Literacy rates in New Zealand present a paradox. Our renowned Reading Recovery programme is an international export success. Yet we are failing dismally to teach our own children to read.

Whether it is rankings in the international literacy league tables, or local assessments by the Tertiary Education Commission (TEC), reading statistics for New Zealand children make, well, bad reading.

In the most recent Progress in International Reading Literacy Study (PIRLS), released late last year, New Zealand's Year 5 students ranked 33rd among the 50 participating countries. And we were last out of English-speaking countries.

A TEC report in 2014 reported similarly grim statistics. Among a sample of Year 11 students with NCEA Level 1, only 49 per cent achieved the international reading benchmark. In other words, more than half of the students were functionally illiterate. If that statistic does not alarm you, then nothing much will.

Reading Recovery's architect, the late Dame Marie Clay, claimed it would "clear out of the remedial education system all children who do not learn to read". How then can our illiteracy rates be so shocking? Especially when Reading Recovery has been adopted in schools throughout the country since the 1980s?

A report from the Education Review Office (ERO) published  in April may point to the answer: for many struggling readers, Reading Recovery does not work.

The ERO report relies on the work of James Chapman and William Tunmer, both professors of education and psychology from Massey University's Institute of Education.

Their research reveals that children from low-decile schools, disproportionately Māori and Pasifika, are less likely than other students to benefit from Reading Recovery.

Massey's Institute of Education is not Reading Recovery's only critic. Research from Macquarie University's Special Education Centre in Sydney found that Reading Recovery does not work well for the children who are most at risk of failing to read. Yet surely that is what remedial programmes like Reading Recovery are designed to do?

The problem appears to lie with Reading Recovery's word-level instructional approach. Rather than decoding words using phonics, Reading Recovery expects children to grasp actual text. This approach is consistent with New Zealand's "whole-of-language" approach to teaching reading in the classroom. But neither works well for children who have not yet mastered the alphabet and lack the phonemic awareness needed to decode words.

In a bid to arrest  its  population's decline in literacy, the Australian federal government is currently pushing states and territories to introduce a mandatory test of phonic skills. The test will measure whether Australian Year 1 students can decode words containing the 44 sounds of the English language.

This will allow teachers to provide support to those children who need it.

In doing this, Australia is following in England's footsteps. In 2006, after an Ofsted review of Britain's declining literacy rates, Labour education secretary Ruth Kelly introduced a "back-to-basics" approach to teaching reading. This made the use of phonics mandatory, effectively replacing the more fashionable "whole-of-language" approach still favoured here.

England's results in the decade since the decision have been impressive. In the PIRLS rankings, English Year 5 children have jumped from 15th to 8th place in the decade between 2006 and 2016.

Unfortunately, New Zealand's Ministry of Education remains resistant to change. In the aftermath of the ERO report, ministry deputy secretary Ellen MacGregor-Reid was reported as saying that the ministry had a "sound and evidence-based approach to teaching children to read" and that Reading Recovery "was successful for the majority of students who accessed it".

But while Reading Recovery may help some children – or even a majority of them – what of those it does not help?

Encouragingly, Professor Chapman's  latest research, published last week in the Australian Journal of Learning Difficulties, reported that many teachers in New Zealand are defying the ministry's longstanding policy, by using phonics programmes to teach reading.

And his study shows that the vast majority of converts report more confident and capable readers as a result.

However, teachers are doing this with one hand tied behind their backs. Professor Chapman's study found teachers had a mixed understanding of literacy-related language structures, meaning their ability to teach phonics effectively is constrained.

The ministry's commitment to Reading Recovery is perhaps understandable. It has invested heavily in it for more than 30 years. Research revealing its shortcomings is no doubt an unwelcome and inconvenient truth.

But there is more at stake here than the ministry's investment in the status quo. The ability to read is the foundation of learning. Ensuring all New Zealand children can read, should surely be the ministry's priority.

England's experience and academic research on both sides of the Tasman suggest the answer to New Zealand's illiteracy problem may not be so paradoxical after all.

Roger Partridge is chairman of The New Zealand Initiative. The role of systemic phonics in teaching reading will be among the topics discussed at the Festival of Education, run by ResearchEd in conjunction with The New Zealand Initiative on June 2 at Auckland Grammar School.