“Unintended Consequences is far smarter and more thought-provoking than most economics written for the general public” - Greg Mankiw, Harvard University, Former Chairman of the Council of Economic Advisors
“…reminds us that inequality sends a signal of what society lacks most, in America’s case, entrepreneurship and risk taking.” - Lawrence Lindsey, CEO, The Lindsey Group, former Director of the National Economic Council
“…a must-read for serious students of economic policy.” - Glenn Hubbard, Dean, Columbia Business School, and former Chairman of the Council of Economic Advisers
“…a very valuable contribution.” - Larry Summers, former Secretary of the Treasury and director of the National Economic Council, president emeritus, Harvard University
“…a comprehensive explanation of the modern economy.” - Julian Robertson, Founder, Tiger Management
“…a fresh argument for the productive value of inequality.” - David Autor, Professor of Economics, Massachusetts Institute of Technology
“Unintended Consequences provides a provocative interpretation of the causes of the global financial crisis and the policies needed to return to rapid growth. Whether you agree or not, this analysis is well worth reading.” - Nouriel Roubini, New York University; Chairman, Roubini Global Economics
“…serious thinking for serious thinkers. …a thought-provoking blueprint for growing middle- and working-class incomes.” - Mitt Romney, former Governor of Massachusetts
“Unintended Consequences represents the most cogent and persuasive analysis of the Financial Crisis to date.” - Andrei Shleifer, 1999 John Bates Clark Medal Winner
“There are an amazing number of good ideas and interesting points made in Unintended Consequences. The thinking underlying it, and the obvious depth of understanding of the author, are very impressive.” - Steven Levitt, coauthor of Freakonomics; 2004 John Bates Clark Medal
“…challenges misconceptions that distort our economic debates.” - Arthur Brooks, President of the American Enterprise Institute
“…challenges misconceptions that distort our economic debates.” - Arthur Brooks, President of the American Enterprise Institute
AI Summary.Census Bureau microdata analysis finds that in Illinois annual out-migration by households earning under $50K fell by 104K while out-migration by households earning over $200K rose by 54K. Across the Midwest, adults with a bachelor's degree or higher now make up 57% of out-migrants, up from around 40%.
Aziz SunderjiHome Economics
Date Posted:
Is Database:
Database
Is Important:
Important
The “Brain Drain” is ongoing in the Midwest. The fraction of adult out-migrants with at least a bachelor’s degree is now 57%, up from ~40% in 2005. “It’s the most educated and highest-earning that are still leaving in droves.”
Are the most educated and highest-earning residents leaving the Midwest?
Core argument: Annual out-migration by Illinois households earning over $200K rose by 54K between 2014–15 and 2023–24, according to the author's analysis of Census Bureau microdata.
Take Illinois, the most populous Midwestern state: my analysis of Census Bureau microdata shows that between 2014–15 and 2023–24, annual out-migration by households earning under $50K fell by 104K. Out-migration by Illinois households earning over $200K rose by 54K over the same period. Across the region, the trend along educational lines is the same: the bachelor’s-or-higher share of adult out-migrants is now 57%, up from around 40% in 2005. It’s the most educated and highest-earning that are still leaving in droves.
Takeaways by Macro Roundup® AI
Annual out-migration by Illinois households earning over $200K rose by 54K between 2014–15 and 2023–24, according to the author's analysis of Census Bureau microdata.
Annual out-migration by Illinois households earning under $50K fell by 104K between 2014–15 and 2023–24, according to the author's analysis of Census Bureau microdata.
Adults with a bachelor's degree or higher make up 57% of out-migrants from the Midwest, up from around 40% in 2005.
Related Articles:
Housing and the American Worker— Wages net of housing costs were once substantially higher in high-density areas than low-density areas for workers both with and without college education, but over the past half-century that urban wage premium has persisted only for college-educated workers, not for those with a high school degree or less.
Where is Standard of Living the Highest? Local Prices and the Geography of Consumption— The authors estimate that geographical differences in cost of living are especially large for low-income US households, and that expensive cities offer college graduates a standard of living similar to affordable cities while offering less-educated households a lower one. A high school dropout household moving from the most affordable to the most expensive commuting zone would see consumption fall 18.5%.
