Early Medicaid Expansion Associated With Reduced Payday Borrowing In California

Various studies have demonstrated that health insurance coverage lowers the medical debt that consumers must take on when they receive health care services.^1,2 Lower medical debts, in turn, can improve consumers’ credit scores and other credit-related outcomes.^3–5 Only a few studies have focused on individual financial outcomes related to the recent expansion of health insurance coverage under the Affordable Care Act (ACA),^6,7 and no studies to date have focused on how health insurance coverage affects the use of alternative financial products.

Research on this topic is especially important given the documented relationship between poverty, medical debt, and bad credit outcomes. People in poverty are more likely to be uninsured and have fewer financial resources to pay for out-of-pocket health care spending.⁸ Poor financial health can permeate all aspects of life, making it difficult to secure affordable housing, purchase homes or automobiles, and manage day-to-day expenses. Importantly, people with bad credit ratings have less access to traditional methods of borrowing, such as credit cards or personal loans.^9,10

One form of borrowing that has been increasingly scrutinized by policy makers is payday loans.¹¹ Payday loans are short-term, unsecured loans that are primarily used by low- and middle-income Americans in states where the loans are legal. In 2012 it was estimated that twelve million Americans take out at least one payday loan annually, with an average of eight loans of $375 each per year and a total of $520 spent on fees.¹² The loans are advertised as a two-week credit product meant to address temporary cash-flow issues. Critics argue that when the loans’ fees are converted into an annualized interest rate, the rates are much higher than those of other financial products—typically over 300 percent—and many borrowers end up in long-term cycles of debt.^12,13 Low-income adults, defined in the 2012 study as those with annual incomes of less than $40,000, were 62 percent more likely than their higher-income counterparts to use payday loans.¹²

Evidence suggests that payday borrowing occurs when access to traditional credit is most limited,¹⁰ which supports the hypothesis that payday loans are perceived as a last resort by consumers.¹⁴ About 16 percent of payday loan consumers report using the loans for emergency or unexpected expenses, while 69 percent report borrowing to pay for recurring expenses.¹² Medical debts could fall into either category, such as when consumers are faced with unexpected financial shocks (for example, an emergency department visit) or when they are balancing recurring medical expenses (for example, for prescriptions) with competing demands like housing and food.

There is early evidence that expansions of eligibility for Medicaid might be an important policy lever for improving the financial stability of low-income Americans.^1,3 The Oregon Health Insurance Experiment found that Medicaid reduced financial strain and improved the credit outcomes of low-income adults, who experienced fewer delinquencies in medical bills and lower amounts of medical debt. Catastrophic medical liability, defined as exceeding 30 percent of annual income, was almost entirely eliminated.¹⁵ Other studies have confirmed that Medicaid expansion improves credit scores and may reduce rates of bankruptcy.⁶ In particular, the Massachusetts health care reform, which expanded coverage in a way similar to the ACA, led to a decrease in bankruptcies and an improvement in credit scores.⁴ Going back further, the Medicaid expansions of the 1990s have been shown to decrease the risk of bankruptcy.³

The fate of existing and future Medicaid expansions is currently unclear, as Congress and President Donald Trump continue to consider repealing and replacing the ACA. As national and state health policy enter a new era of flux, it is critical to have a broad empirical understanding of the costs and benefits of providing Medicaid to low-income adults—especially populations that historically have not been eligible for Medicaid.

We examined the relationship between Medicaid coverage and risky borrowing in the state of California, which was an early adopter of Medicaid expansion through the ACA. Specifically, we compared payday lending in California counties that expanded Medicaid in advance of the ACA’s 2014 expansion to lending in counties throughout the United States (including four in California) that had not yet expanded Medicaid.

