Saturday, September 15, 2018

388. A methodological sleight of hand

Mainstream economics is based on the following three basic assumptions: economic agents are autonomous, have given preferences, and make rational, optimal choices between alternative ways of spending scarce resources.

This is not an ontological claim: good economists readily admit that these agents do not exist. In fact agents are limitedly rational, and routinely do not make optimal choices. The assumption is fictional, according to a methodology of instrumentalism: they serve as instruments for parsimonious, rigorous deduction of explanations of phenomena. It saves the enormous, complicated effort of explaining processes of decision making of individuals, which is the field of psychology, between individuals, which is the field of sociology.

Instrumentalism is not nonsense. In seeing the world one cannot see everything at the same time. Every view is conditioned, enabled and at the same time constrained by a perspective, a way of looking. The justification of counter-factual assumptions ‘as if’ is that they should lead to predictions that can be tested empirically.

The scientific scandal of economics is that this can hardly be done. The reason is twofold. First, predictions affect the choices that produce outcomes, and this can be self-fulfilling or self-defeating.

Second, in an economy there is little opportunity for controlled experiments, where one controls the factors that affect outcomes, next to the factors adduced from the theory. Society cannot be handled as a laboratory. As a result, predictions turn into retrodictions, tested after the fact, in retrospect, on the basis of statistics, in econometrics, where control of other factors, not part of the theory, is exercised with data on them. Great ingenuity has been exercised to develop ever more sophisticated tools for this. However, there still is the problem concerning the ceteris paribus assumption that factors other than those included in the process ‘remain the same’, as well as the causal structure, the logic of explanation. One cannot be sure of that, and when predictions fail one can claim that it is because conditions were not the same.

Also, many relevant variables have not been and sometimes cannot be measured. Measurement being methodologically sacrosanct, research proceeds with what can be measured. This yields the cliché of the joke of the drunk man leaning against the lamppost at night: ‘Why are you standing there, at the lamppost?’ ‘I am looking for my car keys’. ‘Did you lose them here?’ ‘No, but this is where the light is’.

A deeper problem, discussed in the philosophy of science, is that facts are ‘theory laden’: the terms and conditions of observation and measurement are formed according to the forms of understanding: the basic, often tacit, taken for granted, assumptions and meanings, of the theory. That, however, is a problem for all sciences.

Next, and here lies the scandal, economists conduct a methodological sleight of hand. They begin with the admission that their assumptions are not realistic, they cannot perform strong empirical testing, and then, as if the assumptions have nevertheless been proven justified, they deduce recommendations of policies from them as if they are scientific and can be taken for granted.

For that they employ the following, logic, as the core of economics. Going back to Adam Smith, the idea is that of the invisible hand of the market. Each consumer seeks to satisfy its own preferences, this creates demand for products, and when supply cannot cover it, prices rise, which draws producers to that product, until supply and demand are in equilibrium, and thereby scarce resources are directed automatically to where they have the most utility, in satisfying demand.

This is so appealing that, regardless of any lack of empirical tests, policy is invariably slanted in favour of the market. And so the market expands in all directions.          

Another rationale that has been used for the instrumental assumption of rational agents making optimal choices is illustrated by the following metaphor. From Chicago (the crucible of market economics), cars randomly move out in all directions. There are gasoline stations only along a few of the roads. After a large number of kilometres there are cars moving only along the roads with gas stations. It is as if those drivers rationally chose the roads with gas stations. 

The idea is that since markets select out inefficient firms, we observe only efficient outcomes, as if the agents made the optimal choices, so that is how we can proceed, explain phenomena as if choices were optimal.

The response to this has been[i] that if the argument is in fact one of selection, as in evolution, one should conduct the study in that way: model markets as eco-systems. This gave rise to the birth of evolutionary economics. One should see, and model, how efficient markets are in fact, in selecting out suboptimal choices.

However, that becomes complicated, since it requires the modelling of a process, which depends on many subsidiary assumptions, and thereby economics loses its crystalline clarity and rigour that is its pride, the pride of a myth of equilibria of outcomes without an account of how they may or may not be achieved. 

i] I think the first was Sidney Winter, in his PhD study.

Saturday, September 8, 2018

387. The programme of economics

Here I start the series on economics that I announced before.

As in other sciences, in economics there are diverse schools of thought: neo-classical, which is the mainstream, evolutionary, institutional and post-Keynesian. Here, I give an introductory survey. In subsequent items, I will discuss things in more detail. First, I will focus on the mainstream.