The Determinants Of Declining Internal Migration— A 10% increase in wages where people live is associated with 3.5% less migration, while a 10% increase in destination wages is associated with 7.8% more migration; a 10% rise in destination home prices is associated with 2.6% less migration. The authors find wage responses alone would have raised migration rates, but housing-related factors more than offset them.
AI Summary.New US homeowners spent 26% of their income on housing in 2024 versus 20% for existing homeowners, a six-percentage-point gap that is the widest in nearly 40 years. Historically the two groups' costs moved together, with the gap holding at roughly two to four percentage points.
Jess RemingtonAgglomerations
Date Posted:
Is Database:
Database
Is Important:
Important
As of 2024, American homeowners who bought a house within the last year were spending 26% of their income on housing, relative to 20% for existing homeowners – the largest gap in almost 40 years.
How wide is the housing cost gap between new and existing homeowners?
Core argument: New homeowners spent 26% of their income on housing in 2024 versus 20% for existing homeowners, a six-percentage-point gap, the largest in nearly 40 years.
Historically, monthly housing costs for new and existing homeowners have tended to move in tandem. From 1990 through the aftermath of the Great Recession, both groups saw costs rise during booms and fall during downturns, with the gap between them remaining relatively stable at two to four percentage points. That pattern briefly reversed during the Great Recession, when new buyers were able to purchase homes at depressed prices and consequently spent slightly less of their income on housing than existing owners. By 2017, the typical two-point gap had returned. The current divergence began in earnest in 2022. By 2024, new homeowners were spending 26% of their income on housing, compared to 20% for existing homeowners — a six-percentage-point gap, the largest in nearly 40 years. Although new homeowners spent a slightly larger share of their income on housing at the peak of the housing bubble in 2007 (28%), the gap with existing homeowners was smaller (four percentage points).
Takeaways by Macro Roundup® AI
New homeowners spent 26% of their income on housing in 2024 versus 20% for existing homeowners, a six-percentage-point gap, the largest in nearly 40 years.
From 1990 through the aftermath of the Great Recession, housing cost burdens for new and existing homeowners moved together, with the gap holding at two to four percentage points; the current divergence began in earnest in 2022.
At the 2007 housing bubble peak, new homeowners spent a slightly larger 28% of income on housing, but the gap with existing homeowners was four percentage points.
Related Articles:
The Cost of Money is Part of the Cost of Living: New Evidence on the Consumer Sentiment Anomaly— Variation in the University of Michigan consumer sentiment index that inflation and unemployment cannot explain is strongly correlated with growth in consumer borrowing costs, such as mortgage and car loan rates. The authors estimate that including interest payments and homeownership costs in the consumer price index would raise measured inflation from 3% to 9%.
NAR Says the Typical First-Time Homebuyer Age Was 40 This Year, Up from 33 in 2021—but Is This Accurate?— The National Association of Realtors' mail survey, with a 3.5% response rate, puts the typical first-time homebuyer at age 40, while New York Federal Reserve credit-report data for the same period show an average of 36.2 years and a median of 33. The survey underrepresents buyers under 35 by 17 percentage points.
Affordability, Part II— Home ownership functions as the modern marker of full participation in American society, and while most people's incomes have risen faster than the cost of a broad basket of consumer goods, housing has become substantially less affordable over the past generation. The author argues encouraging more housing construction is the most important element of a long-run affordability agenda.
AI Summary.A young K-12 teacher in New York City earns $62,000 versus $37,000 in Jackson, Mississippi, where teacher pay is the nation's lowest. After adjusting for housing, commuting, taxes, and other cost-of-living differences, the New York City teacher is worse off, suggesting salaries are not a source of savings.
Aziz SunderjiHome Economics
Date Posted:
Is Database:
Database
Is Important:
Important
A young teacher in NYC earns $62,000 dollars, compared to the $37,000 earned by a peer teacher in Jackson, Mississippi. Aziz Sunderji finds that “after housing, commute, taxes, and other cost-of-living differences, the NYC teacher is actually worse off.”
Do higher big-city teacher salaries leave teachers better off after costs?
Core argument: A young K-12 teacher in New York City, the country's most expensive school district, earns $62,000, versus $37,000 for a comparable teacher in Jackson, Mississippi, where teacher pay is the nation's lowest.
