House of Debt

4/19/2021

Messy notes

Atif Mian and Amir Sufi explore the causes of the 2008 financial crisis, also known as the Great Recession. David Beim shows in a 2009 study that household debt increased as a fraction of GDP before both recessions in 2008 and 1929. Charles Persons writing in 1930, Frederick Mishkin writing in 1978, and Martha Olney writing in 1999 describe the 1920s as a turning point after which household debt surged and savings dipped. Barry Eichengreen and Kris Mitchener writing in 2003 show that household debt surpassed household incomes in that decade. Peter Temin writing in 1976 showed that household spending dropped. These trends surrounding the Great Depression of 1929 strongly resemble trends surrounding the Great Recession of 2008.

Trends were similar around the world as Reuven Glick and Kevin Lansing show in a 2010 study and as IMF researchers emphasized in another study two years later. Mervyn King in a presidential address prior to the 08 crisis made the same point: that severe recessions follow periods of heavy debt accumulation by households and non-financial businesses. A 2008 study by Carmen Reinhart and Kenneth Rogoff linked recessions in Japan and a few European countries to current account deficits (international borrowing). Oscar Jorda, Moritz Schularick, and Alan Taylor went further to include more countries and show that private debt was closely related to the severity of recessions around the world.

But still, these studies do not establish a causal link (or do they?). Three alternative views make sense of the recession without acknowledging the importance of private debt. The first holds that people's unmet expectations regarding asset prices lead to recessions. This resembles a "fundamental" view of an economy driven by rational actors. The second view is similar but it claims economic actors are irrational, driven by "animal spirits." The third is the monetarist or "banking" view which holds that recessions occur when people find themselves in a sudden credit crunch. Only the monetarist view provides a way out of recessions: increase access to credit. Mian and Sufi promise to show evidence for yet another view that features private debt.

And while insurance protects people against financial loss, mortgages make them more vulnerable to it. Banks typically have senior claims on homes in residential mortgages while the homeowner has a Junior claim. If house prices fall, the homeowner’s equity (their stake in the house) takes the first blow. Since wealthy people own more stocks, bonds, and other non-residential financial assets, they effectively own these senior claims on mortgages through the banks along with the interest payments. A poor man’s debt is a rich man’s asset. Thus wealthier households also have lower leverage. Certain regions in the US saw more severe decreases in house prices and thus higher rates of default. Defaulting on a mortgage leads to foreclosure and hurts one’s credit score, but it also hurts neighboring households as well (an externality) through fire sales. So falling asset prices kick off a vicious cycle: they raise default rates which lead to depressed fire sale prices which lead to even more defaults and even lower asset prices. Mandating judicial foreclosures slows down this process. Of course as households lose their wealth, consumption slumps.

The banking or monetarist view suggests that the collapse of Lehman Brothers and AIG kicked off the recession by constraining credit. However, the year before September 2008 featured a significant decline in household spending on durable goods (auto and home improvement) so it stands to reason that something else happened before lehman collapsed. Instead, consumption and household durable investments drove the recession as they superseded the decline in business investments. So the banking collapse came later. Furthermore, some people argue that a decline in house prices alone causes people to cut back on spending. But this view ignores that different people cut back differently. That is, it ignores that people have different “marginal propensities to consume.” And indeed research shows that households with more debt cut back more on spending.

The fundamentals view can neither explain the decrease in household spending nor can it point to natural correcting forces that justify monetarist policies. A corrective policy of low interest rates suffers from a liquidity trap whereby the zero lower bound on the nominal interest rate leads people to hoard cash, now risk-free because of the zero interest rate. And that's for savers in the positive. Borrowers in the negative will be busy paying off their loans anyway. Another corrective market response under the fundamentals view is that firms would address lower consumption by lowering their prices. This however leads to Irving Fisher's "debt deflation" cycle whereby indebted households facing wage cuts will lower their spending which will induce even more wage cuts.

The fundamentals view posits strong assumptions on employment as well. It assumes wages are flexible enough to fall and people are mobile enough to relocate in pursuit of higher-paying jobs so that unemployment should not rise. However, the data show that wages did not fall and that people did not move. Some studies suggest a skill mismatch, but the evidence there is weak. Others suggest that delayed foreclosures act as a kind of unemployment insurance for people who choose to stay for some time in the houses they can no longer afford. Regardless, unemployment persists.

