Wages are flat because unemployment IS so low

The New York Times had a feature on February 2nd on why wages continue to lag despite the extremely low unemployment rates at the moment.

There is a basic presumption in this article that is no longer valid, however: namely that the unemployment rate today is the same kind of unemployment rate from 10 or 20 or even 30 years ago.

In Wisconsin, the news for some time has been how the state’s unemployment rate and benefit payments to claimants are at record lows. For instance, a January 2018 press release from the Department includes the following observations:

Other indicators of the state of Wisconsin’s economy include:

• Initial UI claims ended 2017 at their lowest level in the last 30 years.
• Continuing unemployment claims ended 2017 at their lowest level since 1973.

See also this October 2017 press release. What is notable here is that this decline is well known and part of an apparent plan.

As previously noted here, this decline is occurring because of Department efforts at making it harder for the unemployed to qualify for unemployment benefits and then disqualifying them for not jumping through some state requirement fast enough or alleging unemployment fraud for nothing more than simple claim-filing mistakes.

But, the data in Wisconsin does not explain what is happening nationally. The National Employment Law Project has already noted how unemployment has changed significantly across the nation the last few years. But, thanks to the efforts of some smart folks in Pennsylvania, national unemployment data is now available in a highly convenient format and which produces eye-catching charts.

NOTE: I cannot say enough good things about this unemployment data explorer. Pretty much any unemployment data currently being collected is now available for quick analysis in a chart. Moreover, you can easily see and download the data being used to create the charts. Excellent work.

This data provides some charts that can compare what is happening from state to state. For instance, separation data (how claims are denied because of a quit or a discharge for which misconduct/substantial fault is found) presents the following set of charts:

The red line in these charts is the national rate. As obvious, this chart shows that there is a great deal of variation from state to state. And, because there have been big changes in the number of claims being filed, this data is somewhat incomplete. See “Employer UI taxes declining because more UI claims being denied” (24 August 2016) for an examination of how changes in the number of claims being filed affect Wisconsin’s claim-filing numbers.

But, the variation among the states also reveals some obvious increases in denial rates in Kansas, Maryland, New Mexico, South Carolina, and Wisconsin. On the other hand, Alabama, Colorado, Florida (a surprise), Mississippi, New Hampshire, New Jersey, and New York show significant drops in denial rates.

This 50-state data gets even more interesting when “other” reasons for denying unemployment benefits are examined:

NOTE: “Other” reasons include, for instance, a claimant not being able and available, not completing the required job search actions for that state, not attending call-in or meeting requirements a state has mandated, or not registering for various state-mandated services. Wisconsin specific data on these issues “other” denial reasons is available here from this prior post on the financial impact of Wisconsin’s substantial fault disqualification. Outside of able and available status, these “other” reasons generally encompass requirements an individual state creates as part of its claim-management bureaucracy for supervising the unemployed.

Again, the red line in this chart is a national average of cases being decided for “other” reasons. As evident here, there has been an increase (and even an explosion in some states) in “other” denial reasons the last few years in Florida, Louisiana, Massachusetts, Michigan, Mississippi, Montana, New Mexico, New York, Ohio, Oklahoma, Pennsylvania, South Carolina, and Tennessee. Indeed, the national trend of these “other” cases increasing over time would probably be much more significant if the extremely large populations of Texas and California were removed from the analysis (Texas has been flat, and California has actually declined significantly). As such, national data is masking significant changes in the availability of unemployment benefits in numerous states.

Of course, this table is simply showing the number of cases being decided for “other” reasons. If all of these cases did NOT lead to a denial of an unemployment claim, then there is essentially no harm, no foul in these cases. But, the actual denial rates for “other” reasons reveal a not-so innocent story.

Where California declined and Texas increased slightly and then plateaued, Delaware, Florida, Iowa, Illinois, Indiana, Louisiana, Massachusetts, Maryland, Michigan, Mississippi, Montana, North Carolina, North Dakota, New Mexico, New York, Ohio, Oklahoma, Pennsylvania, South Carolina (showing shocking jumps from year to year), Tennessee, Utah, Wisconsin, and Wyoming have seen obvious increases in denial rates since 2005.