Study Data And Methods

Data

This study used a novel source of national data on payday loans provided to the authors by an industry trade group, the Community Financial Services Association of America (CFSA). The data set consisted of the universe of payday loans originated by five national storefront payday lending chains with locations around the country. It contained information on over ninety-three million loans, which had been deidentified for research purposes. For each loan, we observed the dates that the loan was made and was due, the outcome of the loan (for example, a default or late payment), and the ZIP code of the payday lender’s storefront. The data set covered all loans from 2009 through the early months of 2014. Appendix Exhibit A1 maps the states included in the data set.¹⁶

Methods

We focused on the state of California, which entered into an agreement with the administration of President Barack Obama for early county-by-county implementation of the ACA’s Medicaid expansion in 2011 and 2012. We studied the early expansions in California, because our data did not offer a sufficiently long time series to study the 2014 expansions and provided relatively little information about loans in other early expansion states. We compared California counties that expanded Medicaid early to counties nationwide that did not do so, including four California counties that delayed expansion.

We aggregated the CFSA data to the county-month level, creating aggregate loan counts, default rates, and other measures of loan volumes and outcomes in each county and month combination. The aggregated data set contained 58,020 county-month observations for the period 2009–13, which covered roughly twenty-four months before and twenty-four months after the California Medicaid expansions. California rolled out Medicaid expansion over 2011 and 2012, and we used the dates of expansion by county provided by Benjamin Sommers and coauthors.¹⁷ These dates are listed in Appendix Exhibit A2, along with county-specific average monthly payday borrowing before to expansion.¹⁶ Appendix Exhibit A3 shows the aggregate study sample statistics.¹⁶ We examined outcomes in the 43 expansion counties in California, using as a comparison group 920 counties in nonexpanding states and 4 California counties that delayed expansion.

Our primary outcomes were three measures of loan volume: the number of loans, the amount of money borrowed, and the number of unique borrowers. We measured unique borrowers in the data each month using the data set’s anonymized borrower identifiers. Medicaid expansions provide health insurance for uninsured adults younger than age 65, so we stratified our outcomes by age and focused on people younger than age 65. Given previous research findings that Medicaid expansions disproportionately benefited those younger than age 50, we further examined the distribution of the number of loans among nonelderly adults by borrower’s age (18–34, 35–49, and 50–64).

Additionally, we thought that we might see higher reductions in payday lending within counties with higher preexpansion shares of low-income uninsured adults. We investigated this possibility by comparing counties with a high share of uninsured to those with a low share. Counties categorized as having a high share were those in the top tercile of the share uninsured with incomes of less than 138 percent of the federal poverty level, according to the 2010 Census Bureau’s Small Area Health Insurance Estimates; counties categorized as having a low share were in the bottom tercile.

Our secondary outcomes were the shares of loans that ended in default, were repaid late, and were rollovers. Rollovers are loans that are taken out at the same time a previous loan is due, which allows the borrower to extend the loan’s duration without repaying the principal—in exchange for paying a finance charge. We identified likely rollovers in the data as loans that began within two days of a previous due date for the same borrower and same lender.¹⁸

For both our primary and secondary outcomes, we used a standard difference-in-differences analysis of county-month outcomes that covered roughly twenty-four months before and twenty-four months after the 2011–2012 California Medicaid expansions. As noted above, we compared 43 California early expansion counties to 924 nonexpansion counties (including the 4 previously mentioned nonexpansion California counties) in the national data set, with standard errors clustered at the county level. We stratified our findings by the age of the borrower—focusing on people younger than age sixty-five, who would have been most likely to be affected by Medicaid expansion. As a sensitivity test (see Appendix Exhibit A7),¹⁶ we examined borrowers older than age sixty-five and used a triple-differences approach at the county-month-age level.

To rule out systemic preexisting time trends that could have undermined our difference-in-differences approach, we estimated an “event study” regression of the effect of Medicaid expansion on the number of loans. This tested the validity of our assumption that payday borrowing would have had similar trends in expansion and nonexpansion counties if none of the counties had expanded Medicaid. The regression included a fixed effect for every county, a fixed effect for every month, and indicators for four six-month periods before Medicaid expansion and three six-month periods after expansion (see Appendix Exhibit A8).¹⁶

Limitations

Our study was not able to directly link individual insurance status to payday borrowing; to our knowledge, the data to do so do not exist.

