Elm Ridge commentary for the second quarter ended June 30, 2017; titled, “Dumb And Lonely.”
“It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so.” – Mark Twain
This was our worst half ever. This quarter and year have been all about energy, which accounted for all of our losses and then some (-12.5% and -24.5% alpha1 respectively). And while the statistical data continues to come in generally as we would have expected, there are enough inconsistencies and one-off data points to scare off investors from fighting against an overwhelming shift in sentiment – and enduring the patience necessary to mend a market that had been absorbing excess supply for the past two years. Investment banks are now tripping over each other to lower their price assumptions,2 and the sector has been the worst performer in the S&P 500 this year, trailing the index by more than 20 percentage points.
“The principal dynamics in the world’s capital markets revolve around a tug-of-war between feeling secure and making money. In the end, the feelings generally win out. A substantial amount of money can thus be made if a value investment manager is willing to spend the bulk of his or her professional life feeling depressed, isolated, and afraid, waiting for the forces of mean reversion to relieve the stress, at which point the manager will sell and use the proceeds to rebuild anxiety. Is it worth it? This question, of course, is philosophical, but the money on the table is considerable, and the question deserves serious thought.” - Lew Sanders
“Don't Give Your Money to Psychopaths, Says Science.” - E-mail Subject Line from Compliance Website
Tumbling oil and oil-linked equity prices, followed by what seems to be the last remaining bulls from both the buy and sell-side throwing in the towel, would lead anyone to question a bullish view of the oil market. And yet we are still there. Imagine that. We are once again off by our lonesome, with many of our readers and almost everyone else thinking we have lost our way, are overly stubborn and disconnected from reality. It may as well be 1999, 2002, 2006 or early 2009.3
Indeed, while we are going to spend most of this missive dwelling on our oil bet, I would hazard a guess – admittedly I can’t demonstrably support this contention – that oil’s performance has as much to do with allocators seeking bond-plus like returns in a low-rate world, as it does with the data themselves. As FMMI research notes:
In the hedge fund industry’s former years the investor base consisted of very-wealthy individuals, willing to take risk in an attempt to earn double-digit returns in a sliver of their portfolio. That’s changed … Institutions have sought refuges from market volatility… Underfunded and feeling vulnerable they believe they can ill afford to fall further behind. They’re looking for uncorrelated returns and want their portfolios… to produce bond-like volatility. For 2017 they’re targeting 7.5% returns with 6% volatility, a fairly ambitious target when the entire corporate bond market yields 3.9%. The rise of that risk-averse group, that now represents two-thirds of the asset base, has transformed the industry’s character.4
We are old school. Most professional investors and allocators cannot afford to fall behind their peers in an attempt to capitalize on what some might see as big changes ahead, as only a select few are given a long enough leash to ride it out. The much heralded shift toward quant investing is just a logical outgrowth from the enduring low interest rate bond-plus investing environment, as machines are optimized for inductive logic to uncover and capitalize on pattern behavior at what seems to be an instantaneous pace. We saw a huge opportunity, albeit with an ill-defined timeline, and filled up while we thought we still had the chance. But to this point, we got the psychology wrong and are still paying the price (and in hindsight, I’m not sure that we would have signed up to pay this price) for taking on idiosyncratic risk in order to garner what we deduce (computers can’t do that) to be some juicy rewards when the market gets too tight to ignore. Then we’ll see just how low-vol the popular funds really are.
Oil’s Well If It Ends Well
“When you're that successful, things have a momentum, and at a certain point you can't really tell whether you have created the momentum or it's creating you.” - Annie Lennox
“Doubt is not a pleasant condition, but certainty is absurd.” - Voltaire
What data are we looking at? Rather than declining, OECD inventories actually rose 50m barrels through May (the last reported monthly). The US oil rig count, up by 440 or 140% since the May 2016 trough, is back to early 2015 levels, and the prevailing story holds that producers are much more efficient and can easily balance the market at less than $50. Meanwhile, demand is supposedly weakening and it is now accepted that inventory levels won’t decline to normal levels by March 2018, when OPEC resumes production growth.
Supply / Demand
We forecast global supply and demand in great detail to inform expected changes in inventory levels and thus the direction we think oil prices are headed. Most importantly, we recognize that nearly all the data can be squishy and massaged to argue almost anything. With that said, you would expect that we have been way off on one or both given our YTD performance. Reported supply is generally the most accurate, so let’s start there.