These other reasons for the most part did not exist until very recently and almost none go back a decade in significant numbers in any one state. The Wisconsin data on this issue, for instance, is telling: cases involving profiling registration requirements were in single or low double-digits until 2015 when they sky-rocketed to hundreds and then thousands. What this last chart reveals is that numerous states have essentially created numerous mechanisms for disqualifying claimants even when those claimants are initially eligible for unemployment benefits.

Given that we are all human and can only take so much abuse before moving on, it is extremely likely that most folks have simply stopped filing unemployment claims because of the obstacles states have placed on their eligibility and not because they have found the jobs they have wanted all along.

So, if states are making it much, much harder to receive unemployment benefits when filing a claim, then the low unemployment rates of today are NOT comparable to the low unemployment rates of yesteryear or even to unemployment rates of a decade or so ago. Instead of creating a question about how low wages and low unemployment rates can co-exist, the low unemployment rates of today may actually be placing a brake on wage growth: the state unemployment policies at issue here increase the supply of individuals looking for any work in place of their missing unemployment benefits. That increase in the labor supply, as a result, creates downward pressure on wages. At least, that is what I learned in labor economics 101.

Advertisements

The actual financial impact of substantial fault

Back in April 2016, I described the confusion about the two versions of the Department’s substantial fault proposals and calculated the financial impact of substantial fault based on that estimate.

But, there is actual data available for determining the financial impact of substantial fault. Wisconsin reports its handling of unemployment claims to the Employment & Training Administration of the United State Department of Labor. This federal agency then makes this data available to the public, and quarterly numbers regarding the number and outcome of non-monetary determinations is available via the ETA 207 series.

NOTE: Non-monetary determinations are those determinations that do NOT involve calculations to determine eligibility based on prior earnings or other kinds of monetary calculations. The data for non-monetary determinations includes determinations regarding discharges, voluntary leaving (i.e., quitting), and determinations regarding claimants’ able and available status, refusals of suitable work, adequate job search efforts, and other eligibility status issues. There is both a short description and a long description of this data.

Accordingly, this data can indicate specifically the kind of impact the substantial fault disqualification standard has on unemployment claims in the state of Wisconsin.

NOTE: The misconduct label for this data is used nationally because historically misconduct was the only disqualification standard used in discharge cases. But, starting in 2014, the misconduct data here for Wisconsin includes both misconduct and substantial fault determinations.

The substantial fault disqualification began to be applied by the Department in initial determinations issued on or after 5 January 2014. See 2013 Wis. Act 20 § 9351(1q) (new misconduct and substantial fault provisions “first apply with respect to determinations issued under section 108.09 of the statutes on January 5, 2014”).

Until the first quarter of 2014, the Department denied on average about 26% of all claimants who were discharged from their jobs. From the first quarter of 2014 until the latest available (the quarter ending June 2016), however, the number of discharge cases being denied jumped to 38.47% of all discharge determinations. This increase nearly doubled the number of denials from before 2014 — a stunning and remarkable jump in the number of claims being denied.

NOTE: The actual data for creating these charts is set forth in a table, WI Separation Data, compiled from the ETA 207 data.

This jump is even more shocking in light of the decline in discharge determinations since the start of 2014.

From 2007 to the end of 2013, the number of discharge determinations averaged 19,462.43 per quarter. Not surprisingly, during the height of the last recession in 2009 and 2010, there were discharge determinations in some quarters that numbered over 21,000 or even 22,000. See Table: WI Separation Data. But, in general the number of discharge determinations per quarter hovered around 17,000 to 19,000. In the first quarter of 2014, however, the number of discharge determinations plummeted to under 14,000. And, the number of discharge determinations has continued to decline since then. From the start of 2014 to June 2016, the Department has issued on average only 12,605.50 discharge determinations per quarter.

