Private credit overhangs and the business cycle
Back in 2012, three economists published a paper via the San Francisco Fed that looked at nearly every advanced economy business cycle from 1870 forward with the object of understanding the role of credit in the business cycle. And a post yesterday by Matthew Klein at the Financial Times alerted me to the paper.
Now, what the economists found, not surprisingly, was that “financial-crisis recessions are more costly than normal recessions in terms of lost output; and for both types of recession, more credit-intensive expansions tend to be followed by deeper recessions and slower recoveries”. I want to discuss this both in terms of endogenous money and in terms of its implications for the present recovery and proposed recovery solutions. What follows is pretty wonky but very important as a thought piece for framing the economic environment.
The paper by Oscar Jorda, Moritz Schularick and Alan Taylor is a tour de force that puts paid to the notion that the neoclassical economic models which ignore banks and credit were deficient in understanding the dynamics that were developing during the global mortgage credit bubble that ended in crisis in 2007. It needs to be stressed that credit is central to understanding business cycles as a result.
Or as the Jorda paper puts it:
Our results do not speak as to the causes of credit accelerations nor can we make strong inferences yet about the net effects of credit booms, these being goals of our ongoing work. Yet our results would generally seem compatible with the idea that financial factors play an important cyclical role.
Importantly, the Jorda paper concludes that post-crisis, “the USA is seen to have performed as could have been expected given the historical outturn for financial recessions following mid- or hi-tercile credit booms”. So, two obvious question come to mind, one for policy makers and one for market forecasters and investors. The first question for policy makers is this: how can these types of crushing financial crises be avoided? And I will have some thoughts on this at the end of this piece. But the more relevant question for this site is the second one for forecasting and that is: what should we expect the economic path going forward to be given the Jorda, Schularick and Taylor analysis? My thoughts are below.
Let’s put this out here from the start. The reason that credit has such an overriding impact on business cycles is because money is endogenous. And by that I mean that credit is created out of thin air to fuel economic activity, independent of base money created by government. Traditionally, the creation process has gone through banks as financial intermediaries. But the reality is that any economic actor can create credit to another actor endogenously. In recent years, shadow banks have been a key area driving credit creation in advanced economies. The key is whether that credit is fungible and transformable or exchangeable freely within the economic system.
For example, in the United States, during the housing bubble, private-label mortgage securitization was a significant part of the credit structure used to fuel the credit bubble. Financial institutions would originate mortgages that ended up being often of dubious quality because the mortgage originators knew they could package the loans into private-label mortgage-backed securities and offload the risk. But, during the bubble, these securities were considered high quality assets, freely tradeable on the open market, rated AAA by the ratings agencies and usable in lieu of government-backed assets as collateral in rehypothecation chains to create even more credit. In short, private-label mortgages were money created out of thin air, but freely traded and transformable into government-backed assets.
Now, in our credit system, the currency of account is the only legal tender in a central bank’s economic domain. The government is the only economic agent able to create this money. And in a fiat currency world, the government can create money in infinite quantities because the currency of account is simply a government IOU. As the British five pound notes say, “I promise to pay the bearer on demand the sum of five pounds”. This government base money, exactly because it is backed by the government, which is the only entity that has coercive taxing power, is considered the risk free and most desirable safe asset in the economy. But government money is only a fraction of the credit structure in any economy. All other money is created by private sector actors and is not of equivalent desirability because of the solvency risk of those private actors. Hence, private-label mortgage-backed securities are not government money. Rather they are private money which was accepted for a time as nearly as safe as government money.
A financial crisis is the result of a discrimination en masse against private money that was once considered on par with government money in favour of government money or forms of private money most similar to government money. Financial distress and bankruptcy in the private sector causes a recognition that the credit structure has become overextended and that a flight to safe assets is necessary. And as a result, economic actors increasingly discriminate between forms of money that used to be considered freely fungible and interchangeable. The central bank, as lender of last resort, is supposed to step in at this point and aid in that discrimination process by exposing the private economic actors with the best asset base and safest credit structures as very much distinct from those economic actors with assets of dubious quality and in distress. That’s what the Bagehot rule is all about, lending freely at a high rate of interest against good collateral favours high quality financial institutions and penalizes bankrupt ones, eventually restoring order and faith in private money. The central bank is in effect telegraphing to the market that it is willing to create a market in private assets to help transform them into government money where private sector financial firms are unwilling to do so out of fear of counterparty risk.
In 2007, the credit structure in the US was so overextended that the Fed’s attempts to restore order failed massively. The originate-to-distribute model of credit in the US had created tens of billions of dollars of mortgage assets that were worthless because they were based on fraudulent loans.
However, the whole advanced economic system had seen a massive increase in household debt in the period leading up to the financial crisis. Denmark, the Netherlands, Spain, Ireland, and the UK in particular, had astronomical household debt levels concentrated in the mortgage arena. The result, then, was financial fragility when the economic shock of financial crisis hit those economies, leading to a debt deflation and massive housing busts.