A young K-12 teacher in New York City—the most expensive school district in the country—earns $62,000. A comparable teacher in Jackson, Mississippi, where teacher pay is the nation’s lowest, earns $37,000. Could salaries be a source of savings, helping to bring school costs here closer to the national average? The chart suggests not: after housing, commute, taxes, and other cost-of-living differences, the NYC teacher is actually worse off.
Takeaways by Macro Roundup® AI
A young K-12 teacher in New York City, the country's most expensive school district, earns $62,000, versus $37,000 for a comparable teacher in Jackson, Mississippi, where teacher pay is the nation's lowest.
After housing, commute, taxes, and other cost-of-living differences, the New York City teacher is worse off than the Jackson, Mississippi teacher despite the $25,000 higher salary.
Teacher salaries appear unlikely to be a source of savings for bringing New York City school costs closer to the national average.
Related Articles:
Where is Standard of Living the Highest? Local Prices and the Geography of Consumption— The authors estimate that geographical differences in cost of living are especially large for low-income US households, and that expensive cities offer college graduates a standard of living similar to affordable cities while offering less-educated households a lower one. A high school dropout household moving from the most affordable to the most expensive commuting zone would see consumption fall 18.5%.
Housing and the American Worker— Wages net of housing costs were once substantially higher in high-density areas than low-density areas for workers both with and without college education, but over the past half-century that urban wage premium has persisted only for college-educated workers, not for those with a high school degree or less.
The Determinants Of Declining Internal Migration— A 10% increase in wages where people live is associated with 3.5% less migration, while a 10% increase in destination wages is associated with 7.8% more migration; a 10% rise in destination home prices is associated with 2.6% less migration. The authors find wage responses alone would have raised migration rates, but housing-related factors more than offset them.
AI Summary.The author argues the housing affordability crisis stems from the sharp decline in 30-year fixed mortgage rates rather than their subsequent rise, since fixed rates let rate dips boost purchasing power, lifting existing-home sales 25% and prices 32% relative to rents. A costly data center buildout is crowding out for-sale housing.
Jason ThomasCarlyle Group
Date Posted:
Is Database:
Database
Is Important:
Important
Jason Thomas argues that high rates driven by the AI buildout and US fiscal policy are “crowding out interest-sensitive sectors like for-sale housing.”
Do falling mortgage rates, not rising ones, drive housing unaffordability?
Core argument: During the pandemic, U.S. existing-home sales rose 25% and house prices increased 32% relative to rents, which the author attributes to the sharp downward adjustment in rates in 2020–22 rather than their post-pandemic rise.
To speak of a rate-induced “affordability crisis” is no exaggeration. But this crisis resulted from the sharp downward adjustment in rates in 2020-22 rather than their post-pandemic rise. With implicit housing costs fixed over the entire term of residency, cyclical dips in 30-year rates boost purchasing power to an extent not observed with 2 floating rate loans. House prices and sales volumes rise predictably in response. During the pandemic, sales of existing homes rose 25% and house prices increased by 32% relative to rents. What the for-sale housing market really needs is a sharp reduction in interest rates. But the U.S. finds itself in the midst of a really expensive data center/supercomputer buildout. The issue is not whether high rates are “crowding out” interest-sensitive sectors like for-sale housing. Clearly, they are.
Takeaways by Macro Roundup® AI
During the pandemic, U.S. existing-home sales rose 25% and house prices increased 32% relative to rents, which the author attributes to the sharp downward adjustment in rates in 2020–22 rather than their post-pandemic rise.
Because implicit housing costs are fixed over the entire term of residency, cyclical dips in 30-year rates boost purchasing power to an extent not observed with floating-rate loans, and house prices and sales volumes rise predictably in response.
The author argues that the for-sale housing market needs a sharp reduction in interest rates, but the United States is in the midst of an expensive data center and supercomputer buildout, with high rates crowding out interest-sensitive sectors such as for-sale housing.
Related Articles:
How America’s AI Boom Is Squeezing The Rest Of The Economy— About a sixth of America's 2% real GDP growth over the past year came from investment in computer and communications equipment and data centres, and one estimate puts the artificial intelligence boom's total contribution to growth at 40%. Average electricity bills rose 7%, attributed in part to data-centre demand, while housebuilding and non-AI business investment slumped and real consumption flatlined.
Rapidly Growing Treasury Supply Crowding Out Other Types of Credit Growth— Roughly half of all fixed income issuance coming to market over the past year has been US Treasuries, meaning investors must allocate a growing share of dollars to financing the government. The argument is that lower government debt would free up more funds for consumer borrowing and business investment.
