Some Overpayments Aren’t Worth Trying To Collect

      Social Security’s Office of Inspector General has issued a report on an investigation into the cost effectiveness of the agency’s efforts to collect small overpayments. Here’s an excerpt.

… Of the 250 low-dollar OASDI [Old Age, Survivors and Disability Insurance] overpayments we reviewed, SSA took actions on 50 (20 percent) that we did not consider cost-beneficial because it sent more notices to the overpaid individuals than required. Since SSA could not provide its average cost to send an overpayment notice, we applied the average cost to collect overpayments as reported in CAS during our audit period, and we did not consider it to have been cost-beneficial to recover these 50 overpayments. Specifically, we estimate SSA spent $14,492 to attempt to recover the 50 overpayments, which totaled $8,129. 

Projected to our population, we estimated SSA spent $4.6 million to recover almost 16,000 low-dollar OASDI overpayments totaling almost $2.6 million. Therefore, we estimate SSA spent about $2 million more than it would recover.  …

Medicare Enrollment Penalty Mistakes

      From a report by Social Security’s Office of Inspector General:

SSA and the Centers for Medicare and Medicaid Services share administrative responsibilities for Medicare. …

SSA enrolls beneficiaries, establishes applicable premium penalties, and collects premiums from individuals who receive Social Security benefits. If an individual signs up for Part B after their Initial Enrollment Period, they may have to pay a late enrollment penalty. 

We identified 101,516 OASI beneficiaries who enrolled in Part B during the 2023 and/or 2024 General Enrollment Periods and whom SSA penalized for late enrollment. We reviewed a random sample of 200 of these beneficiaries. 

SSA employees accurately processed 177 of the 200 Part B applications we reviewed. However, SSA employees did not accurately process the applications for the remaining 23 beneficiaries. As a result, SSA’s systems improperly assessed approximately $24,000 in Part B premium penalties. 

Based on our sample results, we estimate SSA employees accurately processed approximately 90,000 beneficiaries’ applications and did not accurately process 12,000 beneficiaries’ applications. As a result, SSA’s systems improperly assessed about $12 million in Part B premium penalties. 

These errors occurred because SSA employees did not consider Group Health Plan coverage, the impact of U.S. residency and lawful presence start dates on Initial Enrollment Period determinations, deemed enrollment, Exceptional Conditions, Special Enrollment Periods, and equitable relief. …

     What I’ve seen over the years is that you can almost always find a legitimate way to avoid the late enrollment penalty. I don’t know why they even try to apply a penalty. It’s not worth the trouble.

41% Error Rate On Widow And Widower Claims

     From a recent report by Social Security’s Office of Inspector General:

The Old-Age, Survivors, and Disability Insurance program provides benefits to wage earners and eligible family members. The Agency uses the wage earner’s Primary Insurance Amount (PIA) to determine monthly benefit amounts. The eligibility year SSA uses to determine the PIA is usually the year a wage earner attains age 62, becomes disabled, or dies. However, an alternative PIA computation for widow(er)s—the widow(er)’s indexing (WINDEX) PIA—may apply when wage earners die before they attain age 62. 

When a claimant applies for Old-Age, Survivors, and Disability Insurance benefits, the application usually covers all benefits for which the claimant is eligible unless they specifically limit the scope of the application. For example, widow(er)s may limit the scope of the application to only include widow(er) benefits and exclude retirement benefits to maximize future benefits. 

We reviewed 2 samples totaling 120 beneficiaries who, as of November 2023, either did not have a WINDEX PIA (from a population of 54,843 beneficiaries) or were dually entitled to widow(er)s and retirement benefits (from a population of 7,253 beneficiaries).

SSA paid 71 (59 percent) of the 120 widow(er) beneficiaries we reviewed the appropriate monthly benefit amounts. For the remaining 49 (41 percent), we found the following. 

