Conversable Economist

The US spends about 18% of GDP on health care. Other high-income countries spend an average of about 11%, Thus, the Society of Actuaries and Henry J. Kaiser Family Foundation have created Initiative 18/11 to consider ways of holding down US heath care spending. A first report from the initiative, \”What Can We Do About the Cost of Health Care?\” (January 2019), doesn\’t yet offer proposals for action. But it offers a useful sense of what many of the main targets are likely to be of any serious effort to reduce healthy care costs. Here are some of my own reactions and takeaways from the report.

Although the report doesn\’t emphasize this point, it\’s worth noting that 18-11=7, and with a US GDP of about $20 trillion, 7% is $1.4 trillion. If you find yourself wondering why other countries can apparently afford improvements in physical infrastructure or higher support for education or higher social benefits of certain kinds,  part of the reason is that a much lower level of their social resources is going to health care spending. If you wonder why your paycheck seems to go up so slowly, part of the reason is that your employer keeps paying more for you health insurance. My point here isn\’t that a substantial share of US health care spending is wasteful or even harmful, although that does seem to be true. It\’s just that output going to health care has a tradeoff of less output available for other uses–even when some of those other uses might do more to improve health.

When thinking about cutting health care spending, and obvious approach is to look at how the money is being spent, and the bulk of health care spending is on those with multiple chronic conditions. The Kaiser report notes:

\”Remarkably, 86 percent of health care spending is for patients with one or more chronic conditions—conditions expected to last three months … Among the chronic population, people with more than one condition account for 71 percent of total spending. The cost of chronic diseases goes far beyond the direct amounts spent on these diseases. In the United States, seven out of every 10 deaths are caused by chronic diseases each year. There are indirect costs through lost productivity and an unmeasurable loss in the quality of life and the loss of ability to perform activities of daily living, such as bathing and eating. For adults, the most prevalent conditions are uncontrolled hypertension (uncontrolled blood pressure) and hyperlipidemia (high cholesterol and high triglycerides). For children, the most common conditions are allergies and asthma.\”

One way of thinking about chronic conditions is that if they are managed properly (medicine, diet, exercise. whatever is needed),  then health care costs are usually low. But if such conditions are not not managed properly, very expensive episodes of hospitalization become likely.

For example, just looking at noncompliance with taking prescriptions drugs, the report notes:

In a 2011 Consumer Reports survey, one of the leading complaints among primary physicians is that patients do not take the doctor’s advice or follow treatment. For example, although 3.8 billion prescriptions are written every year, more than 50 percent of them are not taken or are taken incorrectly. The cost of noncompliance has been estimated at $290 billion. Also, 125,000 deaths each year are attributed to poor medication compliance.

The report is also bracingly honest about what can be accomplished by going after \”indirect\” costs. The report notes:

It can be hard to draw broad lessons about macroeconomics from the experience of one country alone, or from the experience of one or two recessions. Thus, a group of researchers including Òscar Jordà, Katharina Knoll, Dmitry Kuvshinov, Moritz Schularick, and Alan M. Taylor have been working to compile macroeconomic and financial data for 16 high-income countries going back to 1870.  Alan Taylor provides a readable overview of several puzzles that emerge from the long-run financial data in \”The Rate of Return on Everything\” (NBER Digest, December 2018, pp. 20-23).

(The detailed research behind this short article is available as National Bureau of Economic Research Working Paper #24112: Ò. Jordà, K. Knoll, D. Kuvshinov, M. Schularick, and A. Taylor, \”The Rate of Return on Everything, 1870–2015,\” December 2017. It\’s also available as a Centre for Economic Policy Research Discussion Paper #12509: Jordà, O, K Knoll, D Kuvshinov, M Schularick, and A M Taylor (2017), “The Rate of Return on Everything, 1870–2015.” A short readable overview of the work that is very similar to the version I\’m referring to here, with the same title but listing  Òscar Jordà, Katharina Knoll, Dmitry Kuvshinov, Moritz Schularick, Alan Taylor as the authors appeared at the VOX website on January 2, 2018).

#1 The Housing Puzzle

In general, economists would expect that assets with more risk–that is, more likely to rise or fall over time–will tend to have higher returns on average. From the standpoint of investors, the higher returns are needed to make up for the higher risk. This logic suggests that over the long run, a risky asset with volatile prices like corporate stock should have a higher average rate of return than a less risky asset with less volatile prices like housing. But that doesn\’t seem to be true. The blue line shows returns to housing, while the black line shows returns to corporate stock across the 16 countries in this sample.. Corporate stock is more volatile, but the average rates of return are quite similar.

Why might this pattern hold true? One possibility is that the risks for housing are higher than they at first appear, because it\’s harder to diversify the risks of owning housing, and perhaps also because it\’s harder to buy and sell housing quickly or in incremental chunks when prices change. But these factors don\’t seem nearly enough to explain the housing puzzle.

