Biden-Harris policies and their consequences were no surprise to those paying attention

Milton Friedman used to advise researchers to focus on large policy changes rather than attempting to separate a small change’s signal from the noise. In this sense, the “ambitious” policy agenda of the Biden-Harris administration was expected to be a gift to the research community.

Accepting this gift, since 2020 I have been making forecasts of some of the consequences of those policies. Now is a good time to assess the accuracy of those forecasts, which relate to aggregate labor markets, insurance price controls, and drug price controls.

As promised, Biden and Harris redistribute income with health insurance expansions and student-loan forgiveness, although not necessarily in the Robin Hood direction. They give union bosses more tools for reducing competition in the labor market. They try to regulate the internet as a public utility. They distort healthcare markets in many ways, including a new ban on short-term health insurance plans, and granting selected companies a monopoly on a generic drug. They go significantly further than the Obama administration in terms of requiring the private sector to change behavior in ways the bureaucrats expect to reduce carbon emissions. At great human-capital expense, they enabled teacher unions and blue-state governments to maintain “social distancing” far longer than warranted.

The exhaustive list would have more than 1,000 entries.  Overall, even the federal agencies’ own low-ball estimates of the costs of the regulations finalized 2021-24 are almost $2 trillion.

1. Macro Performance

Four years ago, I released a study with Kevin Hassett, Tim Fitzgerald, and Cody Kallen of the economic effects of candidate Biden’s agenda compared to President Trump’s. Knowing that campaign promises do not necessarily turn into policy, we analyzed several policy scenarios. The scenario closest to the portfolio of policy changes over the past four years we called “capital taxation constant” (CTC). Biden-Harris climate regulations proved to be somewhat more aggressive than represented by the CTC scenario including, for example, a requirement that manufacturers of medical inhalers (sic) either cease production or convince the Environmental Protection Agency that they are earnestly seeking lower-emission technologies. On the other hand, as nonlawyers we did not account for such a high failure rate of Biden-Harris rules in federal courts.

Under the CTC scenario, labor and capital would be 5.0 percent below the Trump baseline in the long run. In tomorrow's Wall Street Journal, we show that a single trend fits the data well from 2017-Q1 through 2021-Q4, except for the first full pandemic quarter. Then inflation hit and employee compensation—and national income more broadly, which isn’t shown in the chart—fell 5 percent behind. To be more precise, the latest data (2024-Q2) show inflation-adjusted employee compensation per adult to be 4.6 percent below the trend.

Arguably, human capital would have fallen somewhat below its trajectory after 2020 due to pandemic behaviors unrelated to Biden-Harris policies. By itself, this would have pulled labor income somewhat below its prior trend for several years. On the other hand, with some of the regulations not taking effect until 2025 and beyond, we have not yet seen the full effect of the Biden-Harris policies. Large language models and other “AI” technologies have been a positive growth effect that was unanticipated in 2020 when we made the forecasts.

We used a closed-economy model (tariffs were modeled like other excise taxes) that imposes a constant labor’s share, a constant depreciation rate, no statistical discrepancy, and equality between GNP and GDP. In reality, labor’s share of national income has been pretty constant, but the national income’s share of GNP has fallen a bit. More significant has been a fall in the ratio of GNP to GDP. Conversely, a real GDP per capita chart would look “better” than our compensation chart, which of course is no consolation for workers.

Economics Lessons from the Kibbutzim

In March 2023, I visited Ma'agan Michael Kibbutz and had a meeting with a former Israeli government official charged with managing the government bailout of about 200 kibbutzim.  Both offered empirical lessons about the challenges of communal living.

 

Background and Description

Ma'agan Michael was one of the few kibbutzim not bailed out.  It is one of the wealthier and more populous kibbutzim, with approximately 2000 residents. (Kibbutzim is the plural of kibbutz, which is described in more detail below but might be briefly described as a voluntary socialist community in Israel.)

 

In an earlier era, this kibbutz and the others followed a more egalitarian model.  For example, children were once raised communally in dormitories and would visit their biological parents for only three hours each day.  This alone was a big sacrifice for families for the sake of ideology, not to mention harmful actions that adults might occasionally take against children knowing that their parents were not around.

 

Now nuclear families live together in their own residence.  Every family receives an income that varies only according to the age and structure of their family.  As of 2023, that was about $3,500 per month.  Dining and laundry are in common. Rent, healthcare, formal education, and those two services are "free" -- not taken out of the $3,500.  A family can purchase and own household appliances for their residence.  Some get a washer and dryer because the community laundry takes too long.

 

Part of the kibbutz is something like a car rental agency.  They have a fleet of cars.  A central office holds the keys.  A member could pick up a car key and use a car for the day.  I believe that this was charged against their $3,500. Much of the accounting is done on computer with a fab system.