• SSA employees did not apply the WINDEX PIA appropriately when they manually processed cases for 11 widow(er) beneficiaries and, as a result, did not pay the appropriate monthly benefits. We could not determine why SSA employees did not appropriately apply the WINDEX PIA for these widow(er) beneficiaries. Based on our random sample, we estimate SSA underpaid 8,618 widow(er)s approximately $50.4 million.
• SSA overpaid one widow(er) because employees used the incorrect PIA.
• SSA employees did not document in the Agency’s system regarding whether they informed 37 widow(er) beneficiaries of their option to receive widow(er) benefits only and delay filing for retirement benefits. Therefore, we could not determine whether SSA appropriately paid these widow(er) beneficiaries despite reminders SSA issued to employees. We estimate 5,367 widow(er)s would have been eligible for $113.8 million in additional benefits had they chosen to delay their retirement claims. …

Critical Payment Errors

      From a report by Social Security’s Office of Inspector General:

An SSA employee initiates a critical payment when an Old-Age, Survivors, and Disability Insurance beneficiary or representative payee alerts SSA of a critical case or special situation when SSA is not paying regular monthly benefits, additional benefits are due, or a beneficiary reports they did not receive a monthly benefit. These include dire need, court orders, legislative mandates, and preliminary and expedited payments. …

We reviewed a random sample of 175 critical payments from a population of 3,549 issued in Fiscal Year 2023 …

We estimate SSA employees accurately processed about 44,000 (62 percent) of the 70,980 critical payments and did not accurately process about 27,000 (38 percent)—making over 28,000 errors when they processed the payments. About 2,800 of these errors resulted in the payments being incorrectly documented on beneficiaries’ records. For the remaining errors, we estimate the following. 

• Field office employees improperly paid approximately 6,900 beneficiaries about $1 million, but processing center employees identified the special situations and deductions from payment calculations during their post-payment review and adjusted future benefits due. 

•  About 8,100 of these errors resulted in SSA improperly paying beneficiaries $12 million because employees did not accurately adjust beneficiary records.

• SSA issued approximately 10,500 beneficiaries a Form SSA-1099, Social Security Benefit Statement, with a benefit total that was over- or understated by about $14 million because employees did not manually adjust records for replacement checks. …

“Financial Hardship And Emotional Distress”

      From Follow-up on Claims Denied Because Claimants Were Not Insured for Benefits, a report by Social Security’s Office of Inspector General:

… This audit is a follow up to our 2016 review of Retirement Claim Denials Because of Lack of Insured Status.  

Generally, to be insured for retirement benefits, a claimant must have 40 quarters of covered earnings and have attained age 62. SSA employees should not deny a retirement claim if the claimant is not insured for benefits at the time of filing but will become insured within 4 months and evidence of the earnings is available.  

We obtained the records of 450,209 retirement claims filed between May 2014 and June 2023 that SSA employees determined the claimants lacked insured status. From this, we identified 4,077 claimants who may have been insured because they had 40 or more quarters of coverage in the year of filing. Of the 100 claimants in our sample, SSA employees denied retirement claims for 43 who alleged lag earnings when they filed their claims, were fully insured, and entitled to retirement benefits but employees did not consider their lag earnings. Despite reminders issued to employees after our prior review, employees denied the retirement claims because they determined the claimants lacked insured status; however, the claimants had lag earnings when they filed their claims.  

Based on sample results, we estimate, from May 2014 to June 2023, employees denied retirement benefits to 1,753 claimants who were insured for benefits. Of these, 1,347 claimants were entitled to over $3 million in retirement benefits.  

Without controls and processes to ensure employees identify, review, and document lag earnings, SSA will continue denying millions of dollars in retirement benefits to claimants who should be receiving them. Depriving retired individuals of the benefits to which they are entitled could have a significant and harmful effect on beneficiaries, including financial hardship and emotional distress. …

     This problem is not limited to retirement claims. I’d say it’s more common in disability claims. 

I Don't Think We Can Trust This Report

      Back in July, Senator Elizabeth Warren asked that Social Security's Office of Inspector General (OIG) perform an audit of Social Security's reported telephone call waiting times. The audit has now been completed and a report issued -- just before Christmas. The report says that the agency's metrics were accurate which is probably technically accurate but whether those metrics are fully meaningful is another question.        

