A thorough and comprehensive analysis of the history and effects of quantitative easing (QE), a policy introduced by many of the world's leading central banks, notably the Fed, Bank of Japan, European Central Bank, and the Bank of England, in the aftermath of the Great Financial Crisis (GFC), 2007– 09, is most welcome. This is exactly what my friend and colleague, Jonathan Ashworth, has now provided in this book. We worked together in the economic research section of Morgan Stanley in London for over half a decade and have jointly authored several published papers. Perhaps he will forgive me if I start by setting out some of my own idiosyncratic thoughts on our recent experience with QE.

Prior to the sluggish recovery from the GFC, the idea that central banks might struggle to raise inflation would have been laughed out of court, regarded as hare-brained. Consider the following syllogism:

In the exercise of QE the central banks involved did expand their assets and monetary liabilities by a large multiple of their initial starting point, often four or five times, in one case returning the central bank to its earlier nineteenthcentury status as the largest bank in the country, i.e. in the UK. Moreover, with the non-banks’ demand for currency growing slowly (another area of analysis where Jonathan and I have collaborated), the bulk of this ended up in commercial banks’ deposits at the central bank, which at one point in the United States expanded by over 100 times their initial value.

So what went wrong? While a few braver souls are beginning to query Premise 1 above, the main blame is usually put on the inability of the central bank to guarantee a commensurate expansion of the broader money supply by raising the monetary base. The money multiplier has collapsed. The commercial banks appear to be in a liquidity trap, where they cannot expand their assets, either because the demand for (profitable) lending is not there, and/or they do not want to buy securities, since the meagre return (at low, or negative, interest rates and a flat yield curve) is insufficient to match the added risk and regulatory requirements.

The Bretton Woods system of quasi-fixed exchange rates that had been in place for much of the period since the end of the Second World War finally came to an end in early 1973. Under Bretton Woods, the value of the US dollar (USD) had been fixed against gold while other currencies such as the British Pound (GBP), German Deutschmark (DM), etc. were pegged against the USD, with only periodic adjustments permitted in order to help countries correct deficits in their balance of payments. Its demise led to the major industrialized countries allowing their currencies to float freely on global financial markets. The move from fixed to flexible exchange rates allowed national policy-makers to shift from having to use economic policy to help maintain the economy's external balance (the balance of payments and the exchange rate peg), to a primary focus on using it to better achieve the economy's internal balance (stimulating/dampening aggregate demand, achieving full employment, etc.).

The main tool used by policy-makers to manage aggregate demand and control inflation since the end of Bretton Woods has been monetary policy, particularly official short-term interest rates. This primarily reflects the fact that changes in interest rates are much easier to enact and typically impact the economy much faster than fiscal policy. Indeed, for obvious electoral reasons politicians are typically averse to raising taxes or cutting government spending to slow growth, while to avoid wasteful spending or what are commonly referred to as “white elephant” projects, it can often take considerable time to plan and start new infrastructure projects when activity needs stimulating. Tax cuts have remained a favourite recession fighting tool in the US, although a number of economists have questioned their effectiveness, suggesting that they are more likely to be saved (Rivlin 2015).4 Meanwhile, countries where the power to enact fiscal stimulus does not lie solely with the executive branch of government, such as the US, may suffer resistance to and/or delays in passing timely legislation.

Should a seller in a prominent position in its market, knowing potential customers will sample its goods before those of its rivals, charge a higher or lower price than its non-prominent competitors? Should a government increase competition or information provision in order to protect customers from being exploited by charlatans selling services of no actual value? Should we allocate our limited memory across all the decisions we have to make or should we devote it to only one of them? A research project in the 1950s at the Graduate School of Industrial Administration at the Carnegie Institute of Technology (now part of Carnegie Mellon University) led to two very different views of how we make decisions, which have very different implications for questions such as these. This project split the Graduate School, and the economics discipline as a whole, in two, but also led to the concept of bounded rationality, the study of which has become one of the most interesting and wide-reaching fields of economics. We explore this field in what follows, which will take us into the realms of mathematical modelling, artificial intelligence and cognitive science. It will also lead to surprising insights about situations as diverse as how businesses should launch new goods, the likelihood of centrist candidates being successful in elections and how governments should protect our interests most effectively. This exploration will enrich our understanding of the everyday situations in which we find ourselves and will at times show what we think has been proved by traditional economics to be incorrect, even harmful.

The schism

The research project focused on how businesses make decisions and was conducted by a team made up of some of the smartest minds in the field. John Muth's interpretation of the resulting data was that businesses make their decisions according to what became known as rational expectations. According to this view, businesses are able to accurately identify relationships and lines of causation between the variables that are relevant to them. They are able to quickly establish the likely effects of any changes in their environment, enabling them to continuously identify their optimal courses of action.