Honey, AI Capex is Eating the Economy— Capital spending on AI data centers is around 20% of railroads' peak spending as a share of GDP and is rising quickly, having already surpassed the dot-com era peak in telecom investment. Unlike long-lived rail infrastructure, data centers are short-lived assets requiring frequent hardware replacement, and the capital is being reallocated from other sectors.
AI Summary.The authors' Young Family Median Multiple, comparing median incomes of 20-34 year-olds to local median home values, finds that virtually every US metro area met a basic affordability standard in 1969 versus just 21% in 2018-2022. New housing completions trailed new household formations by 3.8m units.
Wendell Cox and Lyman StoneInstitute for Family Studies
Date Posted:
Is Database:
Database
Is Important:
Important
Between 1980 and 2023, home ownership for the US 25-35 cohort has fallen from 50% to 25-30%. @lymanstoneky attributes this to high prices, noting that in 1969 the median home cost ~ five years of a young adult’s income vs. nine years today.
Has housing affordability for young families declined across US metro areas?
Core argument: The authors' Young Family Median Multiple, comparing median personal incomes of individuals ages 20–34 to local median home values, shows just 21% of metro areas met a basic affordability standard in 2018–2022, versus virtually all of them in 1969.
We developed a new indicator we labeled the Young Family Median Multiple (YFMM). It compares the median personal incomes of individuals ages 20-34 to median home values in their local area. Whereas in 1969, virtually every metro area met a basic standard of affordability for young people, in 2018-2022, just 21% did, and over half of the nation’s metro areas were rated as seriously unaffordable for young families, or worse. Between 2012 and 2023, 13.4 million new housing units were completed, while new household formations were 17.2 million, for a supply gap of 3.8 million housing units. Only 9.5 million of these new completions were single-family units.
Takeaways by Macro Roundup® AI
The authors' Young Family Median Multiple, comparing median personal incomes of individuals ages 20–34 to local median home values, shows just 21% of metro areas met a basic affordability standard in 2018–2022, versus virtually all of them in 1969.
More than half of U.S. metro areas rated as seriously unaffordable for young families or worse in 2018–2022 under the authors' Young Family Median Multiple.
Between 2012 and 2023, 13.4 million new housing units were completed against 17.2 million new household formations, a supply gap of 3.8 million units, with only 9.5 million of the completions single-family.
Related Articles:
Inflation Adjusted House Prices 1.3% Below 2022 Peak— Inflation-adjusted US house prices stand 1.3% below their 2022 peak. An affordability index built from median income, national house prices, and 30-year fixed mortgage rates shows that over a three-year span, the monthly principal-and-interest payment on a $500,000 house with 20% down rose 75%, as prices rose 19% and mortgage rates went from 3.07% to 6.43%.
The Price of Housing in the United States— Real housing price appreciation in the United States has varied widely across regions, with West Coast cities rising earlier while prices in other regions remained comparable in real terms to decades earlier. The authors find essentially no relationship between real price growth and zoning restrictiveness in the earliest period studied, and progressively steeper gradients in later periods.
This Is How to Fix the Housing Crisis— US homebuilding falls short of demand, particularly in high-demand coastal markets, with per-person permitting running at less than half the level of two and five decades ago. The author argues Washington could unlock construction by conditioning federal highway funds on state housing reform, the same lever used to push states to set the drinking age at 21.
AI Summary.Because restrictive planning laws prevent housing supply from responding to demand in rich countries, researchers estimate that a 1% rise in a city's migrant population raises rents by 0.5–1%, with other studies estimating larger effects on house prices and median rents. Since migration has only a small positive effect on wages, housing becomes less affordable for natives.
Economist StaffThe Economist
Date Posted:
Is Database:
Database
Is Important:
Important
The ongoing wave of immigration to OECD countries has likely put upward pressure on housing costs. @TheEconomist reports on studies finding that a 1% increase in the migrant population of a region raises housing costs from 1-3%.
Does immigration make housing less affordable where building is restricted?
Core argument: A University of Waikato meta-analysis finds that a 1% increase in a city's migrant population raises rents by 0.5–1%, while Umut Unal of the Czech Research Institute for Labour and Social Affairs estimates a 1% rise in migration to a German district raises house prices by 3%.
