Corporate Investment

“Are U.S. Companies Too Short-Term Oriented? Some Thoughts,” by Steven N. Kaplan. June 2017. NBER #23464.

Criticism of U. S. corporations for focusing on short- rather than long-term investment returns dates back to at least 1980 when Harvard Business School faculty members Robert Hayes and William Abernathy wrote in a landmark article, “By their preference for servicing existing markets rather than creating new ones and by their devotion to short-term returns and management by the numbers, many of them have effectively forsworn long-term technological superiority as a competitive weapon” (“Managing Our Way to Economic Decline, Harvard Business Review 58 (July–August): 67–77).

In this paper, Steven Kaplan marshals evidence that, in recent decades, firms generally have not fallen prey to corporate short-termism. First, corporate profits as a percentage of gross domestic product are near all-time highs and have been rising for the last 30 years, suggesting they are doing well at exploiting market opportunities. Second, if existing companies underinvest, then venture capital (VC) investors would fill the gap and be very profitable. But VC as a percentage of the total stock market has fluctuated in a relatively narrow range of 0.1 to 0.2% over time. VC returns are a bit above general stock market returns, but not by much. Private equity funds have a similar record, with a little blip for the internet boom of the late 1990s. That performance suggests firms are not passing up on promising investment opportunities.

Kaplan concludes with other stylized facts that are also inconsistent with the short-termism argument. The internet stock boom of the late 1990s was based on high long-term expected cash flows, which of course did not occur, but the dot-com investors were long-term oriented. Companies are increasingly less profitable at the time of their initial public offering. Amazon and Tesla have high values but no profits (in the case of Amazon, until recently), facts not consistent with short-termism. Some 180 biotech companies went public between 2013 and 2016, and only 4% were profitable. Finally, hydraulic fracturing technology to tap into natural gas reserves—fracking—was developed over recent decades despite long periods of negative cash flows.

—Peter Van Doren

Health Insurance

“Cost of Service Regulation in U.S. Health Care: Minimum Medical Loss Ratios,” by Steve Cicala, Ethan M.J. Lieber, and Victoria Marone. July 2017. SSRN #3007692.

Economists have long argued that traditional rate-of-return public utility regulation reduces the incentives of regulated firms to control costs. If a firm earns a guaranteed rate of return on capital investment, then the firm will be inclined to overinvest in capital because it is protected from downside risk.

In this paper, Steve Cicala and colleagues argue that the same logic applies to a provision of the Affordable Care Act. The act requires health insurers in the “fully insured” market (those insurers who bear financial risk rather than just administer claims for large self-insured employers) to spend 80% of their premium income on medical care and mandates rebates to consumers ex post if this does not occur. This provision was added to the law by consumer advocates and their political supporters who argued that some insurers retain too much premium income, make too much profit, and are stingy in approving coverage of medically necessary procedures.

Much like traditional rate-of-return regulation, this rule creates incentives for insurers to spend more on medical care rather than to reduce premiums. The paper estimates a difference-in-differences model in which firms that spend less than 80% of their premiums on medical expenditures are compared with firms that are in compliance with the rule. In the year before the rule was implemented, 52% of consumers in the individual market were in plans that spent less than 80% of premiums on medical expenditures and thus would not have been in compliance.

The paper concludes that the average effect of the rule on treated firms (those that previously had been spending less than 80% of premiums on medical claims) was to increase their medical expenditure outlays by about 7%. There was no reduction in premiums. —P.V.

Environmental Regulation

“Why Is Pollution from U.S. Manufacturing Declining? The Roles of Environmental Regulation, Productivity, and Trade,” by Joseph S. Shapiro and Reed Walker. August 2017. SSRN #3012564.

Between 1990 and 2000, the real value of U.S. manufacturing output grew by a third while emissions of the six regulated air pollutants (carbon monoxide, nitrous oxide, particulate matter, fine particulate matter, sulfur dioxide, and volatile organic compounds) fell by 35%. From 1990 to 2008, those emissions fell by 60%. Why did this occur?

Three possible explanations have been offered. The first is trade, i.e., the United States offshored pollution-intensive industries, while U.S.-based industry shifted toward less polluting products. The second is improvements in productivity that expand output while reducing the use of polluting manufacturing inputs such as fossil fuels. The third is environmental regulation. The paper concludes that almost all of the reduction in emissions over time stems from changes in emission intensity within vary narrowly defined manufacturing products. And that reduction, in turn, was not the result of trade-induced composition change nor productivity improvements. Instead, stricter environmental regulation resulted in the reduction of emissions. —P.V.