    There are at least three problems with this report. First, there is zero reason to trust OIG. It no longer enjoys any independence. It doesn't report to the Commissioner but it does report to the White House. It is now clear that there can be no OIG reports at any agency which criticize the Administration. The release of such a critical report will be blocked and those who drafted it summarily fired. (Thank you, Supreme Court.) Second, the report admits that 25 million callers to Social Security became so frustrated by Social Security's answering system that they hung up. It didn't add these callers in to the phone answering metrics. If you do add them in you find out that they comprise a whopping 40% of calls to Social Security. That's a lot of frustrated callers. Third, many of the calls were handled by Interactive Voice Response (IVR) rather than a human. How effectively did IVR respond to requests for customer service? The agency reported a big jump in telephone calls this year. How much of that was due to failed IVR? Failed IVR also contributes to customer dissatisfaction. 

     Senator Warren is already accusing the Social Security of lying.

Again, Mr. Commissioner, What Are You Going To Do About These Payment Errors?

      From a recent report by Social Security’s Office of Inspector General:

The Old-Age, Survivors, and Disability Insurance program provides monthly benefits to retired and disabled workers and their dependents as well as the survivors of deceased workers. In general, to be entitled to benefits, a child of a retired, disabled, or deceased worker must: 1) be unmarried; 2) be under age 18, a full-time elementary or secondary school student under age 19, or have become disabled before age 22; and 3) meet certain relationship and dependency requirements. Generally, an SSA employee may appropriately deny a claim when the employee properly completes all necessary actions and determines the applicant has not established the claimant meets the requirements to be entitled to child’s insurance benefits. 

We reviewed a random sample of 100 claims from a population of 21,533 claims filed from January 2019 through July 2023 that SSA employees approved for benefits in July 2023 or earlier and a random sample of 100 claims from a population of 75,424 claims filed from January 2019 through July 2023 SSA employees did not approve for benefits as of July 2023. 

Of the 96,957 claims in our review, we estimate SSA employees correctly denied 37,712 (39 percent) and incorrectly denied 24,555 (25 percent). As a result of employee errors, SSA did not pay these beneficiaries approximately $92.2 million in benefits and delayed paying these beneficiaries approximately $87.7 million in benefits to which they were entitled. 

We could not conclude whether employees correctly denied the remaining 34,690 claims (36 percent). This includes an estimated 28,661 claims SSA employees denied before they appropriately completed all required actions; therefore, there was not enough information in SSA’s records to determine whether Agency employees appropriately denied the claims. …

     This might be tough for the Commissioner. I don’t think it can be solved by laying off employees or intimidating them or demanding they work harder.  It will take analysis of the difficulties in doing this work accurately, coming up with better workflow processes, coming up with better quality control processes and being honest with everyone about the workforce needed to do the job properly. The honesty part will be the hardest thing for this Administration. It’s easier to blame the “deep state” than to do the difficult work of governing. 

Are You Suggesting That There’s A Relationship Between Staffing And Productivity? That’s Crazy Talk!

      From a press release issued by Social Security’s Office of Inspector General:

From FY 2019 to FY 2023, Disability Determinations Services (DDS) productivity, measured as Production Per Work Year, decreased by 21 percent, and average processing time increased by 81 percent from 121 to 219 days. The lower productivity and increase in processing times coincided with the loss of key technical staff, including disability examiners who evaluate disability claims and make disability determinations in accordance with laws, regulations, policies, and procedures governing Social Security Administration (SSA) disability programs.

Social Security Doing A Poor Job Of Resolving Critical Issues

      From a recent report by Social Security's Office of Inspector General:

Objective
To determine whether employees submitted and processed Manager-to-Manager (M2M) requests in accordance with Social Security Administration (SSA) policy.

Background
When a field office (FO), processing center (PC), or teleservice center employee identifies a critical issue for a beneficiary that requires another office’s action, managers can expedite action for the beneficiary by initiating an M2M request. Managers should only use M2M for high priority requests, such as beneficiaries who have terminal illnesses; made homicidal, suicidal, or potentially violent behavior threats; or are in dire need situations, such as facing eviction or homelessness.

Given the critical nature of M2M requests, FO and PC employees must address them within 5-business days or provide an interim reply to the requesting manager explaining the delay.
We reviewed a random sample of 100 M2M requests: 50 in a “resolved” status from June 1, 2021 through June 1, 2023 and 50 in a “pending” status as of June 1, 2023.

Results
SSA employees generally submitted M2M requests according to SSA policy; however, they did not always follow policy when they processed M2M requests. Of the 100 M2M requests we sampled, FO and PC employees did not process 57 requests according to policy.

