The Widowmaker

There was a strange back and forth recently on Twitter after John Aziz pointed out Nassim Taleb's wrong call on treasury yields back in 2010, having said that every human should short US treasuries, and that it was a no-brainer. Taleb is a brilliant thinker, so how did he and other intelligent people get it wrong on sovereign debt? 

One of the hardest concepts to understand in finance is why seemingly insolvent governments can keep funding themselves almost for free. It doesn't seem right, or at least it doesn't fit with expectations. Taleb might have looked at the Euro debt crisis and subsequently painted all sovereigns with a broad brush, but I can't be sure of his process. I know I struggled for about 2 years to really understand why countries like Japan and the US had not collapsed from insolvency - looking at all the different angles, learning the basic operational and legal framework of the Federal Reserve. As counter-intuitive as it felt, in the end I couldn't see how the US would ever need to default, or any other country with its own currency (as the legal and operational constructs are similar).

The fact that this is so counter-intuitive is probably an essential feature of why government bond markets do not face the wrath of bond vigilantes. What I mean by this is that, if everyone understood that the government was not at risk of insolvency, there would be less accountability for wasteful spending - possibly even reckless spending would occur, which would itself put the government bond market at risk of losing credibility and an inflationary episode.

Because most people assume the government's finances resemble our own personal balance sheets, there is a genuine repulsion to high debt levels, and those in government normally try their best to reduce spending when possible, thereby retaining credibility enough that markets never seriously doubt the smooth functioning of this particular market.

This counter-intuitive nature is also probably responsible for so many senior analysts believing that when the Fed scales back its bond purchases, treasury yields will rise - which is of course backwards, because QE has never lowered yields in the first place. The implicit assumption is that the US government is spending beyond its means, and can only do this because the Fed is keeping rates low by buying debt - this is, funnily enough, also backwards even though it sounds correct. If the Fed were to remove its stimulus programs today, far from causing a flight out of treasuries, this would actually cause a flight in to the safety of treasuries, leading to drastically lower yields -  possibly even negative yields for shorter maturity instruments. We have seen this in action since the Fed tapered its bond purchases - yields have stopped rising and have even fallen since the announcement.

To get this backwards, one needs to assume that the natural path for interest rates is up, and that the Fed is keeping them down against what is 'natural'. This tends to play in to a lot of natural resentment felt towards central banks and private banks, so it is often unquestioned. Far more likely is that, because all the factors that caused falling long rates and inflation for decades are still in place, and there are powerful new factors since the GFC that have amplified this; de-leveraging; low demand for credit; high supply of savings; falling inflation or even deflation; Chinese investment in overcapacity - the natural path for interest rates is still down.

This is normal in an era that can be classified as a Balance Sheet Recession (or Great Depression/Great Recession), where the private sector is trying to repair impaired balance sheets and so demand for new loans is anemic. There have been at least 3 of these periods in recent history, and the previous 2 did not see interest rates turn around for decades. The Great Depression is particularly insightful, because even when inflation picked up during World War 2, yields still did not rise (more regarding this later on monetary policy) - again, The Widow-maker was a terrible trade, despite conditions that should have made it successful:

A long term history of interest rates and stock prices. Due to a incompelte data I've had to cobble together several data sets; 10 year UST's; Long term government bonds; High grade railway bonds; Dow Jones Industrial Index; Railway shares. Thus, this is not perfectly accurate, but the series are close enough substitutes to give a reasonable representation.

Understanding of what makes the widow-maker trade such an unrewarding trade involves, among other things:

1. Interest rate cycles.
2. Credit and savings demand.
3. Monetary policy.
4. Asset prices and bubbles.
5. Inflationary vs. dis-inflationary periods.
6. Legal framework, central bank balance sheet construction, and what separates Euro countries from other developed nations.
7. Public and private debt levels.
8. Global balance of payments imbalances (think US and China, or Germany and Greece), and the savings glut/great moderation.

I will need to break this explanation up in to parts, as it is a long one, but well worth the time as interest rates affect valuations across all asset classes. Also, yield movements make a lot more sense when you stop looking at them backwards!

China's Vast Reserves

Macroeconomics and finance suffer from huge gaps in understanding, often arising when reality is counter-intuitive, and a much simpler but incorrect interpretation exists and so is commonly accepted. Today, I'm looking at China's foreign exchange reserves and the idea that they represent accumulated wealth that can be freely spent. Most of the time (in my experience) when people consider this topic, they view China as the powerful dragon guarding vast hoards of accumulated treasure or savings, for use when the economy gets in to trouble:

Disclaimer: Not an accurate representation of China's foreign reserves. 

