Addressing Keith Weiner’s objections to “Gold’s True Fundamentals”

June 27, 2017

A 23rd June post at the TSI Blog described the model (the Gold True Fundamentals Model – GTFM) that I developed to indicate the extent to which the fundamental backdrop is bullish for gold. The GTFM is an attempt to determine a single number that incorporates the most important fundamental drivers of the gold price, where I define “fundamental driver” as something that happens in the economy or the financial markets that causes a significant change in theĀ desire/urgency to own gold in some form. Keith Weiner subsequently posted an article objecting to some of my “fundamental drivers”, which would be fine except that his article contains several misunderstandings of these price drivers and/or how I am using them. The purpose of this post is to address these misunderstandings and provide a little more information on the GTFM’s components.

1. The ‘Real’ Interest Rate

Keith states: “The Real Interest Rate is the Nominal Interest Rate – inflation.” No, that’s not what the real interest rate is, although many people wrongly calculate it that way.

Keith and I agree that it is not possible to calculate the economy-wide change in money purchasing-power (PP), but even if it were possible to come up with a single number that represented prior “inflation” the real interest rate would not be the nominal interest rate minus this number. The reason, to explain using an example, is that the real return that will be obtained by someone who makes a 12-month investment today in an interest-bearing security will have nothing to do with the change in the PP of money over the preceding 12 months. Instead, the real return that will be obtained by this person will be determined by the change in money PP over the ensuing 12 months.

Now, we can obviously never know in advance what the real return on any interest-bearing security or deposit will be, but since the advent of Treasury Inflation-Protected Securities (TIPS) in 2003 it has been possible to roughly determine the real return on Treasury debt expected by the average bond trader. The TIPS yield, which is based on the EXPECTED rate of currency depreciation, is my ‘real’ interest rate proxy.

If there had been a TIPS market in the 1970s then it would probably be apparent that the large gains made by the gold price during that decade were related to a low/falling real interest rate, where the real interest rate is defined as the nominal interest rate minus the expected rate of currency depreciation. In any case, there has definitely been an inverse correlation between the TIPS yield (10-year or 5-year) and the gold price over the past 10 years. Furthermore, the correlation has strengthened over the past 2 years.

By the way, it’s the DIRECTION, not the value, of the TIPS yield that matters to gold and that is taken into account by the GTFM.

The inverse relationship between the TIPS yield and the gold price is far from perfect, the reason being that there are times when other price drivers are more influential. That’s why the ‘real interest rate’ has only a one-seventh weighting in the GTFM.

2. The Yield Curve

There has never been a strong and consistent short-term correlation between the gold price and the yield curve, but near major turning points the yield curve tends to be the dominant driver.

In broad terms, the boom phase of the central-bank-promoted boom-bust cycle is generally associated with a flattening yield curve and the bust phase is generally associated with a steepening yield curve. Gold generally performs better during the bust phase, when the curve is steepening. Somewhat counterintuitively, banks tend to do best during the long periods of yield-curve flattening. This can be demonstrated empirically and makes sense if you understand how the central-bank-promoted boom-bust cycle works.

A major flattening trend in the US yield curve got underway during the second half of 2011 and continues to this day. This flattening trend is associated with a boom, which, in turn, has temporarily helped the banks and reduced the desire to own gold.

3. Credit Spreads

The trend in credit spreads is one of the best measures of the overall trend in economic confidence, with widening spreads (yields on lower-quality bonds rising relative to yields on higher-quality bonds) being indicative of declining economic confidence. Gold tends to do relatively well during periods when economic confidence is on the decline, that is, during periods when credit spreads are widening. I have demonstrated this in the past using charts.

4. The Relative Strength of the Banking Sector

Keith writes: “We haven’t plotted it, but we assume bank stocks will outperform the broader stock market when the yield curve is steeping by way of falling Fed Funds rate. This is when the banks’ net interest margin is rising, and they are getting capital gains on their bond portfolio too. At the same time, credit spreads are narrowing, so the banks are getting capital gains on their junk bonds.