Employment Effects of the ACA

“The Effects of the Affordable Care Act on Health Insurance Coverage and Labor Market Outcomes,” by Mark Duggan, Gopi Shah Goda, and Emilie Jackson. July 2017. NBER #23607.

Health insurance coverage under the Affordable Care Act increased by 4.2 percentage points in states that expanded Medicaid and 2.6 percentage points in states that did not in the first two years of the ACA’s coverage mandate (2014–2015). For households with incomes between 100% and 400% of the federal poverty level, subsidies were available if the households purchased coverage from the health insurance exchanges created by the ACA. Many economists, including those from the Congressional Budget Office, predicted that because those subsidies depended on household income rather than individual wages, second earners in households would reduce their labor market participation to allow their households to qualify.

The authors of this paper found no change in the level or trend of aggregate labor market participation after the ACA. But this aggregate result was the product of two offsetting trends. There was an increase in labor market participation in areas of the country in which the share of people who were uninsured and earning under the poverty line was larger and a reduction in labor force participation in areas in which the share of people who were uninsured and earning between 139% and 399% of the poverty line was larger. “These changes suggest that middle-income individuals reduced their labor supply due to the additional tax on earnings while lower income individuals worked more in order to qualify for private insurance,” the authors conclude. “In the aggregate, these countervailing effects approximately balance.” —P.V.

Minimum Wage

“State Minimum Wage Changes and Employment: Evidence from 2 Million Hourly Wage Workers,” by Radhakrishnan Gopalan, Barton Hamilton, Ankit Kalda, and David Sovich. May 2017. SSRN #2963083.

I summarized some of the recent papers on the effects of minimum wage increases in the “Working Papers” section of the Fall 2015 issue and Ryan Bourne continues that discussion in this issue (“A Seattle Game-Changer,” p. 8). An important component of those discussions was a paper by Jonathan Meer and Jeremy West. They argue that changes in minimum wages do not cause an abrupt change in employment levels, but instead employers respond by slowing their future hiring and hiring higher-skilled workers, thereby reducing overall employment growth.

This paper analyses data on over 2 million hourly employees from over 300 firms for the years 2010–2015. It uses a difference-in-differences regression to compare six states that implemented a large (at least 75¢ per hour) increase in their minimum wage (California, Massachusetts, Michigan, Nebraska, South Dakota, and West Virginia) with states that didn’t pass such an increase.

For the treatment group of states, the fraction of employees earning less than $10 an hour declined 0.7% in the year following the minimum wage increase. But the fraction of workers making $10–$15 an hour increased. The overall result was that total employment didn’t really change. This is consistent with Meer and West in that the effect occurs in the form of slower hiring of the least-skilled rather than termination of existing employees because firing is costly. —P.V.

Airbnb and Housing Prices

“The Sharing Economy and Housing Affordability: Evidence from Airbnb,” by Kyle Barron, Edward Kung, and Davide Proserpio. July 2017. SSRN #3006832.

Airbnb has reduced dramatically the transaction costs of renting housing on a short-term basis. Some critics have argued that this has reduced the supply of housing available for long-term renters, thus exacerbating the housing affordability problem in major American cities.

In this paper, data on Airbnb listings from 2012 through 2016 at the ZIP code level are regressed on Zillow housing price and rental price information. Fixed effects for ZIP code and city-level time trends are included. The authors control for time-varying factors by ZIP code with a variable that measures Google searches for Airbnb at the ZIP code level interacted with number of restaurants and hotels in a ZIP code reflecting underlying tourist demand. The expectation is that landlords in more “touristy” areas will be the most likely to convert long-term rentals to Airbnb and reduce long-term rental supply and increase the rents of remaining housing.

The authors find that a 10% increase in Airbnb listings increases housing prices by 0.65% and rents by 0.38%. The annual rent increase in their data was 2.2% and the average ZIP code experienced a 6.5% annual increase in Airbnb listings. Thus from 2012 through 2016, only 0.25% of the 2.2% increase in annual rent was explained by Airbnb. The Airbnb effect is not zero, but it is small. —P.V.