• For 48 requests, FO and PC employees did not process them timely, resulting in delays in employees addressing critical issues and beneficiaries waiting weeks or months to receive the benefits they were due. 
  

• For 9 requests, PC employees placed them in a “resolved” status in the M2M application before completing all necessary actions to address the requests.
  

SSA managers provided reasons for delays, and we identified control weaknesses that contributed to delays, such as: (1) case complexity; (2) insufficient communication between offices, including no notifications in the M2M application when employees take action on requests; and (3) the absence of controls that prevent employees from prematurely closing M2M requests.  ...

    This is the sort of thing that drives me and other Social Security attorneys crazy We can't get problems resolved.

The Actual Scope Of The Problem And The Sophisticated Efforts To Deal With It

     A press release from Social Security's Office of Inspector General:

Receiving accurate reports of death is a major and ongoing concern for the Social Security Administration (SSA). A recent report by the Office of the Inspector General (OIG) illustrates the need for SSA to continue to improve its death reporting processing systems.

According to the report entitled, “Rejection of State Death Reports,” from November 2018 through October 2022, states throughout the country submitted about 13.7 million death reports to SSA. SSA’s Death Information Processing System (DIPS) accepted about 12.2 million and rejected nearly 1.5 million (11 percent) state death reports. SSA OIG determined that SSA rejected over 1.4 million state death reports that did not pass DIPS verification checks. SSA uses DIPS to verify the death information it receives. DIPS rejects death reports that do not pass its verification checks to prevent posting erroneous death information to SSA records.

It is only after death information passes DIPS verifications that SSA records the information to the Numident (a database that stores information for all Social Security numberholders) and terminates payments to deceased beneficiaries. The verification checks prevented DIPS from posting incorrect or duplicate death information to SSA records for approximately 773,000 of the 1.4 million death reports submitted by states.

However, SSA OIG reported an estimated 702,000 of the 1.4 million state death reports that were rejected contained valid death data but did not pass DIPS verification checks. DIPS rejected most death reports when it detected a verification date submitted by the reporter was different than the latest verification date in SSA’s records. When this occurs, DIPS rejects the report without validating whether the reported death information is correct. This delays posting of the date of death to the Numident and payment records and results in continued payments to deceased beneficiaries until SSA receives and processes the death information.

This issue led to improper payments to beneficiaries of $327 million and could lead to an additional $108 million over the next year if SSA does not add death information to payment records for beneficiaries in current payment status. Moreover, SSA employees must manually process the rejected death records. It is estimated it will require SSA employees to spend 199,000 hours to process this workload, costing $12 million in administrative expenses. OIG made three recommendations to improve the accuracy of the death information in its records to which SSA agreed. See the full report here.

    These things are vastly more complicated than DOGE can imagine.

I Have Been Seeing More Of This Problem In The Last Year

      From a recent report by Social Security’s Office of Inspector General (OIG):

… [W]e reviewed a sample of 274 cases from a population of 1.5 million recipients SSA placed in non-payment status codes from March 2020 through May 2022 after determining the recipients failed to provide requested information or take requested actions. Additionally, we identified 61,176 recipients who were placed into 7 non­payment status codes during periods SSA had prohibited their use. 

Results 

SSA did not act in accordance with its policy and procedures when it processed SSI ineligibility determinations and suspensions based on applicants’, recipients’, or representative payees’ failure to provide information. SSA’s employees did not complete all required steps for 156 (57 percent) of the 274 sampled cases placed in non-payment status, which led to 96 of the 156 recipients not receiving $203,133 in SSI payments they should have received. Projecting these results to our population, we estimate SSA did not follow its policy before it denied or suspended SSI payments for 871,330 recipients. Of these recipients, we estimate 536,203 did not receive $647 million in SSI payments they should have received. …

     It’s way too easy for field office employees to deny on the basis of failure to provide information without making any serious effort to contact the claimant.

It’s Hard To Administer SSI

      From a recent report by Social Security’s Office of Inspector General:

… SSA’s SSI financial account validation process for applicants and recipients who alleged having less than $400 in financial accounts did not always lead to accurate SSI determinations. SSA’s process led it to make inaccurate SSI resource determinations for 27 of the 140 recipients reviewed. Based on these determinations, SSA paid the recipients $130,430 in SSI payments they were not eligible to receive. Based on these sample results, we estimate SSA incorrectly made SSI resource determinations that led to 198,960 recipients receiving $718 million in SSI payments for which they were not eligible because applicants/recipients under­reported their financial account balances by $100 or more. 

