>> if adj-no is to be kept, i'm more on board with Johan's method of "this seems to get used a lot as a descriptor and if you look it up in gg5 and eijiro, they offer adj-no translations, so we should have it as adj-no".

I really don't know why you're saying this given the discussion up to this point.
My previous two comments go into detail about why translations tell us nothing useful.
"金融" is no more of a "descriptor" than "人", or indeed any noun.

reopening and removing adj-no, as it was before i closed, though i think it should remain

@@ -18 +17,0 @@
-<pos>&adj-no;</pos>

temporarily closing to get these entries out of the queue

@@ -17,0 +18 @@
+<pos>&adj-no;</pos>

i guess my point is that i don't we have any evidence that there's a difference between category 1 and category 2 in the minds of most native speakers.  i think it's the result of "this japanese-english dictionary has a PoS called [adj-no], so now i need to seek out some distinction between japanese words in order to make sense of it", even though there's no consistent japanese reference for the PoS existing at all, let alone for different categories of nouns to which it can and cannot be applied.
-
if adj-no is to be kept, i'm more on board with Johan's method of "this seems to get used a lot as a descriptor and if you look it up in gg5 and eijiro, they offer adj-no translations, so we should have it as adj-no".  at least then we have some references we can consult to decide whether or not something should be marked [adj-no]
-
or we could just delete adj-no altogether since the evidence that it actually exists as a PoS is scant, and we're really just using it for our own convenience.

I probably shouldn't have used the terms "category 1" and "category 2"; I only did that for convenience.
What I'm arguing is that only a small proportion of words that take の behave like adjectives. 
I wouldn't say this is "shoehorning Japanese into English translations"; I'm actually trying to forget English altogether. Thinking in English is part of the problem.
---
Structurally, there's no difference between 金融のガバナンス and 牛の飼育 (cattle raising/raising of cattle), but I don't think anyone would say that 牛 is functioning like an adjective here.
The way I see it, it's only because the English adjective "financial" exists that someone would argue that 金融 should be classified as [adj-no].
Now compare 牛の飼育 and 牛の病気. Again, these are structurally identical but because we can translate the latter as "bovine diseases", 牛の suddenly beings to look like an adjective to some people.
If 金融 should be considered [adj-no], then I see no reason why 牛 shouldn't either. The question then becomes "which nouns *aren't* [adj-no]?"
---
Simply put, it's useful to be told that words like 衝撃 can be used in in constructions like 衝撃の事実 because you can't do this with most nouns. (悲哀の事実, for example, does not mean "saddening fact").
However, we don't need to be told that 金融 can be used in constructions like 金融のガバナンス because *all* nouns can be used like this.
I don't know how to explain in linguistic terms the difference between 衝撃の事実 and 金融のガバナンス but I feel it's clear that 衝撃 looks and behaves like a 形容詞 or 形容動詞 here whereas 金融 doesn't.
---
Regarding those meikyo examples, I would say they're all [adj-no], including 鉄. 鉄 describes an attribute of 扉. Contrast with 鉄の生産.

It is possible that in the mind of a Japanese person there is really no difference between Robin's category 1 and 2.
But how could I know?
--------
Here is how I go about these things
金融 is mostly used as part of longer compounds, such as 金融機関
For independent use we have the KM ngrams 
金融の	16790
金融に	 5255
金融と	 3047	  
金融を	 3021
金融が	 1687	  
金融は	 1572	
We see that independent use is about 50% 金融の.
I'd be happy to make it an [n,adj-no]
----
If we remove adj-no from the dictionary, should we use noun glosses for all these words?
That can lead to some awkward glosses.

and i mean this is also in gg5:
金融の　monetary; financial; credit
...
so they don't follow the proposed categorization either.  i think we'd just be making it up without reference

i think that's a good classification, at least as far as it goes for us as non-native speakers/translators trying to turn japanese into english.  what i wonder is:
--
1) is there *actually* a difference in japanese between your category 1 and category 2 that most japanese linguists would recognize, or is that categorization merely the result of trying to shoehorn japanese into english translations? for example, meikyo does not seem to distinguish.
状態・状況・素材などの特性を表す。
「薄幸の人」
「休業中の店」
「鉄の扉」
if i'm following you, you would argue that 薄幸 *should* be an adj-no, but that 鉄 *should not* be an adj-no?  but meikyo does not distinguish.
--
2) if this is not a recognized PoS, should we really be defining our own japanese PoS simply because it helps us translate?
--
3) what are our references for what are 'category 1' and 'category 2' nouns?  if we are the only source that recognizes this, are we just making it up as we go along? are any of us qualified to be dictating japanese grammar without any reference?

I'm strongly in favour of keeping the [adj-no] tag but I think it needs a much stricter definition if it's to be of any use to anyone.

As it stands, the entries that are tagged [adj-no] can be split roughly into two categories.

The first comprises words that function much like 形容動詞 (na-adjectives): 衝撃 (e.g. 衝撃の告白),  地下 (e.g. 地下の水流), 使い捨て (e.g. 使い捨ての紙おむつ), 偽 (e.g. 偽のダイヤ), 厚手 (e.g. 厚手のセーター), etc. (Here I'm ignoring actual 形容動詞 that can also take の).
Some of these are almost always used with の (like 厚手) while others also function as regular nouns (like 衝撃 and 地下).

The second (much larger) category consists of everything that doesn't fall into the previous category — essentially any noun, e.g. 結婚, 俳優, 医師, 下痢, 喘息, 骨盤. 

A large proportion of these nouns have been tagged [adj-no] solely because of Eijiro, which uses [noun]-の constructions to translate thousands upon thousands of (often obscure) English adjectives (e.g. "iatric" - 医師の, "histrionic" - 俳優の).

Unlike the category 1 words, these nouns do not function anything like 形容動詞. "結婚の誓い" can be translated as "marital vow" but 結婚な誓い is nonsensical (unlike 衝撃な告白 which is merely ungrammatical).
What's more, you can't flip the sentence round and preserve the meaning — compare 誓いは結婚 with 告白は衝撃.

When の is used with these category 2 nouns, it really means "of or relating to", and there's no end to the number of nouns for which this sort of usage is applicable.

For that reason, I'm in favour of dropping [adj-no] from *all* category 2 nouns.
If usage of the tag is restricted to words from the first category, [adj-no] actually serves a purpose and conveys useful information to the learner. Otherwise, it isn't really helpful at all.

I'm sure there are cases which don't fit neatly into this 2-category model but I think it's a good starting point. I'm interested to hear other people's thoughts.

I agree with Rene that adj-no is not a real PoS.
And it can often be difficult to decide whether an entry is [n] or [n,adj-no].
In such cases I'm happy to go either way.
--------------------------------
But I still think that adj-no can give useful information.
For instance
[adj-no] = almost only used with の
[adj-no,n] = usually used with の
[adj-na,adj-no] = used with either な or の with the same meaning
[adj-no,adj-na] = used with either の or な with the same meaning

eij
 金融の
【形】
    financial
    moneyed
金融のガバナンス
financial governance
金融のグローバル化
financial globalization
--
etc.

obviously "金融の" could just as easily be translated as "of finance" as it could be translated as "financial". but i'm not sure that [adj-no] is a real PoS at all, and this seems just as valid as some other [adj-no] that are recorded...
--
(this is one of the reasons i'd prefer to get rid of [adj-no].  it's not clear when we should say that something is an adj-no or just a use of "...of [noun]", and the kokugos are useless since they don't recognize adj-no.)

prog

For reverse searches.
Not [adj-no].

@@ -18 +18 @@
-<pos>&adj-no;</pos>
+<gloss>finance</gloss>