In theory, I absolutely disagree. I'm assuming that what you label LCT (never heard the term, but then I'm notorious for not following trends,) is like 4GL, Application Generators, etc. In the Pick world alone, there are numerous such, which long-time, very successful companies are using. My own experience more than bears this out. Forge/Ultimate Update (clones) was the engine behind many very successful, complex applications. The one I have dealt with for the last 12 years, is an html generator that can create very complex browser forms & UI.
There is no question that both these products have reduced mindless, repetitive programming by, say, 60-70%.
The secret is an ongoing development by the authors of the product. Of course if it's static it won't meet new needs. And I'd agree there a lot of simplistic, mostly code generators, which expect you to modify the genned code for complex functionality. This of course voids the reason to use the product...if you change the code, it won't be regenerated correctly if you use the product to make further changes.
AFA reports are concerned, agreed that the standard report generator in any flavor of Pick is limited. But again, there are ways around this. I found that a large number of reports, 60-70% could use the standard generator; my 4GL would just make it easier to parametrize it so you don't have to remember the syntax. Then, an additional percentage of reports was done by massaging the raw data into a temporary work file, then using the standard report generator. This way things like rolling totals and page break issues were performed programatically; the report generator was used as a simple formatter, which it excels in. Then minor mods to the report, not involving data change, would be trivial.
I don't know what LCT tools U2 has, but non-vendors certainly have great ones.
Chandru