 

Some of the members are employed outside the kibbutz, but nonetheless must surrender all of their earnings to the community.  A partial exception would be earnings during periods of sabbatical.  One three-year sabbatical is permitted per lifetime.

 

Forty-five hours per week of work is required from all adults, with some categorical exceptions.  Mothers are required only 37.5 hours per week.  Retirees do not work at all; their Israeli-government pensions go to the community.

 

More than 90 percent of the land in Israel is government land, including the land allocated to Ma'agan Michael since at least the 1940s.  It is a prime seaside property, as you can see in this photo.  They once fished in the Mediterranean but now operate a couple of different kinds of fish farms on their property.  They also have a couple of factories for manufacturing.  The plastics factory they founded in 1963 has annual sales approaching $100 million, which helped them avoid the financial crises experienced by most other kibbutzim.

 

 

The common areas of the kibbutz can be compared to a college.  The dining room resembles the college cafeterias from my own college days, which are not as upscale as modern college dining facilities.  The grounds are less neatly kept than college campuses are (the beach photo is not typical in that regard).

 

Problems with communal living, even on a small scale

 

As Ran Abramitzky discusses in his book The Mystery of the Kibbutz, and labor economists have observed with employment contracts, work effort can be a problem when pay is not tied to performance.  The kibbutz, where members have often known each other since birth, tries to police this by watching each other.  They can expel a member from the kibbutz for poor performance, which is a serious punishment.

 

The kibbutz also looks hard at the work ethic of persons applying for membership.  However, membership in the kibbutz is voluntary, and individuals cannot be compelled to stay.  The most productive members have a significant financial incentive to leave.

 

Perhaps a more serious problem is occupational choice and human capital accumulation.  With uniform pay, members have little financial incentive to excel in their jobs, particularly in roles they find unfulfilling.  While the kibbutz covers the cost of formal education, much of human capital development in conventional labor markets occurs after formal schooling is completed.

 

Too much is democratized on the kibbutzim.  First, there is the challenge of reaching agreement, leading to significant and frustrating indecision.  Second, incentives are lacking to acquire and utilize information relevant to collective decision-making.  An individual could work hard to determine the right answer, but his vote hardly counts (the paradox of voting).

 

Financial decisions are a prime example.  Few kibbutz members have an incentive to learn about, say, present values.  About fifty years ago, numerous kibbutzim initiated projects that seemed viable if the time value of money were ignored, but were big losers from an NPV perspective.  The abundance of failures among those projects was a major reason that kibbutzim would later be bailed out.

 

Equality, including gender equality, was an important ideal of the early kibbutzim and, to a large extent, still today.  Nevertheless, despite the ideology, gender segregation by occupation seems just as strong as it is off kibbutz.  All but one of the laundry employees I saw were female.  All of the tractors and heavy machinery were operated by men.

ChatGPT: Marshall, Hicks, or Cliff Clavin?

Marshall discusses his Laws of Derived Demand in his Principles of Economics.  Specifically, in Chapter VI of Book V.

Marshall listed four “conditions” that result in wage-inelastic derived demand for the corresponding factor of production:

1. “no good substitute being available at moderate price”

2. “stiff and inelastic” final demand

3. “small part of the expenses of production” (i.e., small factor-cost share)

4. Inelastic supply of other factors

Putting aside the fourth law, we can discuss the first three in the context of the Allen-Hicks formula for the own-price elasticity of derived demand, βii, in the two-factor case (see also Chicago Price Theory Chapter 11).

where si is good i’s share in factor cost, εD < 0 is the price-elasticity of demand for the final product, and σ > 0 is the elasticity of factor substitution.  None of these parameters are necessarily constants; they may vary according to the factor allocation.  Marshall’s first law can be understood as a comparative static with respect to σ in the direction of zero, holding constant si and εD.  The second law can be understood as a comparative static with respect to εD holding constant si and σ.  The debated third law is a comparative static with respect to the share parameter, with Hicks at least holding constant εD and σ.

ChatGPT confidently asserts that there are three Marshall’s Laws.  Three is a fair count if we exclude the law that Hicks disputed.  However, by that count, ChapGPT gets them 5/6 incorrect.  Instead, it lists the definition of derived demand, followed by a statement about long versus short run.  The third in its list is perhaps a combination of the disputed third law and the first law.

Mulligan: What are Marshall's Laws of Derived Demand?

ChatGPT: Marshall's Laws of Derived Demand are a set of principles that describe the relationship between the demand for a particular factor of production (such as labor) and the demand for the goods and services that factor helps produce. These laws were developed by economist Alfred Marshall in the late 19th century and are still widely studied and referenced today.

