Commodity Bubble Revisited
By Chris Puplava, July 1, 2009
Today’s article will be the first part of a two part series addressing whether commodities represented a bubble this decade with the benefit of hindsight to look at the correction in commodities over the past year. In April of 2008 I penned a two part series addressing the question of whether or not the bull market in commodities represented a bubble, with links to the two articles below:
In today’s article I address my own points in which I recommended readers pose to someone that referred to commodities as being in a bubble, with the points from the “Show Me The Bubble!” article below:
In the future, whenever you hear someone call the commodity bull market a bubble, simply tell them the following:
- Show me the bubble in inventories!
- Show me the bubble in capital expenditures excluding inflation!
- Show me the bubble in employment!
- Show me the bubble in valuation multiples!
Show me the bubble in inventories!
Perhaps one of the easiest ways to spot a bubble is the overcapacity present in which too much capital was invested and the subsequent fallout in terms of inventories that arises once demand collapses. For example, when the technology bubble in the United States burst in 2000, inventories of computer and electronic products surged as the available supply was far in excess of demand. Rising inventories after a bubble bursts is the first hallmark of a bubble, with another hallmark being subsequent prolonged depressed investment. Inventories typically take years to work through and production remains depressed until inventories are liquidated to bring supply inline with demand. This pattern of surging inventories once a bubble bursts can be seen below when viewing the aftermath of the technology and housing bubbles. Note that the surge in inventory levels after a bubble bursts typically exceeds prior recession levels by a wide margin, showing that the business cycle for that sector was outside normal levels.
Source: Bureau of Census
Source: Bureau of Census
As economies across the globe rolled over into recession one by one the level of demand for commodities began to fall while inventories began to rise. With roughly a year passing after most commodity prices peaked last summer, inventories have indeed risen for base metals with perhaps the greatest increase seen in aluminum and nickel inventories. But the question that needs to be asked is, is the increase in inventories of base metals well above levels seen in prior recessions and economic slowdowns as one would expect with a bubble?
London Metal Exchange Inventories (Metric Tons)
When looking at copper inventories one would have to say that, in fact, copper inventories have peaked BELOW levels seen in prior recessions. Copper inventories not only peaked below levels seen after the 1998 Asian Currency Crisis and 2001 U.S recession, but they did not even reach levels seen during the 1994-1995 U.S. mid cycle slowdown!
As mentioned above, the same can not be said of nickel and aluminum that have exceeded levels seen after the 2001 recession, though both are still below the 1994-1995 U.S. mid cycle slowdown.
Lead inventories remain well in check as do inventories for zinc and tin. While it is probably too soon to pass final judgment as these inventories are rising and show no definitive peak, one would have thought inventories would have surged far more than they have given the absolute collapse in base metal prices.
The likely reason for the relatively muted increase in inventories is a dramatic supply response from base metal producers who have cut production and cancelled billions of planned capital expenditures.
In terms of energy, the spike in crude oil inventories in the U.S. has exceeded the levels seen in last recession, showing the highest levels since the spike after the first Persian Gulf War in the early 1990s.
While the current level of crude inventories is certainly on the high side in regard to the last two decades, we have yet to see a dramatic spike in inventories above prior highs as is typically seen once bubbles peak as was seen with technology and housing inventories. The likely reason for this is it is much easier for the energy sector to curtail production than it is for a homebuilder. Once a housing project has been started, even if it is at the peak of the cycle, home builders have to finish work in progress and sell at depressed prices as they do not want stranded capital. Moreover, it helps that the demand for energy is far more inelastic than is the demand for computers or homes. Thus, the energy sector has a greater ability to bring supplies in line with demand and thus control inventories, with current levels nearly in line with levels seen in 2007 at this time of year.
The graph below shows the normalized growth in inventories to the start of the year for 2009 as well as the last four years with the inclusion of 2000 and the 2001 recession year. The figure illustrates that the rise in inventories in the first half of the year has been relatively in line with the path seen in 2005 and 2008, with the current year-to-date (YTD) growth in inventories now below the rates seen in the 2001 recession as well as 2005 and 2007. The inventory build peaked in May as it typically does, with inventories likely to continue to fall until September when inventories begin their season build.
