By Kimberly Cornaggia, Pennsylvania State University - Department of Finance; John Hund, University of Georgia; Giang Nguyen, Pennsylvania State University - Smeal College of Business and Zihan Ye, University of Nevada, Las Vegas
Over the past 20 years, an epidemic of opioid abuse in the US has taken a terrible toll in human life. In our paper, Opioid Crisis Effects on Municipal Finance, we identify an additional significant and pernicious cost: local communities affected by high levels of opioid abuse find it difficult to raise funds, and when they do raise money, they do so only at significantly higher interest rates than otherwise similar communities with low levels of opioid abuse. This difference is driven by investors’ reticence to supply capital to highly affected areas, rather than a lack of demand for funds from these affected communities. This dynamic threatens to exacerbate the effects of opioid abuse as communities most in need of infrastructure find it more expensive to pay for it.
Granularity and importance of local comprehensive data
Opioid abuse has killed over half a million Americans over the past two decades. It is a national problem, yet opioid abuse is highly localized and varies significantly within each state. Dollar costs of the epidemic fall heavily on the individual counties, cities and townships that face reduced property and sales tax rates, increased mental health and law enforcement costs, and lower labor productivity at local businesses.
Addressing the impact of this epidemic therefore requires comprehensive yet granular data. We collect data from every individual death certificate, every prescribed opioid pill, the universe of municipal bond issuance, and socio-economic, employment and budget data. Our research can be interpreted as a population-level study since we track each opioid death in every county of the United States. Counties with high levels of opioid abuse are relatively small (average total population of 107,887) and more rural, but they are not strikingly different from low-abuse counties with respect to employment, GDP per capita or health care availability (proxied by hospital beds per capita).
Impact on community finances
We examine the effect that opioid abuse has on county and city budgets and identify lower revenues and higher expenditures, especially for local law enforcement. To ensure that the differences we identify are driven by levels of opioid abuse and not local economic conditions, every year we match each high-death county with other counties that have an identical degree of urbanization and that are very similar in terms of income, property values, population and poverty, and then control for changes in GDP, employment and population over time. We calculate the average differences between the high-abuse counties and the closely matched low-abuse counties for property tax revenue, sales tax revenue, total operating expenditures and expenditures on police and protection over a four-year horizon.
We find that property tax and sales tax revenues, which account for nearly 84% of total tax revenues for the average high-death county, fall by 7.4% after four years. This fall represents approximately US$2.4 million less revenue (US$1.6 million in sales tax and US$800,000 in property tax) for the average high-abuse county. High-abuse counties also have significantly higher expenditures, in general and for police and protection expenses. Since local municipalities generally operate with a balanced budget requirement, even small increases in expenditures combined with falling revenues create budgetary pressures.
Results on local community financing
To show that opioid abuse causes higher yields and lower issuance, rather than simply correlates, we exploit variation in the supply chain of prescription opioids. Using this strategy, we find that high opioid abuse counties faced new bond offering yields approximately 17 basis points higher than low-abuse counties. Given the average municipal issuance, this translates into nearly US$300,000 in extra annual interest costs for these predominantly rural communities. Increases in opioid deaths are also associated with sharply lower issuance amounts of new municipal debt. Our results imply a reduction of about US$15 million in annual funding that could otherwise have been raised - an amount roughly equivalent to twice the average price of a new elementary school.
These financial effects are significantly larger when we focus on areas where deaths are caused by illicit opioids (primarily heroin and fentanyl), which could either be attributable to the increased costs of illegal abuse or the fact that in the later years of our sample these types of deaths become more common.
Our results suggest that financing problems associated with opioid abuse are, if anything, increasing, as opioid deaths in 2020 reached a record yearly high of nearly 70,000, with most attributed to illicit fentanyl.
Disentangling opioid effects from local economic conditions
Although we control for a host of bond-specific characteristics, variation across states, and local economic variables, it is possible that the higher yields and lower bond issuance we find are the result of unobserved poor local economic conditions rather than opioid abuse. To disentangle these effects, we exploit variation in the distribution and marketing of the opioid pills most likely to lead to abuse. Essentially, we compare municipalities with very similar economic conditions but that had very different supplies of addictive opioids. We show that higher distribution of addictive opioids through pharmacies with little oversight leads to more local deaths. Similarly, we show that counties where Purdue Pharma heavily marketed OxyContin in the late 1990s show sharply higher opioid death rates. Importantly, neither of these supply shocks is likely to be correlated with the yields charged by investors (in the case of Purdue’s marketing, over 10 years later, on average) except through their effects on opioid deaths. Offering yields are higher and dollar issuance amounts are lower in counties with high opioid mortality rates because they contained “pill mills” (retail pharmacies with limited oversight prescribing abnormally high number of opioids) or were targeted by Purdue’s early marketing of OxyContin.
Supply or demand?
We examine differential effects across investor types in order to test whether the bond market effects of opioid abuse reflect a reduction in the investor supply of capital or a reduction in municipal demand for capital. These auxiliary tests indicate a reduction in the supply of capital.
Investors in municipal bonds tend to invest close to home, because of tax advantages and because they reap non-pecuniary benefits of financing their local community (such as better schools, water treatment and law enforcement capabilities). If the negative financing effects we document are the result of investors’ reticence to supply capital, then communities with more tax-incentivized investors or investors with deeper pockets should see lesser impact, as should those localities with higher rates of homeownership. This is exactly what we document. Bonds primarily purchased by diversified institutions and banks are similarly less affected.
Our results on the opioid epidemic demonstrate the profound effects public health emergencies have on municipal budgets and the ability of cities and counties to raise capital to address these emergencies. Capital is less available and more expensive for precisely those communities that require investment to reverse crisis effects, potentially leading to even greater levels of opioid abuse. Government policies to encourage the supply of capital to affected communities, such as federal bond insurance for municipalities with high levels of opioid abuse, may provide a cost-effective way to offset these negative feedback effects.
Find out more about the PRI’s work on sub-sovereign debt:
ESG Integration in Sub-Sovereign Debt: The US Municipal Bond Market
This blog is written by academic guest contributors. Our goal is to contribute to the broader debate around topical issues and to help showcase research in support of our signatories and the wider community.
Please note that although you can expect to find some posts here that broadly accord with the PRI’s official views, the blog authors write in their individual capacity and there is no “house view”. Nor do the views and opinions expressed on this blog constitute financial or other professional advice.
If you have any questions, please contact us at firstname.lastname@example.org