No, that’s not how it works. Refer to my yield curve comments above for a very brief explanation.

The banking sector will often fare poorly during major yield-curve steepening trends because a banking crisis is often a primary cause of the steepening trend. In any case, this indicator is based on the concept that the investment demand for gold will be boosted by declining confidence in the banking system and reduced by rising confidence in the banking system.

5. The US Dollar’s Exchange Rate

More often than not, the US$ gold price trends in the opposite direction to the Dollar Index. However, there are times when a crisis outside the US causes both a rise in the US$ on the FX market and a large rise in the US$ gold price. The fact that the inverse correlation between the gold price and the Dollar Index can break down in a big way at times is why the US dollar’s performance on the FX market only has a one-seventh weighting in the GTFM. To put it another way, if the gold price always moved in the opposite direction to the Dollar Index then there would be no reason for gold traders to consider anything except the Dollar Index.

6. The General Trend in Commodity Prices

I have included the general trend in commodity prices as indicated by the S&P GSCI Commodity Index (GNX) in the GTFM for the practical reason that there are times when it tips the balance. That is, there are times when a strong upward trend in commodity prices enables the US$ gold price to rise despite an otherwise slightly-bearish (for gold) fundamental backdrop and there are times when a strong downward trend in commodity prices causes the US$ gold price to fall despite an otherwise slightly-bullish fundamental backdrop.

7. The Bond/Dollar Ratio

There are fundamental reasons for the existence of a positive correlation between the bond/dollar ratio (the T-Bond price divided by the Dollar Index) and the US$ gold price, but I currently don’t have the time or the inclination to go into these reasons. Instead, for the sake of brevity I present the following chart-based comparison of the gold price and the bond-dollar ratio. The positive correlation is obvious and is evident over much longer periods than the 3-year period covered by this chart.

gold_USBUSD_260617

I hope the above goes at least part of the way towards explaining the components of my gold model.

Print This Post Print This Post

Gold’s True Fundamentals

June 23, 2017

[This post is a modified excerpt from a TSI commentary published a few weeks ago]

To paraphrase Jim Grant, gold’s perceived value in US$ terms is the reciprocal of confidence in the Fed and/or the US economy. Consequently, what I refer to as gold’s true fundamentals are measures of confidence in the Fed and/or the US economy. I’ve been covering these fundamental drivers of the gold price in TSI commentaries for almost 17 years. It doesn’t seem that long, but time flies when you’re having fun.

Note that I use the word “true” to distinguish the actual fundamental drivers of the gold price from the drivers that are regularly cited by gold-market analysts and commentators. According to many pontificators on the gold market, gold’s fundamentals include the volume of metal flowing into the inventories of gold ETFs, China’s gold imports, the volume of gold being transferred out of the Shanghai Futures Exchange inventory, the amount of “registered” gold at the COMEX, India’s monsoon and wedding seasons, jewellery demand, the amount of gold being bought/sold by various central banks, changes in mine production and scrap supply, and wild guesses regarding JP Morgan’s exposure to gold. These aren’t true fundamental price drivers. At best, they are distractions.

In no particular order, the gold market’s six most important fundamental price drivers are the trends in 1) the real interest rate, 2) the yield curve, 3) credit spreads, 4) the relative strength of the banking sector, 5) the US dollar’s exchange rate and 6) commodity prices in general. Even though it creates some duplication, the bond/dollar ratio should also be included.

Until recently I took the above-mentioned price drivers into account to arrive at a qualitative assessment of whether the fundamental backdrop was bullish, bearish or neutral for gold. However, to remove all subjectivity and also to enable changes in the overall fundamental backdrop to be charted over time, I have developed a model that combines the above-mentioned seven influences to arrive at a number that indicates the extent to which the fundamental backdrop is gold-bullish.

