2

Do Americans View Financial Threats as Important Political Issues?

Chapter 1 put forth a puzzle. Even though we have good reason to expect that people would become active on many political issues that reflect financial threats, there are many cases in which we observe limited action.

In response I proposed that there are heretofore unexamined communicative barriers that make people less willing to spend scarce resources of time and money on these issues. The validity of my argument rests on two key assumptions: that people subjectively recognize these threats and that they actually consider them to be important political issues. I assess these assumptions in this chapter.

I begin by describing in more detail the nature of financial constraints that potentially affect Americans across four major areas: jobs, health care, retirement, and higher education. I expect that some readers will already be familiar with the content of this section, and thus I invite them to skip ahead. The second part of the chapter focuses on the degree to which Americans subjectively perceive these aspects of insecurity in their daily lives. After all, it’s one thing to identify the objective situation facing Americans and to identify ways in which it has changed in recent times. It’s quite another to claim that Americans perceive their situation to be as insecure as the objective indicators would suggest. In the final part of the chapter I examine both whether people see these issues as important political problems and, if they do, who sees them in this way. Do their concerns reflect personal experiences with job insecurity, health care costs, retirement concerns, and the cost of education, or are the concerns more about others’ situations?

SETTING THE STAGE: FOUR AREAS OF FINANCIAL THREAT

Capitalism is often described as a process of creative destruction, involving the never-ending creation of new markets, means of production, goods, and forms of transportation. Joseph Schumpeter famously described this process as “incessantly revolutioniz[ing] the economic structure from within, incessantly destroying the old one, [and] incessantly creating a new one.”1

With such change comes risk. Sometimes people actively decide to take such risks. The midcareer professional who leaves her company to start her own business risks losing a comfortable income and lifestyle. The college student who incurs student loan debt in order to expand future employment opportunities risks facing an uncertain job market. The newlyweds who decide to buy a house risk the possibility of being unable to afford the mortgage if one of them loses a job. All of these activities are part of what Schumpeter called the “fundamental impulse that sets and keeps the capitalist engine in motion.”² In other cases, the very nature of the options available to people entails risk. This is especially the case for people whose only employment opportunities are low-wage jobs with few, if any, benefits. These people constantly risk losing their jobs, having their hours cut, or experiencing a health emergency that would make them unable to work.

From the end of World War II until the mid-1970s, as the United States emerged as the world’s newest economic superpower, households were (on average) in a reasonably good position to weather such risks.³ The American economy was in the midst of a postwar boom as the demand for American-made goods at home and abroad reached unprecedented levels. Involuntary job loss was uncommon, and when it did occur it was typically not permanent. Downturns in the economic cycle (as in 1958, for example) signaled a temporarily reduced need for labor and other inputs into the production process rather than a permanent response to the pressures of international competition. Workers at this time were also far more likely to stay with one company for their entire working lives, which aided unionization efforts and was also thought to provide employers with loyal and more productive workers. Overall, from 1947 until 1977, median family income, adjusted for inflation, more than doubled.4 Economic gains were also more evenly distributed than they had been in the past.

In addition to rising incomes, Americans also benefited from institutions that enhanced their economic security. When it came time for working adults to retire, they could often count on their company to provide a defined-benefit retirement plan. These were, in essence, forced savings plans that provided a secure income source for as long as the worker lived post-retirement. As an added layer of security, starting in the early 1970s the federal government insured these pensions via the Employee Retirement Income Security Act.⁵ Both their prevalence and their insurance meant that individual companies and the federal government would help ensure that workers could enjoy a comfortable retirement for as long as they lived. On top of this, in retirement workers and their families could count on relatively generous benefits via Social Security and Medicare. Social Security, for example, enjoyed near continuous expansion from the 1950s until the mid-1970s.

At this time many goods entailed far less risk in the event of an adverse event. Two prime examples are health care and higher education. Health care was less expensive as costs were lower and employer-provided health programs demanded lower premiums and deductibles. Higher education was also far more reasonably priced, and returning veterans from World War II and the Korean War could benefit from generous GI Bills that financed their college costs. Young adults just starting out were far less likely to be burdened with college debts like they are today. Both of these goods and their associated costs posed less risk to the average household because if a health care emergency struck or a worker experienced job loss, there was a much smaller threat of mounting debt.

Taken as a whole, the decades since the 1970s look altogether different. One aspect of this involves the hyperconcentration of income among these at the very top of the income distribution. Data derived from tax records show that between 1976 and 2007 the top 1% of families in the United States went from receiving 8.9% of national income to receiving 23.5%6. This trend has greatly accelerated in the recovery from the Great Recession, as 95% of the total income growth that occurred between 2009 and 2012 accrued to those in the top 1%.⁷ Broadly similar conclusions emerge using data that take into account a more comprehensive measure of pre-tax income (including, for example, the value of in-kind benefits such as employer-paid health insurance premiums) as well as most forms of federal taxes, as the Congressional Budget Office (CBO) has done since 1979. Between 1979 and 2007, before the start of the Great Recession, the CBO finds that the top 1% of households enjoyed average income gains of over 314% whereas those in the middle 60% of the income distribution’s gains were a relatively paltry 42%.⁸

During this same time period another divergence also occurred. This is the divergence in labor market earnings between those with a college education and those with less education. Between 1980 and 2012, the real earnings among men working full-time with at least a college degree rose between 20 and 56% (with the precise percentage depending upon how much postgraduate education they had completed). During that same time, however, the real earnings of those with a high school or lower educational level fell between 11 and 22%. Among women, while those with the lowest education did not see their real earnings fall over this time period, they did experience extremely modest gains unless they had completed at least some college.⁹

Scholars have offered several reasons for these twin patterns of economic divergence, including changes in the structure of the labor market, technology, globalization, unionization, and government policies that have reduced the progressivity of the tax code, allowed the bargaining position of ordinary workers to erode, allowed executive compensation to skyrocket, and failed to police increasingly sophisticated financial markets.10 But, in addition to widening inequality, another set of changes has also taken place, which involves heightened economic insecurity for many people on the middle and lower rungs of the income ladder. In what follows, I describe some of these growing financial threats in four different areas: the threat of involuntary job loss and reduced work hours, higher health care costs, reduced retirement security, and increased higher education costs.¹¹ My discussion is not meant to be a comprehensive discussion of the causes and facets of these forms of insecurity but instead a way to highlight broad trends over time (for more comprehensive treatments, I direct readers to the citations in each section). Moreover, and underscoring the point made in Chapter 1, this particular group of four issues does not capture every aspect of economic insecurity that Americans might face, but does include the ones I focus on in this book.

Job Insecurity

As Schlozman, Verba, and Brady observe, “we live in an era in which workers, even highly skilled ones, are squeezed by many trends designed to cut labor costs … from the export of jobs overseas to the outsourcing of service functions to the increased use of part-timers and independent contractors.”¹² Here I focus on two of the biggest threats with regard to job security: the possibility of losing one’s job involuntarily and the possibility of losing hours on the job. Because pensions and health insurance are so closely tied to employment status in the United States, these occurrences pose several threats in addition to just lost income.¹³

One of the most comprehensive data sets for studying involuntary job loss over time is the biannual Displaced Workers Survey (DWS), which began in 1984 and is administered as part of the U.S. Census Current Population Survey. The survey asks if workers experienced an involuntary separation from their workplace over the previous three years, in which “involuntary separation” includes any job loss stemming from business decisions such as plant closings, layoffs, or downsizing (i.e., anything other than performance-related reasons). From the perspective of understanding whether job insecurity is greater now than in the postwar period, the fact that this survey did not begin until 1984 is itself worth underscoring. As Arne Kalleberg notes, “That the [Bureau of Labor Statistics] did not even collect data on permanent job loss until the mid-1980s is very telling, as it indicates that it was widely believed that this was not really a systemic, large-scale problem before then.”14

As one might expect, the rate of job loss is cyclical like the unemployment rate—both fall during periods of economic expansion and rise during periods of contraction. Also, as one might expect, both reached high points during the two most recent periods of severe recession: the early 1980s and the Great Recession from 2007 to 2009.¹⁵ In addition, the threat of job loss is lower for people with a four-year college degree compared to those without. For example, 19.4% of those with a high school degree lost their jobs during the Great Recession while 11.0% of those with a four-year college degree lost theirs.