Soda Taxes

“The ‘Soda Tax’ Is Unlikely to Make Mexicans Lighter: New Evidence on Biases in Elasticities of Demand for Soda,” by Mabel Andalon and John Gibson. May 2017. SSRN #2971381.

The tax on soda in Mexico has been hailed as reducing consumption with likely long-term health benefits. The Mexican soda tax is large (9% of pretax average prices). Estimates of the reduction in consumption assume that the tax is passed through to prices and that consumers react to prices only by reducing the amount consumed using the standard (elasticity) estimates of –1 to –1.3. That is, a 1% increase in the price results in a 1% to 1.3% decrease in consumption.

But another response of consumers to a price increase is to change to a cheaper soda brand and keep the quantity consumed constant. For example, before the tax, Coca Cola was priced 15% above Pepsi. After the tax, the Coke price is still 10.8% higher. Thus a consumer could respond to the tax by reducing Coke consumption or simply switching to Pepsi. The authors’ estimate of the true elasticity for soda consumption adjusted for the substitution of cheaper brands is much smaller (–0.2 to –0.3).

According to survey data, average soda prices increased 11.9% between 2012 and 2014. But the average price of purchased soda increased by half that rate, suggesting that consumers purchased lower-priced soda. When the corrected elasticities are used, the 2–4 pounds-per-person predicted weight loss advanced by some academic papers becomes less than a pound. —P.V.

Mortgages and the Financial Crisis

“Credit Growth and the Financial Crisis: A New Narrative,” by Stefania Albanesi, Giacomo De Giorgi, and Jaromir Nosal. August 2017. NBER #23740.

The conventional explanation of last decade’s financial crisis is that credit growth from 2001 to 2006 was concentrated in the subprime segment of the housing market even though there was no aggregate income growth in that group. This unwise allocation of credit to people with poor probabilities of repayment was exacerbated by the Great Recession. The subprime holders of mortgages disproportionately lost their jobs and couldn’t maintain their house payments, according to this theory. The most prominent citation for this view is a 2009 paper by Atif Mian and Amir Sufi (“The Consequences of Mortgage Credit Expansion: Evidence from the U.S. Mortgage Default Crisis,” Quarterly Journal of Economics 124[4]).

This paper argues that Mian and Sufi’s finding is the result of their decision to estimate borrowers’ credit status by using a 1996 ranking of ZIP codes by the fraction of residents below 660 credit score, along with their 1997 individual credit score. Low subprime scores are found disproportionately among young people with thin or nonexistent credit histories. Credit scores grow normally as people age and demonstrate success with paying bills on time. The research design used by Mian and Sufi conflates the normal life cycle growth in credit to those who were young before the boom with poor credit worthiness during the boom.

To avoid that problem, this paper uses individual-level credit scores calculated shortly before mortgage borrowing occurred rather than in 1996 or 1997. The more time-appropriate credit scores revealed that defaults among borrowers with low scores actually decreased during the Great Recession. The fraction of mortgage delinquencies accounted for by the lowest quartile of credit scores dropped from the normal 40% to 30% and the fraction of foreclosures from 70% to 35%.

Instead of credit growth to sketchy borrowers, this paper highlights the role of credit growth to “investors,” defined as those who hold two or more first mortgages. From 2004 to 2007, the share of mortgage balances held by investors in the middle quartiles of the credit score distribution rose from 20% to 35%. For investors, foreclosure rates did increase four-fold for the lowest quartile credit scores, but it increased 10-fold for the other three quartiles. The fraction of “investors” (those with more than one first mortgage) who became delinquent grew 30 percentage points for the lowest three quartiles and by 10 percentage points in the top quartile. For people with just one first mortgage, the foreclosure rate doubled in the lowest two quartiles and barely changed for the highest two quartiles.

The paper reproduces the Mian and Sufi ranking of ZIP codes by the fraction of subprime borrowers in 1999 and finds their result. Borrowers who resided in the top quartile of (subprime) ZIP codes in 1999 exhibited larger growth in per-capita mortgage balances. But the individual borrowers living in those “subprime” ZIP codes who were responsible for most of the credit growth actually were prime borrowers. This result reinforces the rule to avoid making inferences about individuals based on characteristics of aggregations. —P.V.