Rich-world housing markets are often rigid and bound by restrictive planning laws that prevent supply from responding to demand. The lump of labour may be a fallacy but, without more housebuilding, there is effectively one lump of land that must be divided between a larger population. Research supports this intuition. A University of Waikato [meta-analysis] finds that a 1% increase in the migrant population of a city leads to a 0.5-1% rise in rents. Umut Unal of the Czech Research Institute for Labour and Social Affairs estimates that a 1% rise in migration to a German district leads to a 3% rise in house prices. James Cabral and Walter Steingress, both of the Bank of Canada, calculate that a 1% increase in an American county’s population raises median rents by 2.2%. As migration only has a small positive impact on wages, the rise in costs makes housing less affordable for natives. Housing demand depends not only on the number of people in an area but how densely they live. Often new migrants are willing to accept worse conditions than natives.
Takeaways by Macro Roundup® AI
A University of Waikato meta-analysis finds that a 1% increase in a city's migrant population raises rents by 0.5–1%, while Umut Unal of the Czech Research Institute for Labour and Social Affairs estimates a 1% rise in migration to a German district raises house prices by 3%.
James Cabral and Walter Steingress of the Bank of Canada calculate that a 1% increase in an American county's population raises median rents by 2.2%.
Because migration has only a small positive impact on wages, rising housing costs make housing less affordable for natives, and restrictive planning laws in rich-world markets prevent supply from responding to demand.
Related Articles:
Housing and the American Worker— Wages net of housing costs were once substantially higher in high-density areas than low-density areas for workers both with and without college education, but over the past half-century that urban wage premium has persisted only for college-educated workers, not for those with a high school degree or less.
Why Immigration Is an Urban Phenomenon— Because many immigrants send remittances home or plan to return, part of their spending occurs in their origin country rather than where they work, so local prices apply to only part of their budget. The authors argue this makes immigrants less deterred by high local costs, concentrating them in expensive cities.
Importing Shelter Inflation— Shelter costs make up about 30% of the consumer price index, and their monthly inflation appears to be stabilizing above pre-pandemic rates rather than delivering the disinflation forward-looking data had implied. The author suggests this may be related to a surge in migration, estimated at 3.3m migrants in a year.
AI Summary.Government spending and fiscal deficits exceeding 23% and 6% of GDP late in the business cycle have absorbed savings and pushed mortgage rates to 7%, according to this argument. Higher rates have driven existing single-family home sales down 20% from their pre-pandemic peak, roughly the financial-crisis trough level.
Edward ConardMacro Roundup
Date Posted:
Is Database:
Database
Is Important:
Important
Who is paying for Biden’s blowout spending that exceeds 23% of GDP w/ 6% deficits that suck down savings, drive mortgage rates to 7%, and sales of existing homes down to their trough at the depths of the financial crisis—20% below their pre-COVID level.
Are government deficits pushing mortgage rates up and home sales down?
Core argument: Sales of existing family homes are down 20% from their pre-COVID peak, placing them approximately at the trough reached during the financial crisis.
Who is paying the price of the Biden administration's blowout spending? Unprecedented government spending and resulting fiscal deficits--which exceed 23% and 6% of GDP late in the business cycle--have sucked up savings, driven mortgage rates up to 7%, which in turn has driven sales of existing family homes down 20% from their recent pre-COVID peak. Home sales are now approximately at the trough of the financial crisis.
Takeaways by Macro Roundup® AI
Sales of existing family homes are down 20% from their pre-COVID peak, placing them approximately at the trough reached during the financial crisis.
Government spending exceeding 23% of GDP and fiscal deficits exceeding 6% of GDP late in the business cycle have absorbed savings and driven mortgage rates up to 7%, according to the author.
The author attributes the 20% decline in existing home sales to mortgage rates rising to 7% as government borrowing consumed available savings.
Related Articles:
Inflation Adjusted House Prices 1.1% Below 2022 Peak— Inflation-adjusted US house prices are 1.1% below their 2022 peak, based on the Case-Shiller National Index. An affordability measure combining median income, house prices, and 30-year fixed mortgage rates shows the monthly principal-and-interest payment on the same house rose about 80% over three years, as prices climbed 18% and mortgage rates went from roughly 3.07% to 6.81%.