How does one represent a school of thought? Imre Lakatos[i] proposed the notion of a research programme. That has a ‘core’ of fundamental principles, assumptions and directions for research, which must be protected from falsification at all costs, by means of a ‘protective belt’ of subsidiary assumptions that supplement or implement the core principles. When something comes up that falsifies the whole, it is attributed to the subsidiary assumptions, and a replacement is sought there to make the core work better.

Isn’t such tenacity to a core unscientific? There is an argument for it. If something has performed well you will not give it up at the first sign of imperfection. That is an economic argument, but also an epistemological one. All theory is abstraction, imperfect and incomplete. It is by sticking to a programme, and milking it for all it is worth, that one discovers where its real limits lie, and finds indications for improvement. It does constitute a form of conservatism, but if someone is unsatisfied, he can start a new, competing programme. And that is again an economic argument: the argument for competition. As I proposed in this blog: imperfection on the move.

This conservatism can derail into dogmatism, and that happened to mainstream economics. One reason for it is that unlike natural sciences, the falsifiability of economics is dubious. I will discuss that in another item.     

There are two dominant characterizations of economics: optimal allocation of scarce resources, and exchange (through markets). In mainstream economics, the core assumptions are: rational choice by autonomous agents, in the calculation of optimal choice, and the operation of markets that yields equilibrium between supply and demand. The protective belt gives subsidiary assumptions of legal conditions (e.g of ownership), technology of production, infrastructure, the role and availability of information, etc.

Research seeks ‘forbidden events’ (falsifications), to repair and improve the subsidiary assumptions. The process entails what Thomas Kuhn called ‘normal science’, solving puzzles within the programme, such as, in economics, finding yet another, more sophisticated ‘production function’ to model production technology.    

Game theory brought a major transformation, as a tool to model strategic interaction between agents, with the central notion of a Nash Equilibrium: an outcome of interaction that is stable, in that every player wants to maintain its present strategy, as the best in view of the strategies adopted by the others. It is a useful device, but the basic principle of optimal choice remained the guiding principle. It is assumed that the set of strategies players can chose from, as well as the ‘pay-offs’ of combinations of the strategies of players, are given.

A later, more fundamental change, towards ‘behavioural economics’ allowed a relaxation of the principle of optimal rational choice, in allowing for decision heuristics from social psychology that are not substantively rational, not yielding optimal outcomes, though they may be rational in the face of conditions. A methodological advantage was that experiments could be made in laboratory settings, often with students of the researcher. However, it sits somewhat uneasily in the research programme of mainstream economics, as a more or less separate appendage. In a collaboration in this between economists and applied psychologists, a complaint of the latter is that the former cannot desist from forcing the heuristics back into optimal choice.

In economics, a distinction is made between risk and uncertainty. With risk one knows what can happen, and one can then append probabilities to calculate optimal choice, such as the one with highest expected outcome (outcomes multiplied with their probabilities). Under uncertainty, by contrast, one does not know all that can happen: that is not given prior to choice but emerges after choice, in action.

There, economists stand empty-handed, cannot ply their trade of calculation, so they ignore or neglect uncertainty. However, in innovation and relations uncertainty is routine. Radical innovation is uncertain, and the most fruitful relations are the most uncertain: they yield the surprise of novelty that goes beyond present insights. One engages most fruitfully with others who bring in things one could not before have imagined. One does not know in advance even of oneself how one will respond to unforeseeable events. There, also game theory falls short. In technical terms: there is no longer a matrix of strategies of agents and the values of outcomes of their combinations, because those mostly emerge during strategic interaction.  

Here, economics short-changes itself. If the most valuable relations are the most uncertain, incalculable in advance, then shying away from them for that reason foregoes opportunities for value, which is an uneconomic thing to do.

An exception is Keynes, who recognized uncertainty and used it to explain herd effects in the economy that produce booms and busts: if one cannot calculate optimal choice, then it is reasonable to follow choices others are making, especially others who claim that they know what they are doing, even if they don’t. Keynesian economics has not replaced the mainstream, but it remains as a minority stream, in post-Keynesian economics.   

Mainstream economics is focused on optimal outcomes, not on processes that may or may not yield those outcomes. Evolutionary economics recognizes uncertainty and models economics as an evolutionary process of more or less random initiatives, selected by markets and institutions, and transmission of what survives.