Retirement Income and Expenses

“Change in Household Spending After Retirement: Results from a Longitudinal Sample,” by Sudipto Banerjee. Employee Benefit Research Institute Issue Brief #420, November 2015.

“Retire on the House: The Possible Use of Reverse Mortgages to Enhance Retirement Security,” by Mark Warshawsky. Mercatus Center Working Paper, June 23, 2017.

The “retirement funding gap”—the deficit between the amount of money needed to provide for future seniors in retirement and the public and private money actually set aside for them—is the source of considerable angst in the policy world. But not all analyses of this gap are gloomy. While many say this number is large and growing, Andrew Biggs, a senior fellow at the American Enterprise Institute, contends in a September 2017 paper for the Mercatus Center that the number of seniors in poverty who failed to save sufficiently for retirement has been greatly exaggerated. Penn Wharton Public Policy Initiative scholar Jagadeesh Gokhale has noted that the living standards for seniors are generally higher than the rest of society.

A central question in this literature is how much income do retirees need in order to maintain their standard of living. We know the answer is less than 100% of their employment income: when people stop working they no longer have commuting costs, they don’t need to buy work clothes, and they no longer have to eat their lunches out. What’s more, a significant proportion of retirees have finished paying off their homes, leaving them with one less expense to worry about. However, they do have higher health care costs, and these costs tend to increase as they age. What’s more, about one in four people will spend significant time in a nursing home at some point in life, an expense that can be potentially ruinous to a family’s wealth.

In a 2015 paper for the Employee Benefit Research Institute (EBRI), research associate Sudipto Banerjee attempts to determine how people’s spending changes after retirement. He uses data from the Health and Retirement Study (HRS) and the Consumption and Activities Mail Survey—a supplement of the HRS—to estimate patterns in retiree spending. His top-line finding is—as we expect and as is consistent with other research—that spending declines as age increases. In these data sets, annual spending fell 19% between ages 65 and 75, and 34% between ages 65 and 85. However, the decline slows over the latter part of that age range, and then begins increasing around age 85.

Banerjee offers a few explanations for this pattern. Concerning the decrease in spending, one contributor is a decline in the number of people in a retiree-headed household. When a spouse dies or divorces and leaves the household, spending falls. A small but growing proportion of retiree-headed households will also have children at home at the beginning of this span; their subsequent departure contributes to the decline in spending. However, there is a large fixed component to household spending that’s independent of the size of the household. Concerning the late-life rise in spending, it correlates with an increase in out-of-pocket medical spending and is concentrated at the top of the income distribution.

The survey breaks down consumption into seven spending categories: home, food, health, transportation, clothing, entertainment, and “other,” which includes charitable contributions and gifts. From these data Banerjee makes three broad observations.

First, there is a large variation in spending across the country. Consumption is much lower in the South, for instance, than in the wealthier northeast. But on average, spending drops 5.5% in the first two years of retirement and continues to fall in the following years.

Second, housing comprises the largest component of total spending for households, even in retirement. Banerjee notes that households that pay off their mortgage do not necessarily see an immediate reduction in home spending; a common strategy is to channel payments that had gone to a mortgage into various home improvements such as a new furnace, new furniture, and the like.

Third, health expenses increase for senior citizens, although the rate of increase varies quite a bit across the country. Health care spending increases are much higher in the urban Northeast than in the South, a difference that may be driven by the relative shares of the population on Medicaid. On the other hand, transportation costs go down quite a bit as people retire, regardless of the locale.

Banerjee uses these data to offer a richer picture of spending in retirement. One realization: while the permanent income hypothesis (a manifestation of consumer rationality) would suggest that we should strive to keep our spending as constant as possible over our lifetime, the reality is that for most households consumption falls when we retire and our incomes fall, which implicitly repudiates the permanent income hypothesis.

This is not necessarily a troubling phenomenon. In retirement, our expenses fall and our potential leisure time increases, so we substitute more home production for buying things (like meals) from the market. Besides, some people do not reduce their spending when they reach age 65: fully 45% of all households see their spending increase immediately following retirement, and one-third of all households report higher spending five to six years after retirement. This increase is relatively even across the income distribution, suggesting that the increase is attributable to health issues.