NOTE: The total number of determinations being issued by the Department has not declined, however. Prior to 2014, the number of determinations issued per quarter averaged 58,945.25. From 2014 on, the average number of determinations being issued increased to 59,668.60 per quarter. As indicated in the table for WI Non-Separation Data, the number of determinations not connected to separation issues being issued jumped from 46.87% of all determinations per quarter prior to 2014 to 64.01% after 2014. In particular, much if not all of this increase in non-separation determinations concerns an approximately 26% increase in determinations regarding a claimant’s able and available status, a five-fold increase in determinations (from just over 3,000 determinations prior to 2014 to almost 16,000 determinations on average after the start of 2014) over a claimant’s failure to follow the Department’s reporting requirements, and a nearly 100-fold increase in determinations (around 13 cases per quarter prior to 2014 to nearly 1,200 per quarter after 2014) over a claimant’s failure to follow the Department’s job profiling services. In all three of these categories, the percentage of benefit denials has also jumped at least 10 percentage points on average after 2014.

It should also be noted that these non-separation denials generally do not disqualify a claimant for an extended period of time. For instance, a denial of benefits because of failing to report to Department-mandated profiling services or provide requested information is usually cured by reporting for those services or providing the needed information. As a result, the disqualifications from receiving unemployment benefits pursuant to these denials are generally short-term denials. A denial of benefits because of substantial fault or misconduct, on the other hand, lasts 7 weeks at a minimum and requires new earnings of 14X a claimant’s weekly benefit rate in order to re-qualify for unemployment benefits.

This decline in discharge determinations, however, does not indicate that the impact of substantial fault should be discounted in some way. Quarterly reports on each state’s unemployment system from the Employment & Training Administration indicate both the average weekly benefit rate for claimants during the previous twelve months and the average number of weeks unemployment benefits are being received during the last twelve months. The report for Wisconsin for the first quarter of 2015 indicates an average weekly benefit rate of $288.04 for the previous twelve months and an average duration for benefits of 14.8 weeks, leading to $4,262.99 in unemployment benefits at issue. Applying the pre-2014 25.99% denial ratio to the post-2014 12,605.50 discharge determinations that take place on average in each quarter means only 3,276.17 cases would be denied rather than the 4,852.00 being denied with substantial fault in place — a difference of 1,575.83 cases. Multiplying this number of cases by the $4,262.99 of unemployment benefits at issue leads to an amount of $6,717,747.53 per quarter being denied claimants currently under this new substantial fault standard. As substantial fault has now been in effect for ten quarters, the amount of unemployment benefits “saved,” or not paid to claimants, amounts to $67,177,475.32.

It is expected that substantial fault will also, on the whole, lead to employees filing fewer claims because claimants will learn how broad the substantial fault disqualification is and stop filing claims altogether. The data supports this trend. In the second quarter report in 2016, the weekly benefit rate for the last twelve months is $306.43, and the average duration of benefits for the previous year is 13.3 weeks. With these figures, the amount of benefits at issue is $4,075.52. Multiplying this amount by the 1,575.83 average number of cases per quarter denying unemployment benefits to claimants because of substantial fault leads to an amount of $6,422,326.68 per quarter being denied to claimants and a ten quarter amount of $64,223,266.82. As a result, the range of lost benefits because of substantial fault is between $67 and $64 million.

NOTE: The Department’s original estimate of $19.2 million per year, after 2.5 years, amounts to $48.4 million — approximately $15-$20 million less than what the actual data reveal.

So, even as fewer and fewer discharged employees are filing claims for unemployment benefits, the new substantial fault standard that become effective in 2014 is leading to thousands of claimants being denied millions in unemployment benefits.

Employer UI taxes declining because more UI claims being denied

Wisconsin employers are having their unemployment tax rates slashed in 2017 because the fund from which unemployment benefits is reaching ever higher solvency metrics. The Walker administration is heralding this news here and here.

Understandably, there are two possible explanations for what is going on with the state’s unemployment fund. The state’s unemployment funds are positive because either job growth is booming or because fewer folks are claiming benefits despite NOT having jobs.

Is job growth booming in Wisconsin?