#2 The \”Safe Rate\” Puzzle

A lot of theories in finance and macroeconomics start with idea of a \”safe\” investment that pays a low rate of return but also has low risk. A common example would be investing in US Treasury debt, where the risk of default is near-zero. The theory then discusses how the safe assets might be combined with riskier assets. The puzzle is that \”safe\” assets like government debt actually can have quite volatile rates of return, once factors like inflation are taken into account. Here\’s a figure showing international returns on government debt.

For a concrete example, think about US experience since the 1970s. When inflation went way up in the 1970s, it mean that those who were holding government debt paying a low fixed rate were experiencing negative real returns for a time. The nominal rates paid on government debt rose by the early 1980s, but then when inflation declined substantially, those holding the \”safe\” asset for a time had substantially positive real returns for a time. Since then, a combination of declining nominal interest rates and low inflation have meant a steady decline in the real rate of return on \”safe\” assets. In real terms, the \”safe\” rate doesn\’t look all that safe.

Indeed, if you look at the \”risky\” assets like housing and corporate stock, but focus on moving averages over any given ten-year period rather than annual returns, the returns on the \”risky\” assets actually look rather stable.

#3 The r > g Question

If wealthy people can invest and receive a rate of return r, while the economic grows at a slower rate g, then wealth might grow faster than the economy over time (at least if wealthy people don\’t spend all of the returns on wealth), leading to greater inequality of wealth. This was a common interpretation of the work of Thomas Piketty in Capital in the Twenty-First Century on causes of income and wealth inequality a few years back. The findings here are that returns on risky assets like stocks and housing are often twice as large as rates of economic growth, or more.

But interestingly, Piketty himself doesn\’t view this r>g dynamic as central to the processes that generate wealth inequality. In an article her wrote for the Winter 2015 issue of the Journal of Economic Perspectives, \”Putting Distribution Back at the Center of Economics: Reflections on Capital in the Twenty-First Century,\” Piketty commented:

\”[T]he way in which I perceive the relationship between r > g and wealth inequality i soften not well-captured in the discussion that has surrounded my book—even in discussions by research economists. … I do not view r > g as the only or even the primary tool for considering changes in income and wealth in the 20th century, or for forecasting the path of income and wealth inequality in the 21st century. … I certainly do not believe that r > g is a useful tool for the discussion of rising inequality of labor income: other mechanisms and policies are much more relevant here, for example, the supply and demand of skills and education.  …  The gap between r and g is certainly not the only relevant mechanism for analyzing the dynamics of wealth inequality. As I explained in the previous sections, a wide array of institutional factors are central to understanding the evolution of wealth. Moreover, the insight that the rate of return to capital r is permanently higher than the economy’s growth rate g does not in itself imply anything about wealth inequality. Indeed the inequality r > g holds true in the steady-state equilibrium of most standard economic models …\”

What are some of the main factors that affect the rise or fall of wealth inequality over time? Examples would include taxes on wealth, the extent to which wealth is saved or consumed, and even the birth and death rates of the population, which affects how long concentrations of wealth will stay together and how many slices they will be divided into when passed to a new generation. There are questions about the extent to which whether the new fortunes being created in businesses around the globe will displace earlier fortunes, and whether the new fortunes will be long- or short-lived. There are also events of history like World Wars, and events of politics like surges of populist sentiment. For more on these topics, see \”Piketty and Wealth Inequality\” (February 23, 2015), or the four-paper symposium on these issues in the Winter 2015 issue of the Journal of Economic Perspectives.

Simon Kuznets (Nobel 1971) usually gets the credit for doing as much as anyone to organize our modern thinking about what should be included in GDP, or left out. But I had not known that Kuznets apparently argued for leaving military spending out of GDP, on the grounds that it wasn\’t actually \”consumed\” by anyone, but should instead be treated as an intermediate input that supported production and consumption. Here\’s how Hugh Rockoff tells the story in his essay, \”On the Controversies behind the Origins of the Federal Economic Statistics,\” in the Winter 2019 issue of the Journal of Economic Perspectives. [Full disclosure: I work at JEP as Managing Editor.]  Rockoff writes:

Military spending presented another problem. In one of his last discussions of national income and product before US entry in World War II, Kuznets (1941, pp. 19–20) explained that his estimates included “dreadnoughts, bombing planes, poison gas, and patent medicines because they are rated economic goods in our country today,” even though they “might well be considered worthless and even harmful” in a society organized differently. In a footnote, Kuznets (p. 31, fn. 5) used an analogy with private spending to buttress his case for including military expenditures: “If the activities of the private police used by many large corporations are productive, why not those of the municipal police? And if of the domestic police, why not of the international police, i.e., the armed forces of the nation?” During World War II, however, Kuznets (1945) modified his thinking. He argued that military spending should be counted in national product during a time of total war, but it should be excluded during peacetime because military spending was then an intermediary good for producing a flow of consumption to consumers. Other economists, including decisively those at the Department of Commerce, thought otherwise (Gilbert, Staehle, Woytinsky, and Kuznets 1944).