Institutional economics recognizes that markets do not work automatically and require institutions to function, and, more fundamentally, deviates from the assumption of autonomous individuals, recognizing that they develop their knowledge and their preferences in interaction with their environment.   

[i] Lakatos, The methodology of scientific research programmes, Philosophical
  papers volumes 1 and 2, J. Worrall and G. Curry (eds), Cambridge University Press.

Saturday, September 1, 2018

386. The freedom of rules

I will give a series on economics, as announced in the previous item, but here is an intervening item on something else.  

Different forms of freedom were discussed earlier in this blog (e.g. in item 340).The one that appeared odd is the Kantian one, where moral injunctions that constrain behaviour are claimed to set you free. This is odd, at first sight, because rules constrain. At second sight it is not so odd, because rules can also enable. Moral rules set you free from slavery to impulse or emotions.

There is a great variety of rules, and here I focus only on certain kinds, not, for example, legal rules.

There are enabling rules in poetry, in the sense of regular, standard, well-known or recommended forms. Here I focus on poetry because I recently became more involved in it than I was before. I had been writing poems on and off, in batches with long intervals, all my life, about 100 in total, with no ambition to publish them, and now I was ready to pick them up and improve them. For that purpose I followed a course on writing poems, joined a poetry criticism group, and started a separate blog of poems, parallel to this philosophy blog. The title is POETRY PINFOLD (

Here, the ‘rules’ concern rhythm (‘metre’), such as the ‘pentameter’, with five beats to a line, or the ‘tetrameter’, with four beats. Rhyme, in different patterns and forms, as in the classic sonnet. Word choice, depending on both meaning and sound. ‘Show, don’t tell’: don’t explain ideas, but let things happen that exemplify the idea, for the reader to interpret. No sentimentality, and certainly no clichés. Few adjectives, let nouns do their work. Where to end a line: Do you let a phrase run on to its end, at the end of a line, or do you cut it off before, to run on in the next line (‘enjambment’). I was told that such ‘rules’ do not form a constraint, but help poets to set themselves free. Again, I was puzzled: how can this be?

I came to the following analysis. Rules are a form of opposition to impulse: the inclination to do what first crops up in the mind. In social psychology this is know as the ‘availability heuristic’. You pay attention to what is ‘available’, comes to mind as salient. Those are propelled by emotion. They can be threats, or opportunities, or recent experiences, or exigencies of the immediate context of action. This was adaptive, in evolution, to yield a fast response to danger and opportunity, in order to survive, but it can lead to neglect of better alternatives. And a poet should take time. 

Earlier, I had argued that to develop oneself one needs opposition from others, to have a chance of being freed from one’s preconceptions. I proposed that this yields the highest level of freedom.

The principle is wider: one needs outside corrections more in general. In ordinary experience it is things that don’t work that tell you that you are mistaken. This is what I called ‘the success of failure’ (in item 357 in this blog). In science it is facts that contradict theory.

And here, in poetry, there are regularities. They are not imperative: a poem is not necessarily bad if it breaks the rules, in fact often the rules need to be broken to make the poem interesting, to shift an emphasis, or to make a point. In fact, whatever form is taken, almost any exception or irregularity you can think of is tolerated, and is legitimized with a technical term of its own, which makes the whole exercise rather pedantic, a bit of a show-off.

Nevertheless the ‘rules’ can help. They serve to trigger a critical reflection on what impulse gives you: could you have said it differently? The rules set you to ask what form of poem you are using, whether it conforms to some of the classic forms that set the standards, whether you are adhering to those ‘rules’, and what would happen if you tried to do so. It forces you to think of alternatives and indicates directions where to look. They may yield shifts of thought that turn out to be more appealing than what you first had in mind.

Here is a quote from Auden (by Glyn Maxwell) that says it all: ‘Blessed be all metrical rules that forbid automatic responses, force us to have second thoughts, free from the fetters of self’. Glyn Maxwell argues that poetry is rooted in the constitution of the human being. Metre and rhyme are inspired, literally, by breathing in and out, a beating heart, and movement, in step and run. That has been embodied in the full body of poetry, and it is a shame to leave it aside.  

In his book about poetry, Stephen Fry offered something deep that I like: ‘Poetry is concerned with the connections between things …. looking for ways in which one thing chimes with another’.[i] That can be done by metaphor, connections between meanings, or by alliteration and rhyme, connections by sound. The ‘rules’ can offer suggestions for that. 

