Thursday, 29 August 2019

How worried should we really be about an inverting yield curve?

In an interview a few years ago, Elon Musk responded to a question on what had enabled him to innovate so effectively by saying: "I think it's important to reason from first principles rather than by analogy. The normal way we conduct our lives is we reason by analogy. [With analogy] we are doing this because it's like something else that was done, or it is like what other people are doing. [With first principles] you boil things down to the most fundamental truths... and then reason up from there". (He might also have added that, unlike many others, he has not been constrained by the need to generate a positive return on capital, but I digress).

Musk's comment resonated with me immediately, because reasoning by analogy is something we see a lot in markets, and I have never found such arguments persuasive. They always felt dubious to me, even if I could not articulate exactly why. Musk's comment clarified perfectly for me what the fundamental problem with them was. They were not based on first principles.

A clear example of reasoning by analogy in markets is the frequent anxieties expressed about an inverting US yield curve. In cycles past, an inverting yield curve has often (though not always) preceded a recession, so investors view such inversions as particularly ominous. A widely subscribed belief in the inversion's predictive value can and has had market implications - an inverting yield curve contributed to the 15% market crash US markets experienced in December 2018, which proved to be a false alarm. By 2018 year-end, surveys revealed that a majority of investors expected the US to fall into recession in 2019. Instead, the US went on to print 3.3% GDP growth in 1Q19, and markets surged. More recently, the US yield curve has inverted again, which has contributed to another savage decline in US and global markets in August.

While you often hear concerns expressed about an inverting yield curve, it is much rarer to encounter a good first-principles explanation of why it has preceded past recessions and why we should be so concerned. Most investors reason purely by analogy, and are content to note that an inverting yield curve has been associated with past recessions. But as Keynes once said, if you do not know the underlying causal factors driving past experience, you are in no position to know whether present conditions are sufficiently similar to render past experience a useful guide.

One of the most common arguments made is that the bond market is more efficient than the stock market, and the bond vigilantes see recession coming before the equity guys. Baloney. Bond markets are buying 30yr government bonds at negative yields at present. Not very smart/efficient. Equities, being lower down in the capital structure, are also at the 'sharp end of the whip', so to speak, in terms of upside and downside exposure. If any market should sniff out trouble/risk first, it should be the equity market. There is no sound reason to believe the bond market is more efficient than the stock market, in my opinion.

The best first-principles explanation I have been able to uncover as to why a yield curve inversion has preceded many past recessions (and I would love investors to educate me in the comments if they have a more insightful explanation) is that in decades/cycles past, many banks/financial institutions borrowed short term wholesale money, and lent it long. In other words, there were many financial institutions running loan-to-deposit ratios (LDRs) meaningfully above 100%. So long as such institutions were able to borrow in the wholesale market and on-lend with healthy spreads (i.e. with a positively sloped yield curve), they did so, and grew their loan books (and profits) vigorously.

However, when short term rates rose sharply and began to exceed long term rates, the lending spreads of these institutions were sharply squeezed. Even if they had taken measures to protect themselves from interest rate risk on their back book (which sometimes they hadn't), the spreads on their new loan originations sharply deteriorated. This, predictably, resulted in a sharp slowdown in new credit origination, which in turn triggered an associated reduction in aggregate economic demand, and thus meaningfully slowed economic activity - particularly because virtually all recessions in the post-war period were preceded by years of double-digit system credit growth.

However, in the post-GFC period, a lot has changed. Interbank and other short term wholesale funding markets were severely disrupted following Lehman's collapse (whose own collapse was driven as much by a liquidity crisis - its short term funding lines dried up as counterparties lost confidence - as it was losses on the asset side of its balance sheet). The cascading effect on funding markets resulted in liquidity crunches and near-death experiences for many other financial (and non-financial) institutions dependent on short term wholesale money to fund their balance sheets/operations, and necessitated massive Fed intervention/bailouts to avoid large portions of the financial system collapsing.

The legacy of the crisis has been not only a voluntary change in bank behaviour (once bitten, twice shy), but also a significant tightening in banking regulation, including much stricter liquidity coverage requirements. Very few systemically important financial institutions are today running their businesses with any meaningful reliance on short term wholesale funding markets. Today, the vast majority of US financial institutions rely almost exclusively on deposits and (where necessary) longer term wholesale funding instruments, and the system LDR (loan to deposit ratio) in only about 70%. Indeed, the banks have (and have had for a while) excess deposits, and for the major banks at least, many of these deposits cost nothing or close to nothing (well below short term Fed rates).