A number of economists, however, have found Kuznets’s concept of a Peacetime National Income to be attractive. Higgs (1992), for example, argued that the then-current interpretation of the impact of World War II on the American economy, that it created unprecedented prosperity, was reversed when one used Kuznets’s peacetime concept rather than the conventional measure. Higgs even took exception to Kuznets’s decision to include some military durables such as aircraft in investment because Kuznets thought that they could later be turned to peacetime purposes. 

In retrospect, a number of concerns weighed against adopting Kuznets’s concept of peacetime national product. One reason, as Coyle (2014, p. 20) suggests, was the rise of Keynesian economics. In principle, one could use Kuznets’s peacetime version of national product to analyze the macroeconomy, but the conventional measure fit more smoothly into the simple Keynesian model taught to a generation of economics students in Samuelson and other textbooks. Perhaps the most important reason for rejecting Kuznets’s concept, however, was the Cold War. In his famous study of productivity, Kendrick (1961, p. 25) chose to include all defense spending in his estimates of national product partly on the grounds that “national security is at all times [Kendrick’s italics] a prime objective of economic organization.” In political terms, excluding national defense from national product would create the appearance that the government’s statistical agency was siding with the critics of America’s defense budget. Of course, no one was required, as Kuznets had pointed out, to use only one measure of aggregate product. To the contrary, Kuznets thought that it would be best to produce a series of measures, some specialized for one purpose and some for another. But as we have learned, public attention does tend to focus on a single measure of national product, so the decision to ignore Kuznets’s peacetime concept may have had important consequences.

I find myself in agreement with the views of Kuznets expressed back in 1941, that if private security guards and municipal police are in GDP, the military should be, too.

But more broadly, the dispute serves as a useful reminder that GDP includes some categories of expenditures that society would have preferred not to make. For example, GDP includes all measures for home security and corporate security–not just guards but also locks, bars, and electronic measures. In addition, GDP includes cleaning up after pollution spills and natural disasters, although it would certainly have been preferable if such events had not happened in the first place. It would also be socially beneficial if people got more exercise and at healthier diets, and as a result a substantial proportion of health care spending didn\’t need to happen.

For other comments on the relationship between GDP and social welfare, readers might be interested in the well-known comments from \”Robert Kennedy on the Shortcomings of GDP in 1968\” (January 30, 2012). My own sense is that economists are well-aware of the shortcomings of GDP–indeed, probably better aware of the shortcomings than many critics. But economists also point out that on a wide array of dimensions, people who live in societies with higher GDP tend to live better lives. For samples of these arguments, see \”Why GDP Growth is Good\” (October 11, 2012)  and \”GDP and Social Welfare in the Long Run\” (April 6, 2015).

A wealth tax is what it sounds like: a tax imposed not on income, but on wealth. The standard economic definition of wealth includes both nonfinancial assets like real estate and financial assets like stocks and bonds. Thus, a wealth tax doesn\’t care if the value of someone\’s wealth went up or down in the last year/ It is not a tax on the transfer of wealth to others, like an inheritance tax or a gift tax. It is just imposed on the amount of wealth.

In the US, property taxes are a cousin of a part of a broader wealth tax, in the sense that they are imposed annually on the value of a property, whether the value rises or falls. But they are not at true wealth tax in the sense that they don\’t differentiate between someone who own their home debt-free–and thus all the value of the home is wealth–and someone who is still paying off the mortgage, where only the equity you have in your home is wealth. The inheritance tax is also a form of a wealth tax.

Back in 1990, 12 high-income countries had wealth taxes. By 2017, that had dropped to four: France, Norway, Spain, and Switzerland (In 2018, France changed its wealth tax so that it applied only to real estate, not to financial assets.)  The OECD describes the reasons why other countries have been dropping wealth taxes, along with providing a balanced pro-and-con of the arguments over wealth taxes, in its report The Role and Design of Net Wealth Taxes in the OECD (April 2018).

For the OECD, the bottom line is that it is reasonable for policy-makers to be concerned about the rising inequality of wealth and large concentrations of wealth But it also points out that if a country has reasonable methods of taxing capital gains, inheritances, intergenerational gifts, and property, a combination of these approaches are typically preferable to a wealth tax.  The report notes: \”Overall … from both an efficiency and an equity perspective, there are limited arguments for having a net wealth tax on top of well-designed capital income taxes –including taxes on capital gains – and inheritance taxes, but that there are arguments for having a net wealth tax as an (imperfect) substitute for these taxes.\”

Here, I want to use the OECD report to dig a little deeper into what wealth taxes mean, and some of the practical problems they present.

