Investors always have to ask themselves if their notions are timely, but with gold, it's especially critical. Is it just too late in gold's bull market to justify the risk of entry?

You might think that the array of metrics available to financial professionals would help you identify periods of heightened risk. You might think so, but that doesn't necessarily mean they do. Most measures of risk rely upon “lookbacks” over an investment's trading history. Such measures are useful to investors only to the extent that they believe in mean reversion and trend momentum.

### Standard Deviation

Take volatility as an example. This is most commonly expressed as the annualized standard deviation of an investment's returns. Let's say you track gold's price on a daily basis. You can derive the metal's price volatility rather simply with a spreadsheet application such as Excel.

Figure 1 illustrates a string of daily gold prices and shows the Excel arguments you'd use to calculate daily returns and annualized volatility. The volatility calculated here is adjusted for a 252-trading-day year.

The volatility reading implies a roughly two-thirds probability that, at the end of a year, gold's price will fall within a range bounded by values 17 percent higher and 17 percent lower than its mean.

There are a couple of problems with measuring risk this way. First, we have no way of knowing whether a 17 percent volatility is exceptional or modest. There's no benchmark against which we can measure gold's variance.

Second, volatility of any sort — up or down — is treated as risk here. This metric doesn't distinguish between “good” volatility, i.e., rising prices for a long position, and the “bad” volatility of declining prices.

The standard deviation also assumes a normal distribution of prices — a bell-shaped curve. Prices over the near-term, however, can skew radically. An annualized volatility really doesn't offer a good perspective on risk over the immediate investment horizon.

### Beta Coefficient

Modern Portfolio Theory offers us another risk metric that gauges the volatility of an asset against a commonly employed benchmark. The beta coefficient is part of the MPT triad along with Jensen's alpha and the r2 correlation.

Beta, more often than not, uses the Standard & Poor's 500 Composite as the yardstick for measuring an investment's historical volatility. You can generate a beta for COMEX spot gold settlements, for example, by comparing the assets' covariance to benchmark's volatility as shown in Figure 2.

By taking gold's measure against the S&P, it seems that bullion was a low-volatility play in November. That's really a mathematical artifact since gold appreciated while the blue chip index lost ground. Volatility (and volatility-based metrics like beta) tend to shrink in a rising market and balloon in a sell-off.

But who's to say that gold's beta will remain low? Heading into the year's end, gold investors have had plenty of reasons, in fact, to puzzle over their volatility exposure.

The gold market got frothy in November after the ramp-up to the midterm election pushed up bullion's per-ounce price by more than \$260. The ascent was capped with a churning \$80 trading range as bulls and bears punched it out for control of the market. By Thanksgiving, support emerged for spot metal but signals of an intermediate top began flashing.

If bullion's course reverses and heads lower, its beta would likely rise. This arithmetic presents a problem to investors who monitor beta and standard deviation. They may be obtaining precisely the wrong information about gold's impending risk. As the market advances towards a top, both metrics would tend to decline, seeming to signal less risk. It's only after prices break into a downtrend that volatility would start to rise. Unfortunately, the risk would have already been realized by then.

The essential problem with conventional volatility-based risk measurements is that they aren't predictive. Advisors need to keep some boilerplate in mind when it comes to these metrics: “Past performance is not a guarantee of future results.” This applies to standard deviations and beta coefficients in spades. Both measurements are based upon historical movements which may not predict future price gyrations.

### Fear and Greed

Considering the tactical nature of gold allocations for many investors, we need metrics that are more predictive of increasing downside risk: measurements that can provide timely warnings of potential downdrafts.

By taking the measure of the two emotions most commonly associated with gold — fear and greed — investors and advisors can obtain a better sense of when they run to excess. Spikes in cupidity and trepidation often presage market turning points.

Let's dive into the greedy end of the pool first.

Some investors, emboldened by gold's positive returns, will naturally seek more leverage. That leverage is often obtained through gold mining stocks. Mining issues magnify gold's price movements because of the metal's influence on company earnings. Once gold's market price exceeds a company's production costs, further positive price fluctuations go straight to a producer's bottom line.