I use the rules in a relaxed fashion. I will not let them force me to use a word that I don’t like, that does not fit the purpose of the poem, or does not sound right.

As in opposition from others, you listen to it carefully, see if it can improve your ideas, but in the end you follow your own conviction, and that can be something adopted from outside, from the rules.    

i] Stephen Fry, 2005, The ode less travelled, Arrow Books, p.124.

Saturday, August 25, 2018

385. A series on economics

Here I start a series of critical philosophical reflections on economics, as a phenomenon and as a scientific discipline. In earlier pieces in this blog I was critical of current capitalism and globalization, which with their perversities have sown seeds for the present populist uprising. I have been involved in an attempt at a fundamental change of economic theory, in a project of my own and in collaboration with others. I published a book on it in 2014, with the title ‘How markets work and fail, and what to make of them’, and I will tap from it for this series. I will also make extensive use of previous items in this blog, on self and other, ethics and morality, knowledge and truth.
My approach is shaped by my background as a mathematician, a long career in the economics and organization of learning and innovation, and more recent work on philosophy, in this blog, among other things. I have given up, at least for the time being, on the attempt to get economists along in a fundamental change of their discipline. What is needed is too fundamental to expect them to go along. It goes against their fundamental and combined assumptions of rational choice, calculation of optima, and their ‘methodological individualism’: the principle that everything should be deduced from choices by autonomous individuals.

It is no coincidence that in economics my background lay in innovation and learning. One cannot adequately deal with those without accepting radical uncertainty, i.e. uncertainty not only concerning what will happen but also on what can happen. That differs from risk, familiar to economists, understood as knowing what can happen, so that one can attach probabilities and calculate optimal choice, with which economists are very familiar. Among great economists, Keynes was an exception in accepting uncertainty, which gives rise to non-rational, non-individualist behaviour, in ‘animal spirits’. Individuals are only limitedly rational, and far from being autonomous, they are affected, indeed constituted, by social interaction.

I do not think that all of economics should go. First, in particular, I still think that markets are needed, though under some control to limit their perversions. It is a major challenge to sort that out. In particular, while economists focus on competition as the essence of markets, there is also, more than ever, a need for collaboration, and it is a major challenge to sort out how they are to be combined. Here, attention is needed to trust, which it is difficult for economists to deal with.

Second, as a science, a method, for calculating the best choice for spending scarce resources among alternative uses, economics still has its place, when the elements of that are known and calculable: goals, resources, effects and outcomes. Often they are not, and that is where the problems start. Economics has grossly overreached itself in affecting policy where the conditions for its validity are not satisfied.    

Third, there is an economic intellectual toolbox with many instruments that remain useful, indeed indispensable. I mention just a few: effects of scale, effects of scope, decreasing marginal utility, supply and demand, substitution, complementarity, transaction costs, entry barriers, information asymmetry, …..

Fourth, there have been attempts to widen the horizon of economics. One useful and influential extension was the use of game theory to clarify and analyse strategic interaction, with the notions of ‘Nash equilibrium’, and ‘zero and non-zero games’. Another, probably the most fundamental one, was ‘behavioural economics’, where insights were brought in from social psychology, about limitedly rational choice. That is flourishing. However, it is a bit of a foreign substance that while it is widely used it also tends to be rejected or adjusted for its violation of the core principle of methodological individualism, or twisted to fit into the optimising framework, neglecting the radical uncertainty involved.    

Fifth, beside mainstream neo-classical economics, there are unorthodox streams of economics that are not the target of my criticism and yield elements that I use for a new unified theory. These are: several forms of institutional economics that are closer to sociology, neo-Keynesian economics that takes uncertainty seriously, and evolutionary economics that replaces optimisation with evolutionary processes of emergence and is more amenable to the analysis of innovation.  

In this series I will deal with all these things, and more.  What is needed is a new science of society that includes elements from economics to deal with rational choice, sociology to deal with the social constitution and interaction of individuals, (social) psychology to deal with limitedly rational heuristics of choice, cognitive science to deal with learning and uncertainty, and philosophy to deal, in particular, with ethics. I have argued, in the book mentioned above, that for this new science the traditional, deeply rooted liberal utility ethics underlying present economics needs to be replaced by a wider virtue ethics that includes utility but also wider issues of morality of goals, and justice of process and outcomes. In these connections, use can be made of the unorthodox forms f economics indicated above.

