These statistics are lag indicators. They do not reflect nascent phenomena. This means that if something started 4 to 5 years ago (e.g., the Teach Less and Learn More strategy), these numbers won't yet reflect the reality that people like me are living right now.
From my own personal experience, when The Daughter took her PSLE 7 years ago, the system was still not so hugely dependent on parents nor on tuition teachers. The Daughter is in a top class in a top JC. She is in the same class as students who hail from neighbourhood primary schools and not a few of her classmates live still in HDB 4-room flats. Up to 7 years ago, the system worked fine. What happened between 7 years ago and now, is anybody's guess.
What I am sure of is that I am living a PSLE reality today where Little Boy is tested on material he is not taught... where he does not get access through his school to the high quality materials he needs to learn independently (I have to buy material that costs $40 a book)... The PSLE reality today is skills heavy. Yet, the type of teaching does not provide enough skills practice. As a result, parents and private tutors have to come in strongly to support the school's teaching. This greatly disadvantages students from poor homes.
For Little Boy's cohort today to make its way into the type of statistics reported in yesterday's papers, it is likely take a few more years. We should not be waiting for a lag indicator to tell us something is wrong. We should be looking at some lead indicators instead.