Macro Roundup

Trends in global venture capital investment (if restricted to US trends don’t ~ change), “…Based on data from Pitchbook on global venture-capital investments,Figure 5.1 examines the sectors which have seen the most rapid growth in venture capital financing in the 2010s. It lists the total dollar value of all deals reported in 2010 and 2019, categorized by the main industry sectors reported in Pitchbook. Figure 5.1 shows the remarkable growth in the value of venture capital deals over this period, rising more than fivefold from 2010 to 2019. Figure 5.1 also shows that growth was largely driven by increases in investment towards IT software and services, consumer products and services (B2C), business products and services (B2B), and financial services. The figure looks virtually identical if restricted to only U.S. venture capital deals, implying that this is driven by an across-the-board change, rather than due to the composition of deals in countries such as China, which have seen faster growth of VC in recent years. Due to the ubiquity of software, many innovations classified as IT software, B2C, and B2B cut across traditional industry sectors. For example, Uber disintermediated the taxicab business by more efficiently connecting passengers with drivers, and in less than ten years from founding, Airbnb had more listings than the largest hotel chain in the world, despite owning no assets itself. Hundreds of other such VC-backed start-ups serving consumers and enterprises across a range of industries have been financed in the last decade, bringing immense value to their users in many instances, as well as being adopted or replicated across many countries around the world. However, Figure 5.1 and Figure 5.2 also show that investments in three sectors have not kept up with the overall growth: healthcare; IT hardware—comprising communications and networking equipment, computer hardware, and semiconductors; and energy, materials, and resources. As shown in Figure 5.2, the share of investments in these sectors fell from over 50% of total spending in 2010 to below 25% in 2019. Energy, materials, and resources and IT hardware combined accounted for less than 5% of capital invested by VCs in 2019. To some extent, these ebbs and flows of funding across sectors reflect technology life cycles, the huge wave of applicationrelated innovations made possible by the Internet revolution in the late 1990s, and the subsequent rise of cloud computing in the mid-2000s. However, the introduction of cloud computing services in the mid-2000s also had another important effect: it dramatically lowered the cost of learning about the ultimate potential of risky web-based start-ups. Specifically, it allowed those start-ups to rent hardware in small increments from providers like Amazon Web Services, use this to quickly gauge customer demand, and postpone expensive investments to scale up until after learning about the size and nature of demand from consumers.6 This, in turn, led to a disproportionate rise in the number of start-ups that could benefit from such lowered cost of experimentation. The increase in such start-ups is reflected in the changing shares of industries shown in Figure 5.1 and Figure 5.2, and also in the development of crowdfunding, accelerators, angel groups, and other early-stage investors who finance the lower initial capital needs of such ventures and promote effective learning about product-market fit using frameworks, such as the lean start-up model. 7 While technological advances, such as rapid prototyping and the advent of advanced simulation and prediction tools, have also lowered the cost of learning and experimentation beyond software and web-based start-ups, growing academic research has begun to articulate certain characteristics of start-ups that make them a poor match for the venture capital model of financing innovation. Three particularly salient elements include: 1) the longer timelines required to build such companies, 2) capital intensity associated with de-risking these ventures, and 3) the nature of market and technology risk faced by new ventures…”

