Improving parliamentary information processing
4 min read

Improving parliamentary information processing

Improving parliamentary information processing

You can see organizations as information processing units: they eat and process information after which they create all of the products and services which we see around us. Parliaments seem to be especially constrained in their processing capacity, as they are at the center of governmental decisionmaking. I argue for various ways to reduce the amount and scope of decisions which parliaments have to make, e.g. using direct democracy or delegating legislative tasks to other institutions.

Information flow between organizations

Organizations may be coordinating lots of complex activities, but this is not revealed to customers (when things go according to plan). Customers usually end up seeing just a simple interface: if you pay us X we will let you use product Y.

Regardless of how much complex exchange happens within organizations, the markets between them are thin spaces of interaction. Such market interfaces can be described by function signatures. Imagine a company specialized in selling chairs and banana's; you may then be able to:

buy(10 EUR)  -> chair

buy(100 EUR) -> expensive banana

call(1 EUR)  -> refund for your expensive banana

Governments are also providing lots of services, for which either the inputs or outputs are often a lot more complex than your typical company interface:

declare_taxes(personal_details, tax_data)   -> stay_compliant

request_subsidy(personal_details, 10 forms) -> subsidy

customer_service()                          -> help_with_taxes

Information flow within organizations

What about what happens within organizations? Things are actually no different there. In order to create order out of chaos, organisations are often split up into several parts. This is symbolized by the famous 2-pizza team concept from Amazon (see also Stevey's Google rant for amusing commentary):

get_materials_from_marketing(requirements)   -> marketing_text

fancy_website_upgrade(requirements)          -> website

salary_increase(charm, wit)                  -> raise

Large organizations such as governments are split up into many parts at many different layers of abstraction. On a high level, many governments are splitup between executive, legislative and judicial branches. On lower levels, different departments do different things, from marketing to bookkeeping to management. Ultimately, this enables specialization of labour. Just like specialization in a market helps an economy grow, specialization in a company helps a company grow. Specialization of labour creates the increasing returns phenomena which make society better off.

The first exception to the rule

There is one exception to the ever growing specialization of labour: each organization, no matter how big, only has one CEO and C-level board. And every country, no matter how big, only has one parliament.

At first sight, executives seem to utterly lack specialization, but this is not the case. In large effective organizations, executives cannot and do not take a majority of decisions. They don't even set high level strategy. All they can do is enforce cultural values.

At first sight, parliaments seem to utterly lack specialization, and this... seems to indeed be the case. An empirical case is hard to make here, as I'm not aware of good natural experiments comparing the performance of different types of parliamentary models. Most governments in the world follow roughly the same parliamentary model and objectively measuring their "output" is potentially impossible, as every parliament might end up valuing different things. So therefore, let's try to make a logical case.

If you want to double a company's output, you can increase the supply of scarce resources. Some example from most to least fungible: raise money (and divide it over teams), hire more people (and split them into teams), hire better people, improve the culture or set strategy. Finding out which resources are needed is context dependent: different companies have different constraints at different scales.

In contrast, if you want to double a parliament's output, this simply can't be done by throwing more of the most fungible inputs (capital and labour) at it. The following quote beautifully illustrates the constraint which parliament puts on society:

"Every few months or years, a few very specific policy questions win the memetic battle for public attention, but they're rarely the things-which-most-professional-policymakers-deal-with-on-a-daily-basis or the most-important-policy-questions in terms of peoples' welfare, budget, QALYs, relevance to the state of the world in ten years, etc." (source)
parliamentary_discussions(lots_of_information) > legislative_changes

Are we doomed to accept this status quo? Of course not! Parliamentary decisionmaking can be split up formally and informally in at least three ways:

(1) Scale indirect democracy

We can defer decisions to subcommitties of political parties. This technique is already used around the world. Unfortunately, the more this happens, the harder it is to vote for the right representative. There is a flipside of the complexity of information which parliaments have to deal with: we ask people to democratically vote for parties:

vote(ballot) -> representative

That's not a thin market. It is incredibly hard, time consuming and actually practically impossible to determine optimal representatives and it is why we have to invent costly and unscalable deliberative democracy to get a grip on this decision. I can see it work if (1a) people vote for people in subcommitties of political parties as well (1b) sortition is used.

(2) More direct democracy

We can also engage in more direct democracy to decrease the amount of things parliaments need to decide on. People can vote (using sortition where it makes sense) for the agenda of parliamentarians, as well as direct decisions about legislation and budgets.

(3) Take objective decisions out of parliament

Futarchy, also known as "vote on values and bet on beliefs". Even without widespread prediction markets, by taking as many objectively determinable tasks and decisions out of the hands of parliamentarians, they can focus on their job as "CEO": enforce values and build a valuable culture! Legislative changes which do not require strong new value judgements could be delegated to other institutions.