In this series I will deal with all of this and more. 

However, I will not post only items on economics, and will here and there fit in items on other subjects, as they come up.       

Saturday, August 18, 2018

384. Intervention or laissez faire, in East and West

Taoist political philosophy is non-interventionist, libertarian, approaching anarchism. It criticizes Confucian interventionism with ethical rules, civic and familial values and the imposition, the rigidity, of ceremonies. Taoism aims to avoid what it considers to be artificial constructs (wuwei). Human design cannot cope with the richness and variability of holistic nature. Such design is bound to misfire and is in the way of natural processes that are best left to themselves.

This seems analogous to the split, in the West, between socialist interventionism and libertarian liberal laissez faire. However, a fundamental difference is that the latter is based on views not of holistic nature but of freedom for individuals. Those have a craving and see it as their right to exploit nature to their material advantage. And that has dire consequences for the environment.

However, libertarianism does recognize the natural urge in Man for gratification and, especially, self-manifestation (and Nietzsche’s will to power). And in nature there is not only harmony but also brutality in the struggle for survival. Taoism seems hesitant to face those realities.

I side in part with Confucianism and in part with Taoism. Such mixes have also arisen in neo-Confucianism, as I indicated in item 131 of this blog. I also object to the constraining regimentation of Confucianism, which threatens the variety and variability that are inherent in nature, evolution, humanity and society.

I think there is some similarity between Taoist thought and modern evolutionary thought, which I have endorsed in this blog. Like Taoism, the latter also yields a need for restraint of the urge to engage in ‘intelligent design’. 

For example, and in particular, it is odd to try and plan programmes for innovation while the crux of innovation is that it produces things that were unforeseeable (or else it would not be innovation). By planning innovation one obstructs it. So, here I would go along with Taoist thought.

This does not mean, however, that nothing needs to be done. It does not yield laissez faire. It does entail going along with the natural flow of processes, but one may help evolutionary processes of development to proceed, by facilitating and directing the core processes of the generation of variety, selection and proliferation of success. I think that is consistent with Taoist thought: the growth of plants can be enhanced by seeding, watering and pruning.

Similarly, I appreciate the value of markets, to let people do their own bidding in supply and demand, but institutions are needed to enable markets and constrain them in their perverse effects. In the next item of this blog I start an extensive series concerning economics and markets.  

Will human beings act well when allowed to act freely according to natural impulse? In this blog I have argued that human nature is ambivalent in this respect. It harbours instincts of both self-interest and altruism (within limits). Under existential threat self-interest for the sake of survival is the stronger. Cultural means, in an ethics of conduct, and institutional means, in the rule of law, are needed to curtail egotism. Here I side with the Confucian view.

Institutions are needed to stimulate the manifestation and flourishing of positive natural impulse towards fairness, solidarity, and justice. For example, they may be needed to break through prisoners dilemmas where individually people may be willing to act ethically but collectively find that they are unable to do so unless others do so as well. Society in general, and the economy in particular, are rife with such ‘system tragedies’, as I called them. Intervention is needed to allow for escape from the dilemma’s.

In sum, I side with Taoism in restraint of planning of activities, intervention in natural processes, and regimentation of values and conduct, but I side with Confucianism in the need to curtail perverse instincts and solve social dilemmas.

Saturday, August 11, 2018

383. Silly simplicity

Many concepts that play an important role in public debate are ambiguous: they have a variety of meanings, and often it is not clear which is at issue. Think of God, democracy, justice, power, liberalism, identity, trust, love, truth, meaning, good and bad, culture, beauty, art, causality, change, freedom, and more. Discussants bash each other in disagreement, thinking they are talking about the same thing while they are not. These are slips of simplification, of reducing many-sided concepts to simple, reductive caricatures.

Take atheism, for example. That depends on what meaning you assign to God. Some people say Spinoza was not an atheist because he was passionate in recognizing God as the primary principle from which all else flows, the top axiom of the mathematical system of his Ethics, written, as Spinoza said, ‘more geometrico’, in the manner of geometry, with axioms from which successive theorems are derived. He protested against being seen as an atheist, and being punished for it. Yet he was an atheist in not recognizing a personal, providential, loving, caring God, and that was the operative meaning of God to Spinoza’s contemporaries.