These numbers raise the question of how much money people need for medical expenses in retirement. Investment firm Fidelity suggests at least $275,000, excluding long-term care costs. EBRI estimates that savings of $127,000 for a man and $143,000 for a woman would give a retiree a 90% chance of covering all health care expenses in retirement. For context, Banerjee reports that 46% of all senior citizens have had at least one overnight nursing home stay. On the other hand, only 23% of retirees had an out-of-pocket nursing home expense; in other words, half of all people who do have a nursing home stay manage to get out of the home before their 90 days of Medicare nursing home payments ran out.

Banerjee suggest that for a sizeable fraction of the elderly there is a real possibility that an extended stay at a retirement home or convalescent center could be financially calamitous without any long-term care insurance. He avoids making recommendations but one leaps from his pages: we should worry less about retiree income and more about health care costs.

One reason we shouldn’t worry so much about retiree income is that more retirees could make use of reverse mortgages, as Mark Warshawsky explains in a Mercatus working paper. Warshawsky, who earlier this year became the assistant commissioner for retirement and disability policy at the Social Security Administration, joins Gokhale and Biggs in believing that concerns over seniors running out of money in retirement are overstated. Previous research Warshawsky did with Gaobo Pang found a shortfall greater than Biggs, but still one smaller than what is generally assumed.

Warshawsky observes that analysts and the public often overlook the value of retirees’ houses when considering those retirees’ economic resources. This makes little sense because of the large amount of equity that seniors typically have in their houses, and because homeowners can use reverse mortgages to tap this wealth while remaining in their homes. He suggests that, like annuities, reverse mortgages are wrongly considered to be bad deals for homeowners. In this paper, he essentially investigates the broadest possible use of this product under some reasonable assumptions.

Specifically he has in mind a Home Equity Conversion Mortgage (HECM) loan, which is both issued and insured by the federal government. He would like for the government to allow it to be marketed by the private sector, which he believes would greatly boost its popularity.

HECM became a federal program in the late 1980s. While few people availed themselves of it in the 1990s, it took off at the end of the decade. There was no underwriting at first because the intrinsic value of the house made underwriting seem unnecessary, and the Federal Housing Administration did not worry about defaults. As a result, the federal government did take losses. The defaults caused some of the large banks, worried about their reputation, to stop offering the product.

HECMs are available only to seniors over age 62. They can get a lump sum, line of credit, or a monthly income flow, and they cannot borrow more than $640,000 or the value of their home, whichever is less. These are non-recourse loans; borrowers never pay back more than the value of the home when it is sold. They do not have to pay until the surviving spouse dies or moves out of the house.

HECMs are complicated, and the FHA requires counseling (at a cost of $150) prior to any purchase. The interest rate is LIBOR plus a 2.5% lender’s margin. Despite its complexity there is a market for this, Warshawsky has determined. There are 37 million households led by people over age 62, and 80% of them own homes. Just 2% of them have a reverse mortgage. Normally, Warshawsky notes in the paper, people are limited to withdrawing just half of the equity in a house with a reverse mortgage, which minimizes the risk of the lender not getting fully paid back after death.

If we exclude people in the bottom 30% of the wealth distribution, those with a home value under $100,000 (for whom the transaction costs would be too great to make a reverse mortgage worthwhile), and those in the top 20% of the wealth distribution (who won’t need such a product and would benefit more from a life annuity), we are left with approximately 14% of the population of seniors who could potentially benefit from an HECM.

So why aren’t many of these seniors using HECMs? Warshasky notes that a reverse mortgage is relatively expensive, but there are other forces at work. Jonathan Skinner, a Dartmouth economist, says people consider their homes to be a wealth stock of last resort and are loathe to touch home equity except in an emergency. And, of course, people have a bequest motive.

Thomas Davidoff at the University of British Columbia hits upon something else at play in this decision, noting that the ability of the elderly to use their home equity explains why there’s so little demand for long-term-care insurance. Warshawsky’s solution is to tighten up Medicaid eligibility rules for nursing home coverage, which would force more people to buy long-term-care insurance. Freed of that obligation, more seniors would see the advantage of tapping their home equity.

Warshawsky’s message is that the home is a valuable asset for many people and that more of them should tap some of its equity, and we should make it easier and less costly to do so. If we were to achieve that by spurring more families to buy long-term-care insurance, thereby lessening the burden on Medicaid, that would be a win–win outcome. —Ike Brannon