US S&T clusters are more intensive (in terms of population) than international comps“….latest ranking of the world’s top 100 S&T clusters. Year-over-year changes in cluster ranks remain modest, though they are in line with the longer-term trend— namely, faster growth of S&T activity in East Asia and especially in China. Analyzing the S&T intensity of clusters provides a more nuanced perspective of the world’s S&T cluster landscape. In particular, it suggests that many European and U.S. clusters show more intense S&T activity than their Asian counterparts, even though they show lower S&T activity in absolute term…Table S-1.2 presents our top 100 clusters ranked by their S&T intensity. Our measure of S&T intensity is the sum of patent and scientific publication shares associated with a cluster, divided by its population. As can be seen, Cambridge and Oxford in the United Kingdom (U.K.) emerge as the most S&T-intensive clusters. Both clusters host highly productive scientific organizations in relatively small urban agglomerations. Cambridge additionally has a relatively large presence of tech companies—for example, ARM and Nokia—which results in a patent output normally seen in agglomerations with twice the population.8 In the case of 3rd-ranked Eindhoven, the high S&T intensity principally stems from high patenting output. Interestingly, 4th-ranked San Jose-San Francisco illustrates that high S&T intensity does not have to be associated with small size. This cluster hosts a population of more than six million, and it is the fifth-largest S&T cluster in absolute terms (Table S-1.1). Figure S-1.2 compares the absolute and per capita ranks of the 100 S&T clusters in a scatterplot. It confirms, first of all, that there is no obvious correlation between the rankings. There is wide variation in the S&T intensity of both small and large clusters. For example, Shanghai—ranked 9th in absolute size—holds only the 82nd position in the intensity ranking. By contrast, Lund-Malmö is only the 96th largest cluster but occupies the 10th position in the intensity ranking. Another interesting pattern emerging from Figure S-1.2 is that many of the U.S. clusters appear in the upper right corner of the scatterplot—they are large in absolute and relative terms. Important exceptions are New York City and Los Angeles, which rank in the top 20 clusters mainly because of their large size and not their S&T intensity. Many Chinese clusters, in turn, do not exhibit high S&T intensity, which reflects the large populations covered by them.9 One exception is the 4th ranked Beijing cluster, which still shows considerable S&T intensity and has a performance similar to that of Seoul.Interestingly, TokyoYokohama—the top S&T and second most populous cluster—still shows high S&T intensity notwithstanding its large size. Many of the European clusters show above-average S&T intensity, but do not necessarily feature among the top-ranked clusters. This reflects the different agglomeration patterns in Europe, which have resulted in smaller cities compared to North America and East Asia…”

Some highlights from the World Intellectual Property Organization's 2020 GII

They cite “tough tech” can benefit from state funding/or the state creating a market as a customer, “… one setting where start-ups engaged in innovation have been shown to benefit substantially is theU.S. Department of Energy’s SBIR grant program, which has helped start-ups finance the prototyping of new technologies and thereby substantially increase the odds of receiving venture capital.13 This ties in directly to the friction outlined above—where start-ups in some sectors cannot attract VC due to the difficulty they face in learning about the effectiveness of a new technology in the field as opposed to the lab, and hence have trouble convincing investors they can achieve product-market fit and generate sufficient customer demand….Many successful examples of government involvement in the commercialization of tough tech have beenrelated to the government’s (often the military’s) role as a customer.14 A key reason for this may have to do with government contracts substantially reducing market risk through a willingness to pay for early versions of an emerging technology. A large military contract can also help to establish standards and coordinate the direction of technology trajectories. Finally, through their role as customers, governments can even reduce financing constraints via the timing of contract payment. For example, paying part of the contract value in advance can substantially reduce startups’ dependence on external finance.15 This important role of the government as customer is often underappreciated when considering the role that policymakers can play in jump-starting innovation….”

“….Start-ups that share one or more of the characteristics that make them a poor fit for VC investment have sometimes been referred to as tough tech—in reference to the fact that these technologies are often tough to commercialize using venture capital. In many instances, they involve breakthroughs in fundamental science or nascent technologies, which leads to long timelines and substantial technology risk. Such ventures have sometimes also been referred to as “deep tech”. It is important to emphasize that not all science-based ventures are bad fits for VC; indeed, some ventures spinning out of university labs raise substantial venture capital, generate high returns for investors, and solve important problems for the world. Nevertheless, many of the innovations required to solve society’s most pressing problems do not have solutions that fit the timelines and economic constraints of VC investors. In light of these constraints, and the growing sense that there is also a decline in fundamental innovation coming from large corporations,11 there are several elements that policymakers and other stakeholders could consider to help support the commercialization of tough tech….”

Re: Yes they view “tough tech” as areas where VC is underinvesting versus potential social gains. These investment are now largely funded by other sources, a parent company or the state either though direct funding or as a customer.

