Predicting Inflation: Gold versus Bonds

Predicting Inflation

In my most recent market commentary, dated 2/9/10, I discussed how sensitive market prices are to future inflation expectations. To the extent that you can discern whether the market expectations for future inflation is too high or too low, you should be able to beat the market by using tactical asset allocation. During the market chaos of early 2009, long run inflation expectations had fallen to 1.0%. If you were able to predict that the federal reflation efforts would be successful (at least in the short term) and that market inflation expectations would rise to where they are today (2.6%), you could have caught the market bottom and benefitted from the 70%+ run up in the S&P since that time. I was too pessimistic and missed most of the run-up.

In normal times, the market appears to view 2.5% as the natural long-term inflation rate. The Fed claims to view 1.5-2.0% as its desired inflation rate. In my market equilibrium model, I have used 2.25%, but have tended to favor a range of +/- 0.25% with the acknowledgement that market outcomes aren't that precise. My general argument has been that the market is too focused on inflation, and that deflation is the primary threat. Many market pundits, on the other hand, have been proclaiming that the market (particularly the Treasury bond market) is massively underestimating inflation.

There are two primary market indicators for future inflation expectations: (1) the Treasury-TIPS spread and (2) the price of gold.

What bonds are telling us

As of 3/26/10, the Treasury curve looks as follows:

If you compare this curve to the curve at year end 2009, you'll see that the nominal Treasury rates have moved up slightly (the 30-year rose to 4.8% from 4.6%), TIPS rates moved up (the 30-year TIPS rate rose to 2.2% from 2.0%) and inflation expectations fell slightly (fell to 2.6% from 2.7%, yes there is rounding involved here). The TIPS curve is steep and the inflation curve less so, which makes sense. Below is what I consider to be the "equilibrium" yield curve, using my 2.25% inflation rate as a target:

The bond market is basically saying that the Fed will be a little slow to remove accommodation (which is why short rates are below equilibrium), but that long run inflation expectations are well-anchored.

What gold is telling us

The gold market is telling us something different. A rise in the gold price tends to lead a rise in commodities prices, which in turn leads a long term rise in CPI. Gold also has a history of volatility and of overshooting its equilibrium, however.

Below is a chart of the price of gold since the 1950s:

The gold price was fixed at its Bretton Woods price of $35 per ounce from the early 1930s to the late 1960s, jumped to (briefly) over $800 per ounce in 1980, fell to a new equilibrium range of around $375 per ounce from the mid-1980s through 1995, fell again in the late 1990s to $250 per ounce, and then rose during the past decade to over $1,000 per ounce.

If gold wasn't so volatile, we could expect it to rise steadily over time as the Fed promotes positive inflation. It should be noted that during the entire century between the early 1800s and the early 1900s, there was no sustained inflation in the United States as the dollar was convertible into gold at $20 per ounce. It is only since the Great Depression that the government has actively promoted inflation. In fact, because of the productivity of the capitalist system, you would expect prices to fall over time as producers became more efficient, not to rise as tends to be the case.

All things being equal, if you were expecting the Fed to target consumer price inflation of 2.25%, and you expect productivity gains of 1.75% per year, you would expect the gold price to rise by a rate of 4% per year to indicate that the dollar is being devalued accordingly.

Despite all the volatility in the gold price and inflation since the 1960s, the long term trend shows that gold has led increases in CPI, less a producitivity factor.

In the chart above, we can see from the trendline equation that CPI has been falling at a rate of 1.8% per year relative to gold since the 1950s, but that the relationship has been volatile. Gold was undervalued relative to consumer prices in 1970 and 2000, was overvalued in 1980, and was fairly valued in the mid-1990s. In many of my models I find that the mid-1990s was a well-balanced economy, before Robert Rubin's dollar bubble and the following dot com and housing bubbles threw the economy off kilter, to put it mildly.

One more factor to take into consideration is that economy-wide prices don't adjust instantaneously. I've found that it takes about 10 years for the overall CPI to catch up to a sustained devaluation of the dollar relative to gold. The inflation rate for the next 10 years is therefore dependent on whether monetary and fiscal policy are tightened enough so gold falls back in line with the current consumer prices, or whether prices keep rising to catch up with the new equilibrium level in gold.

My calculations of the theoretical gold price and CPI reveal the following:

  • If gold has overshot to the high side as a function of the financial crisis, and the treasury market's inflation rates are correct, then the current equilibrium gold price would be $550-600 per ounce.
  • If gold has found a new equilibrium level at the current $1,100 per ounce, then we should expect a 10-year inflation rate of over 5%.

This is a massive divergence.

  • If the bond market is correct, then bonds and non-precious metal commodities are fairly valued, and stocks are slightly overvalued.
  • If the gold market is correct, then most bonds are overvalued, TIPS and commodities are way undervalued and stocks and even real estate are slightly undervalued.

So what market should we listen to, gold or bonds?

The bond market is more liquid and understands the deflationary risks of high leverage. While the argument can be made that Treasuries are overpriced (yields too low) due to the Fed's holding rates too low and to Chinese manipulation. If nominal Treasuries were being manipulated by structural factors, then I would expect TIPS to be yielding nearly 0% across the curve. As can be seen above, however, the TIPS yield curve is showing positive real yields right around their equilibrium levels.

The gold market is a bit more ideological, attracting those that tend to dislike the government in power. It should be said, however, that the gold market was far more adept at tracking the debt bubble of the 2000s than the Federal Reserve proved to be, right through to predicting the extraordinary measures taken in the bubble's aftermath. The political landscape, on the other hand, is implying that consensus is building to fight the deficit, reform entitlements and to rein in financial system leverage. While that may be wishful thinking on my part, it would be hard to make the case that the wind is shifting in the other direction. My view is that the health care bill will be one of the last acts of giving us things for some time. The next decade will be occupied mostly by our government taking things away.


The answer to the debate is not obvious. Risk management, therefore, should be the primary focus for investors. I'll do my regular quarterly market update in a few days, but the lesson for me would be to overweight TIPS, commodities and cash, and to underweight precious metals. At the very least, it is probably time to trim bond positions back to a neutral weighting (I have been overweight bonds for years now). It also probably makes sense to have at least a neutral weighting in stocks for the time being, if we are to assume that the ultimate inflation outcome likely lies somewhere between what the bond market and gold market is predicting.

I am not a financial advisor. These analyses are conducted for personal enjoyment only.