The Cost of Money is Part of the Cost of Living: New Evidence on the Consumer Sentiment Anomaly— Variation in the University of Michigan consumer sentiment index that inflation and unemployment cannot explain is strongly correlated with growth in consumer borrowing costs, such as mortgage and car loan rates. The authors estimate that including interest payments and homeownership costs in the consumer price index would raise measured inflation from 3% to 9%.
Why Housing May Be In for Another Cost Shock Next Year— The apartment market is projected to shift from renter-friendly to landlord-friendly over the next two to three years as new supply peaks and then falls below prior-cycle levels in every major market. With an estimated 20% of US construction workers undocumented, deportations combined with California rebuilding demand could create a construction labor deficit that raises building costs and rents.
AI Summary.The authors estimate that a 1% rise in a city's manufacturing productivity is associated with 1.5% higher long-run annual earnings, 1.5% higher rents and 2.5% higher home values, lifting purchasing power 0.6% for renters and 1.6% for homeowners. Direct gains favor less educated local workers, while spillovers to other cities disproportionately benefit college graduates.
Richard Hornbeck and Enrico MorettiReview of Economics and Statistics
Date Posted:
Is Database:
Database
Is Important:
Important
A 1% per year rise in a city’s manufacturing TFP in revenue terms between 1980-1990 boosted annual earnings growth by 1.5% per year over 1980-2000. Renters’ purchasing power grew by 0.6%/year, and homeowners’ by 1.6%/year.
Who gains from city productivity growth: renters, homeowners, or distant workers?
Core argument: A 1% rise in city-level manufacturing revenue-based TFP from 1980 to 1990 is associated with an average 1.5% long-run increase in annual earnings from 1980 to 2000, alongside a 4% rise in local employment driven by immigration.
We estimate that local growth in TFPR (TFP in revenue terms) increases the earnings of local workers. A 1% increase in city-level manufacturing TFPR from 1980 to 1990 is associated with an average long-run increase of 1.5% in annual earnings from 1980 to 2000. Local employment increases by 4%, driven by immigration, [and] demand for housing increases. A 1% increase in city level manufacturing TFPR is associated with a 1.5% increase in housing rents and a 2.5% increase in home values. For workers who rent their home, much of the increase in earnings is offset by increases in the local cost of living. For workers who had owned their home, the gains are much larger because they come through both higher wages and higher housing values. We find that a 1% increase in local TFPR increases the purchasing power of renters and homeowners by 0.6% and 1.6%, respectively. [Direct] gains are larger for local less educated workers, [and] productivity growth reduces local inequality, [but] indirect effects of local productivity growth on earnings and housing costs of workers in other cities are economically important and disproportionately benefit college-educated workers
Takeaways by Macro Roundup® AI
A 1% rise in city-level manufacturing revenue-based TFP from 1980 to 1990 is associated with an average 1.5% long-run increase in annual earnings from 1980 to 2000, alongside a 4% rise in local employment driven by immigration.
A 1% increase in local revenue-based TFP is associated with a 1.5% increase in housing rents and a 2.5% increase in home values, and raises the purchasing power of renters by 0.6% and of homeowners by 1.6%.
Direct gains from local productivity growth are larger for less educated local workers and productivity growth reduces local inequality, while economically important indirect effects on earnings and housing costs in other cities disproportionately benefit college-educated workers.
Related Articles:
Where is Standard of Living the Highest? Local Prices and the Geography of Consumption— The authors estimate that geographical differences in cost of living are especially large for low-income US households, and that expensive cities offer college graduates a standard of living similar to affordable cities while offering less-educated households a lower one. A high school dropout household moving from the most affordable to the most expensive commuting zone would see consumption fall 18.5%.
The Economic Geography Of The 2024 Elections— Counties where government transfer programs such as Social Security, Medicare, Medicaid, and unemployment insurance make up at least a quarter of personal income gave Trump 63% of the vote, while counties where transfers account for less than 15% gave Harris 56%. Transfers supplied 18% of total US personal income.
The World’s Rust Belts: The Heterogeneous Effects of Deindustrialization on 1,993 Cities in Six Countries— A study of 1,993 cities in France, Germany, Italy, Japan, the UK and the US finds former manufacturing hubs lost employment on average after each country's manufacturing peak, though a large share fully recovered. After the peak, hubs with a higher initial share of college-educated workers grew significantly faster.
