12.8 — Globalisation and the 2008 Crash

The most important object in the history of globalisation is a steel box.

The shipping container, standardised in the 1960s, cut the cost of moving goods by sea by something like 90 percent.

Before containers, cargo was loaded piece by piece by gangs of dockworkers — a process that could take a week per ship and in which theft and damage were routine. With containers, a ship is unloaded in hours by cranes, the box goes straight onto a truck or a train, and it is not opened until it reaches its destination.

Once moving a component across an ocean costs almost nothing, it becomes possible to make each part of a product wherever it is cheapest. That is the whole of modern supply chains, and it depends on a box.

What globalisation actually was

Four things happened together from around 1990.

Transport and communication costs collapsed — containers, larger ships, cheaper air freight, and then fibre-optic cable and the internet, which made it possible to coordinate a supply chain across a dozen countries in real time.

Trade barriers fell — the Uruguay Round, the WTO from 1995, and China's accession in 2001 (Chapter 9.12).

Capital controls were removed across most of the world (Chapter 9.10).

And two billion workers entered the world economy — China's reform, India's liberalisation, and the ex-Soviet bloc — roughly doubling the global labour force available to capital.

Who gained and who lost

The distributional picture is well documented and it is not what either side of the argument usually says.

The economist Branko Milanović produced the chart that summarises it — global income growth by percentile between 1988 and 2008, sometimes called the elephant curve.

The largest gains, in percentage terms, went to the middle of the global distribution — the emerging middle classes of China, India and Southeast Asia. Hundreds of millions of people moved out of poverty.

The largest absolute gains went to the global top 1 percent.

And the group with the smallest gains sits around the 80th to 90th percentile globally — which is the working and lower-middle class of the rich countries.

In the United States, real median wages for men without a college degree have been roughly flat for decades. Manufacturing employment fell by around a third between 2000 and 2010.

And Chapter 9.12 established the causal split: automation accounts for more of the job loss than trade does, and trade concentrated the losses geographically. Autor and colleagues' work found that American communities most exposed to Chinese import competition had persistent unemployment, lower wages, higher disability claims and higher mortality a decade later.

The gains were diffuse — cheaper goods for everyone — and the losses were concentrated in specific towns. Chapter 9.12 said that configuration produces political backlash, and the second half of this chapter is that backlash arriving.

The 2008 crash

Understanding it requires four mechanisms, and they stack.

Subprime lending

American housing lending expanded to borrowers with poor credit, low documented income, or no verified income at all.

Why lenders did it. Because they did not keep the loan. Under the originate-to-distribute model, a mortgage broker made a loan and sold it immediately. The broker's income was the fee for making the loan, not the interest over thirty years, so their incentive was volume rather than quality.

Securitisation

Thousands of mortgages were pooled and the pool was sliced into tranches with different risk levels — mortgage-backed securities and, built on top of those, collateralised debt obligations.

The theory was sound in principle. Individual mortgages default idiosyncratically; a large diversified pool should be predictable.

And the theory contained one assumption that was false. It assumed defaults were largely independent — that a default in Florida told you nothing about Nevada. That is true in ordinary times and false in a national house price fall, when everything defaults together. The models were calibrated on a period in which American house prices had never fallen nationally.

The ratings

Rating agencies gave the senior tranches of these structures the highest rating, the same as government debt.

And the agencies were paid by the issuers whose products they rated. Internal communications later disclosed showed employees describing structures they were rating as unsound.

The rating mattered enormously because regulation used it. Pension funds, insurers and banks were permitted or required to hold highly rated assets, so the rating determined who could buy it, which determined how widely the risk spread.

Leverage and the shadow system

Investment banks operated at leverage ratios of thirty to one or higher — meaning a 3 percent fall in asset values wipes out the equity.

And credit default swaps — contracts paying out if a security defaults — were sold in enormous volumes without the seller holding reserves against them. AIG sold protection on hundreds of billions of dollars of securities and could not pay when they defaulted.

Much of this sat outside the regulated banking system, in vehicles designed to keep assets off balance sheets.

The sequence

American house prices peaked in 2006 and fell.

Defaults rose. The securities built on the mortgages fell in value. Nobody knew which institutions held how much, because the structures were opaque and the exposure was spread through derivatives.

So institutions stopped lending to each other. The interbank market — which banks rely on daily for liquidity — froze.