The July state jobs report reveals that job growth in Wisconsin continues to be anemic. This report indicates that, initially, in July 2016 5,000 private-sector jobs were added to Wisconsin payrolls. But, June 2016 numbers for private-sector job growth were revised downward, from 10,900 to 5,600. This loss of 5,300 jobs from the June report means that the initial number for July does not even get the state back to what was first reported for June 2016.

Neither does the quarterly data offer any better news. From March 2015 to March 2016, the quarterly data indicates that the state added 37,432 jobs during that time frame. But, this number is a few thousand less than what was reported for the March 2015 to March 2015 time frame in the July 2015 jobs report: 39,652 private-sector jobs.

So, without adding new jobs to the state’s economy, the decline in unemployment claims must be coming from fewer folks claiming unemployment benefits. In two bullet points, the July 2016 jobs report actually acknowledges this development.

• Year 2016 initial UI claims are running at their lowest level since 1989.
• Continuing unemployment claims in Wisconsin are running the lowest in at least the past 30 years.

But, the question remains: if jobs are not being created, why are claims now so low?

Why are unemployment claims so low?

Actual claims data is available from ETA 207, Non-monetary Determinations Activities Report. See DOLETA data downloads generally for UI data. The 207 data series has all determinations issued by a state compiled on a quarterly basis going back several decades until the most recently completed quarter, June 2016.

Here are some charts from that data for Wisconsin starting in the first quarter of 2007 through the second quarter of 2016.

This chart shows that most initial determinations issued by the Department lead to the denial of unemployment benefits. But, starting in the first quarter of 2014, the denial rate for initial determination jumped markedly. Prior to 2014, 59.90% of all initial determinations denied benefits to claimants. Since the start of 2014, 77.45% of all initial determinations issued by the Department have been to deny unemployment benefits. In other words, currently only one of four initial determinations being issued by the Department allows unemployment benefits, and three out of four initial determinations deny unemployment benefits in some way.

Keep in mind that these numbers are based on the initial determinations issued by the Department in regards to a new unemployment claim. In most states, these determinations would consist almost entirely of separation determinations — whether claimants are disqualified because their discharge was their fault in some way or they lacked good cause for quitting their jobs. In Wisconsin, these separation decisions are only a part of what the Department decides. And, increasingly separation decisions are becoming a smaller and smaller part of what the Department does in disqualifying claimants.

Here, initial determination concerning separation issues (i.e., quits and discharges) were around 60% of all initial determinations until 2009, when they declined and hovered around 50% of all initial determinations until the first quarter of 2014. At that point, the percentage of separation initial determinations being issued by the Department plummeted to 40% of all initial determinations. In the last two quarters of 2015, the number of separation initial determinations fell again to under 30% of all initial determinations. So at present, less than 30% of the initial determinations being issued by the Department concern separation issues related to a discharge or a quit. And, since most of these other determinations (and probably all of them given the analysis below) are denying unemployment benefits, many of these probably include some kind of concealment allegation, given the Department’s push to allege concealment against claimants.

In regards to denying claimants unemployment benefits, the Department consistently denied about 26% of all claimants who were discharged from their jobs until the first quarter of 2014.

From the first quarter of 2014 until the latest, however, the number of discharge cases being denied jumped to 38.47% of all discharge determinations. This increase nearly doubled the number of denials from before 2014 — a stunning and remarkable jump in the number of claims being denied.

The magnitude of this jump is seen when it is compared the number of quit denials over this same time frame.

Here, a slight increase in denials occurs in the first quarter of 2014. But, this increase is part of a general increase in denial rates that appears to have started in the second half of 2010. So, while denial rates for those quitting their jobs are high and gradually increasing, there is no sudden or striking shift in denial rates in quit cases at any one point in time.

Now, consider that in the last two years only about 30% of all initial determinations concern separation issues and that only 1 out of 4 initial determinations is allowing unemployment benefits at all. In this light, it appears that the only initial determinations right now allowing benefits are the discharge and quit separation determinations that are NOT denying benefits. Everything else the Department is doing is to deny unemployment benefits to claimants.

