In November, I outlined (S&P $2,100: The Perfect Short) why August’s bounce was representative of the ‘return to normal’ phase in a boom bust cycle (below), as the underlying fears driving the sell-off were viewed as transitory. Now, we see those fears were not transitory and have led participants to question their assumptions, referred to as the ‘twilight’ period of a bubble. This questioning has accelerated over the past month, leading to scrutiny of the underlying fundamentals. It is my view that participants are just scratching the surface and upon the realization of where things truly stand, the S&P will likely be trading at $1,350.
I am operating under the assumption that we are in the beginning innings of a multi-year bear market. I believe the S&P will fall ~30% to $1,350 in 2016, and then trade to $1,100 (-43%) in 2017, as it overshoots to the downside. I see the fair value of the S&P at $1,350 (-30%) by applying a 14x multiple to my FY16 earnings estimate of $96.5 per share.
This is not solely a valuation call, as expensive markets can become more so, but is rather predicated on a perception change among market participants leading to the convergence of price and fundamentals, which have been disconnected due to excessive central bank intervention (illustrated by the difference in price per share and EBTIDA per share below). This pivot leads participants to question their current assumptions, and upon the realization that they are flawed, provides a catalyst for the convergence of the two.
It is important to first address why I believe perceptions have changed, as the remaining analysis could be deemed irrelevant for the time being, or as bulls would say, ‘it doesn’t matter until it matters.’
As I previously described in my write up on gold (When The Music Stops: Why Gold (GLD $111.85) Is Going To $170), all trends are subject to negative events that either weaken or reinforce the trend. For example, the market over the past seven years has moved to new highs following pullbacks, illustrating and reinforcing in participants’ minds that as price goes lower one should buy because the ‘price will move higher.’ Higher prices then reinforce participants’ views and result in influencing their perception on news and data. This changed recently as the S&P failed to break out to new highs following August 24th’s pull back, as we have seen lower highs and lower lowers, shown below.
As you can see in the chart above, the market was rejected from $2,116 to $2,081, which is technically significant as it relates to the monthly long-term uptrend (below) that was broken in August.
This breach and rejection has led the shorter dated 10 month moving average (green) to roll over the longer dated 20 month (yellow) moving average. The prior two instances were in 2001 and 2008 where on average the market fell 46.9%, which would bring us to $1,100. This is a slight overshoot of what I believe is the fair value of the S&P based on my projections of forward earnings, which I will elaborate on below.
Thus far, the failure of price to break to all-time highs has led participants to question their assumptions, and in my view has caused the market to grow increasingly concerned about the disconnect between price and fundamentals. This can be seen through waning equity market breadth and deterioration in the high yield / leverage loan market.
Part of the disconnect is a function of Wall Street analysts’ forecasts. The street is looking for $121 in earnings power for FY16, up from $111 (+9%) in FY15. This implies the market currently trades at 15.9x forward earnings, in line with its historical average. What is not covered in forward estimates is the 30% difference between GAAP (Generally Accepted Accounting Principles) and Non-GAAP earnings, which is the largest since the financial crisis according to Bank of America. Referring to 2008, we see the same lofty earnings estimates, price targets, and sector love (financials & technology).
True earnings power likely falls somewhere in between GAAP and non-GAAP earnings based on your view of certain non-cash expenses, but that said, one should be cognizant of the large variance.
Below is a juxtaposition of high yield distressed sector ratios versus sectors with the largest GAAP and non-GAAP difference. As you can see, the largest blow outs primarily correspond to the largest difference in GAAP and non-GAAP earnings (Note: telecom, energy, and materials).
Source: Deutsche Bank, AAO
The market is saying that fundamentals now matter, while at the same time Central Bank policy has less of a simulative effect on risk assets. This leaves the market very susceptible to a sharp downward repricing, as additional participants come to the realization that their assumptions are flawed.
The primary driver of ‘growth’ and the foundation of many bull theses is the American consumer, as ‘70% of GDP is driven by consumption’. I intend over the next section to disprove the strong American consumer narrative.
I contend that the strong American consumer thesis is flawed. Upon this realization, it is apparent forward estimates of the S&P are meaningfully inflated and will likely decline materially over the next six to twelve months, in turn leading to an analyst downgrade cycle.
The premise of the strong consumer thesis is that a lower unemployment rates will lead to increases in average hourly earnings and thus consumption. In addition, balance sheet improvements due to financial crisis deleveraging will likely lead to additional consumption, as consumers are increasingly willing to take on debt.