The most prominent proposals for a US wealth tax would apply only to those with extreme wealth, like those with more than $50 million in wealth.  However, European countries typically imposed wealth taxes at much lower levels of wealth. Here\’s a table showing how much wealth is exempt from the wealth tax in European countries. Clearly, most countries with such taxes were applying them to wealth well below $50 million.

It\’s interesting, then, that in these European countries the wealth tax generally accounted for only a small amount of government revenue. The OECD writes: \”In 2016, tax revenues from individual net wealth taxes ranged from 0.2% of GDP in Spain to 1.0% of GDP in Switzerland. As a share of total tax revenues, they ranged from 0.5% in France to 3.7% in Switzerland … Switzerland has always stood out as an exception, with tax revenues from individual net wealth taxes which have been consistently higher than in other countries …\” However, Switzerland apparently has no property tax, and instead uses the wealth tax as a substitute.

The fact that wealth taxes collect relative little is part of the reason that a number of countries decided that they weren\’t worth the bother. In addition, it suggests that a US wealth tax which doesn\’t kick in until $50 million in wealth or more will not raise meaningfully large amounts of revenue.

Why do wealth taxes imposed on what seem to be quite low levels of wealth collect so little revenue in various European countries, especially during the last few decades when high-wealth individuals as a group have done pretty well? The answer seems to be that when countries impose a wealth tax, they often typically create a lot of exemptions for certain kind of wealth that aren\’t covered by the tax. Each of these exemptions has a reasonable-sounding basis.  But every exception also creates a potential loophole.

\”Ceiling provisions or tax caps are common features of net wealth taxes. These often consist in setting a limit to the combined total of net wealth tax and personal income tax liability as a maximum share of income. They are used to prevent unreasonably high tax burdens and liquidity constraints requiring assets to be sold to pay the net wealth tax. In France, the wealth tax ceiling (often referred to as the “bouclier fiscal”) limits total French and foreign taxes to 75% of taxpayers’ total income. If the percentage is exceeded, the surplus is deducted from the wealth tax. In Spain, the aggregate burden of income tax and net wealth tax due by a resident taxpayer may not exceed 60% of their total taxable income.\”

Steve Hanke has devoted considerable effort to building up data on hyperinflations during the last century or so. He offers a quick overview of this work in Forbes (January 20, 2019). Below is his list of hyperinflations. When it comes to Venezuela, he writes:

Now, let’s turn to the world’s only current hyperinflation: Venezuela. It ranks as the 23rd most severe. Today, the annual rate of inflation is 120,810%/yr. While this rate is modest by hyperinflation standards, the duration of Venezuela’s hyperinflation episode, as of today, is long: 27 months. Only four episodes of hyperinflation have been more long-lived.

Here\’s the table of all 58 hyperinflations:

For those who want more, here\’s an earlier discussion of \”Hyperinflation and the Venezuela Example\” (April 28, 2016). Here\’s a discussion of \”Hyperinflation and the Zimbabwe Example\” (March 5, 2012). And I offered an earlier discussion of the list from Hanke and Krus in \”A Systematic List of Hyperinflations\” (August 21, 2012).

In the long-run, the average standard of living in an economy is determined by the average productivity of its workers. For example, Paul Krugman started Chapter 1 of his 1990 book, The Age of Diminished Expectations, by stating: \”Productivity isn\’t everything, but in the long run it is almost everything. A country\’s ability to improve its standard of living over time depends almost entirely on its ability to raise its output per worker.\”

Thus, it matters a lot that US productivity growth slowed down back around 2005, even before the start of the Great Recession, and that we don\’t really understand why. The Congressional Budget Office laid out the issues in its recent report The Budget and Economic Outlook: 2019 to 2029 (January 28, 2019). CBO pointed out that productivity growth–which it refers to by its more formal moniker of TFP for \”total factor productivity\”–seems to be higher or lower for periods of some years, with abrupt transitions between these periods.

\”Over longer periods, however, years of comparatively steady TFP growth tend to be followed by rather abrupt transitions to years with steady but substantially different growth. For example, estimated trend growth in TFP remained  relatively strong in the 1950s and 1960s, slowed considerably from the early 1970s to the mid-1990s, and resurged in the late 1990s and early 2000s. Around 2005, a few years before\\ the recession and financial crisis that began in 2007, TFP growth again slowed in many industries and throughout the international economy. In CBO’s estimate, TFP growth in the domestic nonfarm business sector was only about one-third as rapid during the 2006–2017 period as it had been from 1996 to 2005.\”

In the aftermath of the Great Recession, CBO has been scaling back its productivity forecasts. The top line show the CBO productivity forecasts in 2012. The other lines show how the forecasts were reduced in 2014, 2016, 2018, and now in 2019.