The studies included in this book have been selected because of the story that is revealed when they’re knitted together. The hope is that the previous chapters, particularly the third and fourth, demonstrate that incorporating bounded rationality into our economic models leads to three invaluable outcomes in comparison to traditional models based on complete rationality: (1) more accurate explanations of economic phenomena and behaviour; (2) more effective policy recommendations; and (3) wider and richer insights into the outcomes of our behaviour. These are the conclusions that this book has sought to reveal, the supporting findings for which are summarized in Table 5.1.

An interesting insight becomes immediately apparent by looking at Table 5.1. Although the sample of bounded rationality studies examined in this book is small in both absolute and proportional terms, it shows the merit of considering both optimizing and non-optimizing approaches when modelling bounded rationality. Studies that dispense with the optimization assumption and instead model us as seeking outcomes that are simply good enough are particularly adept at providing more accurate explanations of our observed behaviours: of explaining our behaviours that models of complete rationality are simply unable to do. That we tend to congregate around two specific points when playing a three-player Hotelling game; choose to deny charges (or to cooperate with our opponent) when playing a repeated version of the prisoner's dilemma game, and readily contribute to the provision of public goods are just three examples. Studies that maintain the optimization assumption but impose additional constraints on the decision-makers involved, on the other hand, are particularly adept at revealing wider and richer insights about economic situations. That it is better for us to devote our limited memory capacity to one choice amongst the many we face; that centrist candidates are unlikely to be successful in their bid for power if the electorate is polarized, and that businesses are likely to charge lower prices compared to their rivals when they know customers will consider their goods first are again just three examples.

The first general approach to modelling bounded rationality is to dispense with the assumption that decision-makers seek to optimize the outcomes of their decisions, whatever those outcomes may be, and to instead assume they seek to find outcomes that are simply good enough. Herbert Simon coined the term satisficing to refer to such decision-making: an archaic word of Northumbrian origin that simply means satisfying. Simon maintained that satisficing is fundamental to modelling bounded rationality because it vastly reduces the amount of information decision-makers need to process to reach any given choice, making decisions possible with cognitive capacities that are insufficient to make them optimally.

As illustrated in Figure 1.5, there are at least four ways in which satisficing has been modelled within the literature: satisficing search, the use of heuristics, probability-based responding and the use of algorithms. These will be explored in turn in this chapter, although the ways in which they overlap and interconnect with one another make this division somewhat arbitrary in places.

Satisficing search

Consider again the scenario outlined in Chapter 1 and represented in Figure 1.2, in which a consumer has to decide how to spend her budget on bottles of wine and wedges of cheese. As outlined in the introductory chapter, complete rationality implies she knows all possible combinations of the two goods that are available for her to buy; she is able to rank the available combinations in order of how much utility she would gain from each; and she is able to identify from the combinations she can afford the precise combination that brings her greatest utility. Employing a satisficing search process significantly reduces these mental demands. Instead of seeking to identify the single combination that maximizes her utility, she instead seeks to find a combination with which she is satisfied. The first step in doing this is to establish her aspiration levels, which in this case refer to the minimum amounts of wine and cheese with which she would be satisfied. These are illustrated in Figure 3.1 as ALW and ALC. In this case, she is content as long as the combination comprises at least two bottles of wine and one wedge of cheese. She then simply engages in a search process, going from shop to shop, or flicking from website to website, to find a way of buying a combination that meets these minimum criteria with the budget at her disposal.

The history of bounded rationality as a concept and approach to studying economic behaviour, which is briefly discussed in this chapter, is one of vehement rejection before growing recognition and a widening and deepening of interest. It starts with the marginal revolution of the 1870s.

The marginal revolution

The early modern economists, also known as the classical economists, were interested in understanding the ways in which our societies function, including the laws that govern them, the pressures that cause them to evolve over time and their likely eventual outcomes. They adopted the approach of studying the logical outcomes of both our individual behaviours and the relationships that exist between different people within our societies. This led Adam Smith to show that our self-regarding behaviours actually promote the interests of society as a whole; Thomas Malthus to demonstrate that our propensity to reproduce whenever we are healthy enough to do so forever consigns us to living in the most impoverished state; and Karl Marx to assert that exploitation of the working class by capitalists will inevitably result in our economic system being overthrown. It is perhaps no wonder the subject became known as the dismal science.

By the 1870s, dissatisfaction with this analysis was rising along with a desire amongst economic thinkers to establish a much more rigorous analytical foundation for the discipline. The development of such a foundation began in the works of William Stanley Jevons, Carl Menger and Léon Walras in what is now known as the marginal revolution, with the concept of the margin (simply the smallest single unit) being at the heart of their work. This concept opened the door to the model of complete rationality as it enabled economic thinkers to identify the optimal solution to any given problem.