Additionally, although we found no evidence of this, we could not rule out the possibility that state- or county-level changes in the regulation (or enforcement of regulations) of payday loans or other industry changes might have occurred in California in the period 2010–14. However, we tested the appropriateness of our approach in several ways. First, we stratified our models by age group (people younger or older than age sixty-five): Those in the younger group would be beneficiaries of the Medicaid expansion, while those in the older group would not, since they would be eligible for Medicare. Second, we examined how changes in payday lending varied with the share of uninsured people in the county before expansion: We would expect to find a greater reduction in payday lending in areas with higher shares than in areas with lower shares. Last, we conducted an “event study” regression, described above, to assess any preexisting time trends in payday lending. Our additional methodology provided reassuring evidence that our findings were attributable to the Medicaid expansion.

Study Results

The difference-in-differences methodology we relied on compared payday lending before and after California’s early Medicaid expansion in the state’s expansion counties versus nonexpansion counties nationwide. To control for confounding, time-varying factors that affect all counties at particular times (such as recessions, holidays, and seasonality), this approach used nonexpansion counties, in California and other states, as a control group.

Exhibit 1 presents estimates of the impact of Medicaid expansion on the overall volume of payday lending, our primary outcomes; the accompanying table is in Appendix Exhibit A4.¹⁶ We found large relative reductions in borrowing after the Medicaid expansion among people younger than age sixty-five. The number of loans taken out per month declined by 790 for expansion counties, compared with nonexpansion counties. Given a preexpansion mean of 6,948 loans per month, that amounts to an 11 percent drop in the number of loans. This reduction in loan volume translates to a $172,000 decline in borrowing per month per county, from a mean of $1,644,000—a drop of 10 percent. And 277 fewer unique borrowers per county-month took out loans, which represents an 8 percent decrease from the preexpansion mean of 3,603.

SOURCE Authors’ analysis of data for 2009–13 from the Community Financial Services Association of America. NOTES The exhibit shows the results of difference-in-differences regressions of the outcomes, comparing differences over time in county-month data between 43 counties in California that expanded Medicaid early (in 2011–12) under the Affordable Care Act and 924 counties nationally (including 4 in California) that did not expand Medicaid before 2014. There were 58,020 county-month observations. Dollars loaned is measured in thousands. Appendix Exhibit A4 shows the regression estimates (see Note 16 in text). For number of loans, R² = 0.971, and p (associated with a test that the Medicaid expansion had no effect) $<0.001$. For dollars loaned, R² = 0.969, p = 0.001. For number of unique borrowers, R² = 0.981, p = 0.002. ***$p<0.01$ ****$p<0.001$

Exhibit 2 presents the effect of Medicaid expansion on the number of loans in three age categories: 18–34, 35–49, and 50–64; the accompanying table is in Appendix Exhibit A5.¹⁶ The reduction in the number of loans per month was entirely driven by borrowers younger than age fifty (the slight increase among older borrowers was not significant). For expansion counties in California, relative to the nonexpansion counties in California and other states, postexpansion borrowers ages 18–34 took out 486 loans per county-month, compared to a preexpansion mean of 2,268—a reduction of 21 percent. For borrowers ages 35–49, the decline was 345 from a preexpansion mean of 2,715, a reduction of 13 percent. This observed relationship across age categories remained when we examined the number of unique borrowers and total dollars loaned (data not shown).