At face value, this thesis is currently what we are seeing; incomes have risen, the unemployment rate is at the lowest level since 2008, and balance sheets have strengthened; however, this coupled with a material decline in energy prices has not led to increased consumer spending as shown below by unadjusted retail sales ex food and autos. (Note: I have used unadjusted sales data because companies don’t adjust sales). The trend in consumption has been downward and was primarily negative throughout 2015.
Removing volatile gasoline sales also shows that the growth rate of retail sales is trending lower.
It seems this thesis is having a tough time holding weight, as recent increases in average hourly earnings are the strongest since 2009. The answer likely lies in spending demographics, a leveraged and tapped out middle class consumer, as well as declining household wealth.
Source: Bloomberg, BLS
The largest driver of consumption in the United States is the top 20% of earners, accounting for 48% of aggregate after tax income and 39% of spending (below). This is larger than the bottom 60% of earners combined who make up 38% of spending, and 28% of income.
Source: BLS Consumer Expenditure Survey, 2014
According to the BLS, 59% of US employees are paid hourly, likely corresponding to the lower income deciles. Since Q113, wages and salaries on average have grown 4.18%, with the wage component growing on average 2.13%. This means that 1.25% (2.13%* 59%) of the 4.18% growth in earnings was due to wages. Backing into this, we are able to derive that average salary growth was 7.1%.
Since salaries typically correspond to higher income brackets, I am assuming that this cohort has been the primary driver of growth over the past four years, as their incomes have risen materially. This is in line with a study from Emmanuel Saez, an economics professor at University of California at Berkeley, that notes the top 1% captured 91% of new income from 2009-2012.
That said, in referring to the chart on average hourly earnings, it seems that the growth in salaries is declining based on the divergence between the two. This would suggest that higher income households, while still increasing earnings power, are slowing and may be indicative of the slowdown in retail sales. This is important as 51.4% of aggregate spending is done by the top 30%, and in terms of the middle class driving growth, that doesn’t seem to be in the cards.
Below is a breakdown of core expenses by income deciles. I have boxed in orange what I have designated as the middle class (average lower limit annual after tax income between $31,487 and $77,076, 90th percentile $100,098) and chosen the below expenses as core. (Note many components are captured in retail sales; however, no growth in earnings leads to no growth in these sub categories)
The data shows that the middle class spends a material portion of their income on core expenses, while the top 20% of earners are able to retain over 30% of their incomes. The lowest 30% spend beyond 100% of their income on core, but this is enhanced by government subsidies which are not shown. When factoring in additional expenses such as entertainment, education, and household furnishings the middle class on average spends over 100% of its income. The numbers are based on the lower limit of that income range, but even scaling the values right a decile illustrates the same thing.
Source: BLS Consumer Expenditure Survey, 2014
This conclusion is confirmed by recent conference call comments from Paul Polman, the CEO of Unilever, who stated, “GDP growth goes to a very small amount of people and they’re not eating more products. The middle market is disappearing which is the bulk of the business. That’s why you see a lot of retailers struggling.”
The situation gets more ominous when digging into consumer balance sheets. At face value, balance sheets have strengthened since 2013, illustrated below by declining amounts of debt relative to both disposable income and equity. However, this does not provide the full picture due to the effect of lower homeownership, and the increasing portion of new income going to the highest income brackets.
Source: Federal Reserve
Home mortgages made up 69.5% of total liabilities in Q1 2013. This fell to 65.8% in Q315 as there was no growth in mortgage debt and a significant increase in consumer credit. The lack of additional mortgage debt is a result of declining home ownership rates following the financial crisis, which are touching levels last seen in 1995, shown below.
There are a couple implications here as it relates to consumer and balance sheet strength. Mortgages are the largest liability on household balance sheets, and as shown below are primarily held by top income individuals. Since 2013, mortgage debt outstanding has not grown, meaning that household leverage has been relatively tame. That said, typical housing costs associated with a mortgage and leverage have transitioned to rental expenses, which are off balance sheet per se. This means that household leverage declines, while expenses increase. Theoretically, this would keep the ratio of debt to disposable income constant, as decreased leverage is offset by less disposable income from the rental expense, however; this is not the case due to the disproportionate gains of the top 20%. The ratio fails to reflect most of the middle and lower class who have not seen their incomes grow, but have significantly increased their consumer leverage and operating expenses due to higher rental expenses.