Bear Stearns was rescued in March 2008. Lehman Brothers was allowed to fail on 15 September 2008.

The Lehman decision is the most argued of the crisis. Its immediate effect was that a money market fund holding Lehman debt fell below the value of its deposits, which triggered a run on money market funds, which are the funding source for corporate short-term borrowing — so ordinary companies with no connection to mortgages suddenly could not fund their payroll. The crisis moved from finance to the real economy within days.

The response

Chapter 9.9's lesson, applied. Central banks cut rates to near zero, lent freely against a wide range of collateral, and started quantitative easing (Chapter 9.7). Banks were recapitalised with public money. Governments ran large deficits.

Ben Bernanke, chairman of the Federal Reserve, was a scholar of the Great Depression, and it showed. The banking system did not collapse, the money supply did not contract by a third, and the outcome was a severe recession rather than a decade-long depression.

That is a real achievement and it should be recorded.

And the manner of it produced the politics.

Why the politics turned

The banks were rescued. The homeowners largely were not.

Around ten million American homes went through foreclosure. Household wealth losses fell heavily on people who had bought at the peak, and the racial wealth gap in the United States widened significantly, because minority households had been disproportionately targeted for subprime products.

Very few senior bankers were prosecuted. Iceland is the exception, having jailed several.

And bonuses resumed within a couple of years at institutions that had received public support.

Then austerity. From around 2010, most European governments and Britain cut spending to reduce deficits that had been incurred rescuing the financial system. The distributional effect fell on public services and welfare.

And the eurozone crisis followed. Greece, Ireland, Portugal, Spain and Italy faced borrowing costs that threatened default. The structural problem was Chapter 9.10's trinity in another form: a monetary union without a fiscal union, so member states had given up the exchange rate and the interest rate and had no way to adjust except by cutting wages and spending. Greek GDP fell by around a quarter and unemployment exceeded 25 percent.

The political consequences are the dominant fact in rich-country politics since.

Trust in institutions fell. Populist and anti-establishment parties gained across Europe and the United States. Brexit in 2016 and the American election of the same year are the two most-cited outcomes, and there is substantial research finding that support in both cases correlated with local economic decline and with exposure to import competition and austerity.

And the general finding, which this volume has now stated three times. Chapter 9.9 said mass economic dislocation destroys political systems. Chapter 11.6 showed it in Germany. This is a milder version and the mechanism is recognisable: a large group of people concluded that the system was rigged, and were not entirely wrong.

India in 2008

India was affected less than most, for a specific reason.

Its banking system was largely domestically funded, its capital account was not fully open (Chapter 9.10), and its banks had little exposure to the securities involved. The restrictions criticised as backward before 2008 turned out to be protective.

The transmission was through trade and capital flows — exports fell, foreign investment reversed briefly, and growth slowed from around 9 percent to around 6.

And the fiscal and monetary response was large, which supported demand and which contributed to the high inflation and the fiscal deficit of the following years.

What was fixed and what was not

Fixed, partly. Higher bank capital requirements under Basel III. Stress testing. Resolution regimes intended to let large banks fail without public money — untested at scale. Some derivatives moved onto clearing houses.

Not fixed. Institutions are larger and more concentrated than in 2007. The shadow banking sector has grown. Rating agencies are still paid by issuers. And the fundamental problem of Chapter 9.2 — that a leveraged institution with insured depositors and an expectation of rescue has an incentive to take risk — has not been solved and probably cannot be.

Where this shows up in your life

Interest rates were held near zero in the rich world for over a decade, which raised asset prices — houses, shares — which benefited people who owned assets and made entry harder for those who did not. Intergenerational wealth divergence in most rich countries traces substantially to this.

And the practical checklist from Chapter 9.3 applies. In 2006 the arguments were that house prices could not fall nationally, that risk had been distributed so widely that it had disappeared, that the models were sophisticated enough to price it, and that borrowing to buy was sensible. All four are the euphoria markers, and they were being made by people with doctorates.

What the next page covers

Chapter 12.9 covers the two decades since 2001 — the attacks, the wars in Afghanistan and Iraq, what the intelligence actually said and what was claimed, the surveillance apparatus that was built, the rise of the Islamic State, and the honest accounting of what twenty years of the war on terror cost and achieved.