The Pensions Regulator (TPR) will require trustees to report on record-keeping in their scheme returns to raise standards after finding little improvement.
In an announcement, the watchdog said this measure will better enable it to target interventions at schemes that fail to meet their record-keeping duties.
It comes after its survey of around 532 trust-based schemes carried out by OMB Research between January and March 2016 found standards have not improved at the rate the watchdog would like to see. It also highlights the stark gap in standards between large and small schemes.
While there have been significant improvements since 2008 when TPR first set out its perceptions of record-keeping risks, the rate of change has stagnated in recent years.
Executive director Andrew Warwick-Thompson said: "It's disappointing that we are not seeing more schemes taking their duty to keep proper records more seriously. We've made clear what our expectations are and many schemes, across all scheme types, are not meeting them.
"By adding record-keeping measures to the scheme return, we will be able to target our interventions more specifically at those failing in their duties."
Warwick-Thompson added that good record-keeping is "essential" to the good running of a scheme, providing value for money and protecting member outcomes.
"The time to engage is now - if you don't, you run the risk of increased costs, not managing funding or risks properly, and you could even put members' benefits at risk.
"We know schemes are looking to improve members' experience and engagement through enhanced use of technology such as offering self-service access or participating in industry-wide initiatives like the Pensions Dashboard. But all this relies on good data, and good data security."
The report looks at standards on record-keeping of two types of data in member records: common data and conditional data. Common data is used to identify scheme members while conditional data is scheme-specific data to calculate member benefits. However, administrators' understanding of these two terms is not universal, said the survey.
Overall, 41% of administrators had measured their scheme's common data score, while 12% said they did not know what common data is.
Administrators of large and medium schemes were the most likely to have measured common data (87% and 81% respectively), while administrators for micro schemes were the least likely to have done so (18%).
Two-thirds (66%) of members were in a scheme with a common data score of at least 95%.
In-house administrators were less likely to have measured their scheme's common data compared to externally administered schemes (25% versus 67% respectively).
Similar to the 2014 survey findings, around 90% of members were in schemes where common data had been measured, while two-thirds were in schemes that achieved a common data score of at least 95%.
The survey also found 30% of members are in schemes where conditional data is not measured. Just over a third of members were in schemes with a conditional data score of at least 90%, which is in line with the 2014 survey findings.
Also, administrators and trustees perceive conditional data as secondary to common data. Two fifths of administrators said measuring conditional data was not a priority for their scheme.
Almost a quarter of administrators felt trustees treated record-keeping and administration as a low to middling priority (0 to 6 out of 10), which rose to 32% for micro schemes. Meanwhile trustees of automatic enrolment (AE) schemes were perceived to be more engaged with record-keeping than non-AE trustees.
The report also includes findings from qualitative research conducted with trustees and administrators in 2015, which revealed trustees are generally less engaged with administration than other aspects of their duties. While valuations and investments appear to get more attention, administration tends to be viewed as a day-to-day, technical function that could be outsourced or delegated.
Trustees therefore often approached administration in a passive way, with attention only being paid to it when issues arose or special projects or events necessitated greater trustee involvement, according to the report.
The regulator reminded that it expects all schemes to measure the presence and accuracy of their data, and put plans in place to resolve issues. Trustees and managers should be engaging with their administrators to drive this work forward.
Defined benefit (DB) schemes that provide GMPs must revisit and, where necessary, top-up historic cash equivalent transfer values (CETVs) that have been calculated on an unequal basis, a landmark court judgment said last week.
Technology platform PensionSync has partnered with quantum employment pioneer My Digital to help contractors and employers manage pensions as more workers do temporary work for multiple firms.
Capita Pensions has partnered with data technology solutions firm Intellica to tackle the GMP equalisation challenges facing pension schemes.
The Hewlett Packard Retirement Benefit Plan has reappointed EQ Paymaster as its third-party administrator (TPA) for five years.
Schemes and their administrators have rightly received much praise for ensuring that pensions have continued to be paid in full and on time during an unprecedented period of disruption.