Credit was aggressively extended to marginal borrowers nationwide. Almost 5 million households that would not have gotten credit in previous years ended up getting it and then losing it by 2012. And this boom took place despite falling incomes among marginal borrowers. While the US economy did see productivity gains, they were not seen where debt burdens were growing. Others argue that people had irrational expectations or exuberance that caused credit (lenders) to cease on the opportunity and flood marginal borrowers with money collateralized by rising home prices. But looking at mortgage credit growth in cities where the housing supply could grow unencumbered (think: flat, in-land cities like Indianapolis), we see that prices were stubbornly low. Prices only rose in cities where new houses were difficult to build like in San Francisco. But mortgage credit grew in both types of cities regardless of whether prices there were increasing or not. This means that the lending boom fueled house-price growth, not vice versa. But house-price growth in turn fueled household borrowing for expenditures on home improvement and consumption. People thought they were wealthier. It’s called a “housing wealth effect” that rids households of "borrowing constraints" that a rational household would like to rid itself of in case of expected higher income, But this explanation is gibberish if one needs a house and if their income is falling as we saw in low credit score regions where mortgages surged the most. Households proved to be "myopic" or "hyperbolic" in their consumption.

ch 7

The Asian financial crisis of 1998 surrounding the Thailand housing boom is closely linked to the early 2000s boom in the US. The Asian banks had borrowed in US dollars, so when investors fled, the local central banks were not able to serve as a lender of last resort and the IMF stepped in. However, because IMF loans come with strings attached, East Asian economies learned their lesson and began collecting large stock piles of dollars. By building up their foreign reserves or war chests, foreign central banks poured money into the US economy, $100 billion annually between 1990 and 2001.

Local banks in the US, of the sort captured by George Bailey in it’s a wonderful life, are subject to local or idiosyncratic risks. Although the federal reserve would be established after the great depression to prevent runs on the bank, the US government established the department of housing and urban development to promote mortgage securitization through government sponsored enterprises (like Fannie may and Freddie Mac). Local banks would benefit from avoiding concentrated local risk and government sponsored Enterprises would mandate minimum conforming requirements to avoid lending loosely while pooling from across the country and selling claims on them in the form MBSs. Securitization was profitable and There was a lot of money flowing in from foreign central banks. Naturally, private lenders entered the market. They introduced tranched MBSs as they were largely unregulated and thus created a market for non-prime nonconforming conventional loans. Thus securitization and trenching “transformed global capital inflows into a wild expansion of mortgage credit to marginal borrowers.” It was made worse by assuming mortgages in a pool were uncorrelated but there were many willing buyers who didn’t know better.

Studies show that mortgage originator’s new that they could unload poor quality mortgages to investors through securitization at high prices and so they eased lending standards and lent deliberately to weak borrowers. Across the industry of private label securitization, lenders did not use obvious information like FICO scores or the size of down payments and they misclassified whether mortgaged homes were investor- or owner-occupied. Studies also show that supreme mortgages originated in 2006 and 2007 where of substantially lower quality. The average combined loan to value ratio increased, the fraction of low documentation loans increased, and the subprime rate spread decreased.

What was unique about this crash was not the fraud and misrepresentation in securitization. Rather, it was the fraud in combination with the accumulated household debt.

ch 8

In manias, panics, and crashes: history of financial crises, Charles Kindleberger develops the axiom that asset price bubbles depend on the growth of credit. But what is a bubble anyway? Do experiments show that when people have public knowledge about the future payments of a security, that is its exact expected value, They still traded in an experiment at large premiums to its actual price are they severely under price it. Another experiment also shows that if they can borrow to carry out these trades then the price changes or even larger.

So there are indeed some animal spirits but debt plays a big role as well. Optimists in the market borrow money from pessimists Who are willing to lend, and with this leverage they push up the price of assets. But are investors so willing to lend? Debt is a safe asset, much safer than equity because it allows investors to ignore fraud or theft carried out by the borrower or the issuer. The contract says they’ll get their money anyway.

ch 9

Financial institutions like banks have a unique balance sheet. Loans are their assets. So when loans can’t be repaid, the holders of the most junior of the bank’s liabilities, the shareholders, take the first blow. Non-deposit debt holders hold more senior claims on the bank’s assets. The most senior however are the depositors themselves. If a bank’s depositors are threatened, it might face a run, and the threat extends to the country’s payment system. So saving banks makes sense but what added a layer of complexity in 08 was that banks also listed among their liabilities short-term, uninsured money-markets funds. Banks were indeed stressed as evidence by the surge in commercial paper prices around the time Lehman crashed. The treasury stepped in to fund banks against those liabilities and that helped keep the payment system and the banks alive but the treasury didn’t stop there. In late 08, Injections by the US treasury increased bank debt and equity by $130 b. This was far more than was necessary to spur lending. In fact small businesses surveyed didn’t not cite access to credit among their main concerns during or after the crisis. Instead, low sales was the issue. But the banking view against which Mian and Sufi argue in this book promotes bank support rather than consumer support.

ch 10

What does consumer support look like? Policymakers in the Obama administration conceded that they could’ve done more work supporting households after the great recession. But why did financial institutions receive disproportionately more support than households?

