexity is shifted to modeling agent behavior. If it were easy to define individual behavior and preferences a priori, or if it were sufficient to work with some averaged assumptions, would companies that provide all services for free in exchange for collecting massive quantities of detailed personal data about individuals have such high valuations?

Perhaps we are being over-dismissive. How can computerized trading be so widespread if we do not have any financial or economic models that can actually predict outcomes? Computers, after all, work on garbage-in garbage-out and any errors would lead to the wrong answer. The originator of computerized trading realized that he did not have to predict the correct stock price, if such a thing even existed, only the stock price large investors would calculate using their equations. As long as he could calculate this faster than the institutions whose calculations were not computerized, he could make a profit. Today, when not every large institution uses the same formulas, we have front-running where servers co-located with exchanges get data a fraction of a second before others. Large and consistent profits are made via dark pools which allow a fraction of a cent profit based on orders already placed, multiplied by the large number of trades which are conducted at high frequency. It has little to do with technology facilitating free markets and better predictions, but rather using technology to exploit loopholes. Many major financial institutions survive only because they are too-big-to-fail.

1.2 Libertarians and Egalitarians

Model or no model, economic and financial decisions have to be made. In the absence of meaningful predictions from objective models we are forced to turn to philosophical arguments which are not necessarily better founded. Human evolution has led to the conflicting characteristics of compassion and competition, and both help with our survival. The corresponding philosophies of egalitarianism and libertarianism profess similar goals of enabling people to live happy, meaningful lives even as they prescribe divergent means to this end. Libertarians do not wish for productive individuals to limit themselves in terms of wealth generation or to be forced to work for free-riders. Egalitarians argue that concentration of resources could lead to owners of resources effectively being free-riders in perpetuity and so redistribution is essential. The commonality is the undesirability of forcing one to work disproportionately for benefiting another. The differences lie in what makes individuals happy and whose, and which, rights need to be protected (see for instance Bird, 2014; Gourevitch, 2014). This makes the viewpoints incompatible (Arnold, 2014). In other words the debates can never really reach a conclusion. They have been going on for several hundred years.

A related debate persists between systems with centralized control which rely more on cooperation, and free markets which promote competition.

Let us first show the basis for cooperation using a made-up example. Let us say you have a small piece of land where you grow some wheat and some vegetables. Your neighbor does the same. It is not very efficient; different crops need different conditions and some of the land is wasted in separating the wheat from the vegetables. If the two of you were to cooperate, one person could grow just wheat and the other just vegetables more efficiently. Then the two of you together would have more wheat and more vegetables than before. Imagine how much better things would be overall if everyone could cooperate together in this way. Mathematically, it is well known that global optimization over a domain dominates local optimizations over its sub-domains. Centrally planned economies should then outperform decentralized ones.

So why does it not work? Global optimization presumes a single global objective function whose maximization would be in the best interest of everybody. It means both of you like the same vegetables and both of you are happy if total production is increased. There are no issues with how much work each one of you have to put in, and how the total produce is distributed. It really is not easy to find such neighbors. If everyone were that understanding there would be no need to put up any system. In other words, walking shoulder to shoulder works great as long as everyone wants to go to the same place, along the same path, at the same speed, and is also happiest to have everyone else there with them. This does not hold across any large group of people over the long term. Individual objectives could, of course, temporarily become aligned under extreme circumstances. Your neighbor and you would probably be ready to join forces and defend yourselves against a group of outsiders trying to take over your land. Propaganda to convince everyone that they must work for some obscure common good, and ever-present external threats, become increasingly important for the continuing survival of such groups.

Free markets start by acknowledging that people are competitive. If everyone wants the best for themselves, it is not possible to have a central authority dictating what people should or should not do. After all, the central authority is also human and they will dictate what is in their own best interests. Their basic philosophy then becomes let people do what they want, and let them trade with each other without any restrictions. Sounds simple enough till you realize that lack of rules does not mean trade is free. In theory, one could simply take over someone else's goods and walk away if it was possible to get away with it. Worse, one could simply take over their resources and send them away. Ignoring this is ignoring the history of the world. So after we manage to acquire some property, one way or the other, we create rules for protecting it and set up common defenses. Will trade then be fair? Not necessarily. If one person owned some property upstream they could cut off your water supply. If you need water and do not have any, how much would you be willing to pay for it? Free trade means both parties come to the table without any compulsion and decide upon a mutually agreed value. A prerequisite is the absence of any threats to their survival. A negotiation with one party under the gun does not qualify as free trade.