The three laws of derived demand are as follows:

1. The demand for a factor of production is derived from the demand for the product that the factor helps produce.

2. The demand for a factor of production is more elastic in the long run than in the short run.

3. The elasticity of demand for a factor of production depends on its importance in the production process and the availability of substitutes.

Taken together, these laws suggest that the demand for labor, for example, is ultimately determined by the demand for the goods and services that labor helps produce. In the short run, the demand for labor may be relatively inelastic because it can be difficult for firms to quickly adjust their production processes to changes in demand. However, in the long run, firms may be able to find substitutes for labor or change their production methods, making the demand for labor more elastic. The importance of labor in the production process also affects its elasticity of demand: if there are few substitutes for a particular type of labor, its demand may be relatively inelastic.

ChatGPT also confidently discusses the Hicks challenge but refers to a “law” that was not even one of the four.

Mulligan: Didn't Hicks show that Marshall was wrong about one of the laws?

ChatGPT: Yes, economist John Hicks challenged one of Marshall's laws of derived demand, specifically the second law, which states that the demand for a factor of production is more elastic in the long run than in the short run. Hicks argued that this law was not always true and that the elasticity of demand for a factor of production could depend on the specific circumstances.

Hicks pointed out that in some cases, the demand for a factor of production could be more elastic in the short run than in the long run. For example, if a firm has already made significant investments in a particular technology or capital equipment, it may be difficult for the firm to switch to a different production method or substitute other factors of production. In such cases, the demand for the factor of production may be relatively inelastic in the long run, while being more elastic in the short run as firms can adjust their inputs more easily.

Hicks’ critique of Marshall's second law suggests that the relationship between the demand for a factor of production and the goods and services it helps produce is complex and can vary depending on the specific circumstances of the market. Nonetheless, Marshall's laws of derived demand remain an important contribution to economic theory and continue to be studied and debated by economists today.

Next, I asked for sources.  It got the books correct, which is pretty good for ChatGPT (famous for fabricating books and articles) but wrong chapters.  It should be Marshall Book V, Chapter VI.  Hicks’ second edition only has 11 chapters, although perhaps Chapter 12 can be understood as the material that follows Chapter 11 in which case “Chapter 12” is the correct part of the book.

Mulligan: Where did Marshall state the laws?  Where did Hicks challenge one of the laws?

ChatGPT: Alfred Marshall first introduced his laws of derived demand in his seminal work "Principles of Economics," first published in 1890. Marshall's laws are discussed in Book V, Chapter III, where he discusses the theory of distribution and the demand for factors of production.

John Hicks challenged Marshall's second law in his 1932 book "The Theory of Wages." In the book, Hicks argues that the elasticity of demand for factors of production can depend on the specific production technology and market conditions. Hicks specifically critiques Marshall's assumption that the demand for a factor of production is always more elastic in the long run than in the short run. Hicks' critique is discussed in Chapter 12 of his book.

I disputed the citation of Chapter III, so it apologized [not shown here] and cited Chapter VI of Book IV, which purportedly discusses “Distribution of Produce” (the actual title is “Industrial Training”).  More disputes produced more rounds of apology and fake references.  It tended back to Book V, Chapter III (a famous chapter) but with a different fake chapter title at each iteration.

Cliff Clavin was the famous tavern know-it-all in Cheers.  Perhaps Cliff is back, this time behind the curtain of AI spouting lots of jargon in grammatically correct but ignorant sentences. Is Norm somewhere back there too?

[image credit: https://television-b26f.kxcdn.com/wp-content/uploads/2020/08/Cheers-Cliff-Calvin-Screenshot-770x470.png]

Meanwhile, for more than a decade Artificial Intelligence has offered tools that are never self-contradictory and power a good exposition of Marshall’s Laws and much more in economics and statistics.

[Yes, ChatGPT can do the same for coding too: fabricated function options, fake references, etc.]

www.perplexity.ai correctly recited Marshall's four laws, including links to some microeconomics lecture notes. It correctly reported the sources of Marshall's law, and acknowledged that it could not report which chapter. It was incorrect as to which law Hicks challenged.

Recession Time? Don’t Act Surprised

Treasury Secretary Yellen does not see any indicator of an imminent recession.  She isn’t looking.  The normal economic tailwinds have calmed and, as predicted, Biden's economic policies are a significant headwind.

A recession is sometimes defined as a reduction in the number employed nationally for a couple of months.  Other times it is defined as a reduction in real GDP for two quarters or more.

When it comes to predicting events like this, my recursive approach is to first understand where the general trends are heading.  In technical terms, is the economy’s “steady state” above or below where we are now, and how much?  If the trends are strong up, small perturbations around that trend will not make a recession.  If the trends are flat, then even a small negative shock will create a recession by one or more of the definitions.  Which definition will be triggered can be assessed by contrasting employment trends with productivity trends.