In terms of changes over time, prior to the Great Recession the DWS provided little evidence of an upward trend in the overall rate of involuntary job loss (after controlling for the unemployment rate). Yet some have pointed out that there is good reason to believe that this survey undercounts the number of people who voluntarily lose their jobs, as its measure of involuntary job loss does not account for workers who accepted buyouts and/or early retirement in anticipation of future layoffs.¹⁶ In fact, studies that have used data sets other than the DWS, such as the Panel Study of Income Dynamics and the National Longitudinal Studies of Youth, have found evidence of an increased likelihood of job loss, especially among men in their prime working years. In addition, one study also found evidence that certain segments of the workforce, including white-collar and service workers, have experienced increased rates of involuntary job loss since the early 1980s.¹⁷

This increased threat would perhaps be less worrisome if unemployment was short-lived, but increasingly that is not the case. On average, unemployment spells last longer than they used to, and the incidence of “long-term unemployment”—defined by the federal government as lasting six months or more—has increased. Long-term unemployment is especially worrisome for both economic and psychological reasons. Financially, unemployment affects not only workers’ cash flow but also potentially whether they (and their loved ones) have health insurance and pensions. The longer it persists, the more likely it is that the worker loses marketable skills, which makes it harder to return to the workforce. It also affects other aspects of individual well-being, leading to poorer health, an increased likelihood of divorce, an increased likelihood of depression, and even an increased mortality risk.18

What is particularly noteworthy about long-term unemployment over the past few years is, again, its secular increase even comparing similar periods in the business cycle. For example, 44% of the unemployed were considered “long-termers” in March 2010 as the labor market recovered from the Great Recession. This was higher than the previous post–World War II peak long-term unemployment rate of 26% from the early 1980s. Yet even in November 2007, on the eve of the recession, a full 20% of unemployed workers were considered “long-termers”; only half that number were considered “long-termers” prior to the 2001 downturn.¹⁹ This is again not something limited to workers with low levels of formal education. The Pew Research Center recently looked at an even starker definition of “long term” than what the federal government typically uses: those who have been out of work for one year or more. Using this definition, by the end of 2009, 21% of the unemployed with a college degree would be considered “long-termers” (the figure was 27% for those with only a high school diploma).

Unemployment stings more now not only because it is likely to last longer but also because of what happens when the unemployed finally land a job. In the past unemployment was often cyclical in nature: workers who lost their jobs during the contractions of the business cycle could expect to receive jobs that offered similar pay and required similar skill sets once the economy picked up. Today it is far more likely that once those jobs are lost they are gone forever. New jobs are more likely to entail a permanent cut in wages or benefits and the need to engage in costly retraining. For example, Henry Farber examined workers’ experiences in the aftermath of the (short) recession of 2001.20 The recovery began in November and the expansion in 2002 and 2003 featured rapid productivity growth but only a small decline in the unemployment rate. Between 2001 and 2003, 35% of those who had lost jobs had failed to find employment and 13% of those who had been employed full-time were now only employed part-time. Even those who found another full-time job endured a 17.1% earnings loss on average. These patterns were amplified in the aftermath of the Great Recession.²¹

Lastly, it is also important to note that job insecurity is often tied to the type of employment arrangements that workers are engaged in. A key point of distinction is between permanent and contingent employees, with the former group enjoying more security on average. The permanent workforce includes those who enjoy full-time status, relatively high job security, and living wages. They are also more likely to have health insurance and some form of pension plan (though, as noted later in the chapter, the nature of such plans has changed over the past few decades). In contrast, the contingent workforce holds jobs that do not have an explicit or implicit contract for long-term employment and tend to have a minimum number of hours that varies unsystematically (and involuntarily).²² Their situation is thus contingent on the employer’s demand for labor at each moment. Estimates from before the Great Recession indicate that up to one-third of the labor force is employed in contingent work, with women and African Americans over-represented among this group.²³

Health Care Costs

A second major threat to Americans’ financial well-being stems from health care, because of both its increasing cost and its unpredictable impact on family finances (particularly among those without insurance). Although the 2010 Patient Protection and Affordable Care Act (PPACA) addresses some attributes of this insecurity, several aspects remain.

One of the major goals of the PPACA was expanded coverage, particularly for the millions of people who were previously living without health insurance. As of 2010, 28% of working-age adults—over 52 million people—reported that they had been uninsured at some time during the previous year.24 For working-age adults who are not extremely poor, the American health insurance system is based on employer coverage; thus one of the major reasons that so many people are uninsured is because fewer employers are offering it. Such declines are particularly noticeable in the private sector, which in part reflects the erosion of unionization in the private sector over the past few decades. Declines are also visible when looking across educational levels. The proportion of those with a college degree who had employer-provided health insurance dropped from 80% in 1979 to 67% in 2006, while the analogous drop for high school graduates was 70 to about 50%. Declines are also visible along gender lines. The proportion of male workers with employer-provided health insurance dropped from 75 to 58% during this time period, while the percentage for women dropped from 60 to 51%.²⁵ Stepping back and looking at the entire workforce, we find that the percentage of nonelderly workers with employer-provided health insurance (either their own or a family member’s) declined almost every single year from 2000 to 2010.²⁶

The PPACA’s main vehicle for alleviating this problem is the employer mandate, which requires that all employers with more than fifty full-time employees provide insurance coverage. The other primary means for alleviating this problem is the individual mandate, which requires all individuals without employer coverage to purchase health insurance unless doing so would impose a severe financial hardship or violate religious beliefs. Failure to comply with either results in increasingly steep fines. To help ensure that such coverage is affordable, the legislation instituted subsidies and mandated the creation of health insurance exchanges that would provide access to standardized insurance plans. The new law also prohibits the industry’s most egregious practices, such as rescinding coverage after people get sick or refusing coverage altogether based on preexisting conditions.27

While achieving near universal health insurance coverage was certainly an important goal, and a primary focus for Obama and other health care reformers who supported the PPACA, it is only one source of health cost insecurity. The other major part involves the out-of-pocket costs that individuals are forced to pay, either directly if they do not have employer-sponsored insurance or indirectly via inflated premiums if they do.²⁸ Consider the most recent data on this: while real family incomes remained roughly the same between 2003 and 2009, health care premiums in employer health plans increased 41% and deductibles skyrocketed 77%.²⁹ It’s no wonder that the proportion of households’ annual budget devoted to health care rose every year during the decade.³⁰ It is also perhaps not surprising that health care costs were linked to (in the sense of acting as a contributor) to 62.1% of all bankruptcies in the United States in 2007.³¹

To be sure, in addition to expanding insurance coverage and curbing the industry’s worst practices, the PPACA does promise to reduce out-of-pocket health care costs for millions of working Americans (as well as those on Medicare, which I discuss in the next section). At the same time, many critics contend that even after it is fully implemented the reform will likely not go far enough toward cost containment because this goal was not prioritized as much as coverage expansion. Critics are particularly quick to point out that reform did not include either a single-payer approach or, at the very least, a public insurance option in which a publicly run health insurance plan would compete with private plans. Either of these would have significantly reduced health care costs, as judged by the types of discounts that Medicare is able to negotiate with hospitals and doctors because of its size and bargaining leverage.32 To be sure, as of this writing it is too early to know precisely what effect the PPACA will have on health care costs and out-of-pocket expenses for individuals, but it is reasonable to expect that health care costs will continue to weigh heavily on family finances given that the PPACA did not adopt the strongest cost-containment strategies.

Retirement

A third financial threat concerns what happens after workers retire. Here I discuss changes in employer-provided retirement plans, Social Security, and Medicare.