Specifically, for each of the seven fundamental drivers/influences I determined the weekly moving average (MA) for which a MA crossover catches the most trend changes in timely fashion with the least number of ‘whipsaws’. It’s a trade-off, because the shorter the MA the sooner it will be crossed following a genuine trend change but the more false trend-change signals it will cause to be generated. I then assign a value of 100 or 0 to the driver depending on whether its position relative to the MA is gold-bullish or gold-bearish. For example, if the yield-curve indicator is ABOVE its pre-determined weekly MA then it will be assigned a value of 100 by the model, because being above the MA points to a steepening yield-curve trend (bullish for gold). Otherwise, it will be zero. For another example, if the real interest rate indicator is BELOW its pre-determined weekly MA then it will be assigned a value of 100 by the model, because being below the MA points to a falling real-interest-rate trend (bullish for gold). Otherwise, it will be zero.

The seven numbers, each of which is either 0 or 100, are then averaged to arrive at a single number that indicates the extent to which the fundamental backdrop is gold-bullish, with 100 indicating maximum bullishness and 0 indicating minimum bullishness (maximum bearishness). The neutral level is 50, but the model’s output will always be either above 50 (bullish) or below 50 (bearish). That’s simply a function of having an odd number of inputs.

Before showing a chart of the Gold True Fundamentals Model (GTFM) it’s worth noting that:

1) The fundamental situation should be viewed as pressure, with a bullish situation putting upward pressure on the price and a bearish situation putting downward pressure on the price. It is certainly possible for the price to move counter to the fundamental pressure for a while, although it’s extremely likely that a large price advance will coincide with the GTFM being in bullish territory most of the time and that a large price decline will coincide with the GTFM being in bearish territory most of the time.

2) The effectiveness of fundamental pressure will be strongly influenced by sentiment (as primarily indicated by the COT data) and relative valuation (as primarily indicated by the gold/commodity ratio). For example, if the fundamental backdrop is bullish and at the same time the gold/commodity ratio is high and the COT data indicate that speculators are aggressively betting on a higher gold price then it is likely that the bullish fundamental backdrop has been factored into the current price and that the remaining upside potential is minimal. The best buying opportunities therefore occur when a bullish fundamental backdrop coincides with pessimistic sentiment and a low gold/commodity ratio.

Getting down to brass tacks, here is a weekly chart comparing the GTFM with the US$ gold price since the beginning of 2011.

GTFM_blog_230617

A positive correlation between the GTFM and the gold price is apparent on the above chart, which, of course, should be the case if the GTFM is a valid model. If you look closely it should also be apparent that the fundamentals (as represented by the GTFM) tend to lead the gold price at important turning points. For example, the GTFM turned down in advance of the gold price during 2011-2012 and turned up in advance of the gold price in 2015 (the GTFM bottomed in mid-2015 whereas the gold price didn’t bottom until December-2015).

The tendency for gold to react to, rather than anticipate, changes in the fundamentals is not a new development, as evidenced by gold’s delayed reaction to a major fundamental change in the late-1970s. I’m referring to the fact that by the second half of 1978 the monetary environment had turned decisively gold-bearish, but the gold price subsequently experienced a massive rally that didn’t culminate until January-1980.

The GTFM was slightly bearish over the past two weeks, but three of the model’s seven components are close to tipping points so it wouldn’t take much from here to bring about a shift into bullish territory or a further shift into bearish territory. The former is the more likely and could occur as soon as today (23rd June).

Print This Post Print This Post

The Central-Bank Moment

June 19, 2017

Hyman Minsky was an economist who popularised the idea that “stability leads to instability”. According to Minsky and his followers, credit expands rapidly during the good times to the point where a lot of borrowing is being done by financially fragile/vulnerable entities, thus sowing the seeds of a financial crisis. That’s why the start of a financial crisis is now often referred to as a “Minsky moment”. Unfortunately, Minsky’s analysis was far too superficial.

Minsky described a process during which financing becomes increasingly speculative. At the start, most of the debt that is taken on can be serviced and repaid using the cash flows generated by the debt-financed investment. At this stage the economy is robust. However, financial success and rising asset prices prompt both borrowers and lenders to take on greater risk, until eventually the economy reaches the point where the servicing of most new debt depends on further increases in asset prices. At this stage the economy is fragile, because anything that interrupts the upward trend in asset prices will potentially set in motion a large-scale liquidation of investments and an economic bust.