Recent headlines suggest that a faster than expected US shale response and the return of Libya and Nigeria have muted the 2017 rebalancing. While US production has outpaced our bearish forecasts, largely due to a February bounce (winter weather evidently had a greater impact than anticipated), actual monthly data compiled by the EIA is still coming in well below its own oft-publicized initial weekly estimates (that it doesn’t bother to revise).5 And in spite of their recent ramp, Libya and Nigeria have underperformed our forecast by an equal amount, leaving our total global supply forecast largely unchanged.
Demand is a bit more difficult to measure and the IEA data, which is generally used as consensus, has historically been subject to large revisions, averaging over +1mbd per year this decade. The IEA’s initial 1Q demand growth estimate of 900kbd fueled macro concerns, but the agency still expects full-year growth of 1.3mbd. While it is entirely possible that demand is underperforming expectations, it is just too early to tell, and unlikely given the usual pattern of revisions. With that said, the purported 1Q slowdown looks temporary: as India accelerated from a flat 1Q to up 4% since March and the US grew 2% in both March and April. While oil has mostly been a supply story, occasional weakness in demand can push a balanced market into oversupply and vice versa. We still expect demand to grow in the 1.3-1.5mbd range this year and next.
So if supply and demand have generally been within our expectations, why have oil prices recently fallen to as low as $40? Inventories. This is probably the most difficult aspect of the oil market to piece together, with both lagged and incomplete data. While our global supply and demand model shows a deficit of ~1mbd in 2017, it is nearly impossible to forecast where that deficit will draw from to meet daily consumption needs. OECD, and specifically US inventory data, are the most readily available and accurate, and on a year-to-date basis have most certainly disappointed both the market and, to a much lesser extent, our expectations. But these reports do not paint a complete picture, as the widely watched weekly US figures6 only represent a quarter of the global market (and total OECD demand only accounting for half).
While we collect all the inventory data available, it is not enough. Data sources for storage in non-OECD countries, or for floating storage, are not readily available in a timely manner, and in some cases, not available at all. We just cannot get the definitiveness/”visibility” that the market craves. There is both tangible and anecdotal evidence on the forces that have worked to curtail reported OECD draws to date, such as oil held in more expensive offshore storage making its way onshore and the possible destocking of OPEC inventories as its exports have fallen less than production. Until these less visible draws make their way to more accessible data, the market will continue to underappreciate the supply deficit we currently estimate and likely will be blindsided by the pace of draws to come.
But let’s go back to the US. As we noted in our last letter, OPEC members ramped up production pre-deal so that they would calculate cuts from as high a level as possible. And it would take two months for these shipments to reach US ports. Indeed, right on schedule, they showed up in January and February, with inventories building some 55m barrels in the US and more than 90m for the OECD in total. But since that time (through June), US commercial inventories have drained more than 50m barrels and total OECD inventories close to 65m, when seasonality suggests they should be more or less flat over that same time period.
Furthermore, US crude draws have been slowed by the sale of 13m barrels, out of a total planned yearly sale of 16- 18m, from its Strategic Petroleum Reserve (SPR), to fund SPR infrastructure upgrades and NIH Opioid programs (leaving total US inventories down 64m, see chart below). In fact, including the SPR and despite the poor start to the year, we still recorded the largest first half draw since 2003. We have also compiled reports showing another 73m barrel draw from Singapore, Non-OECD countries, China commercial and floating storage that brings the total since February to 150m (table below). On the other hand, Chinese strategic reserves have supposedly built 80m (although anecdotal satellite imagery of Chinese storage tanks suggests little change in inventory YTD). Even so, we think that global inventories have drawn on the order of at least 70m barrels since February or almost 600kbd with still more data from recent months to come.
Looking forward, global oil consumption on average runs 1.6mbd higher in the second half of the year7 (on top of any revisions or rebound after a sluggish start to the year that seems to be improving). Non-OPEC, non-US production should be roughly flat, while Libya and Nigeria will add back 600kbd and the US another 500kbd. Add these to the current 600kbd shortfall and you get more than a 1mbd draw in the second half, which should bring inventories in-line with their five-year average by year-end.
US Shale Production
How could we expect the US to only produce 500kbd more in the 2nd half when: the US rig count has ramped back to early 2015 levels; productivity is benefitting from unprecedented technological improvements; and shale is growing over 100kbd per month?