So I see the financial crisis as the inevitable result of widespread financial fragility due to high levels of household debt that was exposed when a financial panic resulted from the US subprime fiasco.
The foregoing analysis says that the credit structure at any given time is important in determining how the economy will develop going forward. And while, we have seen a huge wave of deleveraging across advanced economies, credit structures are still overextended, particularly at the household level. And in many markets, government is actively encouraging households to re-lever in order to promote short to medium-term growth. The constraints to growth then still exist until the real value of debt relative to income or asset values can be sustainably reduced.
Debt contracts, as we now treat them, are asymmetric in terms of their burden sharing. The creditor creates credit out of thin air at time T=0 and is considered to have fulfilled its obligation at that time. During the rest of the contractual period, it is the debtor that must fulfil its contractual obligation. If the debtor cannot do so and defaults, the creditor falls back on the sanctity of contract principle to try to enforce repayment. But, of course, unpayable debts cannot be repaid and any collective attempt to repay unpayable debts is going to retard economic growth since new credit cannot be created en masse when previous credit is in distress. This is exactly why growth is slow following a financial crisis, which is the result of overextended credit systems.
The whole system is geared toward helping creditors get paid rather than recognizing that creditors may have been remiss in the capital allocation process and should, thus, share more of the burden of the workout period via credit writedowns or restructuring.
I expect the creditor-centric model to continue to predominate and for policy to be oriented toward boosting the ability of debtors to repay. In the US and the UK, central banks have targeted asset prices in order to improve the debt to asset value ratios that, when low, leads to distress and deleveraging. And everywhere in the advanced economies, central banks have lowered policy rates in a successful effort to reduce debt service costs and give debtors enough breathing room to repay debts out of cash flow. However, this debt overhang is principally a household debt problem and the most important factor in household debt situations is wages. A lack of wage growth is the Achilles heel of the present policy approach.
My longer-term forecasts are predicated in large part on what we are seeing on the wage front. You cannot have high growth rates without a credit accelerator. And you can’t get substantial household credit growth without an increase in income to support it. In the US, the lack of wage growth means we will be in a holding pattern for some time to come. I have consistently said I expect growth to be in the 2% range with a high likelihood that it will fall below 2%. Given the horrible Q1 2014 figures, 2014 will shake out exactly this way. Moreover, given the Fed’s move to a tightening bias, we should expect debt service costs to increase, and correspondingly, for household consumption to decrease into 2015.
What you want to watch right now to ascertain the effect on growth is the inventory accumulation now ongoing. Inventories added 1.5% to Q2 2014 growth. And while that is good for GDP, it is not a sign of early or mid-cycle dynamics. Inventory accumulation is a sign of an overestimation of needed future production due to flagging consumption demand and is thus a sign of late cycle dynamics. If there was any sign that government was poised to intervene with policy accommodation, we might be able to think the inventory accumulation was leading to a mid-cycle pause. More likely, given the lack of policy accommodation, we are very late in this cycle and will see other indicators weaken, particularly corporate earnings. That’s bearish for stocks and for risk assets like high yield that are dependent on companies with more levered capital structures having higher earnings and cash flows. Despite the recent backup in government bond yields due to the Fed’s intention to raise rates in 2015, I think the environment is favorable for government bonds. The yield curve should flatten as short rates increase and long rates remain low or move even lower.
Although I am not a policy wonk and more geared to forecasting, I do think there are some policy implications from all of this. First, it is clear that the slowing growth after a financial crisis is due to the deleveraging associated with financial distress in environments where the credit structure is overextended. So, from a policy perspective, the goal should be to sustainably reduce the causes of the deleveraging. The policies being pursued now are geared to releveraging household balance sheets, which increases financial fragility rather than reducing it. And while this can provide growth in the short to medium-term, it is toxic in the long-term.
I like Amir Sufi and Atif Mian’s thoughts here. Their book, House of Debt, suggests that we should embed burden sharing into mortgage contracts so that the mortgage arena is more flexible during a financial panic. The idea is that in a panic or a down market for housing, there would be automatic debt for equity swaps for the lenders. This would reduce financial distress and the potential for debt deflation by reducing the number of defaults and forced liquidations. More people could stay in their homes and banks could potentially ride out transitory market drops without being forced to write down as much debt.
From a macroprudential perspective, we could also build automatic stabilizers into regulatory rules so that regulations like loan-to-value requirements adjust with the business cycle. Top of the cycle mortgage LTV limits would be lower than cycle trough limits. Debt to income and debt service to income ratios for mortgage contracts would similarly be countercyclical. This would act as an automatic stabilizer, an accelerator at cycle troughs and a brake at peaks. Right now, post-crisis, the opposite is occurring. We have had stricter lending requirements. And I would say these stricter requirements are an overreaction to the laxity we saw at the top of the cycle.
Those are all the thoughts I have about policy for now. None of this looks likely to be implemented anytime soon. Instead, we should expect the household debt ratios to remain elevated. And the result will be slower growth until incomes catch up or debt is worked down.