But this isn't really accurate. There are actually quite a few reasons why these reserves are not particularly helpful in a crisis. Today's post is sparked by a discussion over at Macrobusiness, which I identified as a misunderstanding, and I think it's one that a lot of people make, which is to consider these reserves as unburdened assets.

There is presumably an assumption that the Chinese government accumulated these savings by spending less than it receives in terms of net exports, and by generally being prudent and investing this excess cash overseas. This is an easy assumption to make given the media narrative of wealthy China and bankrupt America, but it isn't real or accurate, because China actually has been a net capital importer if we exclude the PBoC's exchange rate operations. The common understanding of the PBoC's reserve position looks like this, but this is not correct:

The best way to understand China's reserve position is to understand how the reserves got there in the first place. From there it will be easier to show that selling those reserve holdings is much less of a threat to the US than China, and that there might not actually be any accumulated wealth at all, because the heavily US dollar denominated side of the PBoC's balance sheet is offset by RMB denominated liabilities.

Accumulating Reserves:

China is unique, in that it runs a current account surplus, and a capital account surplus. Normally a current account surplus will be offset by a capital account deficit as excess income from abroad must invested in foreign countries - a capital account deficit means money is flowing out of the country. So why is this not the case in China if a balance of payments is meant to balance? By including the reserve account of the PBoC, which offsets both surpluses to prevent the currency from appreciating.

Source: Also Sprach Analyst

Because Chinese monetary policy involves a crawling peg, or a peg to a basket of currencies, the PBoC must stand ready to purchase any foreign dollars in order to defend its target rate or band. Both of the twin surpluses, in a normal world, would act to appreciate the RMB, as demand for Chinese exports should push up the currency, as should demand for investing in China. In both cases, foreigners must sell their own currency and buy RMB in order to deal in the domestic market.

The PBoC creates RMB when it buys USD, in the same way as the Fed does when it purchases assets during QE, creating reserves. Because this would be potentially inflationary, the central bank needs to sterilize this money expansion. It does this in two ways: By issuing RMB denominated bonds and by increasing reserve requirements for banks. Issuing interest bearing securities transforms the commercial bank's assets into longer term maturities that are less likely to cause inflation, while increased reserve requirements keep the reserve balances from supporting excess credit expansion.

In both cases, the PBoC's balance sheet is stacked towards USD denominated assets and RMB denominated liabilities. USD assets can be assumed to be US treasuries, or USD deposits but theoretically there can be purchases of private assets if the PBoC really wanted. I'll create an example and say that this reserve accumulation took place in the early 2000's, when the exchange rate was about 8 RMB = 1 USD. Such a balance sheet (simplified) would be a combination of asset-liability matches such as below at the time of the initial transaction:

Fast forward 10 years and the value of both sides of the balance sheet have changed due to asset appreciation and exchange rate movements. For the first balance sheet shown below, the US treasury bond has appreciated in value by say 20-30% due to roll-down and the long term trend of falling rates. This is great for the PBoC. The other side of the balance sheet offsets this though - the RMB has gained 20-30% and the exchange rate is now roughly 6 RMB = 1 USD - but this isn't too much of an issue, maybe in this simplified case the bank has not made a net gain or loss.

The RMB has been allowed to appreciate partially.

There would be an issue with the second transaction though, which represents a larger portion of the PBoC's balance sheet. Here the value of the RMB liability has appreciated by ~25% relative to the USD asset (I've used 30%). The asset being US dollars that don't benefit from the same asset appreciation as US bonds. 10 years later, the central bank's balance sheet might resemble this if marked to market:

This sounds bad, but it isn't all doom and gloom. The PBoC's liabilities didn't really cost it anything to create - central banks have unlimited power to create reserves. Technical insolvency (if this was marked to market) doesn't necessarily mean a central bank can't function, though in order to maintain trust and credibility it isn't the best place to be. We can't really say they are bankrupt, because this institution is the monopoly currency issuer, so they can't run out of money - but that money is the central banks liability, not an asset, which is a little confusing to think of. But given this, aren't these FX reserves essentially free wealth? 