It is true that QE has contributed meaningfully to this excess liquidity/deposits/reserves situation, and one might reasonably attempt to argue that banks will stop lending if they can make 2-2.25% on their excess reserves parked with the Fed if the yield curve inverts. But this ignores the fact that (1) most banks have been (and continue to) lend at rates significantly above the US 10yr (new mortgage rates are currently 3.9% in the US, for .e.g, vs. 1.5% for the 10yr), and unlike Japanese banks, US banks are managed for shareholders and RoE maximisation); and (2) US banks have been sitting on excess deposits/reserves for the entirety of the post-GFC period, and have been unable to find a useful place to lend them, even in a world with a meaningfully positive yield curve. That was almost never the case pre-GFC in periods of economic expansion. That's why LDRs remain so low; excess reserves so high; and credit growth so anaemic (see below). If credit growth is already slow and banks already aren't lending much, then an increased incentive to keep funds parked with the Fed will have limited effect, because that's already the status quo.

This brings me to another important point - the level of pre-existing credit growth. Over the past decade, private sector credit growth in the US (ex fiscal) has been very modest. This is extremely important, because it's hard to have a significant slowdown in credit growth triggered by an inverting yield curve, and the associated aggregate demand ramifications, if pre-existing credit growth has been modest.

It might surprise many people accustomed to reading bearish prognostications about the outlook for the US economy and the level of aggregate system debt, but US household debt to GDP has actually fallen over the past decade, from about 100% of GDP at the time of the GFC, to just 76% at present, and credit growth has continued to lag nominal GDP growth throughout the recovery, including the recent fiscally-induced acceleration (US household credit growth was just 2.3% in the year to 1Q19). The data is easily accessible for those who care to look. Both banks and households - chastened by the GFC - have borrowed/lent very prudently in the crisis' aftermath. This means there is less capacity for a sudden slowdown in credit growth due to tightening lending spreads, as loan growth is already modest. And US consumer spending - let it be remembered - drives about 70% of US GDP.

Corporate credit growth has been somewhat higher, and is therefore more of a risk, but credit growth has still been only a middling 7%, and at 52% of GDP, is not unreasonably high. After a dramatic decline during and post GFC, corporate debt-to-GDP has slowly risen back to its pre-GFC levels, but corporate debt was not the cause of the GFC, and interest rates (and hence debt servicing burdens on corporates) today are significantly lower.

Furthermore, the complaint we have heard most often over the past decade has been that US companies have been buying back too much stock and not investing enough, rather than investing too much, and with a few exceptions (e.g. shale activity), most of the over-investment that has occurred has been localised and equity-funded (VC funded tech). US corporates have been significantly under-investing over the past decade relative to the availability of credit (and their reinvestable profits), and so by extension, corporate investment is unlikely to sharply slow due to a reduced marginal appetite from banks to lend (shale being a probable exception here). Furthermore, even if financial institutions were to start rationing credit to businesses at the margin, falling sovereign yields have created a large class of financial institutions such as life insurers and pension funds that are desperate for yield, and so it will be relatively easy for creditworthy (and maybe also non-creditworthy) businesses to issue bonds at low cost. And this is before mentioning the many alternative 'credit' funds being set up at present to search high and low for any sort of yield. In such an environment, a generalised lack of access to credit to finance investment seems far fetched.

One of the key things that has changed in the developed world in the post GFC period, which many investors are still yet to have wrapped their heads around, is that we have moved from an environment of capital scarcity into a period of capital abundance. This has turned many of the prior economic relationships that existed in decades past (and which still prevail in emerging markets today) on their head. Richard Koo's excellent book The Holy Grail of Macroeconomics has a Yin/Yang model that discusses this in more detail, which I highly recommend (drawing on Japan's now three-decades of experience in having transitioned to an excess-capital position). Sure, when capital is scarce and needs to be rationed, a change in long rates that dramatically slows lending growth from vigorous levels can absolutely be expected to slow the economy. But that is not the world in which we are currently living (in the developed world).

All told, an inverting yield curve could prove to be far less predictive than it has in the past, and there is a decent chance that many of those 'reasoning by analogy' will be surprised by the resilience of the US economy, and its continued refusal to behave as it supposedly ought to. That does not mean there are not other important risks to the economic outlook (I have discussed the risk of a VC blow up in prior blog posts, for e.g., and the trade war also continues to impact corporate investment activity), but in isolation, betting on a recession merely because the yield curve has inverted appears to me to be a risky proposition.

I would love to hear from anyone who could critique this perspective. The anxieties (and algo bot reactions) to an inverting yield curve keep tanking markets from time to time, and if people are wrong about the predictive value of the indicator in the current environment, there is potentially significant money to be made on the long side buying into these sell-offs.