To see this, consider again the individual in the previous chapter who's deciding how many bottles of wine and wedges of cheese to buy for a social gathering, but this time consider how much she should drink on the evening her friends come round. Having a glass of wine can undoubtedly heighten her enjoyment of the evening, perhaps helping her to relax and let go of the stresses of the day, but drinking is subject to the law of diminishing marginal utility illustrated in Figure 1.1. The first glass brings her the greatest additional utility, with the second bringing somewhat less and the third bringing less again.

The modern view of the decision-making process, illustrated in Figure 4.1, is one in which we draw information from our environments but also work with those environments, shaping and using them, to make our decision-making more straightforward. We outsource the storage of certain information and the provision of certain solutions in order to ease the cognitive burden we bear and so reduce the decision fatigue we experience. Consider the “to-do lists” we magnetically attach to our fridges or the hooks we put in our cupboards so we can hang up the keys to the utility meter boxes and the shed. Such uses of our environments are known as Gibsonian affordances and we use them widely to free our minds of unnecessary mental demands. A decision-maker is faced with a problem and so begins a search of her environment for relevant information but also for opportunities that will be helpful in solving it. It is through repeating such a search-affordance-response sequence that she identifies her choice from amongst the possible alternatives. In a sense, the environment holds the solution as well as the problem, and the distinction between the decision-maker and her environment is almost unidentifiable.

As Figure 4.1 shows, there are three elements to our decision-making process: the information we generate from our environments, our ability to process that information in conjunction with our environments, and the possible responses we find. The second general approach to modelling bounded rationality is to maintain the assumption that we seek optimal outcomes and so engage in optimization processes when making our decisions, but with additional constraints imposed at one or more of these stages in our decision-making. These additional constraints and their effects are examined in this chapter.

Constraining information

Consider a time when you’ve gone from shop to shop, or have flicked from one website to another, looking for a particular item. How did you remember the options and their prices as you did so? And how did you make the final choice? And consider the last time you voted in an election or tried to hire the services of someone with much greater knowledge than yourself, such as a therapist or plumber.

We hope that these dialogues have been a useful and fun way to engage with these ideas. To make sure we are all on the same page as we leave the page, we now want to summarize the core arguments that we have made and also to acknowledge what we are not saying.

Our ambition in Angrynomics has been to provide a primer on why the world is the way it is today. We also wanted to do that in a way that neither patronises nor expects the reader to trudge through a scholarly treatise. We also wanted to keep it short. That has costs. Some people may find our analysis partial and our proposals underdeveloped. If so, they have misunderstood the point of the book. Our aim has been to introduce these ideas, not to seek to prove them or make the case that nothing else matters.

For example, we both agree climate change is the most important challenge facing humanity. Indeed, from that perspective what we are writing about is trivia. But we think our decision not to start and end with climate change as the defining issue of the moment is justified on the following grounds. If you don't first do something about the anger, about restoring “economics” over angrynomics, then you can forget about ever getting serious policies to address climate change. We will never get them off the drawing board for fighting among ourselves, while allowing our tribal instincts to obscure our judgements.

Similarly, some may think we give short shrift to cultural explanations of these same phenomena, and indeed we do. But we do so because our framework allows us to do something about the underlying causes of angrynomics. Starting and ending with culture does nothing. It's a counsel of despair. Once you have decided that half your country is made up of people who are entirely different from you, and therefore wrong, it can only end badly. We need to aim higher than castigation and moral superiority as ends in and of themselves.

The parable of the good father turned populist

We first met John in March 2009. He was working as a waiter in an upmarket restaurant in Naples, Florida. We struck up a conversation. John is Irish-American, originally from Boston, he had recently moved to Florida with his family. His girls had grown up, been through college, and had found jobs. Only his youngest was still living with him and his wife. They had moved to Florida looking forward to a peaceful retirement. The kids could visit in the winter months and escape to the sunshine.

Several months prior, John's youngest daughter was diagnosed with a very rare, but life-threatening, heart condition. Eventually, after many hospital visits, John learnt that there was only one specialist with sufficient expertise and the right technology to help his daughter. The heart surgeon was world-renowned and worked for the Miami Dolphins. John sought, and received, his help. His daughter was operated on and made a swift recovery. She is now a healthy 20-something. But John is saddled with over a million dollars of debt. That's why, approaching his seventies, he is still working.

When we first spoke to John, in 2009, he was angry. All over the news were reports of bonuses being paid to executives at a huge American insurance company called AIG that went bankrupt. John had done most things right in his life. He had played by the rules. He had paid his taxes. He was a good husband and a good father. Right at the end of his working life, all his plans were thrown in the air. He accepted that. His daughter's health came first. That was never in doubt. But these guys had cheated. They had bankrupted their firm, which the government – taxpayers like him – had to bailout, or it would, so we were told, destroy the whole financial system. Usually, this would be called extortion. These guys were lucky to escape jail. But worse, they were due to receive bonuses so extravagant that one year's worth would clear all the debts he had incurred to pay for his daughter's medical bills. John is soft-spoken. He is a mild-mannered man. Not one to lose his temper or start an argument.