How bounded is rationality?

Herbert Simon’s concept of ‘bounded rationality’ captures the idea that although individuals intend to behave rationally, they are constrained by limited cognitive abilities. The concept was put forward as an alternative to the assumption of neoclassical economics that individuals make decisions by maximizing utility functions.


It seems to me that maximizing utility is best viewed as a metaphor, akin to the charioteer, elephant and plane metaphors discussed in a previous post, rather than a description of actual human behaviour. It can be a useful metaphor. (I have used it on this blog to suggest that some seemingly irrational behaviours may in fact be rational.) Nevertheless, as James Buchanan has argued:

‘The modern economist who models the individual as choosing among feasible alternative bundles of goods to maximize a utility function that does exist independently of choice itself presents no evidence that such functions actually exist, and if pushed, the economist would agree that “utility” is little more than a rhetorical artifice that is introduced as an aid in explaining choice behavior within an imposed rational choice reconstruction’ (‘The Economics and the Ethics of Constitutional Order’, 1991).

It is possible to argue that all human action is rational in the sense of being purposeful, but once we acknowledge that humans have limited cognitive abilities then we have to acknowledge that they make mistakes. Economists have often recognized that people make systematic cognitive errors by incorporating arbitrary behavioral assumptions reflecting such errors (e.g. money illusion) in some of their models. This raises the question of whether efforts by economists to obtain a better knowledge of the bounds of rationality will enable them to build better models.

Jonah Lehrer’s book ‘How we Decide’, provides a highly readable discussion of the ways in which limited cognitive abilities can affect decision-making in different contexts. I decided to read the book after reading some comments by Peter Boettke on the Coordination Problem blog.

Lehrer’s main message seems to be: ‘The mind is full of flaws, but they can be outsmarted’ (p.250). He implies that a major source of error is failure to think about the kind of decision being made and the kind of thought process it requires. Insufficient reasoning can obviously result in poor decisions when the mind is strongly influenced by emotional urges and impulses. But it is also possible for the mind to choke on excessive reasoning.

It is common to hear of instances when sporting champions choke at a crucial point in a game because they suddenly become self-conscious and interfere with their performance by consciously trying to avoid mistakes. It is also possible, however, for too much analysis to lead to poor decisions in relation to choices that might be thought likely to benefit from analysis. For example, in considering a trade-off between size of home and time required to commute to work some research has suggested that there may be a tendency for people to give to give greater weight to the size of the house the more time they spend deliberating, even if the additional space is superfluous (p.144). People may often make better choices when they use their conscious minds to gather information and then trust their emotions.

Everyone knows that inexperience is a common source of error but many of us fear the unpleasant symptoms of making mistakes. One of the crucial ingredients of successful education is encouragement of children to learn from their mistakes by praising them for their efforts rather than their cleverness (p. 51-3).

The book contains a chapter on our tendencies to be fooled by feelings – loss aversion, the perception of patterns that don’t exist, the tendency to over-value immediate gains relative to longer term costs etc. Attention has previously been drawn to such problems by Dan Ariely in ‘Predictably Irrational’ (discussed here) and Richard Thaler and Cass Sunstein in ‘Nudge’ (discussed here). Lehrer suggests that the best way to avoid such errors is to be aware of them and to check feelings with a little arithmetic (p.244).

It seems to me that one of the most important contributions of ‘How we Decide’ is to draw attention to the errors that result from our tendency to surrender to ‘shoddy top-down thinking’ because self-delusion feels better than uncertainty. This is reflected, for example, in the influence of party affiliations when voters consider complex political issues. Even expert advice can be biased by adherence to frameworks that are at variance with reality. Lehrer’s suggested remedy is to embrace uncertainty, entertain competing hypotheses and to remind yourself what you don’t know.