SOURCE Authors’ analysis of data for 2009–13 from the Community Financial Services Association of America. NOTES The exhibit shows the results of difference-in-differences regressions of the outcomes as explained in the Notes to Exhibit 1, which also give the sample size. Appendix Exhibit A5 shows the regression estimates (see Note 16 in text). For ages 18–34, R² = 0.951, p (associated with a test that the Medicaid expansion had no effect) = 0.001. For ages 35–49, R² = 0.969, $p<0.001$. For ages 50–64, R² = 0.976, p = 0.442. ***$p<0.01$ ****$p<0.001$

Exhibit 3 examines the impact of Medicaid expansion on the volume of payday lending as it varies by the share of low-income uninsured people in 2010. Counties with the highest tercile of low-income uninsured people in 2010 (that is, in the top tercile in terms of the share of uninsured people with incomes below 138 percent of poverty) showed greater declines in payday loan volume in terms of both numbers and percentages, when compared to counties in the lowest tercile of low-income uninsured people. For example, the number of monthly loans per county declined by 1,571 (12 percent) in counties with a high share of uninsured borrowers, versus 362 (10 percent) in counties with a low share. There were comparable differences in the amounts loaned and the numbers of unique borrowers.

Exhibit 4 shows the effect of Medicaid on the payment outcomes of payday loans, our secondary outcomes; the accompanying table is in Appendix Exhibit A6.¹⁶ We found a proportionally large and significant postexpansion increase of 0.5 percentage points in the share of defaults, from a preexpansion mean of 3 percent. There was a marginally significant change in the share of late payments and a significant increase in rollovers, which had a high preexpansion mean (50 percent of the loans) and a postexpansion increase of almost 3 percentage points.

SOURCE Authors’ analysis of data for 2009–13 from the Community Financial Services Association of America. NOTES The exhibit shows the results of difference-in-differences regressions of the outcomes as explained in the Notes to Exhibit 1, which also give the sample size. Appendix Exhibit A6 shows the regression estimates (see Note 16 in text). For share of defaults, R² = 0.451, p (associated with a test that the Medicaid expansion had no effect) $<0.001$. For share of late payments, R² = 0.686, p = 0.060. For share of rollovers, R² = 0.853, $p<0.001$.

It is important to recognize that the interpretation of the effect of expanding Medicaid is less straightforward for the secondary outcomes than for the primary outcomes. Since we observed a decline in overall loan volume, Medicaid expansion could have changed the types of people who took out payday loans. We could not distinguish between the effect on the types of borrowers and a direct effect of on reducing default, late payment, or rollover rates across all borrower types.

Appendix Exhibit A7 presents the results of our sensitivity analyses for borrowers older than age sixty-five.¹⁶ As noted above, we examined payday loan volume stratified for people in that age group as well as conducting a triple-difference analysis of county-month-age (younger or older than age sixty-five). We found small but significant increases in payday volume among the older borrowers. When we used those borrowers as an additional within-state control group, we had triple-difference estimates that were roughly similar, though slightly larger in magnitude, than the difference-in-differences estimates in Exhibit 1. To the extent that the effects on the older population captured unobserved, latent trends in expansion counties, this suggests that our main estimates might be slight underestimates of the effects of Medicaid expansion on payday loan volume.

As mentioned above, the key assumption in the difference-in-differences framework on which we relied is that California’s expansion counties and all of the nonexpansion counties would have shown similar trends in the absence of the expansion. That assumption would be violated, for instance, if California had experienced a uniquely robust job-market recovery during the study period. That said, we are aware of no evidence that the job-market recovery in California was different from the recovery in other states in a way that would affect payday borrowing. But, more important, Appendix Exhibit A8 shows the time trends in numbers of loans both before and after the expansion.¹⁶ Reassuringly, the exhibit suggests that there were no observable differences between future expanding and nonexpanding counties in preexisting time trends, which validates the parallel-trends assumption that underlies our difference-in-differences approach. Specifically, in the twenty-four months before Medicaid expansion, we observed no preexisting differences in the number of payday loans that could confound the estimated effect of Medicaid expansion when we later compared groups. We therefore found no evidence that the parallel trends assumption was violated. In addition, the Appendix exhibit suggests that a negative effect of the Medicaid expansions on the numbers of loans began approximately six months after expansion, which seems credible given that medical needs and medical bills accumulate slowly.