With respect to pensions, two changes are noteworthy. The first is simply their reduced availability, as the share of employees with employer-provided pension plans has dropped from just over half in 1979 to under 43% in 2009. As with health insurance coverage, here again changes are especially noticeable among private-sector workers. Among those with at least a college degree, the share who had employer-provided health insurance coverage dropped from 61% in 1979 to 57% in 2006. For those with a high school degree, the drop was much larger: approximately 50 to 37%. For male workers, the decline was also quite large—from 57 to 44%—while the proportion for female workers remained constant but was still lower (approximately 41%).³³ These patterns are part of a larger trend, which is the decreasing availability of fringe benefits for employees engaged in blue-collar and service-sector jobs and the increasing prevalence of contingent work.³⁴

The second is the changing nature of most retirement plans. By and large, employers have switched from offering defined-benefit plans, in which workers received a steady income in their post-working years, to offering defined-contribution plans (e.g., 401[k] plans) in which benefits are tied to the value of market investments.³⁵ As with any investment vehicle, defined-contribution plans certainly have the potential for fantastic gains. But as anyone who planned to retire in 2008 and 2009 knows, gains are not guaranteed. Moreover, whereas defined-benefit pensions were highly regulated and even insured by the federal government in case a company reneged on its obligations, defined-contribution plans do not carry such insurance. The worst-case scenario involves situations like what happened after the 2001 collapse of energy giant Enron. Employees’ retirement accounts were heavily focused on Enron stock, which became worthless almost overnight as the company imploded.

Just as the pension system has become more insecure for millions of American families, so has the other major form of postemployment income security: Social Security. From its inception in 1935 until the late 1970s, Social Security experienced a relatively uninterrupted period of growth. The set of eligible workers expanded and benefits became far more generous, including routine double-digit benefit increases in the late 1960s and early 1970s. By the late 1970s, however, concerns arose about the program’s budgetary solvency. President Carter’s 1977 decision to sign legislation that increased the Social Security payroll tax and the level of wages that could be taxed ushered in an era in which even AARP concluded that change was inevitable.

Today younger workers can expect to count on Social Security less than their parents. The full retirement age (the age at which beneficiaries can start collecting their full benefits) was raised in 1983, and some observers expect that it could be raised in the future to help finance the Baby Boomers’ retirement.36 By 2010 just 35% of Americans expressed confidence in the future of Social Security, with a predictably stark divide between younger cohorts and those currently receiving benefits.³⁷ Findings like these are particularly worrisome given that Social Security supplies almost three-quarters of the typical household’s retirement income.³⁸

Where do these changes leave us? The effects of changing pensions and less generous Social Security add up. At a time when Americans are living longer, two-thirds of working-age households are at risk of being unable to maintain their pre-retirement standard of living.³⁹ They simply do not have the private savings and pension benefits that they need.

The other major pillar of Americans’ retirement security is Medicare, the public health insurance program for Americans over the age of sixty-five. It is one of the greatest forms of health security afforded regardless of work status, income, or any other factor besides citizenship. It also, for many years, required seniors to incur large out-of-pocket expenses because it did not include prescription drug coverage. That changed with the Medicare Modernization Act in 2003 and then further coverage enhancements that were part of the PPACA in 2010. But because federal law prohibits Medicare from negotiating lower drug prices that are more in line with the actual cost of production, we should expect prices to remain high. In fact, premiums are set to rise over the next two decades.40 Moreover, many elderly Americans remain severely at risk as a result of exploding health care costs that are not covered by Medicare (such as many nursing home charges). They have to either purchase supplemental insurance or, if they are fortunate to have it, use employer-provided insurance. This private insurance benefits only a lucky few, as the share of employees at private firms that offer retiree health benefits dropped from 28% in 1997 to 17% in 2008.⁴¹

Cost of College

The final example of financial threat that I focus on in this book is the increasing cost of college. President Obama recently declared that “In a global economy, putting a college education within reach for every American has never been more important.”⁴² Although college-educated workers can expect to receive higher earnings on average, higher education entails greater risk than in the past. This riskiness stems from the ever-increasing sums that must be invested into obtaining the education, along with the uncertain rewards in a labor market marked by increased job loss and a higher incidence of long-term unemployment.

To be more specific, the cost of attending college has increasingly outpaced inflation since the 1980s. Whereas it outpaced inflation by 4.5% in the 1980s, by the 2000s published tuition and fees at private four-year colleges increased an average of 5.6% per year beyond the rate of inflation. The overall result is that a college education can now cost up to three times as much as it did in the 1980s. This trend is similar for public four- and two-year institutions as well as private for-profit institutions, and the real cost is still increasing even once we account for the tuition support that many students receive.43

The result is an enormous rise in the amount of student loan debt. By the second quarter of 2013, the total balance of outstanding student loan debt had reached $994 billion, which surpassed both total credit card debt and total auto loan debt!⁴⁴ From 2003 until 2012, the proportion of twenty-five-year-olds with student debt increased from 25% to over 40% and the amount of debt almost doubled. Moreover, as students take on ever-increasing amounts of debt, they are also more likely to be delinquent on their loan payments. The share of borrowers that were at least ninety days delinquent was almost 25% in 2008 and had reached over 30% in 2012 in the midst of the tepid recovery from the Great Recession. All in all, as higher education remains a critical gateway for many future employment opportunities, more young people (and in some cases their parents) are being saddled with ever-growing debts as they face an increasingly insecure job market.

Economic Security Index

Individually, each of these four threats to financial well-being is important. Collectively, they add up to a stark portrait that has the potential to affect millions of Americans every year. Recently Jacob Hacker and his colleagues developed one way to summarize these threats, calling it the Economic Security Index (ESI).⁴⁵ The ESI measures the proportion of Americans who have seen their available household income—household income after paying for health care and debt servicing—decline by at least 25% from one year to the next and who lack an adequate financial safety net to make up the shortfall. It thus takes into account not only income increases and decreases that stem from changes in employment status and hours worked but also medical spending (which can rise when income declines if the income decline is due to lost employment and lost health insurance) and liquid household wealth.⁴⁶

As of this writing, the ESI was available from 1986 until 2011. Consistent with the general trends described over the previous few pages, the lowest level of insecurity was observed in 1986, when 14.3% of Americans faced such significant financial swings. And, just as with unemployment and involuntary job loss, the ESI is cyclical. It increased during recessionary periods—to 16.9% during the early 1990s recession, 18.8% during the early 2000s recession, and over 20% during the Great Recession in the late 2000s.

But, there has also been an overall steady increase in insecurity even during robust periods of economic growth. Thus while the proportion of insecure Americans may have declined after recessionary periods, it did not then retreat to its previous low. For example, using the definition put forth by the ESI, a higher proportion of Americans would be considered economically insecure in every single year between 2003 and 2007 than was the case in the years leading up to the recession in the early 2000s. This means that Americans were increasingly likely to experience a household income drop with inadequate wealth to buffer the fall.

SUBJECTIVE PERCEPTIONS OF FINANCIAL THREATS

Several objective indicators support the contention that Americans face no shortage of threats to their financial well-being. But just because the objective indicators point in that direction does not mean that Americans themselves see things that way. Thus we would like to get a better sense of how Americans evaluate their economic circumstances. Are they actually worried about their jobs, paying for health care, saving enough for retirement, and paying for college?

To unpack these questions, I rely upon several data sets. One asks workers to assess their degree of job insecurity. The objective data suggest that workers should be more worried about losing their jobs and especially about finding comparable opportunities should job loss occur. Do they see it that way? Since 1977, the General Social Survey has gauged perceptions of job insecurity using two questions: how likely is it that they will lose their job or be laid off in the next twelve months, and how easy would it be for them to find another job with approximately the same income and fringe benefits. Combining these into one indicator finds that, as with involuntary job loss as measured in the Displaced Workers Survey, perceived job insecurity is cyclical and closely tied to the objective unemployment rate. However, there is also evidence of an upward trend in perceived insecurity over time from 1977 until 2006, once we control for the unemployment rate and sociodemographic changes in the composition of the labor force.47 To be sure, the actual estimate represents less than 10% of the workforce during this time and thus I do not wish to overstate the point. Nevertheless, it is also likely that these perceptions have become more widespread since 2006 because of the Great Recession.

