For both our main and secondary outcomes, we used a regular analysis that is difference-in-differences of results that covered approximately twenty-four months before and twenty-four months following the 2011вЂ“2012 Ca Medicaid expansions.
As noted above, we compared 43 Ca very early expansion counties to 924 nonexpansion counties (like the 4 mentioned before nonexpansion California counties) within the nationwide data set, with standard mistakes clustered in the county degree. We stratified our findings by the chronilogical age of the borrowerвЂ”focusing on individuals younger than age sixty-five, who does have been almost certainly become impacted by Medicaid expansion. Being a sensitiveness test (see Appendix display A7), 16 we examined borrowers over the age of age sixty-five and utilized a triple-differences approach during the level that is county-month-age.
To exclude preexisting that is systemic trends which could have undermined our difference-in-differences approach, we estimated an вЂњevent studyвЂќ regression associated with aftereffect of Medicaid expansion in the wide range of loans. This tested the legitimacy of y our presumption that payday borrowing will have had comparable styles in expansion and nonexpansion counties if none regarding the counties had expanded Medicaid. The regression included a set impact for each and every county, a hard and fast impact for each month, and indicators for four six-month periods before Medicaid expansion and three six-month periods after expansion (see Appendix Exhibit A8). 16
Our research had not been in a position to straight connect specific insurance coverage status to payday borrowing; to the knowledge, the info to do so try not to exist. Also, although we discovered no proof this, we’re able to maybe not rule the possibility out that state- or county-level alterations in the legislation (or enforcement of laws) of payday advances or any other industry modifications may have took place California within the duration 2010вЂ“14. But, we tested the appropriateness of our approach in a number of methods. First, we stratified our models by age bracket (individuals more youthful or more than age sixty-five): Those in younger team will be beneficiaries of this Medicaid expansion, while those into the older team wouldn’t normally, given that they could be qualified to receive Medicare. 2nd, we examined just how alterations in payday financing diverse with all the share of uninsured individuals within the county before expansion: we might be prepared to find a larger lowering of payday financing in areas with greater shares compared to areas with reduced stocks. Final, we carried out an вЂњevent studyвЂќ regression, described above, to assess any preexisting time styles in payday financing. Our extra methodology supplied evidence that is reassuring our findings had been owing to the Medicaid expansion. The difference-in-differences methodology we relied on contrasted payday financing before and after CaliforniaвЂ™s early Medicaid expansion within the stateвЂ™s expansion counties versus nonexpansion counties nationwide. To manage for confounding, time-varying factors that affect all counties at particular times (such as for instance recessions, vacations, and seasonality), this method utilized nonexpansion counties, in Ca along with other states, as a control team.
Display 1 presents quotes regarding the effect of Medicaid expansion in the general amount of payday financing, our main results; the table that is accompanying in Appendix Exhibit A4. 16 We discovered big general reductions in borrowing after the Medicaid expansion among individuals more youthful than age sixty-five. How many loans applied for per declined by 790 for expansion counties, compared with nonexpansion counties month. Provided a preexpansion mean of 6,948 loans per thirty days, that amounts to an 11 % drop when you look at the amount of loans. This lowering of loan amount equals a $172,000 decrease in borrowing per thirty days per county, from the mean of $1,644,000вЂ”a fall of installment loans Georgia ten percent. And 277 less unique borrowers per county-month took down loans, which represents an 8 per cent decrease through the preexpansion mean of 3,603.
Display 2 presents the end result of Medicaid expansion from the wide range of loans in three age groups: 18вЂ“34, 35вЂ“49, and 50вЂ“64; the accompanying table is in Appendix Exhibit A5. 16 The decrease in the amount of loans every month ended up being completely driven by borrowers younger than age fifty (the small enhance among older borrowers wasn’t significant). For expansion counties in Ca, in accordance with the nonexpansion counties in Ca as well as other states, postexpansion borrowers ages 18вЂ“34 took away 486 loans per county-month, when compared with a preexpansion mean of 2,268вЂ”a reduction of 21 %. For borrowers many years 35вЂ“49, the decrease ended up being 345 from the preexpansion mean of 2,715, a decrease of 13 %. This observed relationship across age groups stayed once we examined the amount of unique borrowers and total dollars loaned (information maybe perhaps not shown).