Since 2010, consumer credit has expanded at 7.9% per year. The core components of consumer credit are credit card debt, student loans, and auto loans. Student loans made up 34.4% of consumer credit in Q315, which would likely correspond to lower income households as these individuals are just entering the work force. Auto loans made up 30% of consumer credit, and are also likely on the balance sheets of the middle class and lower income households. This alone means that 64% of consumer debt resides on the balance sheets of the middle and lower class. Credit card debt is an additional 20.4% and likely is skewed towards higher income households due to their greater ability to spend. These numbers are consistent with William Dornhoff’s study, which shows that in 2010 the bottom 90% of households held 72.5% of outstanding debt. Note that this number is likely higher due to declining mortgage balances, a 9.1% average annual increase in auto debt since 2010, and a 10.3% increase in student loans over the same period.
The chart below provides a better picture of household leverage ratios, illustrating non-mortgage debt relative to household income. The data is from 2013, however, we are able to get a rough estimate where things stand based on the inputs. As you can see, in 2013 non-mortgage debt relative to income was approaching or above levels seen in 2007 across almost all deciles. Given that consumer credit has grown on average 6.9% since 2013 with incomes stagnant, I am very comfortable saying that these ratios are likely at highs across all income deciles, albeit a close call for the top 10%. This means that 90% of America has the highest amount of consumer leverage relative to income going back to 1989.
As you can see below, consumer credit as a percentage of GDP is much higher than the financial crisis and at the highest level in history.
This will adversely affect future retail sales as leverage ratios hinder consumers’ ability to take on additional debt. Even if consumers were to add leverage, its affect in stimulating retail sales is diminishing. Below I have shown the year over year dollar change in auto and credit card debt versus retail sales. Apart from the astronomical increase in auto loans outstanding, the most concerning part is the divergence between retail sales. One would assume that a dollar increase in auto or credit card debt would translate into a dollar in retail sales. This is not the case, as the dollar increases in auto and CC are higher than retail sales.
Digging deeper into credit card debt, we can see the two tracked each other very well from Q112 to Q314, but at the end of 2014 there was a clear divergence. To me, this is an indication that consumers are now using their credit card to pay for expenses rather than retail items. This is likely due to the amount of leverage and lack of income growth that the middle and lower class has seen over the past few years. This is very ominous for future retail sales.
So if the middle class is tapped out and rolling over, we again turn to the primary driver of retail sales, the top 20%. A month ago, former Dallas Fed President Richard Fisher appeared on CNBC in what may go down as one of the most candid interviews to date (must watch). Fisher explains that the intent of QE was to create a wealth effect whereby increased asset prices provided a catalyst to spend.
Given that the largest holders of stocks correspond to the highest income brackets (below), the price appreciation of the S&P likely had a simulative effect on retail sales given that the top 30% makes up 51.4% of retail spending. This changed recently as stock prices failed to trend upward, and will likely have the opposite effect as they decline.
While stock prices are falling and creating an inverse wealth effect, we are also getting indications that the high end consumer may be rolling over. On recent a recent conference call, Colin Dyer the CEO of Jones Lang LaSalle stated, “what we did see is something of a cooling in demand at the high-end, in particular, our investment sales markets, a bit more selective purchasing, buyers not chasing risk as much as they might have done earlier in the cycle.”
Burberry had similar comments noting that it was a tougher external environment for luxury than they were expecting.
And finally, Sotheby’s most recent art auction saw sales plunge 50% year over year. The WSJ journal cited the global shakiness as a reason for the decline in values. Tad Smith, the CEO, echoed those comments on a recent conference call stating “customers had gotten more “discerning.”
The last two times the art market rolled over with shipping were in 2000 and 2008. The stock market followed.
Below I have broken out household balance sheet assets, household net worth from 2001-present, and recent quarterly household net worth with QoQ growth. I have boxed non-market related securities (~41% of assets) to show the foundation of household balance sheets. The remaining 59% of assets are subject to fluctuations of market related securities (households could clearly fall in this bracket as well). Within this group, corporate equities, mutual funds, and debt securities make up 24% of assets. Adding in pension entitlements brings market related exposure to 45% of assets. That said, given the material disconnect between the underlying fundamentals and price of the S&P, a further decline will adversely affect household net worth and thus consumers’ propensity to spend.