Why is US productivity growth slowing down? CBO is forthright in admitting: \”[E]xtensive research has failed to uncover a strong, compelling explanation either for the slowdown or for its persistence …\” The report runs through a number of potential explanations, before knocking each one on the head.

Is the productivity slowdown a matter of measurement issues?

\”Even though mismeasurement of economic phenomena is widespread and persistent, measurement issues do not appear to have been substantially worse since 2005 than they were in the past and probably account for at most a small portion of the slowdown.\”

Is the productivity slowdown a result of slower growth feeding back to reduced productivity growth?

\”The slower growth of the labor force and of aggregate demand in the aftermath of the recession resulted in relatively modest demand for capital investment, slow turnover of capital stock, and slow introduction of new technologies in new plants and equipment. Nevertheless, there is little evidence of a backlog of technology that exists but is not raising output and productivity through its effect on capital stock, which suggests that slower economic growth did not feed back strongly into TFP …

Is it a result of less human capital for US workers, either as a result of less experience on the job or reduced growth in education?

\”Highly skilled and well-educated baby boomers are retiring, and the educational attainment of younger cohorts only modestly exceeds that of their predecessors—two demographic effects that could be restraining TFP growth. Higher-skilled workers tend to continue working longer than their predecessors, however, and younger cohorts made especially strong gains in educational attainment during the recession and the ensuing slow recovery. Both developments have tended to improve the average skill level of the aggregate labor force. As a consequence, growth of the estimated quality of the aggregate labor force since 2005 has been only moderately slower than growth over the preceding 25 years, and that slowdown has played at most a minor role in the overall slowdown in TFP growth.\”

Is the problem one of overregulation?

\”Declining dynamism in many industries, possibly exacerbated by increasing regulatory constraints, could be contributing to slower growth in TFP. Regulatory restrictions on homebuilding in denser, high-productivity urban regions could also be slowing TFP growth. Such problems have been developing slowly over time, however, and are difficult to associate with an abrupt slowdown in TFP growth around 2005.\”

Is the scientific potential for long-term innovation declining?

\”Some researchers believe that long-term innovation may be slowing as well and that the economy is `running out of ideas.\’ The costs of research and innovation are increasing, they argue, and the resulting new ideas are not as economically significant as past innovations. Again, no evidence exists of an abrupt change around 2005 connected to such developments. Moreover, other, more optimistic researchers conclude that the pools of potential innovators and the potential market for innovative products are now global, that research tools have greatly improved and communication of innovations has become much more rapid, and that major advances in technology can continue to be expected in the future, though they may diffuse through industry rather slowly.\”

When it comes to productivity growth, the great irony in our public discourse is that it\’s common to hear concerns that there is likely to be both too little of it and too much of it. The concern over too little productivity growth is that without productivity growth we won\’t have the economic strength both to offer job opportunities and rising wages to American workers–along with having the economic strength to devote resources to environmental protection, health and education, assisting the poor, and other issues. The concern over too much productivity growth is that a combination of robots and artificial intelligence will be so ultra-productive that they will greatly diminish the number of jobs for humans.

Of course, the scenarios of no-productivity-growth and the ultra-productivity-growth are not both going to happen. Personally, I\’m considerably more worried about the the problems of slow growth.

I was hired back in 1986 to be the Managing Editor for a new academic economics journal, at the time unnamed, but which soon launched as the Journal of Economic Perspectives. The JEP is published by the American Economic Association, which back in 2011 decided–to my delight–that it would be freely available on-line, from the current issue back to the first issue. Here, I\’ll start with the Table of Contents for the just-released Winter 2019 issue, which in the Taylor household is known as issue #127. Below that are abstracts and direct links for all of the papers. I may blog more specifically about some of the papers in the next week or two, as well.

The concept of the \”gig economy\” clearly captures something real, but it can be hard to measure define in the statistical sense that brings joy to my heart. For example, it clearly refers to those who drive for Uber and Lyft. But does it refer more broadly to all workers with \”alternative work arrangements,\” who are hired for a short-term job with no serious expectation that it will become a longer-term employment connection? Does it cover people who obtain jobs through a temp agency and work for a series of employers, for example?

Conventional labor statistics don\’t draw these kinds of distinctions very clearly, so researchers have looked for clues in less traditional kinds of data. The results are less clear-cut than one might like.

