Will this system provide any value added compared with current democracy system?
The crux of this question is, as always when dealing with politics: to whom? For that reason, let us examine the change in value to various demographic slices. These are not meant to be comprehensive, nor mutually exclusive.
Value for the Ruling Class
For the purposes of this discussion, the 'ruling class' are those who have disproportionately large say in the affairs of government and policy. In the United States currently, this can loosely be defined as people who are, on average, better off, better educated and/or have some means to spend a significant amount of their time embedding their judgment in actual legislative or legal policy.
Generally speaking, on average, people do not give up power they have willingly. Some will choose to do so for philosophical reasons, and some will do so rather than risk something else (life, assets, etc.). But there are only rare occurrences where an enfranchised class simply give up power they have. Therefore we can assume that the current ruling class would not accept a 'filtered' voting system unless one of two things happen:
- They maintain or increase their demographic's control of the political machinery.
- They are being forced by a superior effort to accede control in this manner.
Addressing the latter first: it is hard to imagine what movement would arise that would put as a central tenet a filtering of the voting system. Typically this happens in the reverse: peasant revolts against landed nobles, or minorities marching to be given the vote. It's hard to imagine that a movement would say, "We are here to exclude 'x' group, which includes a large number of ruling class members, and will do 'y' problematic thing if not given this concession." Perhaps I lack imagination, though.
What this means is that it's reasonable to assume the ruling class would have a large hand in designing the details of the system - where, as always, the true devil lays. Specifically, they would have to decide these broad areas:
- Who gets to decide what questions are asked for a given vote.
- How changes in who gets to decide are effected.
- What sorts of questions are appropriate for the test, which the people in charge of coming up with them would have to adhere to. We can assume there would be a large number of such rules. (As an aside: the voting in the US today is fairly complex because a number of edge cases must be addressed at every polling place, such as how blind, deaf or otherwise disabled persons can vote, or how information is communicated to those who don't speak English. Therefore we know the rules around the questions would be quite numerous.)
Let us consider the direction of politics each of these effect.
Who Decides the Questions
For a long time institutions have taken on the personality based on the goals they were originally formed around. For instance, while the Department of Agriculture is ostensibly about growing food, one of it's early goals was to promote the trade of food. Even today 'trade' is one topic on it's website and many of it's policies are actually purely economic in nature, rather than scientific or even very food-related at all. This is because promoting American food was very important to the US economy at the time of the founding of the Department of Agriculture. So the climate in which the 'filtered election board' is created is very important to how it turns out. The reason this is important is because it selects the biases that become 'unseen'. All humans are biased, but we tend to overlook certain biases based on our environment. These unseen biases will enfranchise certain parties over others.
If the climate is one of rampant nationalism, for instance, you might expect an institution that produces questions deeply embedded in American culture, thus confusing recent immigrants. If the climate is one favoring older Americans, it may use references to things that happened prior to the lifetime of young voters, thereby impacting their ability to answer correctly. Even maths questions can be asked using different methods, and if you are unfamiliar with the way a question is presented it can inhibit your ability to answer. Were you taught math using new math or reform mathematics? If the writer of a question comes from a different context (educational, lingual, cultural, etc.) than you it will impact your ability to answer correctly. Further, we know that because people tend to hire people similar to themselves that these biases will linger, even if the political winds drift to different shores.
How Changes in Who Gets to Decide Are Effected
Given an institution of some variety to decide questions, we also have to ask how the members of that institution change. This is a classic governance problem: you can go the route of the Supreme Court and once elected to the Filtered Election Institute you are there for life. This is very good if you happen to be good, and very bad if you happen to be bad. It could be a randomly selected, rotating set of people, in which case most of the institutional knowledge will reside not in the decision makers but in the 'hidden' support staff and tools. This makes detecting problems very difficult and pernicious, but to a degree blunts the amount of damage they can do. It can be appointed by the current ruling power, which has some of the problems of both: as we see in gerrymandering whoever is in power is liable to modify the rules in such a way as to extend their power.
What Sort of Questions Are Appropriate
There is a very difficult nut to crack here, in terms of justice, and it has two faces. The first face is one of aroused bias: as explained here even the order of questions will affect the opinion of the person being polled. If I am asking about Presidential candidates, what I ask about them will matter:
How many years of foreign affairs experience does Guy DeFalt have? will prime the test-taker to think about foreign affairs, rather than some other category.
The second issue is one of indirect filtering. Someone who is very good at critical thinking cannot deduce the founding date of the European Union, nor be able to name the member states. Someone very good at memorization, though, could. There are plenty of examples of savants who might not be able to reason about whether a thing is good who would nonetheless know a great deal about that thing.