Source: Federal Reserve, Bloomberg
We can see how the decline in equity prices has affected household net worth in the bottom right chart above, as QoQ growth was negative. This was the first time since 2011. Referring to the chart in the top right corner, we see that from 2009 to Q315 household net worth grew 59%, equivalent to the rise from 2002 to 2008 before it rolled over. I am not saying this will happen, but given the trends and valuations we are seeing today, I believe there is a high probability net worth is rolling over. This also likely corresponds to the cyclical nature of markets, as credit cycles drive growth and then hinder it. Below is a chart of the last three short term debt cycles and their effect on the MSCI world equity index.
I believe the consumer is nowhere near as strong as the main stream media, Wall Street, and the market believe. This would mean that growth is contingent on non-domestic demand, which is 46% of S&P profits. The odds of this are extremely low, as the primary driver of growth during the past crisis, China, goes through a non-performing loan cycle, hindering global growth.
I am not a China expert, nor will I pretend to be. That said, I am listening to Kyle Bass and his analysis on China given his ability to analyze unsustainable situations. Bass recently wrote a letter outlining his short on the Chinese Yuan and periphery lenders. He noted China’s previous dependency on investment as the primary driver of GDP growth, which has led to China’s banking system to grow from $3T in 2006 to $34T today. This is roughly 340% of GDP.
Here is a quote from the letter making the comparison to the US financial crisis:
“For context, consider what the United States banking system looked like going into the GFC of 2007-2009. On-balance sheet, the US banking system had about $1 trillion of equity and $16.5 trillion of banking system assets (100% of US GDP). If non-banks and off-balance sheet assets are included, it would add another $12.5 trillion to get to about 175% of GDP. US banks lost approximately $650 billion of their equity throughout the GFC. We believe that Chinese banks will lose approximately $3.5 trillion of equity if China’s banking system loses 10% of assets. Historically, China has lost far in excess of 10% of assets during a non-performing loan cycle (The Bank for International Settlements estimated that Chinese banking system losses throughout the 1998-2001 cycle exceeded 30% of GDP). We expect losses in this cycle to exceed prior cycles. Remember, 30% of Chinese GDP approaches $3.6 trillion today. Think about how much quantitative easing (QE) the US Fed had to create in order to entice $650 billion of common and preferred equity into the US banks and prevent a Japanese-style deflationary bust. The Fed had to expand its balance sheet by roughly $4.5 trillion.”
– A non-performing loan cycle is likely not stimulative for global growth
– China’s third, fourth, and fifth largest trading partners (Japan, South Korea, and Taiwan) have seen material deterioration in trade data. Here are their exports to China in January.
– South Korea -21.5%
– Taiwan -19.3%
– Japan -17.5%
– Rcube & Rcube AM’s China momentum indicator, which takes into account bank lending, rail freight, and electricity production is showing a free fall in momentum, with the largest difference between GDP and the indicator since 2008.
This is an ominous sign for Chinese GDP and global growth especially given the extremely elevated levels of sovereign debt to GDP, which will likely rise as growth slows.
The global economy is at a precarious juncture as there is no driver of growth. The Federal Reserve believes that the diver is the US consumer given their comments in the recent FOMC minutes, “the staff continued to project that real GDP would expand at a somewhat faster pace than potential output in 2016 through 2018, supported primarily by increases in consumer spending.”
My analysis above indicates otherwise.
The Fed has lost almost all credibility in my eyes following their decision to hike in December and subsequent verbiage. After proclaiming the hike would be data dependent, they tightened financial conditions into the worst data over the past 4 years, aside from unemployment, which fails to take into consideration those who are no longer in the labor force. In my opinion, this was done to save face and to show they could hike, rather than based on the economic fundamentals which were blatantly deteriorating.
Here are a few slides from DoubleLine Capital.
You can see that going into December there is clear deterioration in GDP based on what has been an accurate predictor, the Atlanta Fed’s GDPNow forecast. What about this says liftoff?
Here is ISM manufacturing. Notice that when the index was at and above current levels (December of 2010 and 2012) the Fed was easing. The Fed decided to hike into this.
The high yield market was getting obliterated.
And here is the Fed’s preferred inflation measure PCE, which screams inflation.
While their actions during the financial crisis are commendable, subsequent easing and the failure to raise after meeting their benchmarks has distorted markets and pushed participants to take on unnecessary risk in the search for yield. The Federal Reserve has created a very addicted drug that is being passed from Japan to Europe and everyone wants a hit, but the markets tolerance to this drug have grown, and the underlying virus (fundamentals) is becoming worse. Unfortunately, for those using, the side effects are unknown as it’s the first of its kind.
