Table 2.

Effect of insurance expansions on total, coverage‐specific, and treatment setting‐specific admissions rate per 100 000 adolescents: Treatment Episode Data Set (1996‐2017)

Outcome:	Overall	Coverage type			Treatment setting
	Total	Private	Public	No insurance	Residential	Intensive outpatient	Nonintensive outpatient
Panel A: All admissions
Sample mean	478.1	113.8	212.6	151.7	77.0	69.3	331.8
Full parity law	124 (18, 231)**	57 (2, 112)**	91 (‒68, 250)	‒23 (‒174, 127)	‒1 (‒46, 45)	26 (‒17, 68)	99 (4, 195)**
Public insurance income eligibility threshold (/100)	60 (‒56, 176)	10 (‒32, 52)	55 (‒20, 130)	‒5 (‒75, 66)	4 (‒17, 24)	40 (7, 74)**	16 (‒81, 113)
Panel B: Noncriminal justice system admissions
Sample mean	247.2	68.5	109.9	68.8	42.1	32.3	172.8
Full parity law	99 (30, 168)***	40 (5, 75)**	55 (‒22, 133)	4 (‒72, 80)	6 (‒15, 27)	17 (1, 33)**	76 (18, 133)**
Public insurance income eligibility threshold (/100)	57 (‒22, 135)	15 (‒18, 48)	45 (‒2, 92)*	‒3 (‒36, 29)	6 (‒7, 18)	18 (3, 34)**	33 (‒34, 99)
Panel C: Criminal justice system admissions
Sample mean	230.9	45.3	102.7	82.9	34.9	37.1	158.9
Full parity law	25 (‒41, 92)	17 (‒5, 39)	36 (‒53, 125)	‒27 (‒111, 56)	‒6 (‒34, 21)	8 (‒23, 40)	23 (‒33, 80)
Public insurance income eligibility threshold (/100)	3 (‒54, 60)	‒5 (‒24, 13)	10 (‒30, 50)	‒1 (‒44, 41)	‒2 (‒11, 8)	22 (0, 44)**	‒17 (‒63, 29)

The unit of observation is a state‐year. The number of observations included in each regression model is 779. Data are unweighted. All models are estimated with least squares and control for state characteristics, state fixed effects, and year fixed effects. 95% confidence intervals that account for within‐state clustering are reported in parentheses.

^***, **, and *statistically different from zero at the 1%, 5%, and 10% level.