Take liberalism. I am a liberal in being in favour, for want of something better, of a liberal parliamentary democracy with freedoms of expression, religion, association, and election, and the separation of powers. I am not a liberal in favouring laissez faire markets. I do not adopt the utility ethics of liberalism. And then, arguing the shortcomings of markets I run the risk of being called illiberal.

Take democracy. Authoritarian leaders call their states democracies because the leaders were chosen in elections that were free, more or less. But democracy requires more: the features of liberal constitutional democracy specified above. But those are never perfect. How democratic is it if leaders can only be elected at great cost, for which they need sponsorship from business, whose interests they then have to protect? Currently there is spreading complaint in European democracies, in populist movements, that ruling elites have distanced themselves from the lives and interests of the people. The claim next is that representative democracy has outlived itself, and forms of more direct democracy should be used, such as a referendum, where the will of the majority is imposed without debate on the rights of minorities.

Take power. People complain about ‘people in power’ in a democracy and at the same time cry out for more effective leaders who have the power to make decisions where inept democracies are caught in indecision. Behind this lies the confusion between negative power, limiting or restraining choice, and positive power, offering new options and more room for choice. The authoritarian leader projects himself as offering the second while falling back on the first, in his imperfection, which yields mistakes that with his power he can hide.      

Take trust. Some see it as control, constraining opportunities or incentives for doing harm, while others see it precisely as something going beyond control, motivating people to do right not because they have no alternative or need the reward, but because they want to, on the basis of ethical or moral conviction, social customs and connections, or friendship.  

As I argue elsewhere (in item 29 of this blog), this wrong-headed perspective on universals is part of what I call the ‘object bias’ in human thought: universals are treated as if they were specific particulars, like objects moving in time and space, with a well-defined identity across contexts. In fact, they are objects of a different kind, with meanings depending on time and place.   

Saturday, August 4, 2018

382. The short term virus

As suggested by Haldane (and others), there is an increasing preoccupation with the fast and short term in finance and investment, knowledge and learning, and organization and work. It happens not only within these fields, but operates as a virus that contaminates one field from the other.

I gave an example in the preceding blog. Fast, short term politics drives out slow, long term knowledge to feed policy. This lures students into shallow knowledge. And a generation later there will no longer be capacity in policy making to absorb deep knowledge, if the use of it ever again becomes viable and attractive.

The media are under pressure to satisfy to hunger for the fast and shallow, they are under financial pressure to replace the expensive, older carriers of deep knowledge by the younger, cheaper novices carrying the faster and more shallow. Weaned on shallow media, the new world waxes more shallow.

The virus operates especially, most spectacularly, infamously, in the financial sector, contaminating the ‘real economy’, in imposing the imperative of quick returns. Profits in the financial industry are increasingly made more with financial products than with production in the real economy. The snake fattens by biting its tail.

We see it also within firms. I have personal experience with Shell International Oil Company. Shell was populated by two cultures. First there was the slow culture of the engineers (often Dutch), in exploring and surfacing oil and gas, refining oil in manufacturing, liquefying gas, and storing, and transporting and distributing them. Then there was the increasingly fast culture of the financiers, investors and accountants (often British).

With a head of the firm from the accounting function (though a Dutchman), the virus broke through. Financial wizardry and blowing the bubbles of share value led the firm into the temptation of overvaluing its oil reserves, by hiding the probabilities of ‘dry’ or commercially difficult to exploit wells. The engineers protested but were overruled.

With rising remuneration, in salary and bonus, for workers in the financial sector, there is pressure to apply them also in other firms. This yields a widening gap in remuneration between the finance people and the engineers, discouraging engineers to come in, or encouraging them to clamber onto the grandstand of finance.

In organization and work the pressure towards fast, short term returns has led to an imperative of flexibility, with short term, maximally flexible labour, in projects rather than jobs, and, with suppliers, spot contracts rather than more durable collaboration. This muscles out the slower process of making investments that are ‘specific’ to the firm, in workers, in training, education, and teams or communities within the firm, and more durable, deep relations with suppliers, in joint development.

People are willing to invest in such firm-specific investments, needed to produce specialties,  high quality, and innovative products, yielding higher and longer-term profits, only if the duration of the job or relationship is long enough to recoup the firm-specific investment. The upshot is that one should go not for maximum but for intermediate, optimal flexibility: enough to prevent rigidity but not so much as to discourage depth.