For example, back in 2015 two prominent labor economists, Lawrence Katz and Alan Krueger, noted that the usual tool for the U.S. Bureau of Labor Statistics in looking at alternative and nonstandards work arrangements is called the Contingent Work Survey (CWS). But this survey has only been carried out occasionally, and in fact had not been done since 2005. So Katz and Krueger tried to carry out a survey in Fall 2015, attaching the questions to a nationally representative survey regularly done by the RAND think-tanks called the American Life Panel.

The results were striking. As I noted at the time, they found:

[T]he percentage of workers engaged in alternative work arrangements – defined as temporary help agency workers, on-call workers, contract company workers, and independent contractors or freelancers – rose from 10.1 percent in February 2005 to 15.8 percent in late 2015. … We further find that about 0.5 percent of workers indicate that they are working through an online intermediary, such as Uber or Task Rabbit … Thus, the online gig workforce is relatively small compared to other forms of alternative work arrangements, although it is growing very rapidly  … A striking implication of these estimates is that all of the net employment growth in the U.S. economy from 2005 to 2015 appears to have occurred in alternative work arrangements.

Well, Katz and Krueger have now revisited the subject four years later, and their latest reuslts suggest that the number of workers in alternative jobs actually has not been growing at all. The more recent paper is \”Understanding Trends in Alternative Work Arrangements in the United States,\” published as a working paper from the National Bureau of Economic Research (#25425, January 2019, an ungated version is here).

The Bureau of Labor Statistics did get funding to do a follow-up of the official Contingent Workers Survey in May 2017. They write: \”The … findings were released in June 2018 and indicate, in seeming contrast to our earlier findings … a slight decline in the incidence in alternative work arrangements from 10.7 percent in 2005 to 10.1 percent in 2017 …\”

 What\’s going on here? Part of the answer seems to be that the labor market was improving from 2015 to 2017, and so more people moved into traditional jobs. Other differences have to do with the extent to which surveys were answered by workers directly, or by others in the family, and whether the 2015 survey for some reason ended up with an oversample of people holding multiple jobs. But bottom line, the current view is that the number of alternative workers may have risen during the Great Recession and its aftermath, but has since declined back to about where it was before the recession.

Anat Bracha and Mary A. Burke provide some additional perspective in \”The Ups and Downs of the Gig Economy, 2015–2017,\” written as a working paper for the Federal Reserve Bank of Boston (#18-12, October 2018).

The particular focus on Bracha and Burke is on data from the Survey of Informal Work Participation (SIWP), which was carried out as part of something called the Survey of Consumer Expectations, which is done on a monthly basis by the Federal Reserve Bank of New York. They added questions about the extent of \”paid informal work or side jobs,\” and whether \”websites and/or mobile platforms were used in finding and/or performing such work.\” The survey was not really intended to include professional freelancers or temp workers, but as with any survey, it\’s not always obvious how respondents interpret the questions. They find:

Considering either our broader or narrower concept of informal work, we find that participation rates did not change significantly on net between 2015 and 2017, while our point estimates suggest that if anything, the participation rates declined during that period. … Furthermore, conditional on participation in the gig economy, between 2015 and 2017 the average hours among informal workers showed unambiguous declines, and the aggregate amount of informal work as measured in FTEs also fell substantially, while average earnings among informal workers were effectively flat. Our composite measures of informal work would appear to contradict recent popular narratives depicting rapid growth in the independent workforce in response to structural changes in the organization of work. However, we also find some supporting evidence for the story of a rising gig economy …[P]articipation rates increased across our surveys for selected technology-enabled jobs such as ridesharing and online tasks, and the average ridesharing hours increased dramatically among drivers, although in absolute terms these technology-enabled jobs still account for only a very small segment of the US population of household heads.

In more detailed analysis, they also find hints that the improving labor market from 2015-2017 was tending to  hold down the amount of informal work. But their bottom line is that while ride-sharing is way up, along with AirBnB and various non-labor income-earning opportunities, overall informal work is not on the rise.

Other authors have also found evidence that while participation in ride-sharing jobs is way up, the evidence for more employment in other parts of the gig economy is weak. For example, Katharine G. Abraham, John C. Haltiwanger, Kristin Sandusky James R. Spletzer presented \”The Rise of the Gig Economy:Fact or Fiction?\” at the annual meetings of the American Economic Association in early January 2019. They point out that some of the main possible data sources on gig economy work are not telling the same story: specifically, the surveys of workers form sources like the Current Population Survey aren\’t telling the same story as income tax data about self-employment. They write:

