Systems and Freedom

In last week’s post I talked about the difference between positive and negative freedoms, how this maps onto the left/right distinction in politics and how that distinction explains the difference in attitudes to the corona measures between Australia and the US. There is another aspect to the difference that I think explains a lot of the current political and cultural malaise that the West finds itself in. Recall that negative freedom is freedom from interference, especially by the state. Negative freedoms are simple. They are simple to encode in law and anybody can understand when they have been violated (although it’s not necessarily simple to achieve redress when they have been violated). From a political point of view, it’s also easy to know when a movement to attain negative freedom has been achieved. Three important historical examples are the civil rights movement, the suffragettes and the equal pay laws. All three were about equality before the law or, to say it another way, the granting of negative freedom to a part of the public that didn’t have it beforehand. Because negative freedom is a binary, a political movement to attain negative freedoms has a fixed goal and it is plain to see when that goal has been attained. In short, negative freedoms have an inherent objectivity. One can disagree about the application of them but not whether the law states that people have them.

The same is not true of positive freedoms. Positive freedoms are inherently more subjective. This makes it harder to get agreement on whether positive freedom exists and to what extent it exists. A useful way to draw this distinction is to compare the equal pay laws to the modern equivalent of the “gender pay gap”. The equal pay law states simply that a man and a woman must be paid the same amount for the same job. In any specific example, it’s easy to tell whether this was done. You just check the pay slips of the people involved. But the gender pay gap is different as it deals with populations of people and averages. The gender pay gap says that women earn less than men on average. As anybody who’s done Statistics 101 would know, an average calculation hides as much as it shows. To properly understand an average you first need to know whether we are talking about the mean or the median. You need to know the standard deviation around the average and you need to know whether outlier values are skewing the average. It may be, for example, that 99.95% of men and woman earn the “same pay” but a small minority of bazillionaires, who just happen to be men, throw out the calculations. Until you know that, any method to address the imbalance risks being misguided.

That’s just the start of the difficulties, however. The hidden assumption in the gender pay gap issue is that women have less positive freedom than men because they get paid less on average. Is that true? The average pay figure is the result of countless personal choices made by real people in the economy. The economy is a system so complex that an entire profession dedicated to studying it still can’t make accurate predictions about how it will behave. To paraphrase John Kenneth Galbraith, the purpose of economics is to make astrology look good. A person’s gross earnings are the result of all kinds of individual decisions including which field to go into, how hard they work at their job, whether they know the market value of their work and how to negotiate for their salary etc etc. Without in-depth analysis, you can’t draw any firm conclusions about why one person gets paid less than another. We can, of course, make some hypotheses. When it comes to the differences between the genders, it seems highly likely that time taken off from work to raise children is going to be the main variable that accounts for why women earn less on average. One might argue that woman are choosing to take that time off work because they value raising children more than working. A counter argument could be that woman only do that because child care is too expensive and it doesn’t make financial sense to do so. Either is plausible. We’d need more research to find out for sure and we’d probably find that both of these things are true. Some women would raise their children themselves no matter how cheap childcare was and some women would prefer to work. These kinds of issues are intrinsic to systems and, by extension, to debates about positive freedoms. The only way to understand a system is to turn it around and look at it from multiple perspectives. As positive freedom is almost always about systems, it follows that issues around positive freedom require such analysis. Simply reducing a whole system down to a single calculation is naïve at best.