Different readers may see different implications for economics coming out of this book. The book has reinforced my support for Oliver Williamson’s view that modes of contracting that make large demands against cognitive competence should be disfavoured (‘The Economic Institutions of Capitalism’, 1985, p.46). Governance structures will fail if they require managers to have unlimited cognitive capacities. Humans tend to be strongly influenced by moral instincts and conventions, but they are also susceptible to temptations. Another implication is that decision-making skills are likely to vary greatly among different individuals. Modellers should be wary of assuming that everyone is equally susceptible to cognitive distortions or that they have equal abilities to learn from experience. Finally, while the book provides plenty of support for the view that the rational voter is a myth rather than a useful metaphor, it suggests to me that the worm – reflecting immediate emotional responses to what politicians are saying – is unlikely to be a good predictor of voter behavior. Focus groups initially gave the thumbs down to some of the most successful shows on television.

Is state sovereignty relevant to resource rent taxation?

The Henry tax review into Australia’s future tax system recommends:

‘Subject to transitional arrangements, the new rent-based tax should apply to existing projects, replacing existing charging arrangements. The allocation of revenue and risks from the new tax should be negotiated between the Australian and State governments’.

The federal government seems to be attempting to ignore this advice in imposing the new tax. It is proposing to reimburse mining companies for existing royalty payments rather than to replace existing charging arrangements. It has decided unilaterally how it proposes to use the additional revenue from the new tax. In selling the tax to the Australian public it is asserting that mineral resources are owned by all Australians, contrary to the legal position of ownership by the Crown, with state governments having constitutional authority for resource management.

The government of Western Australia is threatening a constitutional challenge to the new tax, but the federal government doesn’t seem to be particularly concerned about this. I’m no lawyer, but I imagine the federal government think they are on safe ground in calling the tax a profits tax rather than a resource rent tax.

However, even if the new tax is legal, I think the federal government should be concerned about the viability of their proposal not to reimburse mining companies for any new or additional royalties that might be charged by state governments. Whatever the High Court might decide about the validity of the new federal tax, it is not likely to rule that the imposition of a new tax by the federal government has extinguished the rights of state governments to raise royalty rates.

Are state governments likely to impose additional royalties? Some proposals for higher royalties were already in the pipeline in Western Australia prior to announcement of the new federal tax and it is possible that these charges will be accommodated in transitional arrangements. The state governments review their royalty charges from time to time and I imagine that they will continue to do so. It is quite possible that having read and digested the Henry report a state government could decide to change the basis of their charging arrangements to a resource rent tax and to increase revenues from the resources sector. In considering such a change the state government might note that there is nothing particularly magical about the 40 percent tax rate proposed by the federal government. They might even read in the Henry report that Norway imposes a total tax rate on petroleum rents of 78 percent.

The point I am leading to is that the new federal tax has not extinguished the potential for state governments to raise royalty rates. This remains a potential source of sovereign risk for mining investment in Australia. This consideration is additional to the argument in my earlier post (Does a resource rent tax solve the problem of sovereign risk?) that the proposed application of the new tax to existing mines would lead investors to perceive that they have under-estimated sovereign risks in Australia. Even if the federal government comes up with satisfactory transitional arrangements for the new tax, miners will still need to factor into their calculations an allowance for possible future increases in state government royalties.

In my view the federal government should take another look at the recommendations of the Henry report and seek negotiations with state governments about the allocation of revenue and risks from their proposed resources rent tax.

What metaphors help us to understand the functions of reason and emotion?

Plato argued that we can only be masters of ourselves if reason, the ‘human charioteer’ is able to control the dumb beasts of passion:

‘Now the winged horses and the charioteers of the gods are all of them noble and of noble descent, but those of other races are mixed; the human charioteer drives his in a pair; and one of them is noble and of noble breed, and the other is ignoble and of ignoble breed; and the driving of them of necessity gives a great deal of trouble to him. ... The right-hand horse is upright and cleanly made; he has a lofty neck and an aquiline nose; his colour is white, and his eyes dark; he is a lover of honour and modesty and temperance, and the follower of true glory; he needs no touch of the whip, but is guided by word and admonition only. The other is a crooked lumbering animal, put together anyhow; he has a short thick neck; he is flat-faced and of a dark colour, with grey eyes and blood-red complexion; the mate of insolence and pride, shag-eared and deaf, hardly yielding to whip and spur’ (‘Phaedrus’).