Other surveys include subjective measures of job insecurity in addition to perceptions about other financial threats. One is the Survey of Economic Risk Perceptions and Insecurity (SERPI), a study conducted in March 2009 and then again in September 2009. It included a series of questions that assessed Americans’ concerns about many of the financial threats described earlier.⁴⁸ At a time when the actual unemployment rate hovered close to 10%, the survey found that almost 40% of Americans were either fairly or very worried about losing their job or finding a new one. Moreover, close to 45% reported a similar level of worry about out-of-pocket health care costs and cuts to their health insurance, and just over half held similar worries about financial health during retirement.

The fact that worries ran high as the United States climbed out of the worst financial crisis since the Great Depression is perhaps not surprising. But what about beforehand? Is it just the case that Americans were unusually concerned during the Great Recession or were their worries a continuation of fears they had held for a long time?

One way to assess this is to examine the Kaiser Health Tracking Poll, which tracked subjective perceptions annually between 2004 and 2010.⁴⁹ Among other topics, the Kaiser poll asks about Americans’ worries about losing their job or losing their health insurance coverage. As we might expect, the responses to both questions show a slight uptick in concern in 2009 and 2010 relative to previous years. But even in those previous years a sizable number of Americans expressed concern. In fact, in every year between 2004 and 2008, approximately 40% reported being either somewhat or very worried about losing their job, and almost 50% were somewhat or very worried about losing their health insurance coverage.

Another piece of pre–Great Recession evidence comes from a survey conducted in February 2007 by the Rockefeller Foundation.50 This survey asked about a wider range of subjective perceptions than did the Kaiser poll but lacks the over-time breadth. Here again the data paint a portrait of widespread concern. Among employed Americans, one-third (31%) were at least slightly worried about losing their job. In addition, paying for health care was a big concern regardless of whether people had insurance. Among those with insurance, 44% were worried about being able to afford out-of-pocket health care costs and 36% were worried about being able to afford prescription drugs. The percentages were significantly higher among those without insurance. Lastly, just over half (56%) of all Americans worried about having enough money to retire on, and 65% of those with children expressed concern about paying for college education.

I have stated at several points that the objective trends are by no means limited to people with low income. The Rockefeller data on subjective perceptions echo this point. In general, the prevalence of concerns about job security, health care costs, retirement, and college costs are approximately the same for upper-middle-income Americans (defined as those with family incomes between $58,000 and $92,000) as they are for all Americans on average. Within this upper-middle-income band, among those who were employed at the time of the survey, almost one-third were at least slightly worried about losing their job. Among those with health insurance, almost half (47%) were at least slightly concerned about being able to afford out-of-pocket health costs and 39% were at least slightly concerned about being able to afford prescription drugs. Almost two-thirds (64%) expressed some concern about having enough money to retire on, and almost two-thirds (65%) of those with children expressed concern about being able to afford college tuition.

While it is important to highlight the broad reach of concerns about financial threats, I do not mean to suggest that there are no important differences by income group. Quite the contrary, particularly once we single out low-income Americans and examine differences in concerns held by men and those held by women. Consider the following.

As expected, concerns about financial threats rang loudest among those with low income. For respondents in the Rockefeller survey’s lowest income bracket (defined as those with an annual family income less than $19,000), just over half (51%) of those who were employed expressed concern about losing their job; among those with health insurance 61% expressed concern about paying out-of-pocket health expenses and 55% expressed concern about affording prescription drugs; 66% expressed concern about not having enough money to retire on; and fully 78% of those with children expressed some concern about paying for college. In several respects, however, these numbers are slightly misleading. They understate the depth of concern because low-income respondents were far more likely to report that they were “very worried” as opposed to “fairly” or “slightly” worried. Also, with regard to health care costs, one of the major barriers to affordable health care for this group of respondents is not having insurance in the first place (unless their income is low enough to qualify for Medicaid and they have enrolled in the program). Lastly, although the prevalence of their retirement concerns matches that of the overall sample, it is reasonable to question whether retirement is truly a salient concern for very low-income people who are likely experiencing difficulty with more pressing daily expenses.

Researchers have also used the Rockefeller data set to compare how perceptions of insecurity differ by gender.51 They found that women were significantly more concerned than men about being able to afford out-of-pocket health care expenses and prescription drugs, not having enough money to retire on, and not being able to pay for their children’s college education. These differences between men and women ranged from 8 to 13 percentage points, and all were statistically significant.52

Taken together, the SERPI, Kaiser poll, and Rockefeller Foundation survey all point in the same direction. A sizable number of Americans do, in fact, perceive threats to their financial well-being. And while such concerns certainly became more widespread during and after the Great Recession, they were also present beforehand. If they did not hold those perceptions, then that might be the end of the book. We would just explain any inaction on those issues by noting that there is not much concern about them. Alas, that is not the case.

FINANCIAL THREATS AND POLITICAL IMPORTANCE

Just because people subjectively perceive particular threats to their financial well-being does not mean that they view them as important political issues. They may view these issues simply as problems that are properly dealt with at the level of the household—perhaps people simply need to develop better skill sets as an insurance policy against job loss, or perhaps they need to get smarter about planning for health emergencies and/or retirement. I have yet to demonstrate any sense in which people believe that economic insecurity issues deserve attention (let alone redress) in the public square. Investigating this is the purpose of the remainder of the chapter.

The majority of my data on political issue importance comes from people’s responses to the following question: “What do you think is the most important problem facing this country today?” The question format is open-ended, and thus respondents are not provided with a preset list of possible issues to choose from. Instead, they volunteer whichever one is most important to them. This question, often referred to as the “most important problem” question, appears in several different surveys that stretch back to the middle of the twentieth century. Some facilitate the construction of a time series to show how particular issues (e.g., the cost of health care) have risen and fallen on the political agenda. Others include individual-level questions that permit building explanations of how people come to view certain topics as important political issues. I leverage both types in what follows.

Advantages and Disadvantages of the “Most Important Problem” Question

Given that I will be relying upon the most important problem question for the remainder of this chapter as well as again in Chapter 6, it is prudent to spend some time discussing its advantages and disadvantages. I already mentioned one major advantage: its widespread presence in publicly available data sets. In addition, its format offers several other advantages (as well as some notable drawbacks).

One of its other major advantages is that because it is open-ended, it avoids any concerns that decisions by the survey interviewer are skewing the results. If instead the question included a set of possible issues to choose from (and even if it allowed write-in responses), several concerns would arise. We would worry that respondents might believe that they have to choose one of the options provided, even if the list did not include what they would consider their “real” most important problem. This kind of bias can arise for several reasons. Respondents might not be motivated to think much about the question and might instead just fall back on the options provided. Alternatively they might reason that the interviewer chose which issues to include in the response set because these are the only reasonable options, and so social desirability pressures might weigh on them. Either way, a purely open-ended response avoids these concerns. It allows respondents to “define [their] own issue space by naming the issues that [are] salient to them.”⁵³

Another advantage of this question is that it is highly conservative. Respondents can label only one issue as most important, and so people are required to make trade-offs. This means that it probably provides a lower bound of the number of people who, at some level, consider an issue as an important political problem. There are probably others who find the issue important but it simply does not rise to the level of “most important” for them. An alternative situation would be one in which respondents are asked about issues on an issue-by-issue basis. Questions like this often ask people how important each issue is to them using a five-point response scale with options ranging from “extremely important” to “not at all important.” One potential problem with this format is that it allows people to say that every issue is extremely or very important (and, indeed, this is typically what happens).54 It does not require respondents to put forth an actual political agenda, which would entail trade-offs between issues. Indeed, people can report that an issue is important in the abstract yet, when pressed, quickly indicate that the issue is not very important relative to others. For example, in 1976 94% responded to a survey indicating that unemployment was an important issue, but then when asked to rank its importance relative to nine other issues a plurality (44%) said that inflation was more important and 27% reported that neither inflation nor unemployment was one of the top four issues facing the country.⁵⁵

Having mentioned the advantages of the most important problem question, it is also important to keep in mind possible concerns. One is that responses might be “superficial” and just indiscriminately reflect information that people were recently exposed to.⁵⁶ This possibility seems plausible in light of work on media effects and agenda setting, which finds that people’s responses to this question do strongly mirror trends in news coverage.⁵⁷ Investigating this possibility, John Geer provides some reason for optimism, finding that open-ended responses do not simply reflect recent information in news sources but instead take into account their prior beliefs.⁵⁸ Other work shows that people’s open-ended responses reflect consideration of “substantive” concerns such as personal experiences. People who are personally affected by a problem are more sensitive to news coverage about it (i.e., they exhibit greater agenda-setting effects).⁵⁹ Despite findings such as these, I recognize that many readers might still remain skeptical about how meaningful responses to the most important problem question really are. For this reason, later in the chapter I will provide several pieces of evidence suggesting that economic insecurity concerns are not simply superficial responses reflecting recently heard information.