SSA’s policy did not require that it validate the recipients’ financial account balances because they alleged they had less than $400 in liquid resources. The AFI reports we requested identified 102 of 140 applicants/recipients under-reported their financial account balances by $100 or more. Additionally, the AFI reports showed 28 of the applicants/recipients owned financial accounts of which SSA was unaware. Based on these sample results, we estimate 800,140 applicants/recipients under-reported their financial account balances by $100 or more, with 219,640 failing to report all the financial accounts they owned to SSA. ..

 

Rep Payee Problems

     From a recent report by Social Security’s Office of Inspector General:

SSA did not ensure employees made complete and accurate capability determinations for disabled beneficiaries who previously had payees.  We estimate, for approximately 19,000 disabled beneficiaries who previously had payees, there was no evidence SSA determined the beneficiaries were capable of managing, or directing the management of, their benefits.  In addition, there was no evidence SSA performed proper follow-up development.  SSA paid approximately $887 million in benefits directly to these beneficiaries without evidence to show they were capable of, or were, using the benefits to meet their basic needs.  Finally, we estimate SSA did not properly document capability determinations for approximately 6,700 beneficiaries. 

This occurred because SSA did not have effective system controls to ensure employees properly documented their capability determinations.  Additionally, SSA did not have controls to ensure employees authorized direct payments to incapable beneficiaries in accordance with SSA’s policy and properly performed the follow-up reviews after they made interim direct payment to those 

A 73% Error Rate?

    From Follow-up on the Accuracy of the Social Security Administration’s Manual Billing Process to Collect Medicare Premiums, a report by Social Security's Office of Inspector General:

...  When a Medicare Part B recipient receives a monthly Social Security benefit, SSA deducts the monthly Medicare premium from the benefit. Some individuals’ monthly Social Security benefit is lower than the monthly Medicare premium. SSA bills these individuals for the balance of the premium. Each year, SSA uses its Benefit Rate Increase (BRI) program to generate an alert for individuals whose Social Security benefit is lower than their monthly Medicare premium. SSA considers this alert to be a high priority because SSA must bill beneficiaries the remaining Medicare premium as soon as possible and timely issue benefit statements.

In a 2016 report on the Accuracy of the Social Security Administration’s Manual Billing Process to Collect Medicare Premiums, we concluded SSA incorrectly calculated the Medicare premium owed for 48 percent of the sampled beneficiaries we reviewed. For this audit, we reviewed a random sample of 100 beneficiaries from a population of 111,976 beneficiaries SSA’s BRI program identified in November 2022.

SSA timely processed the cases in our sample; however, it incorrectly processed or calculated the Medicare premium owed for 73 of the 100 beneficiaries we reviewed where their Medicare premium was higher than their SSA benefit payments. This included approximately $147,000 in processing errors where employees did not correctly credit monthly benefits to Medicare billing information and $102,000 in payment errors where beneficiaries were not correctly billed for Medicare premiums, which led to overages and arrearages.

Based on the results for our sample, we project SSA did not correctly update records or calculate Medicare premiums owed for approximately 82,000 beneficiaries, which resulted in approximately $76 million in billing errors and $91 million in processing errors. ...

Report On Priority Cases

    From The Social Security Administration’s Processing of Priority Cases, a report by Social Security's Office of Inspector General (OIG):

... Each year, SSA identifies at least 200,000 (10 percent) of the initial disability applications as priority cases. SSA identifies priority cases for expedited processing through a combination of automated and manual means. Policy requires that SSA develop and process cases identified as priority expeditiously.

We reviewed 668,352 claimants whose initial disability applications SSA selected for priority processing. We also reviewed 153,964 claimants who had initial disability applications that may have been eligible for priority processing.

Generally, SSA properly identified, expedited, and processed initial disability applications that qualified as a priority case. SSA’s selection of cases for priority processing was proper for over 96.1 percent of claimants we reviewed.