Plato’s metaphor seems to have provided the basis for an influential model of human flourishing which puts reason (or rationality) on a pedestal and views the emotions as crude and primitive. In his book, ‘How We Decide’(2010), Jonah Lehrer links Plato’s metaphor to Cartesian philosophy and notes that Freud used a similar metaphor in which the horse (id) provides the locomotive energy and the rider (ego) determines the goal and guides his powerful mount towards it (p.10-12).

However, Lehrer points out that this classical theory is ‘founded on a crucial mistake’: ‘What we have discovered when we look at the brain is that the horses and the charioteer depend on each other’ (p. 13). ‘When we are cut off from our feelings, the most banal decisions become impossible. A brain that can’t feel can’t make up its mind’ (p.15). We might like to think that reason plays a large role in our decisions, but if our feelings didn’t tell us what we like and dislike we would not be able to make decisions.

Jonathan Haidt has argued similarly that Plato’s metaphor ‘may overstate not only the wisdom but also the power of the charioteer. ... Reason and emotion must work together to create intelligent behavior, but emotion ... does most of the work’ (‘The Happiness Hypothesis’, 2006, p.13).

Haidt has his own metaphor, an elephant and its rider, to explain the relationship between the controlled and automatic systems that determine human behaviour:
‘The controlled system ... is better seen as an advisor. It’s a rider placed on the elephant’s back to help the elephant make better choices. The rider can see further into the future, and the rider can learn valuable information by talking to other riders or by reading maps, but the rider cannot order the elephant around against its will. ... The elephant and the rider each have their own intelligence, and when they work together well they enable the unique brilliance of human beings’ (p.17).

Jonah Lehrer uses a modern aeroplane in his metaphor to explain the functions of the emotional brain and the pre-frontal cortex:

‘To sit in a modern airplane cockpit is to be surrounded by computers. ... These computers are like the emotional brain of the plane. They process a vast amount of information and translate that information into a form that can be quickly grasped by the pilot. ... These computers are so reliable that they perform many of their tasks without any pilot input. ... Pilots are like the plane’s prefrontal cortex. Their job is to monitor these onboard computers, to pay close attention to the data on the cockpit screens. If something goes wrong, or if there’s a disagreement among the various computers, then it’s the responsibility of the flight crew to resolve the problem. ... When the onboard computers and pilot properly interact, it’s an ideal model for decision-making. The rational brain (the pilot) and the emotional brain (the cockpit computers) exist in perfect equilibrium, each system focusing on those areas in which it has a comparative advantage’ (p.256-8).

Which metaphor is best? I doubt whether one metaphor is the best aid to understanding of all aspects of human behaviour, but I think the elephant metaphor is better than the plane metaphor from a personal development perspective. It is possible to think of the rider and elephant as being responsible for their future performance as well as for their current performance. As Jonathan Haidt puts it, ‘virtue resides in a well-trained elephant’ (p.160). I don’t pretend know much about training elephants or their riders but it seems reasonable to suppose that they would have a stronger incentive to learn to work together and to improve their performance if they were held jointly responsible for their behaviour.

How can we ensure that parliaments are representative and governments are accountable?

An update of my views on the topic has now been published by "On Line Opinion":



‘In a really equal democracy, every or any section would be represented, not disproportionately, but proportionately. A majority of the electors would always have a majority of the representatives; but a minority of the electors would always have a minority of the representatives. Man for man they would be as fully represented as the majority. Unless they are, there is not equal government, but a government of inequality and privilege: one part of the people rule over the rest: there is a part whose fair and equal share of influence in the representation is withheld from them; contrary to all just government, but, above all, contrary to the principle of democracy, which professes equality as its very root and foundation’ (J S Mill, Representative Government, Chapter 7, 1861).

Some famous person has probably written in support of strong executive government which dominates parliament and is held in check only by periodical elections (as well as an independent judiciary etc) but I don’t know where to find an appropriate quote. Those who have commented on such a system have tended to refer to it disparagingly as an elective dictatorship. However, I think it is possible to defend a system that tends to deliver the governing party a substantial majority of seats on the grounds that it results in more accountable government than a proportional system in which no party has a clear majority. A government that dominates parliament cannot claim that it has not implemented its promises to the electorate because of obstruction by other parties. It has to wear the electoral consequences of its own actions.