This paper provides an overview of what we know and don’t know about the hypothesized surge in the gig economy. There has been phenomenal growth, confirmed by at least three independent data sources, in the number of self-employed passenger drivers since 2013. The pace of this growth illustrates how quickly new technology can affect the labor market. Outside of this sector, however, the picture is considerably murkier. Furthermore, … there has been a growing discrepancy between self-employment rates as measured in core household surveys such as the Current Population Survey (CPS) versus self-employment rates as measured in tax data. Over the past decade, the former show a slight decline whereas the latter show a notable increase. CPS data also have not captured the surge in passenger driver self-employment that is evident in other data. These facts suggest that, to understand the gig economy, the CPS and other core household surveys will need to be augmented by other types of data

Bracha and Burke from the Boston Fed make a similar point about data on alternative work arrangements from self-employment reported on income taxes. They write:

According to a few different studies, the filing of tax forms indicating self-employment, such as the Schedule C, increased significantly in recent decades (Jackson, Looney, and Ramnath 2017, Katz and Krueger 2016, and Abraham et al. 2018), and one study found that the trends were driven by an increase in independent labor rather than business ownership (Jackson, Looney, and Ramnath 2017). Likewise, an analysis by Dourado and Koopman (2015) of 1099-MISC forms, which are used to report income received outside of traditional employment relationships, indicates an escalation in such filings from 2000 to 2015.

One of the prominent studies of what tax data can tell us on the extent of self-employment is \”The Rise of Alternative Work Arrangements: Evidence andImplications for Tax Filing and Benefit Coverage,\” by Emilie Jackson, Adam Looney, and Shanthi Ramnath written as a working paper for the Office of Tax Analysis at the US Department of the Treasury (#114, January 2017). They note that the category of self-employment is quite broad. However, it is also rising fast, and many of the gains in self-employment seem to be among those who have zero in business expenses–that is, they are just providing labor. They write (footnotes omitted):

In 2014, 24.9 million individuals filed returns reporting the operation of a nonfarm sole proprietorship and 16.8 million earned a profit (and paid self-employment tax) from those activities, representing a 34 percent and 32 percent increase, respectively, from their levels in 2001. The almost 17 million self-employed workers represented 12 percent of all tax filers with earnings. … Self-employed individuals engage in a wide variety of economic activities ranging from operating businesses like restaurants, law offices, or partnerships; providing contract or consulting labor; earning platform-based or “gig economy” income; to house cleaning and babysitting services. Many earn income from both wages and self-employment. … Looking at trends over time, we find that essentially all of the increase in self-employment is due to increases in sole proprietors who have little or no business-related deductions, and who therefore appear to almost exclusively provide labor services (i.e. the contractors or misclassified workers). In contrast, the share of filers that were small business owners was essentially unchanged.

It seems clear that more people are receiving income and tax from activities that are outside traditional jobs. But other than ride-sharing jobs, just how to characterize these jobs remains murky, and the question of what rules and regulations might apply to such income-earning activities remains murky, too.  It feels to me as if a shift is happening in US labor markets, in which the expectation of a long-term bond between employers and employees has declined on both sides. But I don\’t yet feel that I understand the details of this shift.

Twice a year the Congressional Budget Office publishes a \”just the facts\” overview of the federal budget picture and the US economy. The latest version is \”The Budget andEconomic Outlook:2019 to 2029 (January 2019). Here, I\’ll focus on the US budget deficit and debt.

Here\’s the pattern of US federal government spending and revenues in the last 50 years. Average outlays during that time were 20.7% of GDP. Average revenues were 17.4% of GDP. Contrary to the widespread belief that US government spending and taxes have over time surged ever higher, to me the more obvious pattern here over the half-century is one of stability. Sure, government spending is higher and taxes are lower than the historical averages during the Great Recession. But during boom times like the late 1990s, taxes are above their historical average while spending is below. When President Trump took office early in 2017, US government spending and taxes were–whether for better or worse–almost bang on their long-run averages.

But under the surface, two changes are going on–one medium-term and one longer-term. The medium-term change is that the usual pattern over time has been that when the US economy is proceeding strongly, with sustained growth and a relatively low unemployment rate, the budget deficits are usually lower, or in the late 1990s even turned into surpluses. But at present, the trajectory is a relatively healthy economy but with larger-than-usual budget deficits.

This CBO figures shows that if one looks back at years when the unemployment rate was below 6%, the average budget deficit has been 1.5% of GDP. But although the current unemployment rate has been substantially below 6% for several years, the projected budget deficits for the next decade are projected at 4.4% of GDP.

The longer-term issue involves the rising share of government spending that is going to support benefits for the elderly. Here\’s an illustrative figure from the CBO on the share of federal (non-interest) spending going to the elderly. Specifically, federal spending on the elderly was 35% of total in 2005, 40% of the total by 2018, and headed for 50% of the total 2029.