What this means is that the 'intelligence' being tested is unlikely to be in line with the 'intelligence' that is needed to make a good decision. Instead, it becomes a lever to select who gets to decide: perhaps someone who has been prepped on how to formulaically respond to the questions, or people who adhere to a particular party line.
How the decision of what questions are appropriate is decided, along with what additional information is provided with the question, will decide who 'seems' smart enough.
Final Ruling Class Notes
Naturally, all of these seem to favor the ruling class, and with most systems that filter out some members of a nation, it tends to entrench the ruling class (even if that class is poorly defined). For this reason American democracy, a 'representative democracy' is often correctly named a 'Republic': the difference being that in a Republic only certain members are allowed to vote. But wait! Isn't everyone in America allowed to vote? The answer, as betrayed by your description, is no:
Everyone of legal age is allowed to vote.
'Legal Age' is one way the 'ruling class' in America maintains it's control. It is assumed that people under a certain age are too dumb/uneducated/biased/controllable to vote. In essence we already filter the vote. And the effects are clear: we tend to vote towards things that favor adults, whereas concerns about early education, child health, and so on must be dealt with indirectly by people who do have a vote but care selflessly enough about those issues to be the voice of children. The effect varies depending on the specific issue, but the notion that you can give 'lip service' to eduction, for instance, is common: use it for good press but never fund it as well as, say, the military or social security. It is hard to think that this state of affairs is really in the full, 100% best interest of children, and so we see an effect on the filtered-out class: their actual best interest is a national second-interest. This specific form of filtering became a hot issue during the Vietnam War draft debate when children unable to vote were nonetheless forced into military service by ruling class (those who could vote). They weren't even allowed to drink legally at the time, and it took the casualties of WWII, Korea and Vietnam to finally force the political machinery to address the filter.
Next we turn to people who vote. Does this system provide value to the electorate? Presumably the idea is that they want a better result, to have their voices heard better by filtering out voices that should not be heard.
The real issue with any filtering situation is determining whether one result is 'better' or not. Philosophically this is a problem: if you decide 'x' result is 'better', and you tune the system to give you 'x' result, then you're effectively removing democracy. The value added here, then, is that the electorate feels that the result they wanted is the result they got, rather than mostly feeling like they didn't get what they want. The value is faith in the system.
Ironically, you have to look not-too-closely at this form of democracy to net this result. A test filtering for critical thinking might undo the value you get from the test by choosing only those who will not easily be duped by a biased system.
It is interesting, too, that there is an implied assumption here: that the nation would make better decisions if only the process by which they made decisions was better. But it assumes the bottleneck is the decision process and not the decision makers: this is akin to the difficulty with testing-based education. Even with the best tests you might find the best decision makers, but you're not guaranteeing an overall improvement in the decisions being made - suggesting the only 'better' results can come from an overall improvement in the quality of voter. (Which is impeded by expense.)
One aspect of this is that it requires a great deal of additional effort on the part of voters. Already in the United States voting levels are very low, and in particular people tend to only vote on big-ticket, high-level elections. If they vote on lower-level elections, it is often only when they are voting at the same time for a higher-level election. Often they don't vote because voting is difficult and time consuming. Sometimes voters only vote on the one thing they care about and leave the rest of the ballot blank (begging the question as to whether they would take the time to do the test first, or stick around and take the test after if that was the ordering).
By requiring voters to 'study up' to vote, and to take additional time to answer questions the system becomes one where, naturally, more people self-select out: it's too much effort. Therefore the value to the voters is that those who do vote are particularly motivated. Via another method, Robert Heinlein asserted this 'limited democracy' in his book Starship Troopers. A system requiring highly motivated voters is likely to be one that has many such artificial barriers in place: a meritocracy whose quality is proportional to the degree with which the barriers are in line with what they are testing for. The specific effects of this could be varied, and interesting to play with.
As a quick aside: it would also greatly increase cost to run a specific polling place and elections in general. Recounts become not only recounts of votes but of filtering tests. The length of the election is potentially increased in close races, and the value to the motivated voters feeds back on itself: the next time you have to be even more motivated knowing that your vote may not be correctly counted even with all the effort you're expending. Each new barrier or filtering test increases the chance of something going wrong or the system skewing in an unintended way. Faith becomes doubly tested.
Another interesting question is what value is presented for non-voters. The system proposed has a particularly curious tangle: the voter doesn't know if they're a voter.
Unfortunately, lacking imagination, I'm hard pressed to come up with even a scant value for the voter here. Already, we exist in a system where we must place great trust in the machinery to count our vote. Votes are often accidentally discarded or lost. Hanging chads and marking disagreements abound - and there are electronic equivalents. In a system where there was both a ballot and a test, and a failing score on the test invalidates the ballot counting and verification of the count becomes problematic. I, as a voter, don't know that if my time was well spent or meaningless. I have no way of verifying myself that the ballot was cast.