No, for a couple of reasons listed below. But more strikingly, if it was this easy to achieve free wealth - just print money and buy up foreign assets - then every central bank would be getting rich. There are negative consequences for this, and it is a blade that cuts both ways - if accumulating reserves bestowed benefits, doing the opposite will bring costs. It cannot be beneficial both ways, or we have found the holy grail of policy prescriptions.

Take from this section that these aren't reserves that have been saved through hard work, but rather purchased by issuing domestic currency liabilities which still need to be honoured at some point - not so different from the way that commercial banks expand their balance sheets in order to make loans: Of course a commercial bank's assets have increased when it loans money to a customer (the loan is an asset), but so have their liabilities (a deposit or wholesale funding) - we wouldn't call a commercial bank $100,000 richer just because it makes a residential property loan for this amount, because it has also increased the other side of its balance sheet to fund this.

A central bank that issues reserves or longer term debt in order to buy US treasuries hasn't become rich, it has simply expanded it's balance sheet without necessarily creating positive equity. If it then gives away those assets (as people assume it will in a crisis), it's net liabilities increase by the same amount (as liabilities have not changed), taking it further into the region of technical insolvency. 

Is the Federal Reserve rich because it has bought trillions in US treasuries by issuing liabilities to the banking system? No, so the logic should be no different for the PBoC. Similarly, if the US Fed gave away it's treasury holdings to the private banks to help them, what would we think of the Fed's credibility? Not much! It would have no assets left and trillions in liabilities. If the PBoC gives away it's assets to recapitalize the banking system, then it's simplified balance sheet would look more like this:

So who pays for this? Well in the end, most likely the government borrows to recapitalize the central bank, meaning transferring the reserves didn't help, it just changed the mechanism of transfer. Now the government will be inclined to inject the central bank with assets - meaning liabilities of some other institution than the PBoC - such as Chinese government debt or similar.

Unwinding the Reserve Holdings:

A problem we run into in assuming the reserves are wealth is that we need to ignore the exchange rate effect of selling down this USD wealth for use in China. Accumulating reserves has only one purpose, which is to manage the exchange rate below where it otherwise would be to benefit the Chinese manufacturing and export sectors. Clearly, making the opposite transaction would involve appreciating the exchange rate and severely hurting this competitive advantage.

For instance, say China sells it's US treasury holdings - now it has an equal value of US dollars, if we ignore the poor price it would receive for unloading a large amount of bonds onto the open market. This isn't very helpful, it just changes its reserve composition towards USD deposits rather than interest bearing bonds. The PBoC already has US dollars in its reserve holdings to begin with, so it's not especially helpful. But, say they went ahead and did this anyway, what would be the effect?

Initially nothing, except a spike higher in US long rates which would probably subside. Now, they still have US dollars, not Yuan. If the banks are insolvent and need recapitalizing, this may be helpful, but it only makes the banks liquid and solvent in USD. For them to address RMB outflows and maturing RMB obligations which makes up the lions share of operations, this still needs to be converted back to the domestic currency - putting upward pressure on the Chinese exchange rate. 

However, a more likely scenario  is a liquidity crisis being the problem (rather than only solvency), as China's inter-bank market freezes up again, more dramatically than it did last year when SHIBOR moved higher. This would be a similar crisis to the sub-prime credit crunch, and it is not far fetched if defaults begin rising in the realm of WMP's, trusts and real estate loans.

In the case of a liquidity crisis, the problem is that institutions won't lend to each other in RMB. Excess USD liquidity does not help the situation unless it is converted back into RMB, again putting upward pressure on the exchange rate.  The point: if foreign reserves are to help a domestic credit crisis, doing so will reverse the currency peg's benefit. This would unfortunately happen during a time when a sharply appreciating currency would be disastrous for the country's manufacturing and exports, who would likely be already suffering some effects from a freeze in lending. I can't stress enough how unhelpful this would be.

Even still, what if they try and use the USD to save the banking system as outlined further above? As Michael Pettis notes:

"In fact there have been rumors for years that the PBoC would technically be insolvent if its assets and liabilities were correctly marked, but whether or not this is true, any transfer of foreign currency reserves to bail out Chinese banks would simply represent a reduction of PBoC assets with no corresponding reduction in liabilities. The net liabilities of the PBoC, in other words, would rise by exactly the amount of the transfer. Because the liabilities of the PBoC are presumed to be the liabilities of the central government, the net effect of using the reserves to recapitalize the banks is identical to having the central government borrow money to recapitalize the banks... "

"Bailing out the banks, it turns out, is conceptually no different than transferring debt from the banks to the central government. China can handle bad debts in the banking system, in other words, by transferring the net obligations from the banks to the central government, and the large hoard of reserves held by the PBoC does not make it any easier for China can resolve any future debt problems. In fact if anything it should remind us that when we are trying to calculate the total amount of debt the central government owes, the total should include any net liabilities of the PBoC, and that these net liabilities will increase by 1% of GDP every time the RMB strengthens against the dollar by 2%. ..."

