Audit Procedures Over Census Data Used in Pension Calculations

Statistical Sample Size & Risk Assessment for Pension Audits

Census Data Audit Testing Attributes (Standard Procedures)

Attribute Field	Verification Source	Audit Objective
Date of Birth	HR Records / ID Documents	Existence & Valuation
Date of Hire	Original Personnel File	Eligibility & Vesting
Gender	Personnel Records	Actuarial Assumptions
Compensation	Payroll Registers / W-2	Benefit Calculation Accuracy
Employment Status	Termination Notices / Pay Status	Completeness of Liability

What is audit procedures over census data used in pension calculations?

Audit procedures over census data used in pension calculations involve the systematic verification of underlying participant data submitted by a plan sponsor to an actuary. This data forms the bedrock of the actuarial valuation, which determines the pension liability and required contribution levels. Census data typically includes demographic information such as names, social security numbers, dates of birth, dates of hire, gender, and salary history.

Auditors use these procedures to ensure that the data is complete and accurate. If the census data is flawed—for example, if birth dates are incorrect or if active employees are excluded—the resulting pension liability calculations will be materially misstated. This is a critical focus for auditors following standards like AICPA AU-C Section 501 and GASB Statements 67 and 68.

Common misconceptions include the belief that the actuary is responsible for the data’s accuracy. In reality, the plan sponsor is responsible for the data, and the auditor must perform independent audit procedures over census data used in pension calculations to provide assurance to stakeholders.

Formula and Mathematical Explanation

To determine the sample size for testing census data, auditors typically use attribute sampling. The formula calculates the number of items needed to conclude with a specific level of confidence that the error rate in the population does not exceed a tolerable limit.

The Basic Attribute Sampling Formula:

n = (R / T) (Simplified for 0 expected errors)

Where:

• n: Required Sample Size
• R: Reliability Factor (based on confidence level)
• T: Tolerable Deviation Rate

Variable	Meaning	Unit	Typical Range
Population (N)	Total plan participants	Count	100 – 1,000,000
Confidence Level	Risk of overreliance on controls	Percentage	90% – 99%
Tolerable Rate (T)	Max errors allowed	Percentage	2% – 10%
Expected Rate (E)	Anticipated errors	Percentage	0% – 3%

Practical Examples

Example 1: Large Municipal Pension Plan

A city has a pension plan with 10,000 participants. The auditor assesses control risk as high and sets a 95% confidence level with a 5% tolerable deviation rate. Using our audit procedures over census data used in pension calculations calculator, the auditor identifies a required sample size of 60 records (assuming 0 expected errors). Testing involves tracing these 60 records back to original birth certificates and payroll files.

Example 2: Private Sector Defined Benefit Plan

A corporation has a frozen pension plan with 500 members. Because the data hasn’t changed much, the auditor accepts a 90% confidence level and a 7% tolerable deviation rate. The calculator results in a smaller sample size of approximately 33 records. However, if any error is found in those 33 records, the auditor may need to expand the sample or qualify the audit opinion.

How to Use This Calculator

1. Enter Population Size: Input the total number of members in the census file provided by the sponsor.
2. Select Confidence Level: Choose 95% for high-risk audits or 90% if internal controls are strong.
3. Set Tolerable Deviation Rate: Define the percentage of errors that would be considered material to the actuarial valuation.
4. Review Results: The calculator immediately provides the “Required Sample Size” and “Max Allowable Deviations.”
5. Sensitivity Analysis: Look at the SVG chart to see how increasing your tolerable rate significantly reduces the audit workload, though it increases audit risk.

Key Factors That Affect Audit Results

• Data Integrity Controls: Strong IT controls over HR systems reduce the need for large manual samples.
• Plan Complexity: Plans with multiple tiers or complex benefit formulas require more granular audit procedures over census data used in pension calculations.
• Actuarial Sensitivity: If a small change in age or salary leads to a massive change in liability, the tolerable error rate must be lower.
• Population Turnover: High rates of hiring and termination increase the risk of “Date of Hire” and “Date of Termination” errors.
• Prior Audit History: If previous audits found many errors, the “Expected Deviation Rate” must be increased, which raises the sample size.
• Type of Plan: Defined benefit plans are much more sensitive to census data errors than defined contribution plans.

Frequently Asked Questions (FAQ)

1. Why is census data testing mandatory?

Standard audit procedures over census data used in pension calculations are mandatory because the actuarial liability is the largest number on many financial statements, and it is derived entirely from this data.

2. What happens if I find one error in my sample?

If the number of errors exceeds the “Max Allowable Deviations,” you must either expand the sample, perform alternative procedures, or ask management to correct the entire population.

3. Does this calculator work for GASB 68 audits?

Yes, it is specifically designed to help auditors determine sample sizes for employer census data testing in cost-sharing or agent multi-employer plans.

4. Can I use a SOC-1 report instead of testing?

A SOC-1 Type 2 report can provide comfort over the *system*, but auditors often still perform substantive testing of the *data* actually processed by that system.

5. Is salary testing part of census data audit?

Yes, “Average Final Compensation” is a key input for pension formulas, making salary verification a core part of audit procedures over census data used in pension calculations.

6. How often should census data be audited?

Typically, it is audited annually as part of the financial statement audit, though full population reconciliations may happen every 3-5 years during experience studies.

7. What is the difference between active and retired member testing?

Active member testing focuses on hire dates and salaries; retired member testing focuses on existence (death audits) and correct payment amounts.

8. What is a ‘Tolerable Deviation Rate’ for most audits?

Most auditors use 5% for a standard “high” reliance level, but this may drop to 2% if the plan is underfunded or under heavy scrutiny.

Related Tools and Internal Resources

• pension-actuarial-valuation: Learn how census data influences the final liability figures.
• gasb-67-compliance-guide: Specific requirements for plan financial reporting.
• audit-sampling-methodology: A deep dive into statistical vs. non-statistical sampling.
• defined-benefit-risk-assessment: Identifying high-risk areas in pension auditing.
• hr-data-integrity-checklists: Best practices for sponsors to prepare for audit procedures over census data used in pension calculations.
• pension-plan-audit-standards: Updated AICPA guidelines for retirement plans.