When viewing the above data one can not say with concrete assurance that commodities were in a bubble as inventories of base metals and oil have not shown the typical path of bubbles whose inventories rise to levels well above prior peaks. Directly related to inventories, was there a bubble in capital investment?
Show me the bubble in capital expenditures excluding inflation!
While the number of active oil and gas rigs showed a dramatic increase this decade, they did not even reach half the level of the 1980s bubble, though industrial production (IP) for mining and oil and gas field machinery did show a dramatic increase since 2003. While IP rose significantly this decade for the energy sector, one has to take the rise with a grain of salt. The IP numbers are nominal numbers and correcting for inflation shows a much different picture. Deflating the IP numbers by the producer price index (PPI) for oil and gas field machinery shows a much more subdued level of real investment, as much of the increase in the nominal industrial production numbers was due to rising prices and not a real surge in production. In fact, the inflation-adjusted IP index for oil and gas field machinery was on par with the levels seen in 1997-1998, while the non inflation-adjusted IP was nearly double the 1997-1998 levels as the PPI roughly doubled over the same period.
Source: Federal Reserve Board, Baker Hughes Inc.
Source: Federal Reserve Board, Baker Hughes Inc.
Source: U.S. Bureau of Labor Statistics:
This dynamic is easy to see when looking at the data for Exxon Mobil over the past decade. Exxon had been dealing with double-digit inflation in terms of its costs which jumped sharply since 2003. To the naive observer, it would have appeared this decade that Exxon was aggressively attempting to increase production levels when viewed by its capital expenditures, though in reality it was primarily paying for rising input costs. The net result was that, while the company had spent large sums of money to increas oil production, its actual production levels have been relatively stagnant this decade and dropped significantly from the peak seen in 2006, this despite rising lifting costs.
As shown above, in contrast to the dramatic increase in capital expenditures seen in the technology and housing bubbles that led to over capacity, that dynamic is completely absent in the energy sector as the bulk of the increase in capital expenditures was merely to cover the rising input costs experienced by the industry. Not only did the level of investment in the energy industry not reach the levels seen during the 1980s bubble, but neither did employment.
Show me the bubble in employment!
Bubbles are marked by an economy shifting too much investment in one area, which can be observed not only in capital expenditures but also employment. The number of mortgage and nonmortgage loan brokers rose more than 600% from 1990 to the peak in 2006, with the technology sector also experiencing a dramatic increase in its employment levels in the 1990s. The bull market in commodities in the 1970s was a true bubble as employment numbers doubled in the span of ten years, with employment subsequently collapsing over the following decade. In contrast to the 1970s bubble, natural resource employment showed a much more subdued increase this decade, rising to roughly 700,000 workers, nothing close to the bubble employment levels reached in the early 1980s.
Even if the commodity market does turn into a bubble, we are not likely to see the levels of employment witnessed during the last bubble as our country has become far more reliant upon energy imports from abroad, with the rise in employment seen overseas rather than in the U.S. A truer picture of analysis in terms of employment would thus have to use world natural resource employment, though this data is unavailable.
Source: DOE, BLS
Show me the bubble in valuation multiples!
Another characteristic seen in bubbles are high valuations for equities related to the bubble, which was completely absent in the prior bull market for the energy sector. Taking the price-to-earnings ratio (PE), price-to-sales ratio, price-to-book value ratio, price-to-cash flow ratio, and the dividend yield, I made a valuation composite for the S&P 500 Energy and Materials sectors, showing the deviation from the historical average going back to 1995. As seen below, the S&P 500 Energy sector did not reach even one standard deviation above its historical average as the bull market in energy was fundamentally driven by earnings rather than speculation. Speculative markets witness an expansion in price multiples that are due from prices rising faster than earnings (fundamentals). The energy sector currently shows great value as its composite valuation fell to nearly three standard deviations from its historical average, with the sector currently the only S&P sector to still have a trailing PE ratio below 10, currently resting at 8.33
While energy showed subdued valuations in the last decade, the materials sector showed more froth as it spent most of its time between one and two standard deviations from its historical average, though the crash last year drove the sector to nearly four standard deviations from its average, placing it as the most undervalued sector going into 2009. While the sector did reach rich valuation levels in 2008 at the peak of commodity prices, its valuation levels at the time did not exceed the levels seen at the last bull markets peak in 2000, and could not be described as a bubble.