However, SSA did not expeditiously develop and process initial disability applications for 11 (6.1 percent) of 180 sampled claimants. Delays occurred because SSA did not always monitor the processing of the 11 cases that were selected for priority processing to ensure they were processed expeditiously. Further, SSA’s policy does not specify overall processing timeframes and/or goals for priority cases. As a result, we estimate SSA delayed case development and processing for 40,844 claimants with priority cases.

I Thought That Job One At The Payment Centers Was To Pay People

     From Reducing Processing Centers’ Pending Actions, a report by Social Security's Office of Inspector General (OIG):

...  SSA met its annual PC [Payment Center] pending actions performance measure goal in 4 of the 6 FYs [Fiscal Years] between FY 2018 through 2023. SSA reported it did not meet its goals in the remaining 2 FYs because of unexpected staff reductions, increased workloads, and less than expected overtime funding it would have used to pay employees to process more PC pending actions.

Although SSA achieved its PC pending actions performance goals in 4 of the last 6 FYs, there was no overall reduction in PC pending actions over those 6 years. In fact, the PC pending actions backlog increased from 3.2 million in FY 2018 to 4.6 million in FY 2023. As the backlog grows, many PC pending actions remain unresolved for long periods of time. From a sample of 139 pending actions, 102 (73 percent) were pending for 300 days or more, with 60 of the 102 pending for 500 days or more.

Delays in processing PC pending actions can lead to higher improper payments, which increased some beneficiaries’ financial burden as they waited longer for underpayments or were charged with increased overpayment amounts. If SSA resolved the PC pending actions we reviewed at the earliest possible instance, we estimate it would have determined approximately 528,000 beneficiaries were improperly paid approximately $534 million. After the pending actions were not processed for 12 months, the improper payment amount for those same beneficiaries rose to approximately $756 million. By the time of our review, many of the PC pending actions had been pending for longer than 12 months, and the improper payment amount had increased to approximately $1.1 billion.

    Notice that it didn't seem to bother OIG that the payment centers were frequently failing to pay benefits for long periods of time. The only thing they seemed to have been concerned about was an increase in overpayments. This in a nutshell is the OIG outlook on the Social Security world -- It doesn't matter when or if you pay claimants what they're due; the only thing that counts is that you not overpay them. That outlook is a prescription for disaster for claimants.

Belated OIG Report On Field Office Waiting Times

Photo taken on May 7 outside Raleigh FO

    From Customer Wait Times in the Social Security Administration’s Field Offices and Card Centers, a report by Social Security's Office of Inspector General:

... In February and March 2023, we visited 76 FOs [Field Offices] and SSCCs [Social Security Card Centers]. During these visits, we observed customer wait times and interviewed office management regarding best practices for reducing customer wait times. We also interviewed SSA staff to determine Agency initiatives to enhance the customer experience and reduce wait times.

During our office visits, we observed the wait times of customers throughout their visit for services. For the 76 FOs and SSCCs we visited, we observed the average wait time for customers before check-in ranged from 5 to 12 minutes, depending on the method of check-in. The check-in method(s) used varied, based on management’s discretion. Once checked in, customers waited an average of 32 to 45 minutes to receive services. We do not assert, and the reader should not infer, that our observations during our visits represent all SSA offices. ...

    Why did it take more than a year after the data was collected for OIG to issue a report?

Detecting Overpayments Due To Marriage

     From Impact of Undetected Marriages on Social Security Administration Payments, a report by Social Security's Office of Inspector General:

... Marriage can impact a Supplemental Security Income (SSI) or Old-Age, Survivors, and Disability Insurance (OASDI) beneficiary’s payment. We randomly selected 1 of 20 segments from the Supplemental Security Record (SSR) and Master Beneficiary Record (MBR). We identified 3 populations with name changes and sampled 250 individuals ...

SSA did not always take the necessary manual actions to properly update 38 of the 250 payment records for SSI recipients or OASDI beneficiaries when there was a name change because of marriage. Furthermore, SSA had not taken manual action on 11 SSI recipients or OASDI beneficiaries who changed their name via the iSSNRC application. We estimate SSA improperly paid 16,631 SSI recipients and OASDI beneficiaries approximately $240.9 million when there was a name change because of marriage. ...

When a person changes their name, SSA systems do not automatically determine whether they are receiving benefits. SSA does not know about a marriage until an individual reports it. ...

