melissa mcinerney workplace safety and the economic impact of workers’ compensation insurance
TRANSCRIPT
MELISSA MCINERNEY
Workplace Safety and the Economic Impact of Workers’
Compensation Insurance
Q1: Protecting Injured Workers vs. Encouraging Economic Development
2
Previous literature found workers very responsive to benefit levels
Increased WC receipt may increase employer costs
States seek to encourage economic development
Concern: employers will choose to locate in states with lower WC costs
MD: Max=$877/week
DC: Max=$1,233/week
VA: Max=$816/week
Propensity to Claim WC May Differ By Earnings Levels
3
Disentangling the Effects of Wages and Benefits on WC Receipt
4
(1) (2) (3)
Basic Model
VARIABLES
Log (Weekly WC benefit) 0.205*** 0.098** 0.028(0.044) (0.046) (0.054)
Benefit elasticity 0.527 0.251 0.0718
Number of Observations 314,839 314,839 314,839
Disentangling the Effects of Wages and Benefits on WC Receipt
Quartic in Earnings
Earnings Spline
Follow the literature: Increase benefits 10%, see 5% more
claims
Allow relationship between earnings and WC
receipt to increase at increasing
rate: Increase benefits 10%, see
2.5% more claims
Allow relationship to vary
by quartile of earnings distributio
n: No statistically significant relationship between increased benefits
and propensity
to claim
Policy Implications
5
Workers much less responsive to benefit levels than previous estimates suggest
Can better protect injured workers (by raising maximum benefit levels) without adversely impacting economic development
Q2: Wasteful Spending and WC
6
Examine specific policy change—privatization of WC in Ohio
7
Date of
injury
1 week
Claim filed with TCM
Receive cash benefits
15 months
Incentivized injuries
excluded from bonus calculation
Predicted Responses After Implementation of Bonus Payment for Private Case Management Companies (Incentivized Injuries)
REDUCE DURATION
TCM insurance cards and
phone calls
REDUCE DURATION INCREASE DURATION
NO CHANGE
Time elapsed
since injury
0
100
200
300
400
500
600
1995 1996 1997 1998 1999 2000 2001 2002
Quarter of I njury
Incentivized Injuries Non- Incentivized Injuries
8
PRE POST
The 99th Percentile in Distribution of Claim Duration Over TimeDays away from work
Observations are weighted by inverse of predicted probability claim has valid return-to-work information.
9
Identifying Assumption: Other than the bonus payment, all changes in duration
during this time period impact treatment and comparison groups in same way
Ways to test identifying assumption:1. Make sure treatment same in period before
bonus payment If groups are treated the same, coefficient estimate on POST1*INCENT
will be zero
2. Examine different comparison groups with more similar groups of injuries
Injuries with same three-digit ICD-9 code Back injuries Bruises Cuts
10
Standard errors are clustered by injury code. Regressions are weighted by the inverse predicted probability a claim has valid return-to-work information. Each regression includes injury, demographic, job, time, and TCM effects. Significance at 10 percent (*), 5 percent (**).
Factors Impacting Days Away from Work (OLS), Medical Only Claims
(1) (2) (3) (4) (5) Sample: All Injuries Common Injury
Sample Back Sprains Bruises
(Contusions) Cuts
POST1 -.325** (.037)
-.281** (.078)
-.462** (.039)
-.296** (.108)
-.229** (.028)
POST1*INCENT -.035 (.043)
-.131 (.086)
-.017 (.067)
-.130 (.082)
-.041** (.016)
POST2 -.286** (.037)
-.255** (.076)
-.282** (.028)
-.366** (.071)
-.236** (.042)
POST2*INCENT -.010 (.038)
-.069 (.065)
-.011 (.031)
-.011 (.049)
-.019** (.005)
N 431,686 116,064 61,860 83,528 118,189 R2 .090 .075 .022 .037 .019 Mean of Dep. Var.
2.525 (1.861)
2.815 (1.994)
3.353 (2.239)
2.542 (1.805)
1.953 (1.411)
p-value .193 .095 .888 .100 .282
In every sample, find case managers reduce duration for minor claims between 1/3 and ½ a day
11
Standard errors are clustered by injury code. Regressions are weighted by the inverse predicted probability a claim has valid return-to-work information. Each regression includes injury, demographic, job, time, and TCM effects. Significance at 10 percent (*), 5 percent (**).
Factors Impacting Days Away from Work (OLS), Cash Benefit Claims
(1) (2) (3) (4) (5) Sample: All Injuries Common Injury
Sample Back Sprains Bruises
(Contusions) Cuts
POST1 -4.461 (8.580)
-8.960 (9.244)
-13.055 (8.874)
-7.289 (11.406)
8.652 (7.024)
POST1*INCENT 4.450 (7.204)
-1.421 (4.797)
1.411 (3.006)
5.030 (7.975)
1.048 (5.183)
POST2 -11.549 (10.140)
-1.591 (18.711)
-3.287 (18.623)
-20.158 (15.602)
2.496 (10.584)
POST2*INCENT 18.296** (5.553)
17.124** (2.795)
19.989** (3.053)
14.484** (5.521)
-.943 (3.975)
N 44,288 11,726 13,645 3,613 2,412 R2 .201 .048 .036 .049 .055 Mean of Dep. Var.
90.409 (189.248)
69.756 (153.251)
78.753 (167.537)
44.487 (100.274)
29.630 (51.302)
p-value .025 .000 .000 .280 .669
Claim duration increases between 2-2.5 weeks for severe claims
This corresponds to 30% increase in probability a claim lasts more than 15 months
12
TCMs reduced duration by about one third of a day for minor claims
Bonus payment generated an increase in duration for severe claims: An increase of about 19 days among cash benefit claims
30 percent in probability spell lasts longer than 15 months
Potential Mechanisms:
Minor claims: reduce duration through infrastructure changes
Severe claims: Increase duration via enrollment in vocational rehabilitation
Findings
13
Overall Effect of Privatized Case Managers (2002)Bonus structured such that maximize profits if extend
duration
Cash benefit claims having an incentivized injury miss 18 more days
~7,000 cash benefit claims having an incentivized injury
cash benefits ~ $59 per day
=> (7,000)*($59)*(18 days) = $7.4m in additional benefits
TCMs used vocational rehabilitation to increase claim duration
265 individuals received voc. rehab. because of bonus
average cost ~ $5,000 per participant
=> (265)*($5,000) = $1.3m in voc. rehab. Costs
Bonus payment may cost state over $8.5m in additional benefits paid
14
Discussion
WC is an important, large social insurance program, and many policy changes were implemented in the 1990s.
Structure of bonus suggests heterogeneous responses by TCM as claim develops over time, including an increase in duration as claims with incentivized injuries approach 15-month threshold.
Evidence is consistent with these anticipated responses; duration falls by 1/3 of a day for minor claims, increases by 20 days for severe claims
Policy Implications
15
Can provide better benefits for injured workers without risking economic development goals
Take great care in contracts write with private companies—may have unintended consequences