The technological revolution does make for a great storyline. A number of US shale operators are trying to comfort investors with reports that they can make solid incremental well returns at $35-40 per barrel (that is, if they do not include any costs other than the drilling and completion costs of that next well), but that too is overstated.
Continually improving technology8 may dramatically lower the break-even at any one location, but the game gets harder as operators are forced to tackle more difficult geology down the road. Sell-side analysts love to show charts depicting average productivity rising at double digit growth rates for the last few years, but what they do not account for is that as producers were forced to retreat to their best acreage, they lopped off the worst performing wells (cutting drilling by some 50%). And it was that factor, more than any other, that helped boost the averages.
For instance, when we grouped the best 2000 wells for the top four basins,9 we found that productivity for this group in 2016 was about 4% worse than 2015 and in-line with 2014. It is the fact that the producers chopped the remaining number of wells by some 65% that allowed the averages to rise by the 25% figure that gets everyone so excited. If I’m a high school principal and want to boost my standardized test scores, nothing would work better than preventing my worst students from taking the test.10
Admittedly, the Permian has improved through any lens in each of the past two years, which is why it accounts for 90% of our US production growth forecast. Adjusting for well lengths (which seem to have plateaued), the 2016 average peak production did jump more than 17% over 2015. However, when we look at the actual data, we can see that the top 750 wells improved at less than a 3% clip, with all the rest declining. Again, it is the fact that the operators slashed 25% from the less-productive well-count that generated almost 85% of the reported improvement.
Similar to what we see as misinterpreted productivity advances, recent production growth claims have been exaggerated as well. While it is common to declare that shale is already growing over 100kbd each month, these claims are based on estimates from the EIA’s Drilling Productivity Report (DPR), which is derived from a severely flawed formula. As activity rises, the DPR overestimates production by taking the last month for which it has compiled production (e.g. March in the latest report) and then uses the change in the rig count since that time to forecast the four months (in other words, both April and May are still estimates in the latest report, even though they are commonly mistaken with actuals). But changes in rig counts only impact production 6-9 months later, so this forecast will be way off in periods when the rig count is changing rapidly. Comparing just the last two DPR reports to actual state data from Texas and New Mexico (encompassing the Permian and Eagle Ford basins), the DPR overestimated growth by a factor of two (an average of 50kbd per month).
These oversights have led the consensus to forecast US shale growth of 1.2-1.5mbd exiting 2017. But, based on current activity and assuming well productivity at current peak levels, we only see US onshore growing closer to 800kbd. Further, as shale takes share, the amount of production prone to higher decline rates increases – requiring higher activity to maintain the same growth rate.
As we move to 2018, a greater number of rigs will be required to offset natural decline rates rather than growing total production. Hence, we expect US crude, driven almost entirely by the Permian, to grow 700-800kbd assuming similar trends in productivity and activity. While a few more Gulf of Mexico projects will begin producing next year, this should be offset by the lack of action in US conventional production, which still accounts for 20% of the US total. And based on our global supply/demand work, this is not enough to balance the market.
Non-OPEC Non-US Production
Non-OPEC non-US sources produced slightly above 44mbd in 2016. This is predominantly conventional, longer-cycle production, driven by projects sanctioned 3-5 years in advance and will not react quickly to swings in oil prices. It is also subject to much slower natural decline rates, closer to 5-7%. Expectations are for 2017 production to decline slightly as we absorb capacity additions from decisions made under much higher prices, mostly in Canada, Brazil and Kazakhstan. Yet looking at international oil rigs (adjusted for Canadian seasonality), activity continues to be bouncing along the bottom since it fell 40% through January 2016, clear evidence the forward curve is too low for these producers. While some folks opine that non-OPEC production will grow because of the large amount of gross capacity coming online from projects sanctioned several years ago, this sort of analysis ignores both base decline rates and the amount of time it takes to ramp new projects toward peak production. From a high level, 2017 non-OPEC production (ex US and Russia) will average slightly below 33mbd of which 6mbd is associated with ramping projects, and another 5mbd from non-crude related production such as biofuels, leaving a 22mbd base that is prone to 5-7% natural decline rates. That means in order for non-OPEC just to maintain current production, it will require 1.1-1.5mbd of gross capacity additions. Yet there are only 800kbd slated to come on in 2018, leaving us with net declines of 300-700kbd. Again, you should keep in mind that these capacity additions are from projects discovered and sanctioned several years ago, so absent a rise in oil prices, the decline in non-OPEC should only get worse from here.