SSA explored the feasibility of using electronic marriage data to determine if OASDI beneficiaries changed their marital status. However, not all states/jurisdictions have a central repository of electronic marriage data, and many do not require, or collect, the marriage applicants’ Social Security number. ...

    There's a lot that Social Security could do to prevent overpayments if it had an adequate workforce. A different Inspector General might want to issue a report detailing how much money is being wasted due to inadequate staffing at the Social Security Administration.

    By the way, there's also the question of whether marriage should have as many effects as it does on Social Security and SSI benefits but that's another topic.

Problems With Retirement Earnings Test

     From The Social Security Administration’s Enforcement of the Earnings Test, a report by Social Security's Office of Inspector General:

The earnings test is a provision of the Social Security Act that requires that SSA withhold payments from beneficiaries who are under full retirement age (FRA) if their earnings exceed a certain limit. ...

SSA did not accurately or timely pay beneficiaries subject to the earnings test. As a result, we estimate SSA:

• inaccurately calculated approximately 47,000 of the 294,000 earnings-test overpayments established in FY 2021, totaling more than $148 million;
• inaccurately paid approximately 9,000 beneficiaries approximately $29 million based on estimated earnings that were more or less than their actual earnings; and
• did not timely pay approximately 176,000 beneficiaries approximately $81 million in monthly benefit increases. ...

    And to think that the retirement earnings test used to apply to all retirees regardless of age.

Improvement In Mail Processing But Some Problems Persist

    From Follow-up: The Social Security Administration’s Implementation of Mail Procedures, a report by Social Security's Office of Inspector General (footnotes omitted):

...  Mail processing at SSA offices is primarily a manual workload. This requires that managers and employees open each mail item, scan the program-related documents into a workload management system, and assign to staff. SSA’s regional offices are responsible for monitoring the status of mail handling in their region ...

In September 2021, SSA issued a Mail Handling BPD that focused on 10 key issues significant to mail processing. ...

We judgmentally selected and visited 87 SSA offices throughout the continental United States. At each office, we interviewed management and observed the mail handling process. ...

SSA offices had improved mail processing since our July 2021 review. SSA offices we visited generally complied with the requirements in the Agency’s Mail Handling BPD.While most offices were meeting the requirements of 6 of the 10 BPD key issues, some offices did not always meet the requirements ...

Of the 87 offices visited, the following offices complied with the timeliness metrics established:

• 86 (98.9 percent) received mail directly from mail carriers,

• 84  (96.6 percent) opened mail within 1 business day of receipt,

• 81 (93.1 percent) processed all Social Security number applications within 5 business days of receipt, and

• 79 (90.8 percent) returned all original primary evidence documents to customers within 3 business days  of receipt and kept a log of the returned documents.  ...

• 73 (83.9 percent) processed their undeliverable returned mail within 5 business days of receipt, as required. However, the remaining 14 offices had undeliverable returned mail over 30-days-old. Moreover, three of these offices had mail dated back to January 2022, and one with mail dated back to September 2019. ... 
  

• 67 (77 percent) complied with the requirements for remittances and returned unendorsed Treasury checks. Specifically, the offices processed to completion all remittances and returned unendorsed Treasury checks within 1 business day of receipt, 16 recorded receipt in the office’s remittance log, and secured 17 checks that could not be processed to completion the same day. The remaining 20 offices had the following issues:
  

• 17 did not process remittances within the required timeframe. Of these, 15 offices had checks that remained unprocessed 3 to 31 days after the 1-business-day requirement. Additionally, two offices had unprocessed checks that dated back to July 2020 and November 2021. The banks could return these checks as non-negotiable since they were not endorsed within 6 months, as required by banks.,
  

• 5 did not use remittance logs, as required. For example, 1 office had 77 unprocessed checks that were unsecured and accessible to office staff. In fact, 1 office had 58 checks dating back to October 2022 with no record they existed. The new remittance manager was unaware these unprocessed checks were locked in a safe because they were not recorded in the office’s remittance log.
  

• 2 did not properly secure the unprocessed checks. For example, 1 office had 77 unprocessed checks that were unsecured and accessible to office staff.

• 36 (41.4 percent) scanned and profiled mail within 5 business days of receipt, as required. The remaining 51 offices, on average, scanned and profiled mail within 10 business days of receipt. ...