Quantitative easing and low interest rates have been the primary reason for the detachment of price and fundamentals. As of yesterday’s close, the S&P trades at 15.9x forward estimates. This is very misleading and fails to capture the underlying economic situation.
Below is the historical and forward estimates of earnings per share for the S&P. Wall Street sees $121 in earnings power for FY16, good for a 9% increase year over year. This is materially inflated for the reasons I outline below.
On the sales side:
1) My analysis above indicates that the middle and lower class is tapped out and likely unable to take on additional leverage. The same holds true for the top 30% of earners given declining rates of income growth and household equity. This means that consumption is likely to be flat to down year over year.
2) The US economy is slowing, with manufacturing deep in a recession and non-manufacturing very likely to follow its lead as it has historically, shown below.
Ken Chenault, the CEO of American Express, also believes this may be the case, as he stated on a recent conference call, “I would say the segment that I’ve been most disappointed in has been the corporate segment…the easiest expense category to cut is T&E that’s the first thing you see…as I evaluate 2015, that was an area that…we saw a pretty consistent decline. And certainly what we’ve seen in my 30 plus years experience with the Company is cut backs in T&E tends to be an early indicator for a slowdown.”
3) CEO confidence has declined materially over the past year and has historically tracked well with real GDP growth. If this is any indication, a decline in GDP is next.
4) On that note, companies have spent the lowest percentage of corporate cash on R&D and CAPEX since 1999, likely due to their lack of confidence in future business conditions. Future growth is a function of investment and innovation. If corporations continue to under invest in the future, it will manifest in future sales and earnings, or lack of.
5) Inflation expectations have fallen off a cliff, meaning that if the same number of units is sold, the price will be lower and thus total sales.
On the margin side:
1) Average hourly earnings continue to increase, as well as salaries, albeit at a declining rate. This does not bode well for profit margins, which are now rolling over.
3) The cost of funding has risen significantly given credit market deterioration.
Based on the points above, it is very hard to see a $10 ($111 -> $121) improvement in earnings for FY16. I am conservatively estimating that earnings growth is flat year over year. However, I am adjusting the $111 to reflect expenses that companies and analysts have added back to inflate numbers.
I have netted out of the $111:
– Acquisition related expenses
– Restructuring charges
– Stock based compensation
– Legal settlement feeds
– Currency adjustments
And have left in:
– Impairments (debatable)
– Debt extinguishment / refi costs
– Pension mark to market
This results in ‘adjusted non-gaap’ earnings of $96.5. Based on this number, the market currently trades at 20x forward earnings, 26% higher than the quoted market multiple. Given that I see minimal to no growth going forward, I believe it is fair to assign a 14 multiple to FY16 earnings estimates, resulting in a $1,351 price target (-30%). That said, taking the numbers at face value still shows that participants are paying the highest premium for growth since 1985, shown below.
I believe the credit markets are already factoring in this analysis, as both high yield and leverage loans are touching levels not seen since 2009. This is very ominous for the equity market given that credit is typically correct.
Based on the above analysis and my view that the market has undergone a perception change, I believe the S&P will trade to $1,350 in 2016, as the domestic and international economic data continue to deteriorate, and participants realize their assumptions are flawed. This will likely feed on itself, resulting in an over shoot to the downside and S&P $1,100, closing the gap in the chart above.
This is also consistent with the 2001 analog I have been following since late last year. I came across a stat that 80% of IPOs in the past few years had no earnings. This happened last in 1999. Digging in deeper to the time period I found:
– Over the past 30 years, the only other time price to EBITDA was as high as it is today was in 2000
– Equal weighted average debt-to-total-assets ex financials in 2001 was slightly lower than today’s levels (lower in 08)
– Material increases in debt and cash flow deficits
– And the markets love of tech both in the public markets, but more so in the private markets today, i.e Uber, Snapchat
“What has happened in the past will happen again, and again, and again. This is because human nature does not change, and it is human emotion, solidly build into human nature that always gets in the way of human intelligence. Of this I am sure.” – Jesse Livermore
In my opinion, the recent bounce from $1,800 to $1,926 was a gift from the market gods to lighten up long exposure and protect yourself for what could manifest over the coming years. For reference, the top in 2000 to the bottom in 2002 was -48%. This also brings you to $1,100 on the S&P.
I believe that due to the short term nature of the markets and the world today, many are failing to look at the bigger picture and envision what the world will look like in two years time, which is what you pay for today. I have tried to provide what I believe it will look like and urge you to do the same.
On the other hand, you could always just believe this guy. But when has he lived up to his word?