The tradeoffs here are a matter of basic arithmetic. If federal spending on the elderly keeps rising, but but there is a goal of keeping total federal spending at about the same level, then other federal spending needs to be slashed. Indeed, overall federal investment has been dropping over time: less for support of R&D, less for infrastructure, less for support of training and employment.  It used to be a half-century ago, that the federal government spent more on investment than on the broad category of paying benefits, but the nature of the federal government has shifted substantially, and over time it has become primarily about paying benefits. As the figure shows, we are rapidly headed toward a situation in which half of all (non-interest) federal spending involves payments of benefits just to the elderly–not even counting benefits paid to those under age 65.

The figure above, showing overall patterns of spending and revenue, show the outcome. Spending keeps rising, driven in the medium-run by the rise in spending on the elderly. Despite the sustained economic growth and low unemployment rate, federal taxes dipped in the last year and will stay relatively low for a situation where the economy is doing fairly well, in substantial part because of the Tax Cuts and Jobs Act passed in December 2017. As this combination of spending and taxes leads to higher budget deficits, more government debt accumulates, and interest payments start climbing, too. For example, the CBO projections show that \”net interest\” is 1.6% of GDP in 2018, but rises to 2.6% of GDP by 2023.

Don\’t skip over that rise in interest payments too quickly. The US GDP is about $20 trillion in 2018. So 1% of GDP amounts to $200 billion which is being spent as a price paid for past borrowing–and thus won\’t be available for spending increases or tax cuts.

Combine these factors, and the US ratio of debt/GDP is headed out of its historical range. There have been spikes in the debt/GDP ratio before, notably in times of war, or during the large budget deficits during the Reagan presidency in the 1980s. But according to the CBO, using current law as its baseline, we are headed well outside those limits in the next couple of decades.

Presumably the answers involve restraints on spending programs for those over-65 or raising additional tax revenue. Those steps need to be phased in energetically even if the only goal is to get off the pathway of an extreme debt/GDP rise. In addition, if you would like to see the pendulum swing back at least a bit–so that the federal government can again take up its role of investing in R&D, infrastructure, and workers–then even greater changes are needed.

This pattern of a rising debt/GDP ratio raises a number of economic concerns. As the CBO report summarizes:

Such high and rising debt would have significant negative consequences, both for the economy and for the federal budget, including these:

• As interest rates continue to rise toward more typical levels, federal spending on interest payments would increase substantially;

• Because federal borrowing reduces national saving over time, the nation’s capital stock ultimately would be smaller, and productivity and total wages would be lower than would be the case if the debt was smaller;

• Lawmakers would have less flexibility than otherwise to use tax and spending policies to respond to unexpected challenges; and

• The likelihood of a fiscal crisis in the United States would increase. Specifically, the risk would rise of investors’ being unwilling to finance the government’s borrowing unless they were compensated with very high interest rates. If that occurred, interest rates on federal debt would rise suddenly and sharply relative to rates of return on other assets.

Economic growth in Mexico presents a puzzle. Mexico has followed many of the standard recommendations that are said to support economic growth. For example, it has prevented a recurrence of the inflationary fevers that used to grip Mexico every few years. Rates of national investment are up. Investment in education and human capital is up. Mexican workers have a  high labor force participation rate. Mexico has signed international agreements to reduce trade barriers. It has done a reasonable amount of privatization and deregulation And the result of all these changes has been slow growth.

 Santiago Levy describes this puzzle and offers his own answer in Under-Rewarded Efforts: The Elusive Quest for Prosperity in Mexico, published by the Inter-American Development Bank (July 2018). Here\’s Levy on Mexico\’s sluggish growth, which actually implies a negative rate of productivity growth in recent decades.

From 1996 to 2015, the country’s per capita GDP growth averaged only 1.2 percent per year. Moreover, this unimpressive figure arguably overestimates Mexico’s performance, as it reflects the fact that because of the country’s demographic transition, its labor force grew more rapidly than its population during these years (2.2 versus. 1.4 percent). In fact, GDP per worker grew on average by only 0.4 percent on an annual basis, far from what is required to create a prosperous country. … 

Over the same two decades, accumulated per capita GDP growth in Mexico was 25.7 percent, less than every country in Latin America except Venezuela. …Over the medium term, growth occurs because the labor force increases (in quantity and quality), because there is more investment in physical capital, and because the productivity of labor and capital (total factor productivity – TFP) increases. Decomposing Mexico’s growth over this period into these three components, one finds that TFP growth averaged only 0.14 percent annually, without any corrections for the quality of the labor force. Considering increases in schooling (that is, taking into account that workers with more years of schooling can potentially contribute more to output than those with fewer years), yields a negative TFP growth rate of 0.53 percent. … [T]he result is that Mexico’s GDP growth has resulted only from the accumulation of physical capital and growth of the labor force. There have been no improvements in efficiency. Thus, by and large the question of why Mexico grows so slowly is equivalent to the question of why productivity has stagnated.