The price of gold is moved by a combination of supply, demand and investor behavior. That seems simple enough, but the way those factors work together is sometimes counterintuitive. Many investors, for example, think of gold as an inflation hedge. That has some common sense plausibility — paper money loses value as more is printed. But the supply of gold is relatively constant. As it happens mining doesn't add much year to year.

Correlation to Inflation

Two economists, Claude B. Erb of the National Bureau of Economic Research and Campbell Harvey, a professor at Duke University's Fuqua School of Business, studied the price of gold in relation to several factors. It turns out gold doesn't correlate well to inflation. That is, when inflation rises, that doesn't mean that gold is necessarily a good bet. (For more, see: The Better Inflation Hedge: Gold or Treasuries?)

So inflation isn't it, what about fear? Certainly, during times of economic crisis investors flock to gold. When the Great Recession hit, gold prices rose. But gold was already rising up until the beginning of 2008, getting near $1,000 an ounce before falling under the $800 level and then bouncing back up and rising as the stock markets bottomed out. That said, gold prices kept rising even as the economy recovered. The price of gold peaked in 2011 at $1,921, and has been on a slide ever since. It now trades at south of $1,200 (as of mid-March 2015).

Erb and Harvey note in their paper "The Golden Dilemma," that gold has positive price elasticity. That basically means that as more people buy gold the price goes up in line with demand. It also means there isn't any underlying "fundamental" to the price of gold. If investors start flocking to gold, the price rises no matter what monetary policy might be. That doesn't mean it's completely random or the result of herd behavior. There are forces that affect the supply of gold in the wider market — and gold is a worldwide commodity market, like oil or coffee. (For more, see: How Can I Invest in Gold?)

Supply

Unlike oil or coffee, though, gold isn't used up. Almost all the gold ever mined is still around. There is some industrial use for gold, but that hasn't increased demand as much as jewelry or investment. The World Gold Council's 2014 figures show that total demand was 3,923.7 metric tons, but only 389 tons was for the tech sector. The rest was investment at 904.6 tons and jewelry at 2,152.9 tons. Back in 2001, when gold prices were nearing all-time lows (at least since ownership of bullion was re-legalized in the 70s), jewelry took up 3,009 tons while investment was 357 tons, and tech was 363 tons.

So one would expect, if anything, the price of gold to drop over time, since there is more of it around. So why doesn't it? Aside from there being more people who might want to buy, the jewelry and investment demand have some clues here. As Peter Hug, director of global trading at Kitco, said, "It ends up in a drawer someplace." The jewelry is effectively taken off the market for years at a time.

Even though in countries like India and China gold can act as a store of value, the people that buy it there don't regularly trade it (few pay for a washing machine by handing over a gold bracelet). Jewelry demand tends to rise and fall with the price of gold, so when prices are high the demand falls relative to investor demand. (For more, see: Gold: The Other Currency.)

Central Banks

Hug says the big market mover is often central banks. In times when foreign exchange reserves are large, and the economy is humming along, a central bank will actually want to reduce the amount of gold it holds. That's because the gold is a dead asset – it makes no return, unlike bonds or even money in a deposit account.

The problem for central banks is this is precisely when other investors aren't that interested in gold either. So a central bank is always on the wrong side of the trade, even though selling that gold is precisely what the bank is supposed to do. As a result, the price of gold falls. (For more, see: What are Central Banks?)

Central banks have since tried to manage their gold sales in a cartel-like fashion, to avoid disrupting the market too much. Called the Washington Agreement, it basically says the banks won't sell more than 400 (metric) tons in a year. It's not binding like a treaty, more of a "gentleman's agreement" – but one that is in the interests of central banks as unloading too much gold on the market at once would negatively affect their portfolios.

One exception is China. The Chinese central bank has been a net buyer of gold, and that could be putting some upward pressure on the price. The price of gold has still fallen, though, so even Chinese buying has at most slowed the decline. (For more, see: Get to Know the Major Central Banks.)

ETFs

Besides central banks, exchange traded funds (ETFs) are now major gold buyers and sellers, such as the SPDR Gold Shares (GLD) and iShares Gold Trust (IAU), which allow investors to buy into gold without buying mining stocks . Both offer shares in bullion and measure their holdings in ounces of gold. The SPDR ETF currently holds about 9,600 ounces while the iShares ETF has about 5,300. These ETFs, though, are designed to reflect the price of gold, not move it. (For more, see: Which Gold ETF Should You Own?)

Portfolio Considerations

Speaking of portfolios, Hug said a good question for investors is what the rationale for buying gold is. As a hedge against inflation it doesn't work well, but looked at as a piece of a portfolio then it's a reasonable diversifier. It's just important to recognize what it can and cannot do.

In real terms gold prices topped out in 1980, when it hit nearly $2,000 per ounce (in 2014 dollars). Anyone who bought gold then would have lost money. On the other hand the investors who bought it in 1983 or 2005 would be happy selling now even with recent price drops. It's also worth noting that the "rules" of portfolio management apply to gold as well. The total number of gold ounces one holds should fluctuate with the price. If one wants 2% of the portfolio in gold, then it's necessary to sell when the price goes up and buy when it falls. (For more, see: How Much Disaster Can Gold Hedge?)

Retaining Value

One good thing about gold: it does retain value. Erb and Harvey compared the salary of Roman soldiers 2,000 years ago to what a modern soldier would get based on how much those salaries would be in gold. Roman soldiers were paid 2.31 ounces of gold per year, while centurions got 35.58 ounces.

Assuming $1,600 per ounce, a Roman soldier got the equivalent of $3,704 per year, while a U.S. Army private in 2011 got $17,611. So a U.S. Army private gets about 11 ounces of gold (at current prices). That's an investment growth rate of about 0.08% over approximately 2,000 years. (For more, see: Does it Still Pay to Invest in Gold?)

A centurion (roughly equivalent to a captain) got $61,730 per year, or while a U.S. army captain gets $44,543 — 27.84 ounces at the $1,600 price, or 37.11 ounces at $1,200. So the rate of return is either -0.02% per year or nearly zero.

The conclusion Erb and Harvey made, though, was that the purchasing power of gold stayed pretty constant. It also seems unrelated to the current price.

The Bottom Line

If you're looking at gold prices, it's probably a good idea to look at how well the economies of certain countries are doing. As economic conditions worsen the price will (usually) rise. Gold is a commodity that isn't tied to anything else, so it makes a good diversifier in a portfolio in small doses. (For more, see: The 5 Best Performing Gold ETFs.)

Want to learn how to invest?

Get a free 10 week email series that will teach you how to start investing.

Delivered twice a week, straight to your inbox.