Any central bank can create reserves and buy things, and recapitalize banks in its own currency, regardless of foreign reserve holdings - central governments can borrow money to do as well this regardless of accumulated reserves. The FX reserves are a Red Herring, that only help to the extent that foreigners have lost faith in the RMB and the PBoC needs to defend the value of the currency, or to buy international commodities during a war like situation. If the PBoC unloads its reserves in this manner it has severely reduced its credibility - it is not only now completely insolvent (to the tune of trillions of RMB), but does not have any reserves with which to defend the value of its currency... And now the central government will be forced to recapitalize the central bank rather than the commercial banks.

If in doubt, come back to the question I posed earlier: Is the Federal Reserve becoming richer and more secure by purchasing treasury bonds with newly issued liabilities (excess reserves)? This is what the PBoC does when it accumulates foreign reserves. In a sense, you could say that China has been running a Quantitative Easing program for the last decade, only because it is targeting the exchange rate rather than domestic employment and the price level, it must buy foreign bonds instead of its own.

A Simple Explanation

Walter Kurtz from Sober Look points out the growing gap between loans and deposits and asks the question of what is driving this divergence. 

Some other commentators have tried to associate this change in growth rates to some nefarious practice by banks, but the reality is that it is mostly an accounting issue related to QE. The explanation is made simpler if you look at the balance sheet operations of the Fed and commercial banks under QE programs. 

From Soberlook:

"... The last one however is particularly intriguing because the $2.4 trillion gap between deposits and loans is a familiar number. The excess reserves in the banking system is now ... also around $2.4 trillion ...

The chart below adds bank reserves held with the Fed to loans and leases - and the gap "disappears" (here we use total reserves vs. just the excess reserves, but the difference is not material to this trend.)...

 

... Coincidence? Perhaps. But if there is any validity to the explanation #4 above, it would suggest that QE, which is directly responsible for the $2.4 trillion in excess reserves, was not helpful (and possibly harmful) to credit growth in the US...."

This is not a coincidence at all. While I think the explanation of #4 is a bit off, at least it gets to the understanding that this is a consequence of reserve balance growth. I don't particularly agree that QE was harmful for credit growth, but that's going off on a tangent. Back to the loan-deposit gap:

A direct consequence of banks acting as intermediaries for QE is that they accumulate excess reserves, as well as creating deposits in the process, because the end point of QE is to swap a privately held bond for a privately held deposit. This differs from another more common form of deposit creation, whereby banks make loans by expanding their balance sheet. Under the normal loan process deposits and loans grow together:

Under QE, the Fed purchases financial assets by issuing newly created reserve balances - ex nihilo.  The balance sheet operations are fairly simple, and at the end of the asset swap the bank customer ends up with a deposit at the bank. For the bank this is a liability, and the attached asset is the reserve balance at the Fed. The bank's balance sheet looks like this:

1. The bank purchases the bond from the private sector by issuing a deposit.

2. The Fed purchases the bond from the bank by creating reserves, which it swaps for the bond.

Because there is no loan attached to the creation of these deposits, there is a buildup of excess reserves, and naturally a gap will grow between the level of outstanding loans and the level of outstanding deposits. Nothing sinister or complex, just a consequence of bank inter-mediation that should exactly match the value of excess reserves.

Shifting to Neutral on US Equities, Better Value Elsewhere in 2014

While I am not prepared to become a bear just yet, I am definitely becoming neutral on US equities through 2014. At the very least a correction is about due, even if there is no catalyst for a panic. I could become a bear later in the year, perhaps Q3-Q4, as I actually see a few risks building at the same time as the Fed will be slowing down it's easing cycle. Until there is a real catalyst for a crash though, I am only prepared to say that I think the index will finish the year +/- 5% from it's current value, roughly flat. My position is that if you want to own equities, there are many other countries trading at bargain valuations because they are out of favour. US equities are now fully priced, but countries in peripheral Europe are still trading well below long term value, and Australia's slow recovery has kept it from becoming overvalued.

Sentiment became euphoric towards the end of last year, when the Fed decided not to taper in September. Analysts, rather than shifting their estimates back 2 months to the December meeting, walked the taper back to well in to 2014, some even making the call that there would be no taper at all. This might have been sound logic at the time, but it did not eventuate - the Fed announced the taper this December. What I draw from the analyst estimates is that most overestimate Bernanke's willingness to support asset prices at any cost, and most are probably already positioned as net long as possible, given the narrative heading into December was that the Fed would keep running its extraordinary easing at 85 Billion per month and equities might go into a blow off top. Now they are still positioned for this outcome, yet the reality has changed, and the market has not reacted negatively yet.

The market did not correct after the announcement to any noticeable degree for several reasons, a few being that:

1. The taper doesn't actually begin until January.
2. The Fed strengthened forward guidance and made it clear that the taper was not the same as tightening, and that rates would stay at zero for as long as needed.
3. Tapering is still easing at 75 Billion per month, not tightening.

I am unenthusiastic about US Equities now specifically because:

1. Valuations are very high, by many measures that I care to watch - US equities are now some of the most overvalued in the world. At a CAPE-10 of 25.6 it is unlikely 5-10 year returns will be decent from here.
2. Sentiment is close to euphoria - it is unfashionable to be bearish, bears are capitulating and there is a widespread belief that the Fed will not allow asset prices to fall (I think this is misguided).
3. Extraordinary stimulus is winding down over the course of this year, the market is not positioned for this.
4. Unemployment benefits are ending for many.
5. Inflation is falling, despite efforts to raise it.
6. China faces a very difficult dislocation as it attempts to "un-repress" its economy, or re-balance away from credit fueled investment growth - this will become more obvious and problematic throughout the year.
7. The recovery, though weak, is maturing after 5 years.
8. Last year was driven mostly by multiple expansion and buybacks.

Crashes Without Rate Rises:

Something I have begun discussing in this blog, and will discuss further, is the idea that in the absence of an active cash rate policy and the move towards balance sheet monetary policy, asset purchase programs have replaced the cash rate as a signalling device. In an inflationary environment, markets normally don't crash without some sort of monetary policy tightening, normally a rising policy rate is a necessary but not sufficient condition, while an inverted yield curve has been, in recent history, a near guarantee of a recession. This precondition cannot occur currently, as the Fed will not raise short term rates, so this is taken by many to conclude that a crash is unlikely. However, when the economy's natural tendency is towards de-leveraging or deflation, as it is now, policy need not be tightened to instigate asset price falls, instead reduced stimulus is enough to spark a downturn if the economy can not yet stand on its own.

The best analog for the current market conditions is the Great Depression. The market crash of 1929 was followed by a deep and painful depression, though by 1937 (as Wikipedia puts it) the recovery appeared to be under way:

"By the spring of 1937, production, profits, and wages had regained their 1929 levels. Unemployment remained high, but it was slightly lower than the 25% rate seen in 1933. The American economy took a sharp downturn in mid-1937, lasting for 13 months through most of 1938. Industrial production declined almost 30 percent and production of durable goods fell even faster."

This was also a period characterized by falling long term rates, low inflation, global recessionary conditions and a general de-leveraging of the private sector with an opposing growth of the public sector balance sheet - which sounds a lot like the current environment. Of course the monetary system is improved and unemployment lower, conversely market valuations are slightly higher and private sector leverage is also still quite high - there are definite differences, so we shouldn't expect an exact replica to occur, but since this is the closest existing experience with the current macro environment it is worth paying attention to.

Opinions vary on the exact cause of the 1937-1938 recession, depending on your political alignment and your school of economic thought. We can't really know for sure, and I don't care to open a can of worms. What we do know is that the cash rate was not raised and the yield curve did not invert - the two normal preceding conditions of recessions were not present:

After a powerful rally off the lows of the early 30's, in 1937 the DJIA corrected 50% - the Federal Reserve was not raising rates and the yield curve was far from inverting.

The point here is to acknowledge that it is not a sufficient argument to say that, because the Fed will not tighten any time soon, this implies asset prices can not crash. These are not ordinary times, so it doesn't pay to look for ordinary analogs and it is possible for a recession/crash to occur without tightening.