The point I am trying to make is that while proportional representation might be a desirable characteristic of a parliament, it is undesirable to have a system of government in which parties go to the polls to seek endorsement of their policies and then, after the election, enter into negotiations to decide what policies the temporary coalition of parties forming the government will actually seek to implement. Parties forming such temporary coalitions tend to blame each other for poor outcomes and electors find it hard to tell who is responsible for what.

Various compromises between proportional representation and elective dictatorship are possible. One possibility is the reinforced proportional representation system used in Greece under which the party which wins the largest number of seats in parliament is allocated additional seats so that it more likely to be able to form a majority in its own right. Leaving aside the obvious point that it is difficult to envisage that Greece’s recent economic performance could have been much worse without this reinforcement of proportional representation, an arbitrary adjustment to numbers of seats seems somewhat inelegant (if not undemocratic).

Another possibility is to have a bi-cameral system with the government being formed in the lower house, elected on the basis of a system that usually produces workable majorities for a governing party or stable coalition (e.g. single member electorates) and an upper house, acting as a house of review, elected using proportional representation. As recent events in the UK show, single member electorates cannot always ensure that the party winning the largest number of votes is able to govern by itself (or even to form part of the government for that matter). But single member electorates have a reasonable track record in producing stable and accountable governments. This system has the added advantage of allowing voters to vote for a person to represent their locality rather than for a party (or party list).

Luke Malpass and Oliver Marc Hartwich have recently advocated a bi-cameral system, such as I have just described, to replace the single chamber proportional representation system in New Zealand (CIS Policy Monograph 109). This is also the system that we have in Australia.

So, does the Australian system provide the best possible compromise between a representative parliament and an accountable government? I don’t think so, because it gives too much power to the upper house. The Australian Constitution contains a sensible procedure to resolve a deadlock between the upper and lower houses of parliament – a joint meeting of both houses – but joint sittings can occur only after a double-dissolution election.

I think the requirement for an election to resolve deadlocks between the two houses of parliament tends to work against accountable government because it enables governments to blame obstruction in the Senate for failure to implement policies. Before going down the double-dissolution path governments have to consider the possibility that they will lose such elections or be returned to power with more obstructive upper houses than they had before. Although there are half a dozen occasions in Australian history when governments have brought on double-dissolution elections, they have been defeated on about half those occasions. A joint sitting of both houses of parliament has occurred on only one occasion.

Given the difficulty of amending the Australian Constitution it seems that we will have to continue to live with the adverse consequences for government accountability of the requirement for elections to resolve deadlocks. We can, however, take some solace from the fact that the election requirement has the virtue of providing a test of the extent to which governments have the courage of their convictions. The value of such a test has recently been highlighted by the current government’s decision not to trigger a double-dissolution election on the bill to establish a carbon emissions trading system in Australia.

Postscript 1:
After reading a post by Tim Harford I have been reminded of Kenneth Arrow's impossibility theorem. Rather than wondering where to find a quote from some famous person supporting elective dictatorship I could have quoted Kenneth Arrow to the effect that whatever electoral system you use you will always end up with some form of dictatorship (although some forms of dictatorship are worse than others). The inference that I think should be drawn from Arrow's impossibility theorem is that markets are usually better than politics in producing outcomes that are beneficial for everyone.

Postscript 2:
Joseph Schumpeter qualifies as a famous person who emphasized the value of strong executive government. In Capitalism, Socialism and Democracy (1942) he wrote: 

"It is in fact obvious not only that proportional representation will offer opportunities for all sorts of idiosyncrasies to assert themselves but also that it may prevent democracy from producing efficient governments and thus prove a danger in times of stress. But before concluding that democracy becomes unworkable if its principle is carried out consistently, it is just as well to ask ourselves whether this principle really implies proportional representation. As a matter of fact it does not. If acceptance of leadership is the true function of the electorate's vote, the case for proportional representation collapses because its premises are no longer binding. The principle of democracy then merely means that the reins of government should be handed to those who command more support than do any of the competing individuals or teams".

I wrote more about Schumpeter's views of democracy here.