Yes, the rebalancing has been slow and choppy at times. This is to be expected when the market we are analyzing is close to 100mbd, and only a few hundred thousand – or a few tenths of a percent – can completely change its complexion. But the fact remains that global oil inventories are indeed normalizing, and prices remain well below what is needed to incentivize enough supply to meet demand long-term. We check, and re-check our work daily, and every indication is that 2017 will see a massive reduction in global oil inventories, and a likely return to historical averages by year-end – the same inventory levels that supported prices above $90 (well above our forecast) for several years. We do not need heroic assumptions to get there either, with our own expectations that OPEC and Russia will return to pre-cut levels in 2018, and grow from there. We are confident that depressed investment will continue to bite into global supply, and the massive uptick in shale activity will lead to a reversal of efficiencies. With demand expected to grow 1.3-1.5mbd, and lack of new conventional oil fields, we see room for US supply growth of 1mbd+ annually for the foreseeable future, while still maintaining a balanced oil market. Granted, given the cross-currents produced by somewhat murky statistics, we can’t sketch out an exact timeline (and hence other investors’ reticence to stand in the way of current momentum). But we see no other way to meet these growth requirements outside of much higher oil prices.
Where We Are – Portfolio Statistics
As we are still forced to wait for our ship tanker to come in, we kept the long portfolio basically intact during the quarter, trading around a few positions. Poor performance has cut the size of our energy holdings to just over 40% of capital, with better results bringing us to the same level in financials (with 22% in the monolines), just over 10% in industrials and one-off stakes in materials, retail and tech. On the other side, we entered four new shorts while dropping eight and adding to a bunch of others. We currently have 23% of capital in cyclical oversupply (from 30% last quarter), 21% in secular competition (24%), 25% in the excess supply of an asset class (so-called “safety” in this cycle, 21%), and 17% in excess supply of a heretofore successful investment strategy (e.g., levered platform/roll-ups, 14%).13
Turning to the statistics that we typically report, low near-term energy expectations are still producing the rare condition where the Street P/E’s of our longs exceed those of our shorts. But we can again repeat that our estimates paint a remarkably different picture. Meanwhile, the short-to-long ratio for the Elm Ridge P/NEPS (our primary internal measurement of the overall potential return in the portfolio) has bounced back to the levels we’ve only seen just before 2016’s 30% 10-month run and at the very depth of the financial crisis.
Wired Differently Anecdotes, Take 32
Sometimes you have to suffer a little bit in your youth to motivate yourself to succeed in later life. Do you think if Bill Gates [had fun] in high school, do you think there'd be a Microsoft? Of course not.
You got to spend a long time in your own locker with your underwear wedged up your ass before you start to think, “I'm going to take over the world of computers! You'll see. I'll show them.”- Lazyboy, “Underwear Goes Inside the Pants,” 2004
We think we understand why we’re here. We saw an opportunity and failed to grasp that prevailing allocator interests (market psychology will never be our strong suit) would keep other investors from straying too far out on a ledge. Their story reads better than ours – at the moment. If you’ve thrived in the current investing regime, why leave the bandwagon early and risk everything on a bet that all is about to change?
And yet, while this incentive structure might govern most potential investors, it does not apply to those who own – or have built up the ability to act like owners of – the capital that they manage. These are the people with whom we partner.
Of course, no letter would be complete without the obligatory, “we’re wired differently so we think differently, looking for trend changes just a bit (admittedly not a bit on this episode) too early. But we’ll be right in the end, you’ll see.” Somehow devoid of common cognitive biases15 that most humans have adopted in coping with a complex world, I have shamefully passed these traits on to my son. Take a gander at the following text exchange, after he read a couple of books on value and behavioral investing in anticipation of a summer internship.
Him: I’m reading about behavioral investing and it asked me these questions, which I am supposed to get wrong but definitely didn’t cuz I’m not stupid. Does that mean I’d be a great behavioral investor?
Me: It means you’re wired like me and will be way too early to spot the things that go wrong with what almost everyone else believes. You better get used to looking like a dumb $#!& for a while before you don’t.
One has to weigh these things.
Him: Well you’re smart.
But also fat.
So I’m not sure how I feel about that.
Me: You will therefore need to work out more than me.
Him: Fair Point.
See the full PDF below.