The overall score is driven by three measures, “….As every year, three indicators are used to measure the quality of innovation. First, the quality of local universities is measured through the average score of the top 3 universities in each country in the QS university ranking (indicator 2.3.4). Second, patent families filed in at least two offices (indicator 5.2.5) are used as a proxy of the internationalization of local inventions. Third, the H-index (indicator 6.1.5), which is the number of citations that locally produced research documents receive abroad, is used to assess the quality of scientific publications….”I

“…Despite its nonlinearity and the overidentified estimation with more target moments than parameters,our DE model captures both the volatilities of, and to a large degree the correlations between, real variables, beliefs, and financial outcomes in the firm data. To assess the role of overreacting expectations on macro outcomes,we also estimate and simulate a rational expectations (RE) model with θ fixed at 0, which does not match predictable forecast errors at the firm level…. Figure 4 plots the required TFP shocks, along with their implications for the growth of credit, output, investment, and corporate profit forecasts. The red lines connect the pre-crisis and crisis outcomes generated by the DE model, while the green lines present the data. In the top left panel, theDE model by construction perfectly matches the pre-crisis and crisis spreads. Remarkably, though, the bottom left panel shows that the pattern of TFP growth needed to account for the increase in spreads is virtually identical to the TFP growth observed during the period. This overlap is an untargeted feature and shows that in the DE model a moderate 1.5% reduction in TFP growth is able to produce the large observed increase in the average spread….”

New Shleifer, in the spirit of his last book (which largely mirrored the Minsky/Kindleberger view of boom bust cycles) argues that credit cycle basically work the same way - investors are too optimistic, so credit and investment overexpand. Subsequently, beliefs cool off, credit markets tighten, real activity declines, and default rates increase. Both inflated expectations and managers disappointment in those expectations play a key role in the cycle.

Using TFP (Total Factor Productivity) news as it’s impact as evidence"...We assess whether belief overreaction to standard TFP shocks may produce observed boom-bust credit cycles. We incorporate diagnostic expectations into a workhorse business cycle model with heterogeneous firms and risky debt. A realistic degree of diagnosticity, estimated from the forecast errors of managers of US listed firms, produces significant fragility of the economy during good times. This helps account for countercyclical credit spreads and for credit cycles at both the firm and aggregate levels. Lax financial conditions predict future increases in spreads, low bond returns, and investment drops. Spread increases of the magnitude observed during 2008-9 obtain from modest negative TFP shocks. Our results indicate that diagnostic expectations may offer a realistic and parsimonious way to produce financial reversals in conventional business cycles models...."

Core finding echo a wise man's belief that "it's never as good as you think it is, it's never as bad as you think it is", "...present novel evidence based on microdata that directly connects expectations to credit spreads, bond returns, and investment at the firm level. We first show that managers’ expectations of their firms’ profits overreact to current conditions: they are too optimistic in good times and too pessimistic in bad times. In turn, excess optimism about a firm’s future profits measured from expectations data predicts a one year-ahead increase in the firm’s credit spread, low realized returns on the firm’s bonds, and a decline in its investment growth. Overreacting beliefs appear to be directly linked to firms’ financial and real cycles...."

The Tax Foundation reports on impact of the TCJA on individual incomes taxes for 2018 (first tax year after bill passed)

Y/Y impact on Federal revenue, "Taxpayers reported $11.6 trillion in adjusted gross income (AGI) on 144.3 million tax returns in 2018, the year the Tax Cuts and Jobs Act (TCJA) took effect....While the number of returns and amount of AGI rose,total income taxes paid fell by $65 billion. Altogether, taxpayers paid $1.5 trillion in income taxes in 2018, down about 4 percent from 2017. The average individual income tax rate for all taxpayers fell from 14.6 percent to 13.3 percent..."

That burden was overwhelmingly borne by high earners. Interesting the 1% income share fell slightly (-.1pp) whereas their share of Federal income taxes rose by 1.6pp. Some more key factoids, "... The share of reported income earned by the top 1 percent of taxpayers fell slightly, to 20.9 percent in 2018 from 21 percent in 2017. Their share of federal individual income taxes rose by 1.6 percentage points to 40.1 percent... Since 2001, the share of federal income taxes paid by the top 1 percent increased from 33.2 percent to a new high of 40.1 percent in 2018....In 2018, the top 50 percent of all taxpayers paid 97.1 percent of all individual income taxes, while the bottom 50 percent paid the remaining 2.9 percent..... The top 1 percent paid a greater share of individual income taxes (40.1 percent) than the bottom 90 percent combined (28.6 percent)....The top 1 percent of taxpayers paid a 25.4 percent average individual income tax rate, which is more than seven times higher than taxpayers in the bottom 50 percent (3.4 percent)...."