A good proxy for investor sentiment toward gold producers is the price trajectory of the Market Vectors Gold Miners ETF (NYSE Arca: GDX). GDX tracks a benchmark portfolio of 30 large-cap mining companies that began life as the AMEX Gold Miners Index. As of Nov. 30, GDX had notched a 28.8 percent gain for 2010.

Even more leverage can be obtained through investments in junior mining stocks. Pre-production companies — those engaged in the exploration and development of gold properties — are akin to venture capital stakes when compared to the “blue-chip” quality of the revenue-generating producers in the GDX portfolio. Investments in junior mining issues are based on the prospect of finding exploitable deposits and/or becoming an acquisition target.

The Market Vectors Junior Gold Miners ETF (NYSE Arca:GDXJ) mirrors a 60-issue index of juniors, and has gained 57.1 percent over the first 11 months of 2010.

The degree of greed among gold-focused investors can be divined by comparing the prices of the two mining stock ETFs in a ratio. When the GDXJ fund was launched last year, its price was roughly half the then-current share value of GDX. By the end of November, the ratio had risen to 68.5 percent. The ratio rises whenever the buying enthusiasm for the junior portfolio outpaces that of the producers' fund.

The ratio essentially tracks gold's price; by charting the ratio's dips and peaks (see Figure 3) side by side with gold, potential turning points in the gold price trajectory can be spotted. A rising ratio accompanied by a flat or rising gold trend, for example, indicates the potential for higher prices in the immediate future. That's precisely what happened in late October when a sell-off preceded a spike in bullion's price beyond the \$1,400 level.

Conversely, a downturn in the ratio's momentum can indicate impending wobbliness in the gold trend. Other technical patterns, such as head-and-shoulders formations, wedges and pivot points, may also emerge in the ratio's track though they may not be discerned in gold's graph. These can be predictive of changes in gold's trend.

Now for the fear part …

You're probably familiar with the CBOE Volatility Index, or VIX, which measures the markets' anticipation of price instability. VIX, often referred to as the “fear index,” distills a volatility expectation from S&P 500 index options.

There's also a CBOE Gold Volatility Index (GVZ), which is derived from options on the SPDR Gold Shares Trust (NYSE Arca: GLD). GVZ measures the market's expectation of 30-day volatility in gold prices. GVZ's problematic for holders of gold or gold proxies in that it's based upon calls and puts. Holders of gold wouldn't mind upside volatility reflected in call prices; it's only downside volatility manifested in put prices that's fear-inducing.

For that reason, a more direct sense of investor apprehension can be obtained by restricting our field of vision to puts only. We can calculate a daily cost index for GLD puts by looking at a single strike price — one that's 10 percent out of the money. The strike price may vary from day to day as GLD's value fluctuates. We may use a \$122 strike one day, for example, and a \$124 strike the next.

To derive each day's reading, we use the cost of the out-of-the-money GLD put — one which has at least four weeks to run before expiration — as the numerator in a ratio employing a 10-percent out-of-the-money put of similar length on the SPDR Depository Receipt (NYSE Arca: SPY).

Suppose, for example, GLD closes at \$135.37 when SPY ends the day at \$123.76. We'd use the cost of a \$122 GLD put and that of a \$111 SPY put in today's ratio. Let's say that's 33 cents and 37 cents, respectively. Today's cost ratio would thus be calculated as 89.2 percent.

By normalizing our starting date's value at 100, we can index daily variations in put costs as illustrated in Figure 4. Plotting the daily index reveals gold investors' willingness to pay up for insurance. Like any insurance market, increased competition for coverage, i.e., heightened awareness of risk exposure, drives up policy costs. The cost index will tend to peak in anticipation of imminent price breaks.

Notice that's precisely what happened on Nov. 10 as the cost index reached a zenith at 165 just before gold toppled \$53.

Just like the “greed index” shown before, the “fear index” can be analyzed from a technical standpoint to identify areas of support, resistance and congestion as well as to isolate reversal or continuation patterns.