Lastly, although we might reasonably assume that an issue like the economy is always an important issue to citizens, it would be labeled as the most important problem only if it’s not healthy. Given this consideration, although I will typically refer to responses to the most important problem question as a measure of issue importance, I do so with the implied understanding that this is shorthand for “an important issue that is perceived as the most important problem.”60

The “Most Important Problem” from 1952 to 2012

Having covered preliminary considerations, I now dive into the key substantive question: To what extent do Americans see economic insecurity issues as important issues? The first way in which I address this question is the simplest, just focusing on the proportion of Americans who stated that an economic insecurity issue was the most important problem facing the country (and, in particular, to what extent that number is above zero).

I begin by leveraging the extensive time series afforded by the Gallup Organization, which has asked the most important problem question for decades. I use responses collected over a time span of sixty years—from January 1952 to January 2012—which provides a rich data set covering the period of heightened insecurity as well as two decades prior.⁶¹ This data set has 274 surveys. The data from 2001 to 2012 are monthly, while earlier surveys were conducted three or four times per year. My discussion of the Gallup data focuses on three major facets of economic insecurity: jobs/unemployment, the cost of health care, and the cost of retirement. The cost of retirement includes occasional mentions of pensions, but otherwise it refers to Social Security and Medicare. Note that the fourth form of economic insecurity, the cost of college, is not covered here. While general mentions of education commonly appeared in the Gallup time series, it is not always clear whether the respondent meant the quality of education (e.g., the quality of K–12 public schools) or higher education costs. This stood in contrast to health care, for example, in which respondents specifically mentioned (and Gallup specifically coded) “cost of health care” as an issue. Thus, to be conservative, I have chosen not to include the data on education here. Later in the chapter I present issue-importance data from a different source that more precisely coded education costs.

Figure 2.1. Percentage of Americans reporting that unemployment is the most important problem facing the country (Source: Gallup poll). Gray bars indicate periods of recession.

Unemployment

Figure 2.1 presents the Gallup results for unemployment. Each dot represents the percentage of people in a given survey who responded that jobs or unemployment was the most important problem facing the country.⁶² To help situate pieces of the time series, I have also included gray bars that indicate periods of recession.⁶³

At the most basic level, the data in Figure 2.1 clearly support the claim that unemployment is an important issue for a sizable number of Americans, particularly since the 1970s. This is the first major takeaway point. Moreover, looking specifically at the mid-2000s, we can see implications of people’s concerns about losing their jobs that were reflected in the Kaiser and Rockefeller surveys discussed earlier. Even though the economy was growing, concerns related to unemployment ran high.

A few other patterns stand out. The first is the high degree of fluctuation in the data that is clearly tied to the business cycle. Both during and immediately following economic contractions, as businesses lay off workers to adjust to reduced demand, the proportion of people who mention jobs or unemployment as the most important problem rises significantly. Moreover, these fluctuations are not particularly wild. From one survey to the next, even early in the time series when there are relatively long periods between surveys, the proportion of people mentioning jobs or unemployment is typically only incrementally larger or smaller than the one before (with the direction of the change reflecting the survey’s location in the business cycle). The major exceptions to this pattern of incrementalism are the large changes during and immediately following the recession in the early 1980s, one that was widely considered at the time to be the worst since the 1930s and in which the peak unemployment rate exceeded 10% for the first time since the days of the Works Progress Administration.

The Cost of Health Care

Figure 2.2 displays the proportion of respondents who said the cost of health care was the most important problem.⁶⁴ The results—in particular the differences in the time series between early and later—are striking. Almost no one mentioned the cost of health care prior to 1980 in any survey, and the only times it was mentioned during this period were not during recessions. This is presumably not because no one fell ill prior to the 1980s. Instead, this issue simply did not rise to the top of the political agenda in people’s minds.

By the time the debate over the Clinton health care plan was heating up in the early 1990s, however, views had changed. A sizable proportion of Americans reported that health care costs were the most important problem throughout the 1990s and 2000s, regardless of whether the economy was growing. Again focusing on the mid-2000s, and situating these data relative to the Kaiser and Rockefeller studies, we see that people’s concerns about these costs were also reflected in the Gallup surveys. The number of Americans labeling health care costs as the most important problem steadily increased during the economically robust years leading up to the Great Recession. This is consistent with the objective trends mentioned earlier. Health care costs, while certainly tied to the availability of insurance, which itself is tied to employment, impact American families even among those with insurance and even when the economy is growing.

Figure 2.2. Percentage of Americans reporting that the cost of health care is the most important problem facing the country (Source: Gallup poll). Gray bars indicate periods of recession.

Retirement Security

As a final look at the sixty-year time series of Gallup data, Figure 2.3 reports the proportion of people who said that some aspect of retirement security was the most important problem facing the country. In most cases this means that they mentioned Social Security or Medicare, though some respondents mentioned pensions as well.⁶⁵ A few patterns are noteworthy. The first is that people were less likely overall to place retirement security issues at the top of the political agenda than either job security or health care costs. Second, focusing not on levels but on changes over time, we see a similar pattern as we saw with health care costs. Retirement security issues were almost never mentioned until the 1980s, but since then they were mentioned more frequently and the likelihood that they were mentioned was not as closely tied to the economic cycle as was unemployment. The pattern is tied to politics to some degree, though not entirely. On the one hand, the most mentions of retirement security occurred near President Bush’s reelection in 2004, which was precisely the time at which he proposed a partial privatization of Social Security (that would allow Americans to divert some of their Social Security taxes to individual accounts). But on the other hand, we do not observe a spike in concern about retirement security in the early 1980s when President Reagan initially proposed reductions to Social Security spending (including reducing benefits for early retirees, delaying cost-of-living adjustments, tightening disability requirements, and reducing future retirees’ benefits). Nor do we observe a spike in the lead-up to or aftermath of Congress’s adoption of amendments to the Social Security Act in 1983 that reduced benefits and raised taxes.

Figure 2.3. Percentage of Americans reporting that the cost of retirement is the most important problem facing the country (Source: Gallup poll). Gray bars indicate periods of recession.

Taken together, the Gallup data paint a reasonably consistent picture. Since the 1970s, Americans have routinely mentioned three of its facets as the most important problem facing the country. Even using this highly conservative measure of issue importance in which people are able to mention only one issue, it is clear that three of the core issues of economic insecurity have quickly become highly important in many people’s minds.

As a final word on the Gallup data, I want to underscore that my goal here is not to offer an airtight explanation of the over-time trends. While the data are consistent with people being more likely to mention economic insecurity issues over time, the patterns are also consistent with people being less likely to mention other kinds of issues. As Christopher Wlezien finds, just as the objective changes in economic insecurity were occurring, there was also a steep drop-off in the number of people saying that foreign policy–related issues were the most important problem (as the Vietnam War and then the Cold War drew to a close).66 I cannot use these data to separate out the possible explanations, nor is it my aim to do so, as the existence of such alternative explanations does not threaten the key point I wish to make. The mere fact that a sizable number of people are labeling these issues as most important suggests that, at some level, they would like to see the government address them first and foremost. It does not mean that people were entirely unconcerned about them during the 1950s or 1960s, but at the very least it does suggest that people found them to be less important than other issues at that time.

For Whom Are Financial Threats Politically Important?

Knowing that many Americans place economic insecurity issues at the top of the political agenda is a key piece of the story, but it is only the tip of the iceberg. We would also like to dig deeper and see whether such issue importance is systematically related to people’s personal experiences with insecurity. Such an investigation helps sort out whether people’s concerns about these economic insecurity issues are more sociotropic in nature—about people “out there”—or whether they hit closer to home.