Note for this analysis adjusted gross income doesn’t count transfers outside of taxed social security benefits"....AGI is a fairly narrow income concept anddoes not include income items like government transfers (except for the portion of Social Security benefits that is taxed), the value of employer-provided health insurance, underreported or unreported income (most notably that of sole proprietors), income derived from municipal bond interest, net imputed rental income, and others...."

"...Erica York at the Tax Foundation? This week she reports...The new data shows how taxes changed in the first tax year after passage of the Tax Cuts and Jobs Act (TCJA) in December 2017..... The share of reported income earned by the top 1 percent of taxpayers fell slightly, to 20.9 percent in 2018 from 21 percent in 2017. Their share of federal individual income taxes rose by 1.6 percentage points to 40.1 percent....Since 2001, the share of federal income taxes paid by the top 1 percent increased from 33.2 percent to a new high of 40.1 percent in 2018....."

James Freeman, "How Trump Taxed the Rich,"Wall Street Journal, February 5, 2021, https://www.wsj.com/articles/how-trump-taxed-the-rich-11612556008

Just an FYI. CBO is apparently underwhelmed by the monopsony arguments. They updated their analysis of the economic impact of hiking the federal minimum wage to $15 per hour and continue to find strong employment effects. -1.4mm workers (.9%) at the same time the number of people in poverty would be reduced by almost 1mm (.9mm). Note this is stronger than their last analysis which estimated employment would fall by 1.3mm workers (that change was driven by a decision to use mean as opposed to median)

Key takeaway on employment effects, "....Taking those factors into account, CBO projects that, on net, the Raise the Wage Act of 2021 would reduce employment by increasing amounts over the 2021-2025 period. In 2025, when the minimum wage reached $15 per hour, employment would be reduced by 1.4 million workers (or 0.9 percent), according to CBO’s average estimate. In 2021, most workers who would not have a job because of the higher minimum wage would still be looking for work and hence be categorized as unemployed; by 2025, however, half of the 1.4 million people who would be jobless because of the bill would have dropped out of the labor force, CBO estimates. Young, less educated people would account for a disproportionate share of those reductions in employment...."

CBO Staff, "The Budgetary Effects of the Raise the Wage Act of 2021," Congressional Budget Office, February 2021, https://www.cbo.gov/system/files/2021-02/56975-Minimum-Wage.pdf

The impact on output, would reduce capital stock versus the counterfactual, slowing productivity, “….Effects on Real OutputRaising the minimum wage would slightly reduce real GDP, primarily because of reduced employment.However, CBO incorporated the assumption that the bill would not change nominal GDP from the amounts in the agency’s baseline budget projections, following a practice that is conventional for its cost estimates. As the next section describes, CBO’s estimates of the effects of the bill on the GDP price index (and other price indexes) were adjusted to be consistent with that assumption. In addition to its effect on real output through employment, the bill would cause the stock of capital goods to be smaller than it would be otherwise.Capital goods are assets that businesses use to produce goods and services; they include tools, buildings, vehicles, machinery, and equipment. Some businesses would invest in capital goods to replace workers. Other businesses, however, would be discouraged from constructing new buildings or buying new machines if they anticipated having fewer employees to use them. On average, over the 2021-2031 period, real investment would be slightly lower than it would be if current laws did not change, CBO estimates. That reduction in investment would reduce workers’ productivity and lead to further reductions in their employment. The higher minimum wage would also shift income toward lower-income families, which tend to spend a relatively larger fraction of their income. As a result, the total demand for goods and services would increase for several years, boosting overall real output. After that initial boost, however, CBO expects that the economic effects from increases in demand would disappear….”

In terms of pay btw 2021 to 2031, the cumulative pay of affected people "would increase, on net, by $333 billion—an increased labor cost for firms considerably larger than the net effect on the budget deficit during that period.... That net increase would result from higher pay ($509 billion) for people who were employed at higher hourly wages under the bill, offset by lower pay ($175 billion) because of reduced employment under the bill."

Henry Olsen, “Findings From A YouGov National Survey Of 2020 Trump Voters,” American Enterprise Institute, February 5, 2021, https://www.aei.org/events/donald-trump-and-the-future-of-the-gop/