Of course, issues like the gender pay gap get forced onto the procrustean bed of politics where catchy slogans and simplistic solutions are all that can fit into the limited bandwidth of public discourse. No doubt the limitations of political debate do play a role in the oversimplification of complex issues but the problem is deeper than that. I have spoken a number of times about systems in other blog posts with particular reference to the systems thinking and cybernetics movements of the 20^th century. The fact is that humans are really bad at understanding systems and even worse at building them. The systems that we rely on in day-to-day life were not designed and then implemented by wise philosopher kings. Rather, they are the results of centuries or even millennia of trial and error. Look at the history of any successful company (a company is also a system) and it’s almost guaranteed that the company also went through a period of experimentation in its early days. The founders thought they had a great idea in market A but accidentally stumbled onto a goldmine in market B. I have worked in building IT systems from scratch for more than a decade now. I can say from experience that most people have no idea what they are doing and most things that get produced are of no value. This isn’t an indictment of the people involved. Building systems is hard. The failures are also not necessarily a problem if you assume that each project is itself just an experiment which you can learn from. The problem is that learning very rarely does take place. Rather, corporations and government shovel money into various projects and hope that something good results. Politicians and upper managers in corporations are expert at taking credit for things that look to be working and backing away from things that are visibly failing. The organisational structures of bureaucracy and the public service are set up to ensure that nobody must take responsibility. Most projects are structured in such a way that nobody knows how to understand the outcome because there are no criteria to do so. I’ve been in some funny meetings where the question of “defining success” was talked about and nobody knew or wanted to know how to define it. If you define success, then you also define failure and that is what everybody is trying to avoid more than anything. Accordingly, in large corporations and political parties, trying to determine whether a change to a system produced any tangible benefit is nigh on impossible. Systems are already hard enough to understand but when you add in human psychological and political considerations the task becomes insurmountable.

Which brings us back to positive freedoms which inevitably require changes to systems. Even assuming you have a grasp on an issue in a multi-dimensional manner, you are altering a system and so you need to know whether your changes had a benefit and what, if any, unforeseen side effects were caused by the change. The testing of the system is often as much work as the changes to the system itself. One of the key insights of the systems thinking pioneers was that natural, inherently complex systems cannot be understood by reductionist science. Rather, they can only be understood by applying multiple perspectives in a heuristic fashion. Heuristics are fallible, incomplete but useful ways to view a system. The problem from both a scientific and political point of view is who gets to decide which heuristics really are valuable and which are not? If you aim to consciously change a system and then measure the results, you can apply your own heuristics to gauge the effect or you can open channels for feedback. The former option risks missing important types of data about the system and the latter usually sees you receiving a barrage of unintelligible information; noise. At this point we get into the problems of information and testability which are a big driver of why we don’t test the changes made to systems. Testing costs time and resources. The choice to only measure one or two things is a psychological, economic and political one. All these drivers lead to oversimplification and, therefore, misunderstanding of the system. A political movement to enhance positive freedom faces all these challenges in amplified form because politics is as much about being seen to do something good as actually doing something good. Furthermore, changes to systems often have side effects and the side effects might just be voters who resent the fact that they have been disadvantaged by a change. What this boils down to is that the work required to enhance positive freedom is far more complex than the work required to enhance negative freedom and the results of such work are going to be far more ambiguous, which is itself a political problem.

The way to approach systems is to construct cross functional teams, start small, make iterative changes and ensure you set up feedback loops with the real world. You need to set aside time and resources to learn. This means you will need periods of evaluating what has been done rather than doing more. You need to be prepared to admit failure but also able to change tack if an opportunity arises that was not foreseen at the beginning. These are all things that corporations and governments are incapable of doing. The feedback loops they adhere to are the long and imprecise ones of opinion polls, elections, profit and loss reports and annual general meetings. Governments get thrown out of office and companies declare bankruptcy while the systemic effects of decisions made along the way are never known. In the domain of politics, movements to enhance positive freedom mostly fail or at least aren’t seen to be successes. This in itself causes frustration especially given our culture’s innate desire to “progress”. That frustration itself builds pressure to do more. Accordingly, positive freedom has morphed into authoritarianism as the bureaucracy does the only thing it knows how which is to force changes onto the world rather than set up a symbiotic relationship where learning and adaptation is possible. In recent decades, we have run into the problems of systems and we still don’t have the culture or the organisational structures in places to deal with systems properly. Who would increase positive freedom must first learn how to work with systems.

Addendum: Just came across this fascinating post on the supply chain problems in the US. It’s indicative of where we are as a society that a real systemic problem, which I would have thought is an existential threat if allowed to continue, goes unsolved while the news is full of faux-moral issues.