What these numbers reveal is that most folks applying for unemployment benefits are being denied those benefits, that essentially the only folks qualifying for unemployment benefits are those laid off from their jobs by their employers, and that numerous denials of unemployment benefits have nothing to do with separation issues. These non-separation initial determinations most likely are part of the Department’s program integrity efforts and most likely lead to charges of unemployment concealment, especially under the Department’s new strict liability standard for concealment.

So, unemployment claims and benefits are at record lows in the state because the state is making it difficult to impossible for claimants to receive benefits and charging the few that collect unemployment benefits with unemployment concealment. Essentially, employers are paying unemployment taxes for a benefit almost no one is using. Pretty soon, folks will start calling for eliminating the unemployment system entirely, as who wants to pay a tax that does nothing.

UPDATE (14 Sept. 2016): Fixed links so that a click on a chart brings up a full-sized version.

The financial impact of substantial fault

A document available on this blog is cited by the Appeals Court in Operton v. LIRC, namely the original Department proposal for substantial fault — D12-01.

The appeals court observes at n.5 on p.6 of its decision that this document does not quite match the version of D12-01 supplied by the Commission in its briefing. Even though both versions are dated 24 October 2012, the copy produced by the Commission has an actual number for the fiscal impact of the proposed addition of substantial fault and the changes to misconduct — $19.2 million per year. The original D12-01 document introduced at the 27 November 2012 Advisory Council meeting only stated that the fiscal impact was yet to be determined. From my records of the Advisory Council meetings, it appears that the Department made this revision to D12-01 at the 21 February 2013 council meeting.

Obviously, the Department added this fiscal impact information without otherwise noting this change. Certainly, this number reveals a staggering impact on Wisconsin claimants when UI data from 2013 is considered. In the fourth quarter of 2013, the average weekly benefit claimants received in Wisconsin that year was $276.14, and those unemployment benefits lasted 15.9 weeks on average (see p. 64 of the data report). Multiplying these numbers together leads to a total benefit amount received of $4,390.63. Divide this number into the proposed $19.2 million fiscal impact from substantial fault, and 4,510 claimants end up being disqualified under these changes in unemployment. Each year.

Concealment charges and collections

As noted in February of this year, concealment cases jumped to 2.79% of all unemployment benefits paid out. For the three previous years, concealment cases had always been under 2% of all unemployment benefits paid.

NOTE: a 0.8% increase is remarkable and unprecedented. If such a change in the unemployment rate or GDP took place from one year to the next, that change would receive headlines across the state.

Because 2015 collection data was not available at the time of the February post, I estimated how much concealment money the Department was taking in for 2014 when compared to previous years. That estimate indicated that concealment monies being collected from claimants was rising markedly.

Now, in 2016, the Department’s updated fraud report is available. This report, furthermore, has data for both how much concealment is being assessed and how much concealment monies are being collected.

In regards to the concealment being assessed, there is good news and bad news.

As seen in this chart, over-payments assessed in 2015 declined to 2.21% of unemployment benefits paid out. Still, the concealment charges are over 2% of benefits, so concealment cases remain abnormally high relative to 2011, 2012, and 2013, when the benefits at issue were two to three times greater than the benefits paid out in 2015.

The story with the Department’s collection efforts, on the other hand, is pretty much bad news (except for the impact on Department coffers arising from these collection efforts).

Here, the concealment over-payments collected in 2011 were under 1% of the benefits paid out that year. In 2012 and 2013, the concealment over-payments collected were under 2% of the benefits paid out. In 2014, however, the concealment over-payments jumped to nearly 3%, and in 2015 concealment over-payment collections climbed to 3.42% of the total benefits paid out that year. In other words, in 2015, out of every $100 claimants received, the Department took back $3.42 in concealment over-payments.

These collection numbers also reveal how concealment collection is a growing part of the Department’s collection efforts. Whereas assessments in 2015 show a slight decline in concealment/fraud assessments to non-fraud assessments — 121.10% in 2014 to 112.69% in 2015 — collections for fraudulent over-payments relative to non-fraud over-payments continued to increase in 2015 — going from 116.52% in 2014 to 140.11% in 2015. In other words, nearly two out of every three over-payment dollars collected in 2015 were for concealment. The Department is becoming VERY good at debt collection and, essentially through concealment, discounting the benefits being paid out to claimants. The problem is that claimants who apply for unemployment benefits are not aware of this “discount” on their benefits until charged with concealment six to twelve months after first filing their unemployment claims.