The primary value here, actually, seems to be for those who would foment anger at the system: they could easily point to the system and say, "For all you know, you voted and they did what they wanted anyway!" We can easily imagine a Nigel Farage, post a 'Remain' vote saying, "Well, I just can't imagine all these impassioned people having voted 'stay': the test is rigged, the system is rigged!"
And, of course, in this proposal it is rigged. But it seeks to create a larger class of non-voters, on the assumption that 'intelligent' voters would vote 'better'. Without knowing if your vote counted (if it was 'good enough'), it's hard not to feel like you were in the 'stupid' class of people if what you wanted didn't win. If things start going wrong, how can you not feel angry and powerless? That sense of powerlessness is useful for those who would make marked, large changes to the system. People act funny when they suspect they might not be at the cool kids table: that chaotic behavior is rarely good for stability (or that matter the ruling class).
As indicated, a lot would depend on how this system was implemented. There are certainly modifications or implementation details that could be utilized to enhance value for some people and reduce damage for others. But the nature of those details would affect the resulting 'world' that arose. Certainly it would change how power flowed from the electorate to the decision makers, and the notion of a quantum vote (did it count?) is intriguing in it's own right. The outcomes are very hard to predict, but it would certainly seem to hurt the disenfranchised and help the ruling class: the question then is if the hurt is less than the total value to the system. (Which, indeed, is a very broad question.) It seems fairly obvious why filters have been removed over time rather than added: inclusivity leads to less anger in the general populace. (Remember, the US originally only allowed land owning men to vote: filters have changed greatly over the years!) Based on what drove the removal of those filters we can see what we would be (eventually) getting by putting them back into place.
But to directly answer your original question: would there be value added by filtering a vote based on a test? Yes, some value would be added for some people. The specific people would depend on the details of how and what precise filter was added. Value would be taken away from other people, already likely under-represented in the system. An 'educated' vote is likely to come at the cost of a populace with less faith in that voting system. In the short term, the best interests of those making the change are likely to be best served. In the long term, it's likely to be a less stable system.
Addendums and Asides
- To bring Brexit back into this: those who voted to Leave often did so because they felt they lacked a direct representative voice in the EU politics, and were being screwed by the ruling British class and the EU bureaucrats. The EU was not benefitting them directly. While it may have been bad for the system, they used this vote to voice the fact their votes elsewhere seemed to not be counting - and by 'not be counting' I mean that their best interests were not being served. In this manner the referendum was an indirect vote on the quality of the system: the voters know something is wrong, even if they don't know the immensely complicated economics of the EU-British situation. When they're asked about the EU, they voice something is wrong by voting for change. Simple votes are rarely about a direct, singular, easy-to-understand, up-and-down issue.
- You also posit
80% of the populace must pass. It's unclear whether this is decided 'for that election' (i.e. the top 80% of the test takers have their vote counted) or 'for future elections' (in which case the questions might eventually drift to 'so easy to answer anyone can do it'). In either case it begs the question as to whether voters should care about the test: if more than 20% of the voters answer no questions then they all equally have a chance to have their completely unvalidated vote being counted. In fact, depending on how you 'score' the test, their votes may count more than someone who tests badly, or someone who answered test questions randomly. This is the question as to whether a non-answer is penalized more or less than a wrong answer. This is a big deal, and has plagued college entrance exam tests for years, but in the case you posit it suggests the 'smart' thing for voters to do is universally decide not to take the test.
- A security flaw also arises if the test must be passable by 80% of the people. In this situation you can probably game the system by answering just enough of the questions correctly, and then the remaining questions can be used to encode who you are. This is a problem because it means that third parties can positively identify which vote was yours, and determine whether you voted as they wanted you to. California has roughly 40M people (the most populous state), meaning that with as few as eleven 'optional' five-response questions you could easily positively identify any single voter. (Can you ascertain the voter has enough knowledge with fewer than 14 questions?) Meaning people could be coerced into voting. I'm sure there are even more efficient ways (i.e. fewer questions needed to uniquely id an individual) to do this.
- The proposed system also suggests that the questions are topical: that is, they relate to the vote. The Filtered Election Institute would have to know an immense amount, or outsource expertise, to determine good questions to test for competence in any given area. Note that the current educational system struggles to do just this, and their sole mandate is educating and articulating the differences in education between people. The FEI would be a very interesting institution because the shortcuts it chose would have profound impacts. Great fodder for a dystopian novel there.