It would be reasonable, perhaps even intuitive, to expect a link between personal life and political judgments. After all, personal life is something that we experience every day, and it would seem to be a natural (and cognitively manageable) foundation for one’s political beliefs. As it turns out, however, evidence linking personal experience and public opinion is remarkably scarce. Looking at people’s opinions on policy matters, examples abound in which the link is not as strong as we might expect: insurance coverage is a poor predictor of support for national health insurance,67 employment status does not exert a large impact on people’s support for a government jobs program,⁶⁸ one’s likelihood of being drafted had only a weak relationship with one’s beliefs about the Vietnam War, and whites’ beliefs about busing to achieve racial integration were not affected by whether they were parents or nonparents,⁶⁹ homeowners or renters,⁷⁰ or lived in a neighborhood that would likely be affected by busing.⁷¹ To be sure, there are some notable counterexamples. Homeowners were far more likely to support California’s Proposition 13, a 1978 ballot initiative in which California voters slashed property taxes and restricted annual increases.⁷² In addition, in their 2009 surveys of Americans’ experiences with economic insecurity, Hacker, Rehm, and Schlesinger found that people’s experiences with unemployment and unexpected health care expenses (including losing one’s insurance) significantly predicted their support for risk-buffering policies.⁷³ These latter findings support the general contention that the relationship between personal circumstances and policy opinions is limited to situations in which the policy is easily framed in terms of those circumstances, either from the perspective of the citizens themselves (e.g., because it offers clear benefits or imposes tangible costs)74 or from elite speakers with incentives to use strong personal experience frames.⁷⁵

But what about people’s beliefs about the most important problem facing the country? Here scholars have found that unemployment status, proximity to retirement, and race are all linked to people’s beliefs about issue importance.⁷⁶ Thus there is good reason to suspect that personal experiences with insecurity could be quite impactful. But an investigation of a wider range of economic insecurity experiences and issues is necessary before reaching any firm conclusions.

To conduct such an analysis I rely upon the American National Election Study (ANES), a survey that includes at least one thousand interviews with a random sample of Americans during and after each presidential election. Several of the ANES questions have appeared over many years, including an open-ended most important problem question similar to Gallup’s.⁷⁷ Here I focus on the ANES cumulative data file, relying in particular on data from the 1980–2008 time period. This is obviously far more restricted than the Gallup data, though several considerations motivated this decision. The major one has to do with data limitations: while the ANES asked the most important problem question prior to 1980, other questions necessary for my analysis were not asked earlier.⁷⁸ In addition, starting in 1980 means that my analysis approximately matches the time period associated with the objective rise in economic insecurity on the indicators mentioned earlier in the chapter and, as shown in the Gallup data, the time period when we are more likely to observe people mentioning these issues as the most important problem facing the country. In total, I have data on Americans’ perceptions of the most important problem facing the country in seven presidential elections: 1980, 1984, 1988, 1992, 1996, 2000, and 2008 (2004 is excluded because the most important problem question was not available that year).

As a first cut with the ANES data set, Tables 2.1 and 2.2 simply list each of the issues that people found to be the most important one facing the nation; 2008 appears in a separate table because at the time this book was written the raw data for 2008 were publicly available but had not been coded by the ANES. I thus assembled a research team to code the 2008 responses separately.⁷⁹ Both tables include all issues that were mentioned by at least 1% of the respondents.

TABLE 2.1
Most important political issue, 1980–2000 presidential election years (percentage of respondents mentioning issue, only including issues that were mentioned by at least 1% of respondents)

Issue	Percentage
Economic Issues
Economy (general mention)	6
Unemployment	12
Cost of Education	3
Social Security/Medicare	2
Cost of Health Care	3
Cost of Housing	2
Poverty	2
Welfare	2
Inflation	7
Recession/Depression	1
Government Budget (incl. deficit and spending)	16
Taxes	2
U.S. Foreign Trade Balance	1
Natural Resource Issues
Conservation of Natural Resources	1
Pollution	1
Public Order Issues
Abortion	1
Narcotics	4
Crime/Violence	3
Moral/Religious Decay	4
Foreign Affairs Issues
Foreign Relations/Foreign Affairs	1
Latin America	1
Iran	2
Preventing War	3
Obligation to Take Care of Problems at Home First	1
Disarmament	3
Nuclear War	2

TABLE 2.2
Most important political issue, 2008 (percentage of respondents mentioning issue, only including issues that were mentioned by at least 1% of respondents)

Issue	Percentage
Economic Issues
Economy (general mention)	45
Unemployment	7
Cost of Health Care	2
High Prices (incl. gas)	1
Cost of Housing	1
Recession	1
Stock Market	1
Government Budget (incl. deficit and spending)	5
Taxes	1
Social Issues Immigration	1
Public Order Issues Honesty	1
Foreign Policy Issues
War	13
Troops	1
Terrorism	3
International Image	1
Functioning of Government Political Cooperation/Partisanship	3
Corruption	1

A few patterns are noteworthy. Economic issues were by far the most commonly mentioned type of issue. The percentages of people mentioning specific issues such as unemployment, the cost of health care, and retirement were approximately the same on average as in the Gallup data. The key exception to this was in 2008, in which the plurality of respondents encapsulated their most important problem in terms of “the economy” as opposed to something more specific. This is perhaps not surprising in light of the fact that during the fall of 2008 when respondents were being interviewed, the events in the news included some of the largest one-day drops in the history of the stock market, unprecedented bank failures and government bailouts, and daily information about workers losing their jobs, losing their health insurance, and losing their retirement savings. In addition, many of the respondents were no doubt personally experiencing these events and not just hearing about them in the news. It all added up to a gloomy picture in which it was likely hard to pick just one economic problem—everything related to the economy seemed to be going wrong at the same time.

Are Political Concerns Indicative of Personal Circumstances?

In addition to the most important problem question, the ANES includes a wide variety of socioeconomic and political information about each respondent. I begin by investigating who finds the issues most important by focusing only on those who placed unemployment, health care costs, retirement, or college costs at the top of the political agenda. My goal here is to try to approximate how likely it is that rhetoric about the issue could be self-undermining for these people—that is, how likely it is that it reminds them of a financial constraint they are currently facing or could be facing in the future. For example, a tenured professor who states that unemployment is the most important problem facing the country is probably doing so from a vantage point of concern for others, given the extraordinarily high job security that tenured professors enjoy. The situation would be different for a member of the contingent workforce who, even if he is currently employed, may still feel highly threatened by the possibility of unemployment. In some cases, the ANES does not include personal experience questions that allow us to know for certain whether individuals are facing a form of insecurity in their personal life. It does, however, include measures of other attributes of people’s economic situation that are likely indicative of such experiences.

With that in mind I argue that the likelihood that rhetoric on a given “most important issue” reminds them of a personal financial constraint depends upon two things. The first is if they have actually had some experience that is indicative of that aspect of economic insecurity. For example, if someone reports that they have recently delayed medical or dental care, then it is likely that the cost of health care is something that affects them personally. The second is even if they have not recently faced such a shock (or if the ANES does not include data on such shocks), is it likely that a negative economic shock would materially affect their financial well-being? The ANES includes objective and subjective indicators of this latter situation. The objective one is annual household income, which reflects the fact that people with middle to low incomes are typically less able to cope with economic shocks than are those with higher income.80 The subjective one is whether people perceive that their financial situation has gotten worse over the previous year. The idea here is that how people respond psychologically to their objective situation depends upon their reference point, and so an economic shock might be magnified when their situation has already gotten worse.⁸¹

With these considerations in mind, I define individuals as most likely to be facing a given aspect of insecurity in their personal life if they have household income that is below the 67th percentile, report that their household finances are worse now than they were a year ago, and/or have had some relevant personal experience with that issue of insecurity. To be frank, using the 67th percentile cutoff as an indicator of susceptibility to “being reminded of a financial constraint” may strike some readers as quite high and I admit that it is not an ideal measure. But two considerations motivated this decision. First, the data from the ANES cumulative file include very large income buckets, such that the next lower one would have meant a cutoff at the 33rd percentile, which would be too low. The second reason is that, based on the data on subjective perceptions presented earlier in the chapter (in the Rockefeller data set), it seems reasonable to assume that people up to and including the 67th percentile of the income distribution could be reminded of a financial constraint. Indeed, it is worth underscoring that the range of data for upper-middle-income Americans from that survey included those who were even above the 80th percentile.