Discussion

Medicaid expansion has improved access to high-quality health care, increased the use of outpatient and inpatient medical services,^15,19 and improved the personal finances of low-income adults by reducing the number of medical bills subject to debt collection and by improving credit scores.¹ This study adds to the existing evidence of the benefits of Medicaid expansion by demonstrating that it decreased the use of payday loans in California.

Previous research showing that Medicaid expansions led to substantive reductions in medical debt suggested that we might find a reduction in the need for payday borrowing following California’s early expansion. Indeed, our primary results suggest a large decrease (11 percent) in the number of loans taken out by borrowers younger than age 65, and an even larger decline (21 percent) among those ages 18–34. We observed a slight increase in borrowing for those older than age 65, which we found surprising. We also found the reduction in payday borrowing to be concentrated among those younger than age 50, which is plausible given that half of new Medicaid enrollees in California in 2012–14 as a result of the expansion of eligibility for adults were younger than age 40, and almost 80 percent were younger than age 55.²⁰ Previous research has also suggested that younger adults are the primary beneficiaries of Medicaid expansions.²¹

We were unable to identify precisely how and for whom Medicaid reduces payday borrowing. To our knowledge, there are no data that directly link payday lending to insurance status. One possibility is that although a relatively small share of California residents (roughly 8 percent of the low-income population)²² gained coverage, the coverage gain may have been disproportionately bigger in the subset of low-income California residents likely to frequent payday lenders. Thus, the observed magnitude of declines in loan volume could simply be driven by a large change in borrowing for county residents who gained coverage. There is previous evidence that California’s early Medicaid expansions decreased out-of-pocket medical spending by 10 percentage points among low-income adults.²² Another possibility is that the Medicaid expansion affected many more people beyond those who gained coverage directly. Household members of people who gained Medicaid coverage may have also decreased their payday borrowing.

Regardless of the merits of payday lending, a decline in loan volume attributable to Medicaid is a positive policy outcome.

Payday loans are of particular policy import because they are a controversial financial product, outlawed in many states and tightly regulated in several others. These loans would be severely restricted under new rules proposed by the Consumer Financial Protection Bureau.²³ Proponents of payday loans have argued that they are an important resource for people with bad credit ratings who would otherwise not have access to cash in dire circumstances or who would accrue even higher fees through bank overdrafts or informal loans. However, evidence has shown that at least some payday borrowing results from behavioral biases, and some consumers would be better off avoiding these loans. Such behavioral biases may lead consumers to make mistakes when budgeting, be overly optimistic about their ability to repay loans in the future, or focus on short-term financial needs rather than the long-term consequences of high-interest borrowing.^24,25 Regardless of the merits of payday lending, a decline in loan volume attributable to Medicaid is a positive policy outcome and supports previous research on the spillover effects of Medicaid on financial health.^1,3,6–7

Reductions in medical expenditures should also theoretically make it easier to pay back payday loans. Yet the outlook for postexpansion borrowers in our study was more mixed. While there was a slight reduction in the number of loans per borrower, we observed marginal increases in late payments and significant increases in the shares of defaults and rollovers. There are several potential mechanisms for these increases that warrant further study. First, the substantial reduction in payday volume we observed could have a corresponding influence on the composition of the remaining borrowers or on the characteristics of their debts. Specifically, people who borrow because of medical expenses may be sociodemographically different than people who borrow for other reasons (for example, job loss). Second, a trending loss of revenue in the industry could exert influence on the business model, leading to riskier lending practices if payday lenders detect a significant decline in loan volume.

Conclusion

For people younger than age sixty-five, Medicaid expansion in California was associated with significant declines in the average number of payday loans per month, the amount borrowed, and the number of unique borrowers. This decline in payday borrowing did not appear to be due to a preexisting trend. It was concentrated in young adults, was not observed among people ages sixty-five and older, and was more pronounced in areas that had a higher share of uninsured people before the expansion—which is consistent with the view that the Medicaid expansion caused the reductions in payday borrowing. These findings add to the previous literature on the benefits of Medicaid in improving the financial health of low-income Americans.