Another hallmark associated with bubbles is the relative weighting in a particular sector in the S&P 500 which rises dramatically and then sharply falls after the bubble bursts. The technology sector’s weighting in the S&P 500 rose nearly five fold from a low of 7% in the early 1990s to more than 32% at the peak in 2000. After the technology bubble burst the sector’s S&P weighting was cut in half in less than two years as the market cap of technology shares imploded relative to other S&P sectors.
The financial sector showed a smaller bubble than the technology sector, rising nearly five fold since the low in 1984 of roughly 5% to a high of roughly 23% a few years ago, benefitting from the rampant accumulation of debt in the economy over the same period. Like the technology sector, once the financial bubble burst its weight in the S&P fell more than 50% in just over two years. In both cases, the relative ground they gained in over a decade was significantly wiped out in the first year after the bubble, and by more than half after two years.
The bubble in the energy sector in the 1970s was a true bubble, perhaps even greater than the technology bubble as its weighting in the S&P 500 rose nearly 15% in 3 years between 1978 and 1981 while the financial and technology sector took roughly 8 years to match the same feat and the financial sector took two decades to increase its S&P weighting by 15%. Like the other two bubbles, once the energy bubble burst in 1981 the sector’s weight in the S&P 500 fell 30% from its peak weight of roughly 30% to 20% in the first six months and fell by half in just over one year, and by two thirds in five years. In contrast to the last energy bubble, the decline in the energy sector’s weight in the S&P has given back very little since its peak weight of 16.2% in June of last year, quite different than the declines seen in previous bubbles.
As mentioned above, once bubbles burst the bubble sector’s weighting in the S&P 500 experiences a sharp decline in the first two years followed by a mild recovery for about a year. After the brief respite, the sector’s weight is followed by a slower two year decline to new lows and then another two year period where the weighting remains relatively constant. This typical path is displayed in the figure below that shows the aftermath of the 1980 energy bubble, the 2000 technology bubble, and the financial bubble that burst in 2007. Focusing on the financials, the sector appear to be following the path seen after the energy bubble peak in 1980 and the technology bubble peak of 2000, with the sector potentially rising through the remainder of 2009 before resuming its secular decline into the next decade. Continued weakness in the sector isn’t hard to envision given the ongoing commercial and residential real estate losses, corporate defaults, as well as a consumer that has come to grips with the notion of saving.
The energy sector never reached above a 20% weighting in the S&P 500 as the last three bubbles did, and did not experience the same decline in loss of S&P weighting. It’s not hard to understand why the sector did not give back much of its weighting in the S&P over the last year as one simply has to look at valuation levels. The sector reached a trailing PE ratio of 5.8 in March of this year, corresponding to an earnings yield (inverse of the PE ratio) of 17.24%, nearly seven fold greater than the 2.5% yield on the 10-yr UST at the time. Given its cheap valuations it’s hard to see the sector’s weight in the S&P 500 fall much further, particularly so when it is the most undervalued sector of all ten S&P sectors. Because the prior run up in the energy sector was entirely fundamentally driven by earnings and sales which kept valuation ratios from reaching frothy levels, there was not as much investment excess to ring out like the 1970s energy bubble or the technology and financial bubbles.
Show Me The Bubble!
As shown from the analysis above, there was no bubble in inventories, no bubble in real capital expenditures, no bubble in employment, and certainly no bubble in valuations. Perhaps the final proof that commodities did not represent a bubble is that the energy and basic materials sector did not witness the over investment by retail and institutional investment investors with an S&P sector weighting north of 20% as in other bubbles. The energy sector has given back little ground in terms of its relative weight in the S&P 500, which is not the hallmark of a bubble.
While commodities have been shown in the analysis above that they did not reach bubble status, was the prior outperformance of commodities a single cyclical bull market event or are they in a secular bull market, destined to lead in the next cyclical bull market as the sector outshines once more? Next week’s article looks at this very point, determining whether or not the fundamental drivers that drove the commodity sector to outperform other asset classes in the last bull market are still intact.
© 2009 Chris Puplava