My measure of personal experience differs depending upon which issue is at stake. For unemployment, it is whether or not someone is currently unemployed or, among those who are currently employed, whether they were laid off or had their hours reduced in the previous six months. For health care, it is whether individuals have had to delay medical or dental treatment in the previous year because of the cost. For retirement costs (including concerns about Social Security and Medicare), it is whether or not the respondent is at or close to retirement age. For education costs, it is whether or not the respondent has children under the age of eighteen (and, if so, how many). Note that, with one exception, my focus is on particular experiences rather than more general demographic characteristics. In the one case in which I use a demographic characteristic as my focal measure of personal experience (retirement costs), there are clear institutional reasons to infer what a self-interested response would look like.82

Table 2.3 contains two data points for each issue. The first notes the percentage of respondents from 1980–2008 that reported it as the most important problem. The second reports the percentage of those people who are likely facing that form of insecurity in their personal life (given the measure just outlined). The pattern is clear. In every case except one, the overwhelming majority of respondents who report that a particular issue is the most important appear to be facing it themselves. And even in that one exception (the cost of education), it is still the case that the skyrocketing cost of college reminds a sizable majority of people mentioning the issue of a personal financial constraint.

These results provide an initial indication that the political is personal when it comes to economic insecurity issues. The people who place an insecurity issue at the top of the political agenda are almost always likely to be facing it in their own lives. This also means that any attempts to mobilize them on the issue will likely entail reminding them of a precarious aspect of their financial situation. As I will show in later chapters, this attribute turns out to be critical.

TABLE 2.3
Issue concerns and personal experiences

Jobs/Unemployment (All Respondents)
Percentage mentioning this as most important	10.7
Percentage of those people facing it in personal life	94%
Jobs/Unemployment (Respondents Currently Working)
Percentage mentioning this as most important	10.8
Percentage of those people facing it in personal life	98%
Cost of Health Care
Percentage mentioning this as most important	3.1
Percentage of those people facing it in personal life	90%
Cost of Retirement (All Respondents)
Percentage mentioning this as most important	2.8
Percentage of those people facing it in personal life	100%
Cost of Retirement (Respondents Less than 65 Years Old)
Percentage mentioning this as most important	2.6
Percentage of those people facing it in personal life	100%
Cost of Higher Education
Percentage mentioning this as most important	3.3
Percentage of those people facing it in personal life	70%

Note: These data come from the 1980–2008 ANES for all years in which the “most important problem” question was asked and in which the other data are available. The measurement of “facing it in personal life” is as described in the text.

Does Personal Experience Affect Importance?

Understanding whether the people concerned about the issues are likely to be facing them in their own lives is a critical piece of the story. But another important aspect is whether personal experiences actually help explain why people find issues to be important in the first place. In other words, are people who are experiencing a facet of economic insecurity more likely to believe that it is the most important problem than those who are not facing it? For example, are people who have delayed health treatment more likely to report that the cost of health care is the most important problem facing the country than those who have not delayed treatment? And are people who have had their hours reduced or are currently unemployed more likely to report that jobs/unemployment is the most important problem than those who have not experienced either?

In addition to its substantive importance, such an analysis is useful from a methodological point of view. Earlier I mentioned that one potential concern with the most important problem question is that it might yield superficial responses that simply reflect the most recently heard information. One way to examine that is to look at over-time stability. Joanne Miller and her colleagues did this recently and found remarkable stability in citizens’ political priorities over a multimonth period.83

Yet another way to assess the “superficiality” of responses is to see if people’s experiences with insecurity actually affect their political priorities. My empirical strategy for doing so is twofold. First, I test for bivariate relationships between personal experiences with economic insecurity (being unemployed or not, delaying health care or not, etc.) and the proportion of people reporting that an issue is the most important problem facing the country. This is useful as a first cut at the data because it is, at base, my primary interest. Yet in some cases this first look at the data may prove unsatisfying. We might observe evidence of a relationship but nevertheless be concerned that it is really the result of some other factor(s) that might be causing people to have certain experiences and to label certain issues as most important. Or even if we are not concerned about such confounds, we might wish to estimate the effects with heightened precision by taking into account other sources of variation in people’s political attitudes (especially when we have measures of individual-level attributes that might be responsible for that variation). Thus I present both bivariate and multivariate results for each of the four issues. If the results are consistent, then they should provide readers with greater confidence than either by itself. To the extent divergence arises, I leave it to readers to decide which provides a more compelling explanation.

Two concerns motivate my choice of control variables in the multivariate models. One is the fact that social and political identifications can drive people’s perceptions of issue importance. Whereas personal experiences reflect the extent to which an issue connects with one’s material well-being, these broader identifications capture the fact that our perceptions of political priorities may be tied to reference groups as well as longstanding values that guide political choices. To account for this, I include a set of demographic variables as well as measures of the respondent’s political ideology and partisanship. Partisanship is particularly important given that many of these issues concern social welfare, an area in which the Democrats are often perceived to have an advantage.84 Thus not only are elite Democrats more likely to focus on them, but citizens who identify as Democrats might also be more likely to state that they are the most important problem. I also control for aspects of the media environment.⁸⁵ In this case, the ideal data would include information about the specific sources of political news that each respondent was exposed to as well as the content of those sources in the time period leading up to the interview. Such detailed data are simply unavailable in the ANES over the full time period that I am examining. Instead, I include measures of media consumption that gauge how much the respondent gains political information from newspapers, magazines, radio, and television. I also include a dummy variable for each year to acknowledge that some campaigns are more “about” issues of economic insecurity than others. Such differences are reflected in the broader media environment (and thus in any given year the baseline likelihood of mentioning some facet of economic insecurity might be higher or lower).⁸⁶

The Importance of Jobs and Unemployment

First I consider whether personal experiences with job insecurity increase the likelihood that people report unemployment as the most important problem. The ANES includes several relevant questions. One is whether people are currently unemployed; 5.9% of respondents across the years of my data set placed themselves in that category. Two other questions are geared toward currently employed respondents, including whether in the previous six months they have been laid off or had their working hours reduced; 11% of currently employed respondents reported being laid off and 12% reported having their hours reduced in the previous six months.

Figures 2.4A–C contain bar graphs showing these three attributes and the proportion of people who said that jobs or unemployment was the most important problem facing the country.⁸⁷ In two cases—unemployment and hours reduced—we have strong evidence that negative experiences in the labor market increased the likelihood that people identified jobs or unemployment as the most important problem (z = 5.54, p < .01, and z = 4.27, p < .01, respectively). We do not, as of yet, have similarly strong evidence for the effect of being laid off (z = 1.25, p = .21).

Next I look at whether such bivariate relationships hold once we control for demographic and political attributes.⁸⁸ The full model results appear in Appendix A (in Table A.1) and include several different specifications.⁸⁹ With respect to personal experience, the substantive results largely confirm the bivariate relationships. Both being unemployed and having had one’s hours at work reduced in the previous six months significantly increased the likelihood that people placed unemployment at the top of the political agenda. Those who were unemployed were 2.8 to 4.1 percentage points more likely on average to report that this issue was most important (depending upon which model specification I used). In 2008 this would have translated into 6.5 to 9.5 million voting-age Americans.⁹⁰ The corresponding number among those who had their hours reduced was 2.3 to 3.6 percentage points, which would have translated into 5.3–8.3 million voting-age Americans. One way to assess the potential political significance of these numbers is to compare them with the membership size of well-known citizen groups that advocate on broad-based issues. One example of such a group would be Common Cause, one of the nation’s most active and well-known citizen groups working across multiple issue areas such as money in politics, elections, and government accountability. As of 2013, Common Cause reported that it had “nearly 400,000 members and supporters” and was “financed primarily by the dues and contributions of its individual members.”⁹¹ Set against a figure like this, the potential political significance of numbers in the multiple millions becomes clear.