So how do we ensure people are assured of their survival without placing any restrictions on economic activity? Everything is possible in theory, and so we take the easy way out. We simply assume that this happens. Magically.

Free-markets theory starts by assuming that everyone obtains their survival needs regardless of the happenings in the markets, characterized by "manna falling from heaven." We can then prove mathematically that no one would be worse off, and at least one person would be better off. Of course it is reasonable that free trade applies to trading surplus goods against surplus wealth.

If manna were to fall from heaven no one would care for economics or finance. There is no way to separate economic activity that affects survival of individuals from that geared towards surplus or non-essential production. Restrictions on the former necessarily imply those on the latter. Once such restrictions are in place, an analysis of why the system fails to deliver degenerates to debates over whether the problem is free markets or the restrictions placed on them.

If the underlying assumption is not satisfied then it follows logically that we cannot conclude free markets are good for everyone. Supporters of free markets argue that if everyone worked hard for their own benefit, they would all somehow get enough to survive, and then after this state is achieved all trade would indeed be free. Without any proof as to whether this always will happen in the real world, it is simply kicking the can down farther. If producers are entirely free to set prices, an individual may be unable to support themselves if the prices set by the market are greater than what they can afford. Others could be forced to make unfair trades for their goods or labor to ensure survival, implying greater profits for those who already have a surplus. Those with greater wealth can also potentially tilt the playing field, such as by repealing labor-protection laws or legalizing front-running, which again increases disparity. Assuming that compassion or personal charity can help satisfy the assumption is unrealistic, wealth not being positively correlated with compassion.

Of course we can always kick the can a little farther down by assuming everyone can borrow unlimited quantities of money at zero cost. Reality again does not allow for this. Perhaps then people can borrow any small amount of money, at some interest, and start a business that will generate enough income to support them and pay back the lender? Perhaps not everyone has attempted to get a business loan.

Finally we can always try to kick the can sufficiently far into some undetermined future with the assumption, more likely the hope, that things will return to rationality and everyone who works hard will not just survive, but thrive. Survival in the meantime remains a day-to-day issue for many, and no one has yet survived the long term.

The failure of the most extreme form of central control has left free markets as the dominant philosophy, but that does not make it true. It does not necessarily make it false either. We have to look at the results we have obtained in the current situation where we are trying to implement free markets without dropping manna from heaven, amid accelerating technological developments.

1.3 Outcomes

We have to perform our own ceremonies the best we can determine how. With cheap and abundant computing power we are naturally inclined to use mathematical models to guide policy at various levels. Outcomes in terms of realized results have been mixed.

1.3.1 The good

We cannot deny the benefits that current socioeconomic systems have brought to our lives. Work is easier and safer, with drudgery and risk increasingly automated away. Doctors can look up the latest medical information and even patient records online from anywhere, and telemedicine and remote surgery provide access to world-class medical care in far-flung areas. Genetic medicine can literally absolve us of the sins of our forebears. Computers are better, faster and cheaper, as is access to online information. Coursework from the best universities is available online. We can browse vast libraries online, for free. Entertainment is endless and we can talk to people anywhere in the world instantaneously. We have polluted the planet along the way, but information technology is perhaps the first technology that can help us reverse environmental damage. It can make devices more efficient and facilitate use of alternative energy sources. Our cars and trucks will soon drive themselves, faster and more efficiently, without our intervention. Trains run without engineers and aircraft can fly themselves. This will remove human error, the single largest factor in most accidents, of course once errors in the codes and automation equipment have all been fixed. Virtual reality and telecommuting can make a significant proportion of travel redundant. Robots can farm, cook, serve the food and clean up afterwards. Cities and homes will be smart, perhaps smarter than us.

(Continues…)

---

Excerpted from "Basics First"
by .
Copyright © 2016 Sema Dube and Manu Dube.
Excerpted by permission of John Hunt Publishing Ltd..
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.

---