ACKNOWLEDGMENTS

An earlier version of this article was presented in the Health Policy and Management Brown Bag Series at the Mailman School of Public Health, Columbia University, New York City, February 8, 2017. This work was supported by the Russell Sage Foundation (Award No. 94-16-02). Any opinions expressed are those of the authors alone and should not be construed as representing the opinions of the foundation. The authors thank Sherry Glied and Katherine Baicker for their helpful feedback on earlier drafts of this article. The data were generously provided by an industry trade group, the Community Financial Services Association of America, for the purposes of this project. The group imposed no restrictions on the conclusions of the research beyond preserving the confidentiality of the underlying data.

NOTES

• 1 Finkelstein A, Taubman S, Wright B, Bernstein M, Gruber J, Newhouse JPet al. The Oregon Health Insurance Experiment: evidence from the first year. Q J Econ. 2012;127(3):1057–106. Crossref, Medline, Google Scholar
• 2 Hadley J. Sicker and poorer—the consequences of being uninsured: a review of the research on the relationship between health insurance, medical care use, health, work, and income. Med Care Res Rev. 2003;60(2 Suppl):3S–75S; discussion 6S–112S. Crossref, Medline, Google Scholar
• 3 Gross T, Notowidigdo MJ. Health insurance and the consumer bankruptcy decision: evidence from expansions of Medicaid. J Public Econ. 2011;95(7–8):767–78. Crossref, Google Scholar
• 4 Mazumder B, Miller S. The effects of the Massachusetts health reform on household financial distress. Chicago (IL): Federal Reserve Board of Chicago; 2015 Jul. (Working Paper No. 2014-01). Google Scholar
• 5 Dobkin C, Finkelstein A, Kluender R, Notowidigdo MJ. The economic consequences of hospital admissions [Internet]. Cambridge (MA): National Bureau of Economic Research; 2016 May [cited 2017 Aug 11]. (NBER Working Paper No. 22288). Available for download (fee required) from: http://www.nber.org/papers/w22288.pdf Google Scholar
• 6 Hu L, Kaestner R, Mazumder B, Miller S, Wong A. The effect of the Patient Protection and Affordable Care Act Medicaid expansions on financial well-being [Internet]. Cambridge (MA): National Bureau of Economic Research; 2016 Apr [cited 2017 Aug 11]. (NBER Working Paper No. 22170). Available from: http://www.nber.org/papers/w22170.pdf Google Scholar
• 7 Gallagher EA, Gopalan R, Grinstein-Weiss M. The effect of health insurance on home payment delinquency: evidence from ACA Marketplace subsidies [Internet]. 2017 Jul 20 [cited 2017 Aug 11]. Unpublished paper. Available for download from: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2922260 Google Scholar
• 8 Institute of Medicine. Hidden costs, value lost: uninsurance in America. Washington (DC): National Academies Press; 2003. Google Scholar
• 9 Dobbie W, Goldsmith-Pinkham P, Mahoney N, Song J. Bad credit, no problem? Credit and labor market consequences of bad credit reports [Internet]. Cambridge (MA): National Bureau of Economic Research; 2016 Oct [cited 2017 Aug 11]. (NBER Working Paper No. 22711). Available for download (fee required) from: http://www.nber.org/papers/w22711 Google Scholar
• 10 Bhutta N, Skiba PM, Tobacman J. Payday loan choices and consequences. J Money Credit Bank. 2015;47(2–3):223–60. Crossref, Google Scholar
• 11 Servon L. Are payday loans harmful to consumers? J Policy Anal Manage. 2017;36(1):240–48. Crossref, Google Scholar
• 12 Pew Charitable Trusts. Payday lending in America: who borrows, where they borrow and why [Internet]. Washington (DC): Pew Charitable Trusts; 2012 Jul [cited 2017 Aug 11]. Available from: http://www.pewtrusts.org/~/media/legacy/uploadedfiles/pcs_assets/2012/pewpaydaylendingreportpdf.pdf Google Scholar