Figure 2.4. Personal experience and issue importance (with 90% confidence intervals marked)

Lastly, holding constant demographic and political factors does not change the fact that among those currently employed we have no evidence that enduring layoffs changed people’s political priorities. One possible reason why this type of negative experience might not leave such a strong political footprint is that these people were actually employed at the time of the survey. They may therefore view any labor market adversity as no longer applying to them (as compared with, for example, people who have had to put off medical care, who might still very much be experiencing illness and/or uncertainty about their health).92

The Importance of Health Care Costs

Turning to the cost of health care, the personal experience variable of interest is whether respondents had to put off medical or dental treatment during the previous year due to cost concerns.⁹³ Across the data set, almost one-third of respondents (29%) said that this applied to them. And, as shown in Figure 2.4D, 4.7% of them reported that health care is the most important problem as compared with 3.8% who did not have to delay care (though this difference did not reach conventional levels of statistical significance; z = 1.28, p < .20).

What happens if we estimate the relationship holding constant other social and political factors? The full multivariate results again appear in Appendix A (in Table A.2).⁹⁴ Unlike the experience of being laid off, I find that once I hold constant other demographic and political factors it becomes clear that delaying care does affect political priorities. People who delayed care were on average 1.2 percentage points more likely to say that the cost of health care is the most important problem facing the country, which would have translated into approximately 1.9 million voting-age Americans in 1992.⁹⁵ This is a sizable number. Again using a well-known citizen group simply as a point of comparison, the Sierra Club (the nation’s oldest and arguably most influential grassroots citizen group working on environmental issues) reported 1.4 million members as of 2013.96

The Importance of Retirement Security

Next I consider whether people who are closer to retirement are more likely to say that issues related to retirement security—in particular Medicare and Social Security—are the most important problems facing the country. This includes people who are worried about whether or not they will receive promised Social Security benefits and whether Medicare will sufficiently cover their medical expenses as they grow old. Figure 2.4E displays various age categories and the likelihood that the respondent’s top political concern was financial well-being in retirement. This probability increases steadily during the working years and then levels off precisely at the age when people start to retire and become eligible for Medicare and Social Security. If they are not already concerned about the financial security of the aged by the time they retire, they do not (on average) become more concerned at that point.

One potential issue with drawing firm conclusions about the relationship between age and retirement concerns is that poor and middle-class senior citizens rely upon Social Security and Medicare much more than those with higher income. Among seniors in the bottom two income quintiles, Social Security makes up over 80% of their annual income. For those in the top quintile, the percentage is only 18%.⁹⁷ These differences might affect people’s political priorities because they signal the degree to which people are dependent upon government transfers for their economic livelihood. Bivariate results that just look at proximity to retirement might therefore be missing a critical piece of the story. To account for this, in the multivariate models I include an interaction between age and income so that the effect of growing older may be estimated for people at different income levels.⁹⁸ For a given age, I would expect the largest differences to emerge among people close to retirement age.

Figure 2.5. Predicted probability of mentioning cost of retirement as the most important problem (by age and income). Income levels are defined as follows: “low income” means having household income up to the 33rd percentile, “middle income” is between the 33rd and 67th percentiles, and “high income” is above the 67th percentile. These predicted probabilities come from Model 3 in Table A.3 (which only includes people under the age of sixty-five).

The full model results appear in Table A.3, including separate models for all respondents versus just those below the retirement age of sixty-five.⁹⁹ Figure 2.5 displays the substantive results of interest: the predicted probability that a respondent mentioned the cost of retirement as the most important problem based on her age and income. For the purpose of the main text, I focus only on people below retirement age because most proposals (both currently and historically) to alter Social Security and Medicare do not affect benefits to current retirees (for example, the cuts to Social Security signed into law by Jimmy Carter, the bulk of the cuts signed into law by Ronald Reagan, and more recent privatization proposals).

The results in Figure 2.5 provide additional evidence consistent with the basic bivariate pattern but also add some important nuance. Among those with low to middle income, there is evidence of a positive relationship between age and political priorities. But no such evidence emerges among those with the highest income. Comparing across income levels, this difference turns out to largely reflect the political priorities of those nearing retirement. Individuals in their twenties and thirties, who face far more pressing concerns than a far-off retirement, appear to be equally unconcerned about retirement costs regardless of their income. But among those who are on the verge of retirement, strong income differences emerge in which those at the lower end of the income spectrum exhibit heightened concern.

These results are also significant from a substantive point of view. For example, among those near retirement age, 2–4% of respondents (depending upon their income) are likely to view retirement security as the most important problem facing the country, which would translate to 1 to 2 million Americans over the age of sixty-two.¹⁰⁰ On the one hand, that number might strike some readers as not particularly large in the realm of retirement issues. AARP, after all, boasts approximately forty million members. But it is important to keep in mind that AARP is able to recruit so many people because of the insurance discounts and other valuable member-only benefits that it offers (that are typically unrelated to its political activities). Perhaps a more apt comparison would be a well-known advocacy organization such as the National Abortion Rights Action League (NARAL) that, although its focus is not retirement issues, does appeal to potential members primarily with reference to collective political goals. According to its website, NARAL has “more than one million members and supporters.”101 This number is in the same ballpark as that reported in the results in Figure 2.5, suggesting that these numbers are potentially important from a substantive point of view.

The Importance of Education Costs

The final aspect of economic insecurity that I examine is the cost of education.¹⁰² An ideal measure of personal experience in this domain would include a respondent’s amount of student loans and/or current tuition bills. Unfortunately the ANES does not include such data. Instead I rely upon a different variable that reflects the likelihood that a respondent would be faced with college costs (either now or in the not-too-distant future): the number of children he or she has under the age of eighteen. Although one might argue that I should restrict attention to those who have high-school-age children, the general awareness of the skyrocketing cost of college along with the increased marketing of college savings plans means that even parents of young children routinely receive messages related to the cost of college (of the “the sooner you start saving, the better” variety). Against this information background, using the number of children as a measure of personal experience seems reasonable.

Figure 2.4F contains a bar graph showing a respondent’s number of children and the political priority of education costs. On average, the proportion of people placing the cost of education at the top of the political agenda is lower than that for the other economic insecurity issues, though it is above zero regardless of family size. But personal experience does not appear to have much effect here, as there is no evidence of differences in the likelihood that people politicized education based on the number of children they have. The results from the full multivariate model appear in Table A.4, but the predicted probabilities based on one’s number of children appear in Figure 2.6. Here, too, there are no statistically significant differences even after controlling for other social, political, and media factors. Following the logic of the previous section, I also investigated the possibility of an interactive effect. Perhaps having more children matters only among people who do not have particularly high income, and thus it is misguided to look for an effect without taking that important attribute of people’s lives into account. Although it is intuitively reasonable, I did not find any evidence for such an interactive relationship.

Figure 2.6. Predicted probability of mentioning cost of education as the most important problem (by a respondent’s number of children) (90% confidence intervals are marked). These results are based on Model 1 in Table A.4.

Upon reflection, one possible reason why there might be little evidence linking personal experience to issue importance in this domain is that it is actually upper-middle-class families—those reporting annual incomes between $95,000 and $205,000, but also those who tend to have fewer children—that have witnessed the largest increase in student loan debts in recent years.¹⁰³ These are the parents who are more likely to send their children to expensive schools but are also less likely to receive large amounts of financial aid. While it is certainly possible that the effect of personal experience might appear stronger with a more ideal measure that took into account the size of such tuition bills, for now the data I have do not provide evidence of a relationship.

CONCLUSION

When it comes to citizens’ social and political attitudes, American politics is often viewed as reflecting something other than personal experiences. In the realm of economic insecurity, especially when it comes to unemployment, health care costs, and retirement, the data presented in this chapter suggest otherwise. Using the Gallup data over a sixty-year timespan, I found that job insecurity, health care costs, and retirement insecurity were often mentioned as the most important problem. Using the data from the ANES, I delved deeper and found evidence that such concerns are highly likely to reflect financial constraints that people are personally facing. I also found evidence that personal experiences with financial threats can affect people’s perceptions of issue importance. Taken together, this set of findings is useful for showing how economic insecurity can leave a clear footprint in American public opinion when it comes to issue importance. The next step is to unpack the behavioral consequences.