So, what I wrote in February 2016 remains valid, especially since the definition of concealment has now been changed formally to make claimants liable for any mistakes they make on their weekly claim certifications:

These numbers show a sudden increase in 2014 in concealment cases and this increase accelerated in 2015. In this light, the Department’s push to change the definition of concealment is part of an agenda to expand the scope and reach of concealment. The Department countered in its testimony before the committee that an intent to conceal is still required under its proposed changes to the definition of concealment. The proposed language, numerous posts on this blog, a Commission memorandum, and Kevin Magee’s testimony at the public hearing belie the Department’s assertions. Mistakes are increasingly being charged as concealment by the Department, and Commission review applying the actual concealment standard is the only way to fight these kind of charges.

Essentially, concealment is becoming the modus operandi of the Department’s efforts in administering the state’s unemployment law. Anyone who makes a mistake is at risk of a concealment charge from the Department, and the Department wants to change unemployment law to reflect this practice.

The declining market for unemployment benefits

Claire McKenna and Rick McHugh of NELP describe how unemployment benefits continue to be artificially low across the nation in 2015. Their key finding:

Using the latest data, we find that the recipiency rate in 2015 remained at a record low, with just over one in four jobless workers (27 percent) receiving UI benefits in 2015.

Their measure for a recipiency rate actually shows Wisconsin as above-average in the nation at 36%. As noted in their discussion of their methodology, there are several ways to measure recipiency rates. Their measure, for instance, does not account for penalty weeks (such as those where over-payments are being recovered because of concealment). Given the push for alleging concealment in Wisconsin, this 36% recipiency rate in Wisconsin is probably too generous.

Filing for unemployment? It’s a trap!

Thanks to an information request from one of the members of the Advisory Council, concealment data for the 2014 calendar year is now available. Of 21,694 initial determinations that led to appeal tribunal decisions in 2014, fully 11,040 were initial determinations that found claimant concealment. That is, nearly 51% of the initial determinations in 2014 concerned (and found) claimant concealment.

Of these 11,040 initial determinations, however, only 470, or 4.25% of the total, were appealed. Appeal tribunals overturned 216 or 46% of these 470 concealment appeals and affirmed 254 of these concealment cases. In 2014, the Labor and Industry Review Commission heard 196 concealment appeals and only affirmed 34 appeal tribunal decisions. The Commission overturned nearly 63% (123 cases) of the appeal tribunals that found concealment, and the Commission remanded 20% (23) of the 2014 concealment cases that reached it for additional evidence. That is, only 92 (34 affirmed by LIRC and 58 never appealed to LIRC) concealment decisions out of 470 appeals — i.e., 20% of the concealment appeals — were actually confirmed as concealment after review of some kind. So, while very few concealment cases are appealed, those that are appealed are usually overturned either by the appeal tribunal or the Commission.

And, given what has happened in the concealment cases the Commission has overturned — see, e.g., O’Neill v. Riteway Bus Service, Inc., UI Hearing Nos. 15600518MW and 15600519MW (28 May 2015) (“ALJ placed the burden of proving concealment on the wrong party. The ALJ stated that it was the employee’s burden to prove that there was no concealment. This is incorrect. As the commission and the department have stated for decades, the burden to establish that a claimant concealed information is on the department.“) (emphasis in original) and Dabo v. Personalized Plus Home Health, UI Hearing Nos. 14609522MW and 14609523MW (16 April 2015) (“The employee, as a non-native English speaker, missed the ‘did you work’ part of the multi-part question. It is a common mistake, one long acknowledged by the department.”) — it seems that many of the cases that are alleging concealment do not contain actual concealment.

Since less than five percent of concealment determinations are ever appealed, however, the Department has had a relatively unchecked hand in charging claimants with concealment. Unemployment claims, then, have essentially become a vehicle for alleging concealment against claimants. As they say in a galaxy far, far away: