Frustrated by shortcomings of available tools I have created an indicator yielding timely and exact probabilities for US recessions.
Please find the report here (freely available): Lighthouse Macro Report: Introducing LRPI
Frustrated by shortcomings of available tools I have created an indicator yielding timely and exact probabilities for US recessions.
Please find the report here (freely available): Lighthouse Macro Report: Introducing LRPI
Tracking flows in and out of major exchange-traded funds (ETF) can reveal investors’ risk appetite.
The 23 largest ETF’s by assets under management represent approximately 50% of all US ETF assets.
In September, the largest inflows in percent of assets were registered by domestic equity (SPY), gold mining stocks (GDX) and high yield bonds (HYG):
Source: IndexUniverse.com, own calculations. Note: we eliminate the performance effect on AuM, hence end-of-month AuM will differ from actual numbers (which include performance effect).
Largest outflows relative to assets were seen by Inflation-protected Treasuries (TIP), Nasdaq (QQQ) and 2x inverse S&P 500 (SDS).
We smooth out monthly noise by grouping ETF in asset classes and looking at flows cumulative over the last twelve months:
You will notice the dollar amount flowing into international equities coming close to the one for domestic equities, and risky bonds attracting almost as many dollars as “safe” bonds. This is an expression of investors’ risk appetite. In “risk-off”-times, investors seek to repatriate investments (rather domestic equities than international) and prefer the safety of Treasury bonds over high-yield or junk bonds.
By looking at growth in percent of assets, the picture becomes even clearer:
Risk appetite is then calculated on the basis of various ratios (all on 12-month rolling data):
% risky bonds: net flows into HYG, JNK relative to TLT, TIP, BND, LQD, MUB
% international equities: EFA, EEM, VWO versus SPY, VTI, IVV, DIA, QQQ, IWR
non-investment grade versus investment grade: HYG, JNK versus LQD
emerging markets equities versus developed markets equities: EEM, VWO versus EFA
Short equities versus long equities: SH, SDS versus SPY, VTI, IVV, DIA, QQQ, IWR.
While US investors are less keen on international relative to domestic equities, choices within the international equity asset class and short versus long point towards a high risk appetite.
Within fixed income, risk appetite, while not at record levels, is rising since May/June as investors flock towards riskier bond ETF’s and corporates rather than Treasuries.
Lakshman Achutan, ECRI (Economic Cycle Research Institute) made a recession call for the US on September 30, 2011 (and confirmed it multiple times since then).
Gary Shilling, titling his August letter “Global Recession”, says “We are already in a global recession.”
However, equity markets don’t think so, with the S&P 500 trading less than 10% away from a new all-time high. Only one side can be right. Could this be a repeat of October 2007, when the S&P 500 hit new all-time highs mere six weeks before the “Great Recession” began? Are so-called leading indicators, as used by the Conference Board, still reliable?
According to Mish Shedlock, these are the CB’s 10 leading indicators:
As seen above, the stock market (#7) is a terrible indicator of future economic performance. While the slope of the yield curve (#9) was a very good leading indicator in the past, it would be difficult for the yield curve to invert in the current environment. Long bond yields would have to go negative as the Fed promised near-zero interest until 2014, and possibly longer.
Usually, central bankers ‘hit the brakes’ (raise interest rates) for fear of inflation induced by strong demand and brisk economic growth. The yield curve inverts, a recession follows, and inflation recedes.
But what if the economy went into recession with short-term interest rates being near zero already (as seen, for example, in Southern Europe)?
Over the past 50 years, all recessions were easily explained by central bank action and oil price shocks:
The following chart shows how the Fed increased interest rates ahead of each of the last 9 recessions. Black line: absolute level of Fed Funds rate; blue line: increase in %-points from the prior post-recession low. Right-hand scale for absolute data, left-hand scale for percentage changes (negative absolute numbers merely for better formatting):
The same logic applies to the crude oil price (log scale):
I have looked at many indicators from every angle. Some have to be smoothed to cancel out short-term “noise” in order to prevent false signals (used mostly 3-months moving average). Some data do not give good signals unless you look at decline from recent peaks. Other data need to be trend adjusted (number of miles driven, for example, benefits from rising number of cars and population).
The following indicators have been tested for false positives (calling for a recession when there was none), missed recessions, the confidence I have in to work in the future, possible lead time, what it said in H2 2011, the current likelihood of a recession and the importance (1-3) I assigned to it for a weighted overall recession probability:
Not all recessions are equal, and no single indicator is perfect. It is therefore hard to draw the ‘trigger’ line. Set it too low, and you will miss some recessions. Position it too high, and you will get a lot of ‘false positives’ (indicator signals a recession when there is none).
I have drawn boxes shaded in red. The upper end of the box is the highest level necessary to catch all recessions; the bottom end is the lowest setting necessary to avoid all false positives. The taller the box, the less confidence the indicator deserves. If the read-out is somewhere in between those bands, recession probabilities are assigned.
At the risk of boring you with details, here are the remaining charts:
Want to build a house? Need a permit! Any decline in permits of 25% or more from prior peak and you can bet on a recession. Missed the one in 2001 though. 2011 was a close call. Absolute level still below the lows of 1990/91 recession. Population was 250m then compared to 315m today.
UoM Consumer Sentiment: One false positive (2005), one miss (1981). 1980-82 were back-to-back recessions, so let’s not be too harsh about that. Declines of 25%+ indicate recession. 2011 was a close call, but currently out of the woods.
If you run a business you need electricity. Sure, weather has an impact (electricity use in the US peaks in summer due to air conditioning), but this thing seems to work. If electricity usage drops by 1% or more, it’s a recession. Limited historic data, but no misses and no false positives. Currently a close call.
The CB’s Consumer Confidence is similar to the UoM Sentiment. Two false positives (1992, 2003), but it did catch all recessions including the ones in 1981/2 and 2001 (difficult for a lot of other indicators). 2011 was a “close call”. Currently no red flag. Would have to decline to 45 to trigger recession call in 2013.
Retail sales (ex food services) fell annualized 2.7% over the three months to July. Falling retail sales for three consecutive months indicated a recession 27 out of 29 times since the data series began (1947), according to Gary Shilling. Year-over-year, the chance is still positive, but add a few weak months and the economy would be in the tank.
Before you fire employees you reduce their working hours. A drop in average weekly working hours in the manufacturing sector of 2% or more indicates a recession. Except for 1996. According to this indicator, the US economy is currently sailing smoothly.
The Institute for Supply Management (ISM) regularly asks company executives about orders, sales, inventories etc. A level of 50 indicates no net balance for participants seeing stronger and weaker business activity (economy stagnates). The current level (48) points to a slight contraction. However, the decline from prior peak is not yet large enough to raise recession alarm. One false positive (1989).
Unemployed people usually don’t drive to work. But the US population increases approximately 1% per annum, so traffic increases constantly.
If total miles driven grow less than 0.1% versus its own trend, you are likely to be in a recession.
The 2001 recession was missed. This indicator says we had a recession in 2011 (which is theoretically possible – we might not know it yet). Is the prolonged decline in miles traveled since 2007 due to online shopping? According to data from the Census Bureau, electronic shopping and mail order houses amounted for 5.9% of retail sales (ex autos) in 2004, increasing to 9.0% in 2011. However, additional trucks are needed to deliver the additional volume, partially negating fewer trips to the mall by individuals. The number of non-farm employees was approximately the same in 2011 and in 2004, eliminating reduced employment as a source of the decline.
The steep drop in 2011 is puzzling. Mish Shedlock recently looked at possible reasons for the decline in gasoline consumption.
Defense and aircraft orders are lumpy and distort trends, so we exclude them here. “Medium” confidence in this indicator due to limited historic data. If you set the trigger at 0%, we had a false positive in 1998 and are currently in recession.
Electricity production should be linked to economic growth. This indicator, unfortunately, had many false positives (1983, 1992, 1997, 2006), so confidence is “medium”. Setting the trigger below -0.5% would eliminate false positives, but make you also miss 1990/91. Regardless, electricity production suggests we are in a recession as I type.
Finally, the least confident indicators. First, the ISM manufacturing supplier deliveries. The current reading of 48.7 suggests a mild contraction. Multiple false positives (1985, 1989, 1995, 1998, 2005) muddy the water.
Cars need gas, and gas needs to be delivered to gas stations. Inventory effects are unlikely because of high turnover. “Low” confidence because of false positive (1996) and limited historic data. Can the recent decline be explained by online shopping?
SUMMARY: Established leading indicators incorporate questionable input. While there is no perfect indicator, a combination of the ones tested above, weighed by accuracy, confidence and timeliness should produce a good reading.
The higher-confidence indicators say that 2011 was a “close call”, but we are currently not in a recession. However, a lot of lower-confidence indicators are showing readings consistent with a severe recession.
It is entirely possible we narrowly avoided a recession in 2011 and are heading towards another one in 2013. Currently, however, combined indicators suggest only a 10% likelihood of being in a recession at this point (September 2012).
The US economy is experiencing the slowest recovery of the past 50 years (both from an employment perspective as well as from GDP growth). After looking at total credit market debt outstanding I concluded this was due to a decline in the marginal utility of additional debt (see this post). Some readers have mistaken me for an advocate of more quantitative easing. I should have been more clear about this: I do not think printing money is an appropriate remedy for economic woes. Enabling the government to finance fiscal largess at artificially low rates is unlikely to cure problems related to excessive debt levels.
A pdf version of this post can be found at Scribd
Before you trash me in the comments, hear me out.
It started off with Ray Dalio’s “beautiful deleveraging”, which inspired this post.
Since the financial crisis, the Fed has increased its balance sheet from $900 billion to $2.9 trillion (red line in below chart). The difference is $2 trillion (or 13% of GDP).
When the asset side of the Fed’s balance sheet grows, so must liabilities. The Fed’s liabilities consist mostly of money in circulation. So we can assume that $2 trillion in additional money has been pumped into the economy.
Or has it?
When the Fed buys bonds, it does so from “Primary Dealers” (21 global financial institutions). They hand over the bonds and get a corresponding credit on their account with the Fed. The Primary Dealers might then purchase some other securities with that money (which then gets credited to another bank’s account with the Fed).
And that’s where the buck stops. Three quarters of the money “printed” never make it into the economy. They remain as excess reserves (reserves in excess of banks’ minimum reserve requirements, blue line) in accounts at the Fed.
Hence, of $2 trillion additional money, only $500 billion (yellow line) ended up outside the Fed. Why? Banks could use those reserves for lending, but there is no demand for additional loans (from customers with sufficient debt bearing capabilities).
So if the money can’t find its way out of the Fed – how is money created then? What is money?
To understand, we have to take the example of buying a car.
In the US, literally nobody purchases a car with money form a savings account. The ability to purchase a car depends on the availability of credit. No credit, no car.
Credit availability depends on issuance of debt. Take a look at debt outstanding by ABS (asset-backed securities) issuers over the last 30 years:
ABS Credit market debt outstanding fell from $4.5 trillion at the peak in April 2007 to $2 trillion. That’s a decline of $2.5 trillion. This is money not available for purchases. It dwarfs the $500 billion pumped into the economy by the Fed.
Debt is money. The amount of debt outstanding controls the amount of money available for purchases, and hence for the size of the economy.
In addition ABS issuers there is debt by households, non-financial and financial corporations as well as the government sector. By adding them up you get the big picture: the total credit market debt outstanding (TCMDO):
TCMDO is the blue line, on a log scale. The red line is the change in the annual growth rate of TCMDO, measured from the prior post-recession peak growth rate. You will notice that every recession over the last 60 years, with the exception of 1970, coincides with a slowing of the growth rate by at least 2%-points. The red triangle depicts the 1987 crash, which followed a period of serious slowing in the rate of TCMDO growth.
Up until 2009, total credit market debt outstanding has never declined. The ratio of TCMDO to GDP continued higher and higher, at accelerating speed:
Has debt-to-GDP, or the debt-bearing capability of the US economy, hit a ceiling?
Look at how little additional GDP (blue area, below) we obtained in comparison to ever increasing amounts of additional debt (red area):
The dotted black line is the marginal utility of debt (right-hand scale). Think of it like this: how much additional GDP do you get out of one dollar of additional debt (in %). In 1992, for example, you get $0.30 in additional GDP for every additional dollar of debt.
Problem: this marginal utility of debt has trended lower and lower over the years, and actually reached zero in 2009.
Meaning: you can add as much debt as you want, and it still won’t give you any additional GDP.
To repeat: no amount of additional debt seems to be able to get economic growth going again.
That is a dramatic revelation. We might have reached the maximum debt-bearing capability of the economy. If true, no growth is possible unless debt-to-GDP levels fell back to sustainable levels (in order to restart the debt cycle). This could take years.
At this point, the only way to reset the debt cycle is to get rid of debt.
Ray Dalio correctly describes the three options available:
1. Austerity: this would be painful and take quite some time (the Europeans are going down this path)
2. Restructuring: requires write-downs and losses for bond investors (which are not being allowed to happen for fear of systemic risk)
3. Printing money: Inflation. Better yet: hyper-inflation. You have to destroy the value of debt fast enough before debt service costs, due to rising interest rates, drive the government into insolvency.
In the US, (1) and (2) are not happening. That leaves (3).
As shown above, the amounts needed for the Fed to be able to create inflation are much, much higher than what we have seen so far. And it is not guaranteed to work. Destroying the trust in the value of a fiat currency is a dangerous experiment with mostly adverse consequences.
Some of my clients like to challenge my (admittedly gloomy) views, forcing me to think – which isn’t such a bad thing to do.
It started off with Cam Hui’s “A Dalio explanation of Evans-Pritchard’s dilemma“. After laying down his strategy on winning the game of Monopoly, Dalio goes on to model the economy onto the board game. So far so good.
Then, Dalio is quoted in a Barron’s interview, describing the current phase of the U.S. deleveraging experience as “beautiful”. He goes on to explain the three options for reducing debt: austerity, restructuring and printing money.
“A beautiful deleveraging balances the three options. In other words, there is a certain amount of austerity, there is a certain amount of debt restructuring, and there is a certain amount of printing of money. When done in the right mix, it isn’t dramatic. It doesn’t produce too much deflation or too much depression. There is slow growth, but it is positive slow growth. At the same time, ratios of debt-to-incomes go down. That’s a beautiful deleveraging.”
That sounds pretty good and makes sense. Or does it?
“It is flawed economic thinking to target nominal wealth. Stock prices are not real wealth until realized gains are taken. More importantly, stocks are based on the underlying value of the assets they represent. Pushing stock prices up does not make the companies more profitable. So hoping that people will spend more of their current income because of a false price appreciation in the market is a misguided policy.”
Looking at quarterly changes in sector debt visualizes it nicely:
“If the region [Fairfield county] were a country, it would be the world’s 12th-most unequal in terms of income, ranking just below Guatemala.”
While Mr. Dalio’s narrative reads well, it doesn’t stand up to common sense. Unfortunately there is lingering suspicion his views on government spending are a mere ploy to advocate fore transferring even more debt from “his” sector onto taxpayers, while at the same time transferring taxpayers’ money to his firm via tax breaks.
US Bond market:
Our investment thesis can be summarized as follows:
There are only two solutions to the Euro-zone crisis:
In the above map you will notice the red “demarcation” line between A- and non-A-rated countries. It is the divide between the fiscally responsible North and the profligate South & East.
Germany and France are number 1 and 2 EU members by GDP, and together account for 36% of the EU’s economic power (29% of population). But France is culturally closer to Italy and Spain; it historically had a weak currency, and frequently suffered devaluations in the European Exchange Rate Mechanism (the managed floating rate mechanism before the introduction of the Euro). It is clear that as the Franco-German axis breaks, the European project is dead.
CONCLUSION: Sooner or later, something has got to give. There are no good outcomes. Deleveraging of excessive debt levels hurts, especially if the country seen as imposing those conditions is doing relatively fine. A Euro-exit by Germany, possibly together with the Netherlands, would be the least catastrophic “solution”, allowing the rest of Europe to regain competitiveness by weakening their Euro. However, such an outcome is unlikely, as Germany would be seen as the one ending European integration (and, in addition, losing all control over the fiscal policies of its neighbors).
A picture says more than a hundred words, so I wanted to present in graphical terms what happened at the Swiss National Bank over the last few quarters.
Unfortunately, the SNB does not provide foreign currency positions including derivatives on an absolute basis, but here are the unhedged figures:
Central banks have tried to “manage” currencies in the past. Sooner or later, market forces win. As all other major central banks keep printing additional Euros, Dollars, Yen etc., the SNB looks prone to lose this game. A run on the Swiss Franc could lead to a further increase in prices of Swiss government bonds. Swiss equities however would decline, at least measured in Swiss Francs.
“We never pre-commit” is the standard answer the ECB (European Central Bank) has for anyone asking about future interest rates (journalists regularly waste one of their two permitted questions at ECB press conferences).
Why should central bankers not indicate future interest rates decisions? Because market participants will price that information accordingly, so the announcement would have no or even a counter-intuitive effect on markets. This could potentially undermine the central bank’s credibility.
If central bankers want to keep their ability to influence asset prices, they need to keep markets guessing about future actions.
What on earth then is the Federal Reserve (“Fed”) doing?
The FOMC (Federal Open Market Committee) published where every member sees interest rates for the next three years and when they expected the first interest rate increase (see below):
What the heck?
This not only takes any element of surprise away; it also puts “outliers” under the spotlight and enhances groupthink (who wants to be the one that leaned out of the window and was completely wrong?).
Forget all the official statements. Forget anything you read in the “financial press”.
The Fed is printing money (“quantitative easing”) in order to achieve two goals:
1. Finance the government deficit
2. Cheapen the dollar (to help exports).
But how can you justify printing money when inflation is running at 3%, especially when you adopt an inflation target of 2%?
The Fed came up with a great plan: forecasting inflation of 1.6-2.0% for 2014 (at or below its new target) enables it to claim there is “danger of missing the target”. Hence printing more money is “required”.
In order to paint a picture of a weakening economy, the Fed starts out with elevated estimates for GDP growth, only to be able to reduce them at an opportune time.
The Fed now sees 2012 growth at 2.2-2.7%, down from 2.5-2.9% in its November forecast.
How Bernanke & Co. are able to predict inflation in 2014 when they couldn’t see a bursting housing bubble six months ahead is anybody’s guess.
The other benefit of forecasting (close to) zero percent interest rates into 2014 is the following: the Fed “sits” on the yield curve, depressing it as far out as possible by making markets price in available information. Euro Dollar futures have reacted and are now pricing in rates of less than 0.15% until the end of 2014:
This translates further into longer-dated bonds, depressing their yields. The Fed makes it easier for the government to pay its debt. If you pay zero interest on your debt (and assuming you can roll it over indefinitely) it has absolutely no effect on your budget. It is as if the debt didn’t exist (see How Dr. Ben Copperfield makes trillions disappear ).
All the pseudo-scientific yada-yada on economic theory are just hollow bones thrown to journalists and pundits to have something to “chew” on and write about.
The only thing that matters is the monetization of more and more government debt, and how to sell it to the public.
Paul Krugman would argue that despite all the “quantitative easing” inflation has not really picked up. At zero percent interest rates, money has no preference – there is no opportunity cost of just “lying around” without interest. Investing money for 4 years for 0.15% return is not “riskless return” – it’s “return-less risk”. Perversely, the Fed has created a situation where raising interest rates would probably lead to inflation. It is boxed into ZIRP (zero interest rate policy) for infinity.
Things will get serious once the Fed adopts a policy called N-GDP targeting. Instead of inflation, the Fed will try to “target” nominal GDP. If real GDP growth is zero, the nominal GDP growth will be made up entirely of inflation. Debt is a nominal unit, and it is supported by nominal GDP. In order to keep the ratio between GDP and debt halfway bearable, GDP must be inflated. It is a tax on everybody holding dollars, since the value of those will decline.
Meanwhile, the Japanese are resorting to stealth interventions to break the Yen’s strength. Currency wars have gone from “cold” to “hot”. The Fed’s printing of dollars is forcing other central banks to purchase them and selling their own currency in the hope of stemming their own currency’s rise. This makes them involuntary buyers of Treasury bills and bonds, making it easier for the US government to finance its deficit.
This, of course, will end badly. When the public loses confidence in paper currencies, barter trade and new, gold-backed currencies will be the only solution. “There is not enough gold” will the skeptics say, but it is only a question of price.