Four important trends are worth considering: organic productivity growth, organic population growth, recovery from the pandemic recession, and new public policies affecting productivity,  population, or employment.

Organic trends

Given that recessions are defined in absolute rather than per capita terms, population growth is normally an economic tailwind.  However, annual adult population has fallen from a bit above one percent 1980-2018 to about 0.4 percent.  Illegal immigration is a wild card here because we do not know how many are immigrating, what fraction are adults, and whether and how those adults will be economically engaged.  With that caveat, we now are in a situation where even a small negative shock that would not have caused a recession in the one-percent population growth era will now.

Recovery from the pandemic was also a tailwind.  It someday will continue to lift employment, but at the moment it looks like employment has recovered as much as it can given the serious health problems encountered during the pandemic, including but not limited to self-destructive substance abuse habits that are not complementary with productive employment.  Some of these people will show up on payrolls but how reliably they show up for work is another question.  Diabetes, liver disease and heart disease have gotten out of control since 2020.

Workers lost skills and capital laid idle during the pandemic.  These are recovering, although their recovery will not be fully recognized in the growth data.  GDP and productivity levels were exaggerated during the pandemic as many goods were unavailable or low quality in ways not captured by the national accountants.  For example, public school teachers stayed home from school but the national accountants assumed that they were as productive as ever merely because they continued to get paid.  As they get back to traditional teaching, this will not be officially recognized as economic progress for the same reason the pandemic regress was never acknowledged. 

Crime has gotten bad, especially in big cities where productivity is normally the highest.  Consumers and businesses are avoiding big cities, which is a cost (“excess burden”) beyond the crime statistics because the whole point of the avoidance behaviors is to keep from being one of those statistics.

Fitzgerald, Hassett, and I predicted in 2020 that Biden’s economic agenda would reduce the levels of full-time equivalent employment per capita by 3.1 percent and real gdp per capita by 8.5 percent.  If that level effect were spread over five years, that would be 0.6 percent per year and 1.7 percent per year, respectively, as shown in the Table as an addendum panel.  That by itself makes a recession likely in one of those five years.

Regulatory Policy

Our analysis of Biden’s agenda distinguished regulation from capital taxation from labor taxation.  His regulatory agenda seems to be going ahead as we expected.  The good news is that Biden’s nomination of David Weil to the Department of Labor was rejected by the Senate and Biden was slow to fully mismanage the National Labor Relations Board.  But we did not anticipate that Biden’s DOL would disrupt labor markets as much as it did with its mask mandates.  Sticking with our original estimate, it looks that Biden’s regulatory agenda is reducing employment by 0.2 percent per year (of five years) and real GDP by 0.7 percent per year below the organic trends.  See the Table’s top panel.

Of particular concern over the next few months is the reliability of the electric grid and air travel.  Snafus of this type are already built into our regulatory analysis but these examples put more texture on the economic reasoning that links the marginal regulations with poor economic performance.

Capital Taxation: Inflation Sneaks In

Biden’s Build Back Better bill would implement much of the capital taxation we envisioned in 2020.  The good news is that the bill has not yet passed, and passage of its capital tax elements are not imminent in some other form.  The bad news is that inflation is taxing businesses without any Congressional action (recall Feldstein and more recently Hassett on the effect of inflation on the cost of capital), while it appears that Biden will let temporary provisions in the 2017 TCJA expire.  With capital taxation during the Biden administration increasing about half of what we expected, it would reduce real GDP by about 0.4 percent per year over five years.

Speaking of inflation, higher Fed Funds rates are already showing up in mortgage rates.  In effect, the Federal Reserve is introducing a tax (or cutting a subsidy) on structures investment, which is likely to send at least that sector into a recession.  Socially responsible (a.k.a., woke) investing is also skewing the allocation of capital.

Combining capital taxation and regulation, the headwinds in the Biden economy are 0.25 percent per year for employment and 1.1 percent per year for real GDP.

Labor Taxation: Direction Unclear

Labor taxation is an interesting wild card here.  Marginal tax rates on work were cut sharply when the $300 weekly unemployment bonus expired last summer.  That effect has played out already.  But I expect that Congressional Democrats, and even some Republicans, will expand unemployment benefits if anything resembling a recession were occurring.  That could easily and quickly reduce employment by one percent, if not more.  On the other hand, various federal health insurance subsidies are about to expire.  If they do (without resurrection), that will encourage work.

Bottom Line

Overall, a recession is highly likely with so many headwinds and so few tailwinds.  A recession is more likely by the GDP definition than the employment definition.  The depth of the recession depends on how much Congress destabilizes things by further adding to the already large federal portfolio of programs for the unemployed and poor and further adding to tax burdens.