This description of the process is largely correct, but rather than drilling down in an effort to find the underlying causes Minsky takes the route of most Keynesians and assumes that the process occurs naturally. That is, underpinning Minsky’s analysis is the assumption that an irresistible tendency to careen from boom to bust and back again is inherent in the capitalist/market economy.

In the view of the world put forward by Keynesians in general and Minsky in particular, people throughout the economy gradually become increasingly optimistic for no real reason and eventually this increasing optimism causes them to take far too many risks. The proverbial chickens then come home to roost (the “Minsky moment” happens). It never occurs to these economists that while any individual could misread the situation and make an investing error for his own idiosyncratic reasons, the only way that there could be an economy-wide cluster of similar errors at the same time is if the one price that affects all investments is providing a misleading signal. The one price that affects all investments is, of course, the price of credit.

Prior to the advent of central banks the price of credit was routinely distorted by fractional reserve banking, which is not a natural part of a market economy. These days, however, the price of credit is distorted primarily by central banks, and the central bank is most definitely not a natural part of a market economy. Therefore, what is now often called a “Minsky moment” could more aptly be called a “central-bank moment”.

I expect the next “central-bank moment” to arrive within the coming 12 months. I also expect that when it does arrive it will generally be called a “Minsky moment” or some other name that deftly misdirects the finger of blame, and that central banks will generally be seen as part of the solution rather than what they are: the biggest part of the problem.

Print This Post Print This Post

The limitations of sentiment, revisited

June 12, 2017

In a blog post in March of this year I discussed the limitations of sentiment as a market timing tool. I wrote that while it can be helpful to track the public’s sentiment and use it as a contrary indicator, there are three potential pitfalls associated with using sentiment to guide buying/selling decisions. Here are the pitfalls again:

The first is linked to the reality that sentiment generally follows price, which makes it a near certainty that the overall mood will be at an optimistic extreme near an important price top and a pessimistic extreme near an important price bottom. The problem is that while an important price extreme will always be associated with a sentiment extreme, a sentiment extreme doesn’t necessarily imply an important price extreme.

The second potential pitfall is that what constitutes a sentiment extreme will vary over time, meaning that there are no absolute benchmarks. Of particular relevance, what constitutes dangerous optimism in a bear market will often not be a problem in a bull market and what constitutes extreme fear/pessimism in a bull market will often not signal a good buying opportunity in a bear market.

The third relates to the fact that regardless of what sentiment surveys say, there will always be a lot of bears and a lot of bulls in any financial market. It must be this way otherwise there would be no trading and the market would cease to function. As a consequence, if a survey shows that almost all traders are bullish or that almost all traders are bearish then the survey must be dealing with only a small — and possibly not representative — segment of the overall market.

I went on to write that there was no better example of sentiment’s limitations as a market timing indicator than the US stock market’s performance over the past few years. To illustrate I included a chart from Yardeni.com showing the performance of the S&P500 Index (SPX) over the past 30 years with vertical red lines to indicate the weeks when the Investors Intelligence (II) Bull/Bear ratio was at least 3.0 (a bull/bear ratio of 3 or more suggests extreme optimism within the surveyed group). An updated version of the same chart is displayed below.

The chart shows that while vertical red lines (indicating extreme optimism) coincided with most of the important price tops (the 2000 top being a big exception), there were plenty of times when a vertical red line did not coincide with an important price top. It also shows that optimism was extreme almost continuously from Q4-2013 to mid-2015 and that following a correction the optimistic extreme had returned by late-2016.

Sentiment was at an optimistic extreme late last year, at an optimistic extreme when I presented the earlier version of the following chart in March and is still at an optimistic extreme. In effect, sentiment has been consistent with a bull market top for the bulk of the pastĀ four years, but there is still no evidence in the price action that the bull market has ended.

Regardless of what happens from here, four years is a long time for a contrarian to be wrong.

IIbullbear_120617

Print This Post Print This Post