• 13 Pew Charitable Trusts. From payday to small installment loans: risks, opportunities, and policy proposals for successful markets [Internet]. Washington (DC): Pew Charitable Trusts; 2016 Aug [cited 2017 Aug 11]. Available from: http://www.pewtrusts.org/~/media/assets/2016/08/from_payday_to_small_installment_loans.pdf Google Scholar
• 14 Lusardi A, Schneider DJ, Tufano P. Financially fragile households: evidence and implications [Internet]. Cambridge (MA): National Bureau of Economic Research; 2011 May [cited 2017 Aug 11]. (NBER Working Paper No. 17072). Available from: http://www.nber.org/papers/w17072.pdf Google Scholar
• 15 Baicker K, Taubman SL, Allen HL, Bernstein M, Gruber JH, Newhouse JPet al. The Oregon experiment—effects of Medicaid on clinical outcomes. N Engl J Med. 2013;368(18):1713–22. Crossref, Medline, Google Scholar
• 16 To access the Appendix, click on the Details tab of the article online.
• 17 Sommers BD, Kenney GM, Epstein AM. New evidence on the Affordable Care Act: coverage impacts of early Medicaid expansions. Health Aff (Millwood). 2014;33(1):78–87. Go to the article, Google Scholar
• 18 Priestley JL (Kennesaw State University, Kennesaw, GA). Payday loan rollovers and consumer welfare [Internet]. Unpublished paper; 2014 Dec 5 [cited 2017 Aug 14]. Available for download from: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2534628&rec=1&srcabs=2160947&alg=1&pos=5 Google Scholar
• 19 Sommers BD, Blendon RJ, Orav EJ, Epstein AM. Changes in utilization and health among low-income adults after Medicaid expansion or expanded private insurance. JAMA Intern Med. 2016;176(10):1501–9. Crossref, Medline, Google Scholar
• 20 Research and Analytic Studies Division. Medi-Cal’s historic period of growth: a 24-month examination of how the program has changed since December 2012 [Internet]. Sacramento (CA): California Department of Health Care Services; 2015 Aug [cited 2017 Aug 11]. (Medi-Cal Statistical Brief). Available from: http://www.dhcs.ca.gov/dataandstats/statistics/Documents/New_24_Month_Examination.pdf Google Scholar
• 21 Fangmeier J, Udow-Phillips M. The ACA’s coverage expansion in Michigan: demographic characteristics and coverage projections [Internet]. Ann Arbor (MI): Center for Healthcare Research and Transformation; 2013 Jul 30 [cited 2017 Aug 11]. Available from: http://www.chrt.org/publication/acas-coverage-expansion-michigan-demographic-characteristics-coverage-projections/ Google Scholar
• 22 Golberstein E, Gonzales G, Sommers BD. California’s early ACA expansion increased coverage and reduced out-of-pocket spending for the state’s low-income population. Health Aff (Millwood). 2015;34(10):1688–94. Go to the article, Google Scholar
• 23 Consumer Financial Protection Bureau. Payday, vehicle title, and certain high-cost installment loans: a proposed rule. Federal Register [serial on the Internet]. 2016 Jul 22 [cited 2017 Aug 11]. Available from: https://www.federalregister.gov/documents/2016/07/22/2016-13490/payday-vehicle-title-and-certain-high-cost-installment-loans Google Scholar
• 24 Leary JB, Wang J. Liquidity constraints and budgeting mistakes: evidence from Social Security recipients [Internet]. 2016 Mar [cited 2017 Aug 11]. Unpublished paper. Available from: https://drive.google.com/file/d/0B-xlEFaWMd6IREl1T2ZMWlBxekU/view Google Scholar
• 25 Burke K, Leary JB, Wang J. Information disclosure and payday lending in Texas [Internet]. 2016 May [cited 2017 Aug 11]. Unpublished paper. Available from: https://drive.google.com/file/d/0B-xlEFaWMd6IczNhQzRSRmJfSjQ/view Google Scholar