The authors point to two market failures. Microeconomic failures arose because securitization abstracted the relationship between homeowners and lenders thereby impeding renegotiation. Mortgage servicers who served as intermediaries found it too difficult and expensive to untangle mortgages, identify holding parties and write down principal amounts even though all parties involved had incentive to do so given fallen house prices. One reason services didn’t do this was the fear of strategic default by other homeowners who also sought to renegotiate their terms.

Macroeconomic failures related to the zero lower bound and nominal rigidities prevented the economy from adjusting to the large decline in consumption. Studies showed and economists later argued that debt forgiveness would indeed have quelled the recession by boosting household spending.

Both the 1819 levered-losses crisis as well as the 1930s financial panic featured debt forgiveness that made both debtors and creditors better off. People were defaulting on their mortgages and so the contract initially signed between services and investors needed to be scrapped anyway. Cram downs were an alternative proposed later even by the IMF. These involved modifying mortgage principals under chapter 13 bankruptcy without securing the creditors’ consent. The household with deal with a federal bankruptcy trustee. A study from 2013 shows that such debt forgiveness causes an increase in people’s income and employment probability, which helps the entire economy.

Useful as it was proven to be, forgiveness generally of course also raises questions of moral hazard. However, homeowners were not sophisticated individuals who took advantage of naïve lenders because they understood house prices were artificially inflated. Also, the decline in house prices was beyond the control of any individual homeowner. The question is not whether homeowners were to blame for the crisis, nor was it whether government should intervene. Both homeowners and creditors were culpable and the government spends money anyway. It was a question of how to distribute losses. Distributing losses “between both debtors and creditors is not only fair, but it makes more sense from an economic perspective.”

ch 11

Could monetary or fiscal policies could be thought of as alternatives to debt restructuring. A central bank could expand the monetary base by purchasing securities from the non-bank public or directly from banks. Doing so would cause inflation and effectively transfer wealth from creditors to debtors who have a higher marginal propensity to consume as discussed above. The US government did this however. Banks were swimming in reserves and credit was available shortly after September 08, but people were not borrowing. Lower interest and mortgage rates did not work either because “more debt is not the solution to a debt problem.“ Faced with the zero lower bound, some argue that the mere expectation of inflation would simulate a negative interest rate and drive spending. To satisfy inflation expectations, however, a central bank would have to oppose its mandate of taming inflation.

Alternatively, fiscal stimulus operating through taxation could work and would certainly be more effective than austerity measures. But the most effective approach would be to directly address the household debt overhang by restructuring debts or by arranging for mortgage cram downs in bankruptcy. These more direct approaches would preserve incentives in the market place and quell populism. Sadly, the aftermath of such crises has been proven to feature political polarization so any policy would be difficult to pass as the years after the 08 crisis have shown.

ch 12

Just as indebted college students or people in training might graduate into a weak job market, homeowners can find their equity wiped out in the aftermath of a house price bubble. Flexible contracts would help both students and homeowners faced with such contingencies. Mian and Sufi propose a shared responsibility mortgage (SRM) where the lender offers downside protection to the borrower (by linking mortgage payment schedules to a local house price index) and the borrower gives a 5% capital gain to the lender on the upside.

SRMs Would guarantee borrowers and lenders at least the same percentage of home equity and thus protect the population from large increases in wealth inequality. By preserving household wealth, SRMs would also protect the entire economy against foreclosures and further drops in house prices. They would protect against sharp drops in aggregate spending and job losses. Of course lenders would also suffer by transferring wealth to borrowers, but recall that lenders generally have a lower marginal propensity to consume than borrowers. The multiplier effect of this automatic, intervention-free stimulus further protects society against recessions. Finally, their unique structure of risk sharing incentivizes risk aversion among lenders and more disciplined borrowing among debtors.

SRMs do not exist however because the US government itself pushes fixed rate mortgages and because tax policy incentivizes the use of debt. Seeing the negative externalities of high leverage, governments should be more willing to discourage rather than subsidize the use of debt. Certainly, debt solves the costly moral hazard problem by disconnecting the lender from the borrower’s outcome thereby providing a “safe” asset. But some contingencies are beyond the control of the borrower and if lenders demand safe assets, the government should supply it. One study shows that crises are preceded by private sector provisions of “safe” assets when safe government debt is in short supply. Noted economists have argued for the use of more equity-like instruments as opposed to debt to shift towards a more robust system.